Deliverable D2.1
Working papers on Assessment Methods
Project: PRACTICE
Prevention and Restoration Actions to Combat Desertification.
An Integrated Assessment
Coordinator: V.R. Ramón Vallejo.
CEAM Foundation, Spain
Grant Agreement no.: 226818
1
This Deliverable includes four Working papers on Assessment Methods and
Indicators to Evaluate Actions to Combat Desertification. The Working
papers have been prepared by the PRACTICE Expert-Assessment Board
(partners: UA, ABDN, CEAM, UTRIER, and CMCC), in the framework of WP2
(Task 2.1) and have been presented for discussion and refinement to the
PRACTICE Workshop on Assessment Methods and Indicators to Evaluate
Actions to Combat Desertification, held in Sardinia in October 2010.
PRACTICE (October 2010)
[email protected]
2
Table of contents
Preface
I.
Coupled Human-Environment Indicators for the Assessment of Dryland
Restoration
Stephen Whitfield and Helmut J. Geist
II.
1. introduction
2. Developing a human-environment framework
2.1. Ecosystem services for defining dryland degradation
2.2. Baskets and trade-offs in ecosystem services
2.3. The value of ecosystem as an indicator of resilience
2.4. Multiple stakeholders and multiple values
2.5. Value as a problematic concept
3. Dryland ecosystem services
4. Monitoring the value of ecosystem services
4.1. Valuing ecosystem services
4.2. Food, forage, fibre and fuelwood
4.3. Carbon sequestration
4.4. Water provision and conservation
4.5. Cultural heritage and landscapes
5. Conclusion
6. References
Ground-based Methods and Indicators for the Assessment of
Desertification Prevention and Restoration Actions
Susana Bautista and Ángeles G. Mayor
1. Introduction
2. Evaluation approaches
3. Selection of attributes and indicators for evaluation
3.1. Evaluation frameworks
3.2. Number, domain and scale of application of the assessment
indicators
5
7
9
9
10
11
13
14
14
15
22
22
24
28
29
30
31
32
41
43
43
46
46
47
3.2.1. Evaluation approaches based on few holistic
indicators
48
3.2.2. 3.2.2
Evaluation approaches based on large
suites of indicators
50
4. Common indicators for assessing dryland management and
restoration
51
4.1. An evaluation approach based on the comparative
assessment of the provision of ecosystem services
51
4.2. Common indicators for assessing dryland management and
restoration
5. Conclusions
6. References
52
55
56
3
III.
IV.
Remote Sensing Based Options for Assessment of Mitigation and
Restoration Actions to Combat Desertification
Achim Röder, Marion Stellmes, Joachim Hill, Thomas Udelhoven
1. Introduction
2. Coarse scale
2.1. Data sources
2.2. Diachronic analyses
2.3. Vegetation Indices
2.4. Time series analysis using hypertemporal satellite archives
2.4.1. Preprocessing
2.4.2. Assessment of seasonal components
2.4.3. Transitional components
2.5. Integrated interpretation concepts
2.5.1. Mapping functional vegetation types
2.5.2. Syndrome-based mapping approaches
2.5.3. Integrated land condition assessment
3. Fine scale
3.1. Data sources and time series setup
3.2. Standardisation requirements
3.2.1. Geometry
3.2.2. Radiometry
3.3. Derivation of vegetation-related information
3.4. Fine-scale temporal analysis
3.5. Land use / land cover classification
3.6. Integrated interpretation concepts
4. Plot scale
5. References
A Methodological Framework for Assessing and Valuating Ecosystem
Services and Mitigation Strategies against Desertification
Laura Onofri, Andrea Ghermandi, and P.A.K.L. Nunes
1. Introduction
2. Some taxonomy
3. Total economic value and ecosystem services economic tools
3.1. A survey of economic valuation techniques
3.2. Advantages, disadvantages and applicability of ecosystem
services valuation techniques
4. Valuing drylands ecosystem services and anti-desertification policies
4.1 Evaluation of anti-desertification policies through socioeconomic indicators
4.2. Desertification policy assessment: A cost-effective meta
indicator
5. Conclusions
6. References
62
64
67
67
69
70
72
73
74
77
77
77
78
79
80
80
83
83
84
88
92
94
96
97
98
115
117
118
119
121
124
127
129
136
137
138
4
Preface
There is a consensus on the need for the evaluation of management actions
to combat desertification, which ultimately can provide essential inputs for
decision-making (see, for example, UNCCD 20091). It is increasingly
recognised that the assessment of land condition must be based on both
biophysical and socio-economic attributes (MA 20052), which also applies to
the evaluation of restoration and management actions (SER 20043, Zucca et
al. 20094). The participation of stakeholders and the incorporation of local
knowledge in the assessment of environmental problems and potential
solutions have also been increasingly demanded by international institutions.
PRACTICE addresses these needs and demands by linking the evaluation of
management and restoration actions with knowledge exchange and social
learning through a participatory process. PRACTICE assumes the mutual
human-environment interactions in land-use change and simultaneously
considers both biophysical and socio-economic attributes (Fig. 1).
Figure 1. PRACTICE framework for the assessment of actions to combat desertification.
1 UNCCD. 2009c. UNCCD 1st Scientific Conference: Synthesis and recommendations. Note by the
secretariat.
ICCD/COP(9)/CST/INF.3.
Bonn:
UN
Convention
to
Combat
Desertification
(http://www.unccd.int/php/document2.php?ref=ICCD/COP%289%29/CST/INF.3)
2 MA (Millennium Ecosystem Assessment). 2005. Ecosystems and Human Well-being: Desertification
Synthesis. World Resources Institute: Washington, DC.
3 SER (Society for Ecological Restoration, Science & Policy Working Group). 2004. The SER Primer on
Ecological Restoration (http://www.ser.org/).
4 Zucca, C., Bautista, S., Previtali, F. 2009. Desertification: Prevention and restoration. In R. Lal (ed.),
Encyclopedia of Soil Science, Second Edition. Taylor & Francis Group, New York, US.
5
PRACTICE protocol is based on (1) key common indicators that represent
overall human-ecological system functioning, (2) site-specific indicators
identified by local stakeholders that are relevant to the objectives and the
particular context conditions, and (3) stakeholder perspectives.
The selection of the key common indicators and assessment methods has
been based on a series of working papers (WPs) produced by the PRACTICE
Expert Assessment Board within the framework of PRACTICE Work Package 2
(WP2. Assessment methodology). The WPs were presented and discussed and
further refined during the PRACTICE Workshop on Assessment Methods for
Prevention and Restoration Actions to Combat Desertification (PRACTICE
Deliverable D2.2), held in Sardinia in October 2010.
This report presents the refined versions of the working papers, which define
the conceptual framework that underlies the assessment methodology (WP I),
and review and propose ground-based bio-physical indicators and
assessment methods (WP II), remote-sensing approaches (WP III), and socioeconomic analysis to assess prevention and restoration actions (WP IV).
Susana Bautista
PRACTICE WP2 Coordinator
6
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
I. Coupled Human-Environment Indicators for the
Assessment of Dryland Restoration
Working Paper for the PRACTICE Project
Stephen Whitfield
Prof. Helmut J. Geist
7
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Front cover photo source: Focx Photography (Flickr)
CC: BY NC SA
Stephen Whitfield and Helmut J. Geist (August 2010)
Department of Geography and Environment
University of Aberdeen
Elphinstone Road
Aberdeen
AB24 3UF
Scotland, UK
[email protected]
8
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
1 Introduction
A study of the human-environment system is a fundamental component of EU
PRACTICE project, which, through WP2, aims to ‘develop integrated
evaluation tools to assess practices to combat desertification’ (PRACTICE,
2009) and apply and evaluate these tools within long-term monitoring sites
(LTEMs). The project has a unique opportunity to build on the progress made
by research groups such as REACTION, DESIRE, WOCAT and the Dryland
Science for Development (DSD) Consortium, by synthesizing a large body of
theoretical and case study literature on the human-environment aspects of
dryland management and develop tools that will facilitate the application of
this knowledge in the practical assessment of dryland management and
restoration.
This paper will outline a conceptual framework for the PRACTICE approach
from a human-environment perspective. The overarching concept of
ecosystem services is used as a basis for describing the aims of dryland
restoration and discussing indicators and methods for monitoring the progress
of restoration projects.
2 Developing a human-environment framework
The monitoring and feedback components of development projects are
often left neglected due to constraints on resources and short-sighted project
planning (Dipper, 1998). However, because of the dynamic and adaptive
nature of socio-ecological dryland systems (Holling, 1973; Folke, 2006),
monitoring and evaluating restoration efforts are essential for implementing
strategies that successfully adapt to changing conditions and stakeholder
needs. Indicators of sustainability provide a simple metric for evaluating the
success of a restoration project or mitigation action and are therefore
necessary for facilitating the monitoring and learning components of a
project, which is often neglected, particularly where resources are limited.
Indicators are most effective when they are directly linked to pre-defined
aims of a project, such that a change in indicators represents measurable
progression towards (or regression from) a desired end point or along a
desired pathway.
Dryland restoration projects might set-out with any number of specific goals
such as increasing the populations of specific vegetation species (Gao et al,
2002); reducing soil erosion (Sivanappan, 1995); increasing agricultural
productivity (Zaal and Oostendorp, 2002); improving access to water (Cotula,
2006); or conserving cultural landscapes (Moreira et al, 2006). Under the
umbrella of ‘dryland restoration’ might be included reforestation projects,
9
Department of Geography and the Environment
S.Whitfield and H. J.Geist
land-use conversion, protected area designation, water conservation
strategies and property right regime shifts (Bainbridge, 2007). As such, any
discussion of restoration as a whole needs to be equally broad in scope, and
indicators must be devised in such a way that they can be interpreted in line
with specific project aims and be adaptable to the changing dynamics of
interacting social and ecological spheres.
In this chapter, a brief overview of the large body of literature on socioecological system dynamics will be linked to a discussion of the merits of using
the value of ecosystem services as an indicator for monitoring dryland
restoration.
2.1
Ecosystem services for defining dryland degradation
‘Ecosystem services’ is a utilitarian device for describing the valuable
functions of an ecosystem to humans and as such it fits well within a
framework designed to be responsive to the changing dynamics of a socialecological system. Daily et al (1997) recognised that human societies derive a
variety of goods from natural ecosystems, a concept that has subsequently
been widely used in putting a value on ecosystems (e.g. Costanza et al, 1997)
and evaluating the impacts of environmental change (e.g. Worm et al.,
2006). The Millennium Ecosystem Assessment (MA) advanced the concept of
ecosystem services by providing a framework for categorising ecosystems
services and drawing links between ecosystem services and human wellbeing (MA, 2005). The MA defines ecosystem services as:
‘The benefits people obtain from an ecosystem, including provisioning
services such as food, water, timber, and fibre; regulating services that
affect climate, floods, disease, wastes, and water quality; cultural
services that provide recreational, aesthetic, and spiritual benefits; and
supporting services such as soil formation, photosynthesis, and nutrient
cycling.’ (MA, 2005: xiii)
These benefits contribute to a holistic understanding of the well-being of
stakeholders in the ecosystem, which represents not only income and
material needs, but also health, social relations, security, and freedom of
choice and actions (MA, 2005). The well-being of different stakeholders in an
ecosystem depends in part, but not entirely, on the relative cultural,
economic, social, and spiritual benefits that each derive from the
environment. The dependency of livelihoods on the services provided by
ecosystems is greater in drylands than in any other ecosystem (MA, 2005) and
the marginal properties of these ecosystems is such that the balance
between sustainability and degradation is a fine one (Geist and Lambin,
2004).
Defining dryland degradation as ‘a persistent reduction in the basket of
services provided by a dryland ecosystem over an extended period’ (an
10
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
adaptation of the MA 2005 definition), and therefore understanding effective
management (restoration) as being represented by a long term increase in
the relative value of ecosystem services’ has several benefits:
• A dryland ecosystem provides a range of services. By focusing on a
basket of services (DSD WG1), rather than individual components of a
dryland ecosystem, adaptation strategies in land management can be
considered. The Millennium Ecosystem Assessment (2005) recognised
that trade-offs between services may be strategically made in
adapting to changing pressures and maintaining well-being – it
therefore fits comfortably with the concept of resilience.
2.2
•
Considering interactions and trade-offs between ecosystem services is
a useful way of defining and identifying stakeholders at all scales as
those who derive direct or indirect benefits from services provided.
Reynolds and Stafford Smith (2002) explain that the unsustainable use of
ecosystem services at the local level can place increased pressure on
ecosystems (and affecting stakeholders) elsewhere. For practical
purposes, however, it is necessary to define the boundaries of an
ecosystem at the local scale and work outwards (Fisher et al, 2009).
•
It is a common language that can be applied in land management
from the perspectives of multiple stakeholders at multiple levels.
•
By concentrating on a persistent reduction (or long-term increase), the
definition makes a clear distinction between long-term trends and
short-term fluctuations, which create noise in the observation of trends
in ecosystem development (Scholes, 2009). There is a consensus within
literature on dryland degradation monitoring and sustainable
management assessment that attention should be primarily focused on
slow variables (Reynolds and Stafford-Smith, 2002).
Baskets and trade-offs in ecosystem services
Interactions between ecosystem services are complex and often not welldefined (MA, 2005; van Jaarsveld et al, 2005), particularly where they result
from human interactions within the ecosystem. The MA (2005) recognised that
trade-offs between services may be strategically made by stakeholders in
adapting to changing system pressures and maintaining well-being. There is a
wealth of literature on local environmental knowledge in drylands that
highlights examples of such adaptive strategies (see Reed et al, 2007).
Furthermore, ecosystem services are strongly interlinked; the exploitation of a
particular ecosystem service can have implications for the whole basket of
services (De Fries et al., 2004), a change to the system of resource
management may increase certain ecosystem services whilst degrading
others (Reyers et al., 2009), and as such, a sustainable system may be one
that moves between several stable states (or baskets) (Folke et al, 2004). It is
11
Department of Geography and the Environment
S.Whitfield and H. J.Geist
for these reasons that it is necessary to take a holistic approach to monitoring
drylands from an ecosystem services perspective, considering a whole
‘basket’ of ecosystem services rather than individual services.
The resilience alliance (Holling, 2001) combined several social-ecological
system dynamics – threshold effects and non-linearity (Scheffer et al., 2001),
adaptation through feedbacks (Walker and Meyers, 2004), legacy effects
and time lags (Liu et al, 2007) – in describing the sustainability of a system in
terms of resilience.
The concepts of resilience and sustainability are
somewhat interchangeable, in both cases we can consider a resilient, or
sustainable, system as one with the capacity to avoid going beyond critical
thresholds (tipping points) or recover from surpassing non-critical thresholds
(turning points) through adaptive strategies. This concept might be extended
to describe dryland restoration as actions to rebuild the adaptive capacity of
systems that have passed critical thresholds.
Table 1: A description of social-ecological system dynamics illustrated with examples
from the literature
Socio-EcoSystem Property
Thresholds
Description
Human-environment examples in
the literature
A point reached in a linear Kinzig et al (2006) identify a soil
trend that causes a shift in salinity threshold between productive
agricultural land and non productive
the system state
land in the Western Australian
wheatbelt
Non-linearity
Dependent factors within a
system that do not respond to
each other with proportional
magnitude
Koch et al (2009) describe nonlinearity in coastal protection (an
ecosystem service) provided by
marshes, mangroves and coral reefs.
Adaptation
through
feedbacks
A response to a system
change that acts to increase
(positive
feedback)
or
decrease (negative feedback)
its effects – i.e. adaptation
through exploiting ecosystem
services
Gordon (2007) explores feedbacks in
the management of socio-ecological
systems in the Great Barrier Reef
catchments, describing links between
upstream land
management, soil
erosion,
downstream
sediment
deposition and reef degradation
Lambin and Meyfroidt (2010) describe
socio-ecological feedbacks that slow
down deforestation in Vietnam
Legacy effects
and time lags
The effects of a defined
system
event
continue
beyond, or may not be
realised until sometime after,
the event.
Liu et al (2007) show how the
introduction of a keystone species in
the Wisconsin Highland Lake District
continues
to
restructure
fish
populations for decades afterwards
12
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Case studies of adaptation in drylands show how land users or dryland
stakeholders substitute the utilisation of certain services for that of others in
order to maintain or improve their well-being in response to changing system
properties (e.g. Enfors and Gordon, 2007). Adeel and Safriel (2008), for
example, describe the adaptation of a community at the edge of the
Cholistan Desert in Pakistan, who have adapted to a reduction in ground
water availability by switching their predominant income source from cattle
grazing to aquaculture (breeding carp), which, through controlled conditions
and careful management of surface water, requires less water investment.
Adaptation through ecosystem service substitution also takes place in
response to all manner of system changes, from climate change to
population growth and from cultural change to market fluctuation (Adger
2000). A community in the Dana Biosphere Reserve, Jordan, for example,
have adapted to unsteady market conditions for their food produce by
switching to production of a non-food, high-end product – olive oil soap, and
on the basis of this unusual industry they have also attracted additional
tourism (Adeel and Safriel, 2008).
Importantly, within resilience thinking, a system is understood as being mobile
between steady states rather than having just one. In the language of
ecosystem services, several significantly different compositions of services
within the basket may result from sustainable land management (SLM) and
systems are capable of moving between baskets through adaptation to
system processes (Carpenter and Brock, 2004; Janssen and Carpenter, 1999).
Restoration is essentially about building the resilience of a system, utilising its
ecosystems services in such a way that it doesn’t reduce the capacity of the
system to adapt by altering its basket of services, or reduce the well-being of
those with a stake in the ecosystem.
2.3
The value of ecosystem services as an indicator of resilience
The value of a broad basket of ecosystem services represents a strong
indicator of the resilience of a dryland system as it reflects the potential
pathways and substitution options for achieving an increase in the well-being
of stakeholders (MA, 2005). Without prolonged and intensive data collection,
which would contradict the rationale behind employing indicators, it is very
difficult to identify and assess all of the ways in which each stakeholder
derives different components of well-being from different services,
furthermore, it is unrealistic to expect to weigh-up the relative importance of
these different components. In this respect, valuing a basket of services is
somewhat crude, but although it may smooth over some of the intricacies of
the human-environment interactions within the system, it does strongly reflect
the overall capacity of the ecosystem to contribute to the well-being of its
stakeholders.
13
Department of Geography and the Environment
S.Whitfield and H. J.Geist
The value of ecosystem services derived by stakeholders is partly indicative of
the dependence of their well-being on the ecosystem, and the changing
value of ecosystem services reflects a changing component of the well-being
of stakeholders that is related to the environment. Whereas a biophysical
approach might focus solely on the quantity of services provided (e.g. the
amount of carbon sequestered by an ecosystem), a human-environmental
perspective, which is concerned more with how ecosystem services translate
into human well-being, necessarily focuses on the value of services to
stakeholders.
2.4
Multiple stakeholders and multiple values
Different stakeholders, at different scales, exist within socio-ecological systems
of different properties (Schwilch et al, 2009). The combination of countless
social, economic, cultural and environmental factors, operating at multiple
scales, which creates a local context for dryland degradation is such that
each individual socio-ecological system is unique (Warren, 2002; Schwich et
al, 2009) and is uniquely experienced by the range of people with a stake in
it. This presents a problematic reality for the appropriateness of using generic
tools in describing desertification and sustainable dryland management. A
holistic assessment of dryland management requires multi-scale and multistakeholder analysis (Reyers et al, 2009). At the local scale, sustainable
management can often appear as degradation because of a lack of
understanding about how land managers prioritise and strategically substitute
different ecosystem services, particularly those that represent cultural services
(Jollands and Harmsworth, 2007). Incorporating local environmental
knowledge within an assessment of dryland management is imperative for
understanding how degradation is defined and experienced by those whose
well-being is most directly linked to the services provided by a particular
ecosystem (Danielsen et al, 2008). However, there is a significant challenge in
attempting to monitor the value of ecosystem services across multiple
stakeholders at multiple scales, particularly where there are significant
differences in the benefits that different stakeholders receive from ecosystem
services, such that a given scenario might result in an increase in value of
ecosystem services for some stakeholder groups, but a decrease in value for
others.
2.5
Value as a problematic concept
The mainstream approach to linking ecosystem services and well-being has
been socio-economic or monetary valuation. The advantages of putting a
monetary value on ecosystem services – those of comparability, quantitative
analysis, and the potential avenues it opens for sustainability through
payment (weak sustainability) – have been heralded and embraced within
14
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
the literature (Reid, 2006; Costanza et al, 1997; Daily et al, 1997) as
vociferously as they have been criticised by those who argue against the
intrinsic failings of economic methodologies (Toman, 1998) and the moral and
philosophical problems of putting monetary values on nature (McCauley,
2006). An economic valuation of the ecosystem has its limits, particularly with
regards to considering the non-utility value of the system, for example its
existence value and long-term sustainability (Spangenberg and Settele, 2010).
Value statements need not be expressed in monetary terms, qualitative
description of value or semi-quantitative scoring and ranking systems (see
DEFRA 2007) can generate informative and comparable information. It is
suggested that monetary valuation of ecosystem services is avoided within
monitoring and assessment because of the tendency it creates within
participants to revert to a prioritisation of individual and market-related
benefits. Discursive and qualitative approaches allow greater scope for
participants to express values that relate to the sustainability of the system
and the needs of future generations, and language-based analysis of
discussions provides much richer information on the opinions of stakeholders.
3 Dryland ecosystem services
The Millennium Ecosystem Assessments’ (MA) ‘Desertification Synthesis’ report
brings together work from the four working groups of the MA in an attempt to
answer the following questions:
How has desertification affected ecosystems and human well-being?
What are the main causes of desertification? Who is affected by
desertification? How might desertification affect human well-being in
the future? What options exist to avoid or reverse the negative impacts
of desertification? And how can we improve our understanding of
desertification and its impacts? (MA, 2005)
As such it represents a comprehensive summary of global dryland systems
and provides a useful foundation on which to build on in an attempt to define
indicators for the assessment of dryland restoration projects based on the
value of ecosystem services. The MA outlines the following list of the key
dryland ecosystem services, presented in accordance with its services
framework:
15
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Table 2: Ecosystem services from drylands categorised in accordance with the MA
schema. Source: Millennium Ecosystem Assessment (2005) Ecosystems and Human
Well-Being Desertification Synthesis
Provisioning
Service
Goods produced or
provided by
ecosystems
•
•
Provisions derived from biological
productivity: food, fibre, forage, fuelwood,
and biochemicals
Fresh water
Regulating Service
Benefits obtained
from regulation of
ecosystem
processes
•
•
•
Water purification and regulation
Pollination and seed dispersal
Climate regulation (local through
vegetation cover and global through
carbon sequestration)
Cultural Service
Nonmaterial
benefits obtained
from ecosystems
•
•
•
•
•
Recreation and tourism
Cultural identity and diversity
Cultural landscapes and heritage values
Indigenous knowledge systems
Spiritual, aesthetic, and inspirational
services
Supporting Service
Services that
maintain the
conditions for life
on Earth
•
Soil development (conservation,
formation)
Water conservation
Nutrient cycling
•
•
The basket of ecosystem services outlined by the MA is comprehensive, but in
many cases is too large to translate into a practical assessment tool. The
ecosystem services listed can be categorised in a number of ways, but
depending on the vagueness of definition, this basket could reflect anything
from four to twenty-six categories. Clearly, the more narrowly defined an
indicator is, the more consistently it can be applied across research projects
and locations; however the universality of indicators is compromised by
specificity. Defining an assessment tool that is universally applicable, but
practical, will require a limited suite of indicators that are well-defined, but
contextually adaptable and representative of the broader basket of
ecosystem services provided by the dryland system. The challenge taken up
here is to select between three and five services that combined can
adequately and universally reflect the full basket of ecosystem services
provided by drylands and the variety of linkages that are made between this
basket and the well-being of stakeholders.
In this chapter a wide-reaching literature review is drawn on in determining
and defining those ecosystem services which, in combination, best reflect a
common basket of ecosystem services provided by a dryland system of
unspecified social, cultural, economic and biophysical properties. It is
important to bear in mind is that the relative importance of each ecosystem
16
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
service within the basket differs (spatially and temporally) across socioecological systems and across stakeholders, and as such what is presented
here can only represent a guideline that should be altered and adapted to
suit specific case studies.
The following criteria are used here as a basis for narrowing down the
ecosystem services to a practical number for universal assessment purposes,
they describe the desired properties of the services that should be selected.
•
Interpretable and of measurable value at the local scale – such that
stakeholders and researchers alike can draw clear conceptual links
between the ecosystem function and the well-being that is derived
from it
•
Identifiable within peer-review case-study literature as being key
services in all major global dryland regions – particularly as being
relevant in both more- and less-economically developed regions
•
Indicative of long term change rather than short-term variability
(however, this is more of an issue for the valuation methodology;
addressed in the next chapter)
And in combination, the group of services selected should:
•
Cover the three broad categories of ecosystem services with direct links
to human well-being: provisioning, regulatory and cultural (supporting
services have indirect links with human well-being and their valuation
can be methodologically problematic)
•
Conceptually link to all four of the categories of human well-being that
are directly related to ecosystem services: security, basic material for
good life, health, good social relations
The following table summarizes literature, which reflects the potential of each
of the ecosystem services identified by the MA to be used as indicators of the
full basket of services. The categorisation of these services reflects that of the
MA.
17
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Service
Food, fibre
Provisioning and forage
Services
Summary of available literature
• A service of fundamental importance in all dryland regions
and the main source of income or subsistence for the
majority of land users (Zhen et al, 2010; Kerven et al, 2007;
Russelle, 2007; Schulz et al, 2010; Luck et al, 2003; Pereira
et al, 2006)
• Interpretable and adaptable at local scales in accordance
with local and seasonal crops/livestock/land use (MA, 2005;
Kerven t al, 2007)
• Closely linked to three fundamental components of human
well-being: security (Nyariki et al, 2002); basic material for
good life; and health (Corvalan et al, 2005)
Fuelwood
• Wood is harvested for fuel in almost all dryland regions, but is
of particular importance to the livelihoods of local land users
in less economically developed regions (Zhen et al, 2010;
Kerven et al, 2007; Schulz et al, 2010; Dubroeucq and
Livenais, 2004; Tovey, 2008; Bautista et al, 2010)
• Ranges from subsistence harvesting ( Zhen et al, 2010) to
large scale commercial production (Kakuru et al, 2004)
• Quantities and species used are locally and temporallyspecific
• Closely linked to two fundamental components of human
well-being: basic material for good life (Dovie et al, 2004);
and health (Corvalan et al, 2005)
Biochemicals
• Some commercial opportunities but a sparse literature cites
few examples of biochemical extraction in drylands (e.g.
Guaseelan, 2009)
Fresh water
• A fundamental component of two components of human wellbeing: basic material for good life; and health (Corvalan et al,
2005)
• Typically low volumes of accessible fresh water in drylands
means that it can represent a critical determinant/threshold in
system properties (Huber-Sannwald et al, 2006; Reynolds et
al, 2007) but is of little commercial significance in the majority
of drylands
• May be of importance to a wide spectrum of stakeholders
(Jiang et al, 2005; White et al, 2002)
18
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Regulating
Services
Water
purification
and regulation
• Because of the critical importance of freshwater in almost all
dryland ecosystems, the capacity for purification of water is a
key service (Carpenter et al, 2006), particularly where water
used in agriculture contaminates groundwater (Scanlon et al,
2005)
• Complicated process that depends on many interacting
system properties (climate, vegetation, soil structure,
geology, water extraction technology, irrigation)
• May be of importance to a wide spectrum of stakeholders
(Jiang et al, 2005; White et al, 2002)
• A fundamental component of two components of human wellbeing: basic material for good life; and health (Corvalan et al,
2005)
Pollination and
seed dispersal
• An important ecosystem service in all major dryland regions
(Kerven et al, 2007; Proctor et al, 2002; Bennett et al, 2009;
Reyers et al, 2009; Safriel, 2006)
• In certain ecosystems the vast majority of vegetation may
rely on seed dispersal by wind, this limits the usefulness of
the value of pollinating services as an indicator of dryland
health (Standish et al, 2007; Benayas et al, 2008)
• Closely linked to two components of human well-being: basic
material for good life (MA, 2003; Luck et al, 2003; Balmford
and Bond, 2005); and security (Lal, 2002)
Local climate
regulation
• Very little literature on local climate regulation by dryland
ecosystems, probably because the hydrological cycles
associated with these systems are relatively unresponsive
(Hope et al, 2004)
• Wind breaks offered by structural vegetation may be an
important service in drylands at risk from wind-induced soil
erosion (Mukhopadhyay, 2009; Masri et al, 2003)
• A complicated ecosystem service to monitor because of the
difficulty of untangling local, ecosystem-related climate
change from broader scale climate variations and trends
(Stohlgren et al, 1998; Herrmann et al, 2005)
Slope
stabilisation
• Only relevant in mountainous and hilly dryland regions
• Value of the ecosystem service is closely linked to the value
of the predominant land use (provisioning services), in few
cases there may be a risk to lives, property and infrastructure
associated with slope instability in drylands (Dregne, 2002;
Wenlan et al, 2006)
• Slope stability depends strongly on climate events as well as
the stabilisation that is offered by vegetation structure
(Christanto et al, 2009)
19
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Regulating
Services
Cultural
Services
Carbon
sequestration
• The capacity for carbon sequestration and storage of dryland
ecosystems is significant (Lal, 2002; 2004), but depends on
the vegetation composition and soil properties, as such its
significance as an indicator of dryland health is highly
spatially variable (Lal, 2002; Kerven et al, 2007; Asner et al,
2003; Glenn et al, 1993; Carretero et al, 2007; Lal, 2004;
Bautista et al, 2010)
• Implications for stakeholders on a global scale, and because
of the fragile nature, and climate dependence, of drylands it
will be of value of local land users (Dietz et al, 2004; IPCC,
2007), however valuation by land users will rely on their
ability to make complex conceptual links between carbon
storage and climatic change at multiple scales
• Connected to all four components of human well-being:
security, basic material for good life, health, good social
relations
Recreation
and tourism
• Only of significant contemporary value where a tourism
industry is established, although this may also be thought of
in terms of potential value
• The commerciality of the tourist industry may depend on
levels of investment and as such it may be of significance to
a wide range of stakeholders (from large corporations to
small informal entrepreneurs and including tourists
themselves) at multiple scales (White et al, 2002; Portnov
and Safriel, 2004)
• Significant primarily as a source of income and as such it has
indirect links to all four components of human well-being:
security, basic material for good life, health, good social
relations (MA, 2005)
Cultural
identity and
diversity
• In the majority of dryland ecosystems the livelihoods and
lifestyle of local land users is intrinsically linked to the land
and as such the ecosystem makes up some component of
their cultural identity (Kerven et al, 2007; Moreira et al, 2006)
• In practical terms the value of this component of an
ecosystem service may be very difficult to separate from the
value put on other components such as provisioning services
• From the premise that cultural identity changes with the
ecosystem, it becomes problematic to objectively value one
identity relative to a previous (or to several alternative)
identity(ies) (Kumar and Kumar, 2008).
20
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Cultural
Services
Supporting
services
Cultural
landscapes
and heritage
values
• A cultural landscape in dryland ecosystems is created by a
history of land use (however long or short) and as such is a
concept that is applicable within almost all dryland socioecological systems (Jiang, 2003; Kerven et al, 2007; Moreira
et al, 2006; Schultz et al, 2010; Pereira et al, 2006)
• Because a cultural landscape is not static, this value should
not be interpreted as the value of conserving ‘traditional’
practices. It is a value that is difficult to conceptualise, but
perhaps should best be presented as the social-ecological
system allowing users to continue to exercise choice in
respecting cultural practices and self-defined tradition.
• It is a component of the ecosystem that should be defined
and interpreted by local stakeholders
• Closely linked to three components of human well-being:
security, basic material for good life, and good social
relations (MA, 2003)
Indigenous
knowledge
systems
• Only appropriate in those dryland in which the land users
consider themselves indigenous
• Knowledge is often passed down by land users through
demonstration and practice, this ecosystem service
represents the capacity of land users to continue to practice
land use in a self-defined ‘traditional’ way (Alverson, 1984);
thus it is affected not only the biophysical state of the
ecosystem, but also the institutional , political and economic
factors that affect the freedom of choice of land users (Geist
and Lambin, 2004)
• It is a component of the ecosystem that should be defined
and interpreted by local stakeholders
• Closely linked to three components of human well-being:
security, basic material for good life (Berkes et al, 1995;
Thomas et al, 2000), and good social relations (MA, 2005)
Spiritual,
aesthetic, and
inspirational
services
• Very sparse literature on spiritual and aesthetic services
specific to dryland ecosystems (e.g. Upton, 2010). The
relative importance of this ecosystem service will vary widely
across different dryland locations
• It is a component of the ecosystem that should be defined
and interpreted by local stakeholders
• Closely linked to three components of human well-being:
security (Butler, 2006; MA, 2005), basic material for good life,
and good social relations (MA, 2005)
Soil
development
• Nitrogen content, salinity, soil depth, soil crust may all act as
key constraints/supports of dryland ecosystem function.
• Particularly closely linked to the productivity of land
Water
conservation
• The primary constraint on dryland system functioning
• Indirectly linked to all components o human well-being and a
key constraint on all of the services listed above in most
dryland environments
21
Department of Geography and the Environment
S.Whitfield and H. J.Geist
In accordance with the criteria outlined, it is suggested that for a dryland
system of unspecified social, cultural, political, economic and biophysical
properties, the dryland ecosystem services basket focused on should include:
the provision of food, fibre and forage, and fuelwood, carbon sequestration,
water provision and conservation and cultural landscapes and heritage
values. However a review of literature and participatory research within a
particular dryland socio-ecological system may reveal that certain ecosystem
services are of particular significance (or insignificance) to that context and
the basket of services used should be altered accordingly. Appendix A
presents a more detailed review of literature on two particular dryland
contexts (the Kalahari and the Mongolia rangelands) and illustrates similarities
and differences in the compositions of the basket of ecosystem services
between the two.
The following chapter takes this basket of representative ecosystem services
and considers in more detail how its value can be measured and monitored
as an indicator of successful dryland restoration/degradation mitigation.
4 Monitoring the Value of Ecosystem Services
This chapter will outline some of the methodological issues involved in
measuring the value of each of the ecosystem services within the indicator
basket and it will make suggestions about how trends in these values might
be most effectively monitored. For each ecosystem service an outline of the
key stakeholders is given.
4.1
Valuing ecosystem services
A participatory approach to stakeholder valuation of ecosystem services is
advocated here. We suggest that valuation should be done in a deliberative
way, bringing together groups of stakeholders and experts such that they
might share their knowledge and facilitate the adoption of alternative
perspectives by stakeholders. Such an approach targets consensus about the
value of ecosystem services, although in many cases this will be an ideal
rather than a reality. The public nature of ecosystem services and the
realisation that sustainability value can only be captured through an
understanding of the values of non-human agents and future generations,
means that it is necessary for those valuing ecosystem services to be able to
achieve a perspective that looks beyond the utility that they derive as
individuals (Owens 2000). Through engagement in a deliberative process that
targets knowledge-transfer and consensus-based decision making, small
groups of citizens are able to base judgements of value on communal
benefits (Ostrom, 1990; Owens 2005). Deliberation involves the presentation of
alternative viewpoints, including those of ecological experts, to a group of
stakeholders, such that they might become better informed about the
22
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
functions of the ecosystem and the ways in which utility is derived from it.
From informed positions, stakeholders debate and hold each other to
account for the opinions that are expressed. With an emphasis on learning,
justification and logic, deliberation targets a consensual or democratic
outcome (Baber and Bartlett 2005; Owens 2000). Pellizzoni (2001) highlights
the civic virtue of a deliberative model; he explains that the opportunity to
engage with alternative perspective and the responsibility that comes with it
encourages individuals to be better citizens that are less prone to acting
strategically to their own individual end. A civic valuation of ecosystem
services has greater potential to be sensitive to the values of future
generations and the inherent values of ecosystems (Parfit 1984), particularly
where a case for sensitivity can be made within the deliberative process and
stakeholders are able to learn more about the ecological functioning of the
system through it.
Combining the multiple valuations of multiple stakeholders requires sensitivity
to the social context of the environment being studied. It may be appropriate
to place a social weighting function in combining the results of multiple
stakeholders, such that the values of those whose livelihoods are most closely
linked to the ecosystem services (i.e. local land users) are considered a more
important determinant of sustainability than the values of those with an
indirect stake in the ecosystem. This issue is contentious and must be resolved
in a way that is appropriate to the specific location.
Tools available to the researcher concerned with monitoring, and particularly
modelling, ecosystem services are becoming increasingly sophisticated.
InVEST is a computer model that has been developed as part of the Natural
Capital Project to map the distribution and value of ecosystem services within
defined boundaries. Parameters within the model are capable of simulating
the ecosystem services of carbon sequestration, storm peak management,
soil conservation, pollinators, water quality and the market value of
commodity production. It is capable of running scenario projections based
on a relatively small suite of input variables and is therefore an ideal model to
use for evaluating the supporting and regulating service components of
dryland ecosystems. The outputs of models may be utilised in stakeholder-led
discussions and valuations.
Whether or not a computer model is used, understanding and defining the
parameters of the socio-ecological system is an important component of
monitoring restoration efforts. Defining and mapping the boundaries of the
social-ecological-system is a necessary first step, such that a frame of
reference exists consistently for the long term and interdisciplinary monitoring
process. Defining boundaries should be a participatory process involving
discussions with key local informants (village elders, wildlife service
administrators, local councillors, etc.) about important political, social, cultural
and biophysical boundaries and should make use of remotely sensed maps
(vegetation/soil properties). Categories for defining boundaries will inevitably
depend on the characteristics of the social-ecological-ecosystem being
23
Department of Geography and the Environment
S.Whitfield and H. J.Geist
studied. Identifying the ecosystem services that are derived from the drylands,
and understanding what drives the decision of local people to make use of
certain services in a particular place and time, is something that might be
best achieved through close engagement with local stakeholders. The
intricacies of social-ecological relationships at this local level do not lend
themselves to quantitative definition, but aim here should be to qualitatively
describe the socio-ecological system in terms of ecosystem services, with
specific attention paid to thresholds, feedbacks, adaptation strategies and
critical determinants of change. It is important to note that the dynamics of
the socio-ecological system are changeable over time, and as evaluations
are carried out periodically in monitoring management, it may also be
appropriate to alter the parameters of the model.
One of the advantages of monitoring the value of ecosystem services over
time is that it is based on relative rather than absolute values. As such, the
significance of some of the issues over the accuracy of the economics are
avoided. In fact, a purely comparative application of the value of ecosystem
services may not require values to be expressed in monetary terms, but rather
can be done through a retrospective or prospective consideration, and nonmonetary description, of the changes in value over a given period. Such an
approach, of course, is not immune to its own methodological pitfalls,
particularly with regards to its accuracy and vulnerability to
romantic/optimistic/pessimistic bias.
The following discussion will describe both absolute and relative approaches
to the valuation of each of the considered ecosystem services. Combining
the two approaches would provide the most reliable and accurate
monitoring tool, but the extent to which either of both approaches can be
applied will depend on the scope and capabilities of the research being
undertaken. The DEFRA 2007 report, ‘an introductory guide to valuing
ecosystem services’, describes a participatory approach to valuing
ecosystem services which is advocated here.
4.2
Food, forage, fibre and fuelwood
The first step in measuring the value of provisioning services, such as food
production, is to become familiar with all of the ways in which the land is used
for cultivating, gathering and grazing and how this production translates into
meeting the dietary needs of local people and, in some cases, contributes to
the economic activity of land users. This is an ecosystem service that is
potentially highly reactive to changes in the social-ecological-system; a
balance between commercial and subsistence activity, for example, may
alter in response to market or climatic changes, as may agricultural activity
and the diets and dietary requirements of local people (Mortimore and
Adams, 2001; Barbier, 1996; Watts, 1987).
24
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Monitoring the value of this ecosystem service is somewhat data-intensive,
records of quantities of the full range of food products harvested, collected,
hunted, gathered, cultivated and reared for a representative sample of land
users need to be combined with information collected on dietary
requirements and social and cultural connections to food. This is best done
through the application of participatory research tools (such as Rapid Rural
Appraisal described by Chambers (1994)). Remotely-sensed data can
provide useful triangulation and a broad perspective on trends in agricultural
trends by offering a visible and quantitatively comparable description of crop
health.
Where there is a market for crops being produced, even if land users are
purely subsistence farmers, the market value of what is being produced on
the land can be a good measure of the social value of the ecosystem service
(De Groot et al, 2002), because the alternative to producing the food is
usually to purchase it. There are problems with such an opportunity cost
approach though, particularly where local people’s food comes through a
number of pathways (e.g. hunting, gathering, agriculture, market), as is most
commonly the case in drylands (e.g. Owuor et al, 2005). It may be
inappropriate to base calculations on a direct and linear relationship
between quantity and value in production, above certain thresholds (e.g.
production above the dietary requirements of local people in subsistence
communities) the increase in value of production will begin to decline with
increases in production. Food production may well come at a cost to other
ecosystem services, particularly where it has involved the input of chemical
fertilisers and pesticides, the clearing of trees/vegetation, or irrigation. These
marginal costs may not be reflected value of food. Using simple market
values, or even expressing values in monetary terms, is problematic in cases
where local people are completely disconnected from the market economy
and therefore have little concept of the value of money (Hein et al, 2006).
The alternative is stakeholder led description of the value – use an impact
pathways approach whereby stakeholders are asked to assess the severity of
trends in a semi-quantitative way (Defra, 2007).
It is necessary to find a methodological approach that is context-appropriate
and in most cases this will involve a combination of participatory tools, a mix
of qualitative and quantitative descriptive tools and a consideration of both
market values and meeting local needs in outlining a series of aim for
restoration in terms of food production and determining some metrics of
value (e.g. ability to meet local dietary needs; market share; balanced diets)
using such metrics allows for consistency in the valuation process and are
themselves consistent with the aims of the restoration.
Measuring the value of food production requires long term monitoring,
because of its sensitivity to short term and seasonal variation. Averaging out
annual food production is likely to provide the most appropriate unit for
monitoring trends, ideally (but depending on data constraints) this production
25
Department of Geography and the Environment
S.Whitfield and H. J.Geist
should be mapped across the ecosystem in order to identify areas of greatest
change.
Similar to measuring the value of food production, the value of fuelwood will
depend on the quantities harvested and the needs of the local people,
which again may range from subsistence to commercial. The need for
fuelwood is similarly sensitive to short-term and seasonal variations, thus
requiring annual averaging, and the value may well be reflected in the
market price in those locations where fuelwood is available to buy. There are
a couple of challenges in measuring the value of fuelwood, both of which
require an approach that is context-specific and participatory. The first is
access and property rights, these can be quite complicated for non-timber
forest products such as fuelwood, and it may be the case that there is dispute
over the legality of taking deadwood from privately owned property or
licences to extract fuelwood from communal property – this is particularly the
case with fuelwood, as it may be taken as dead or live wood, which may fall
under different legislation. Legislation often, especially in drylands, will
stipulate a maximum level of extraction of fuelwood, although again whether
or not this includes collecting deadwood may be contentious. The
importance of legality in the extraction of fuelwood depends on the context
and the strength of legislation within it. If one was looking solely at the
ecosystem properties, then the quantity of fuelwood production would be a
sole consideration, but in the social-ecological-system, legislation and social
norms are an intrinsic component of the system, and as such legislatory limits
may represent a threshold in the value of the commodity – the social cost of
illegal harvesting might well be factored into calculations above this threshold
and where fuelwood is harvested commercial an influx of illegally harvested
fuelwood in the market will affect the market value of fuelwood.
The second complexity comes in considering the opportunity costs of
fuelwood. Living trees can be assigned a whole host of alternative values (as
timber, as carbon sinks, as cultural heritage, as wildlife habitats, as
reproductive agents etc.). Different trees will be more or less valuable as
different ‘service providers’ and the wood harvested as fuelwood may be
done so because of it burns well or it may selected because it does not make
good timber. In almost all cases, fuelwood could be valued as an alternative
commodity. Ecologists would argue that even deadwood has important
ecological value and should therefore not be considered a free commodity.
The reality is that once fuelwood is harvested, chopped up, and put up for
sale in the market, its market value is unlikely to reflect all of its alternative use
values (especially those for which there is no market). Therefore basing
valuations of fuelwood solely on market values, rather than on social value,
may not adequately reflect the true value of this important ecosystem
service.
Again, a mixed-method approach to drawing up aims for restoration in terms
of fuelwood is an appropriate strategy here. An alternative metric of value for
26
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
fuelwood might be the ability to meet local needs through sustainable
extraction.
Measuring levels of fuelwood extraction can also be effectively achieved
through combining participatory research tools with remotely sensed data.
Although the quantities collected can be determined through simple
household surveying, participatory research helps to improve understanding
of how wood is selected for fuelwood, where it is taken from, and whether it is
felled or gathered. Participatory approaches can target information about
whether the needs of the local people are met by the ecosystem service and
this service is being utilised in a sustainable way.
Key literature:
Description
Stakeholders
Reference
Details dietary requirements, trends in
food production quantities and market
values, household consumption
patterns, and sources of food and
forage in Southern Africa.
Subsistence
farmers;
commercial
agriculturalists;
national and
international
markets; local
consumers
Van Jaarsveld et al (2005) Measuring
conditions and trends in ecosystem
services at multiple scales: the
Southern African Millennium
Ecosystem Assessment (SAfMA)
experience. Philosophical Transitions
of the Royal Society of Biological
Sciences 360(1454): 425-441
An insight into the challenges faced by
commercial agri-business in South
Africa’s semi-arid southwest, including a
changing climate, competing demands
for water and the trend towards
intensive-input farming.
Commercial
agriculturalists
Archer et al (2009) Climate change,
groundwater and intensive commercial
farming in the semi-arid northern
Sandveld, South Africa. Journal of
Integrative Environmental Sciences
6(2): 139-155
A consideration of the trade-offs
between scarce water consumption and
commercial food production in Tunisia’s
arid agricultural lands. Discusses
dietary requirements, demographic
change and food security.
Food and water
consumers at the
national scale
Besbes et al (2010) Changing Water
Resources and Food Supply in Arid
Zones: Tunisia. Chapter 7 in SchneierMadanes and Coure (eds.) Water and
Sustainability in Arid Regions Bridging
the Gap Between Physical and Social
Sciences
A special edition of the journal
Ecological Economics, entitled
‘Ecosystem Services and Agriculture’
considers the potential and challenges
of incentivising the provisioning of
ecosystem services, through markets
that reflect their full value, for the overall
welfare of society.
Local land users,
but drawing links
to the welfare of
national and
global society
Swinton et al (2007) Ecosystem
services and agriculture: Cultivating
agricultural ecosystems for diverse
benefits. Introduction to a special
edition of Ecological Economics 64(2)
A synthesis of perspectives on the links
between fuelwood collection and the
degradation of dryland forests and soil
in Nigeria, and the discourses that
define them.
Local land users,
neighbouring
ecosystem
stakeholders,
national and
international
community
Cline-Cole, R. (1998) Knowledge
claims and landscape: alternative
views of the fuelwood -- degradation
nexus in northern
Nigeria? Environment and Planning D:
Society and Space 16(3) 311 – 346
27
Department of Geography and the Environment
S.Whitfield and H. J.Geist
4.3
Carbon sequestration
Both soil and vegetation may represent large stores of carbon in a dryland
ecosystem that can be reduced through deforestation and soil erosion.
Valuing this ecosystem service will rely on complex scientific computations
based on vegetation species, soil structure, dead organic matter, moisture
content (and much more) (Nelson et al, 2009; Daily et al, 2000). Remote
sensing again plays a useful role in monitoring many of these key variables,
and increasingly sophisticated models, such as InVEST, can be utilised in
computing carbon storage based on simplified data input needs.
Carbon sequestration is one of the ecosystem service values that the InVEST
model is designed to map, and this provides a useful tool for monitoring
changes in the relative carbon sequestration capacity of the ecosystem over
time with minimal data input requirements. The key parameter on which this
model is based is and LULC map and the input variables required for running
the model are a categorisation of carbon pools (in accordance with InVEST
categories), a description of harvest rates, and biophysical information on the
carbon storage properties of predominant vegetation. The output maps of
the InVEST model provide a useful tool for stakeholder- and expert-led
discussion of trends and policy options.
Because drylands are marginal and vulnerable environments there is a direct
cost
associated
with
climate
change,
although
it
may
be
unreasonable/unreliable to expect local land users to draw conceptual links
between vegetation at the local scale and large scale climatic trends,
instead these links might be left to the researcher to conceive and
communicate based on a participatory study – if it is appropriate. Again
there is also a social cost issue in relation to alternative uses of wood (e.g.
fuelwood and timber...). Because of the breadth of stakeholders in the
carbon sequestration services of a dryland ecosystem, however, less
emphasis should be placed here on local users valuing the ecosystem service
(in comparison with things like food and fuelwood). A quantitative and
globally interpretable metric for valuation, such as working in measurable
quantities (e.g. Mg/ha) and monetary values, is more appropriate here.
28
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Key literature:
Description
Stakeholders
Reference
Based on the study of semi-arid agroecosystems in Sudan, they calculate that
converting marginal agricultural land to
rangeland will restore soil carbon content to
80% of savannah levels within 100 years and
argue that this could be facilitated and
incentivised by regulated markets
Local land users
Olsson and Ardö (2002)
Soil carbon sequestration in
degraded semiarid agroecosystems – Perils and
potentials. Ambio 31:471-477
Considers US semi-arid rangelands form a
ecosystem services perspective and argues
that market incentives for maintaining carbon
sequestration services is a key strategy for
sustainable management
Local land users
Discusses the complexities and institutional
and social challenges of accessing and
profiting from carbon markets for low-income
producers in Africa’s arid lands.
Local land-owners
Perez et al (2007) Can
carbon sequestration
markets benefit low-income
producers in semi-arid
Africa? Potentials and
challenges. Agricultural
Systems 94(1): 2-12
Explains the biophysical process of carbon
sequestration in dryland soils and the effects
of land-use change and degradation on
carbon storage capacity, and considers
market opportunities provided by Kyoto’s
CDM and World Bank’s BioCarbon Fund
Local land users and
local communities
Lal, R. (2009) Sequestering
carbon in soils of arid
ecosystems. Land
Degradation and
Development 20(4): 441-454
4.4
International
community
National and
international
institutions and
markets
National and
international
community
International
institutions and
markets
Havstad et al (2007)
Ecological services to and
from rangelands of the
United States. Ecological
Economics 64(2): 261-268
Water Provision and Conservation
Water provision and conservation benefit stakeholders indirectly, through
supporting almost all provisioning, cultural and regulating services. As a
consequence it may be seen as an indicator that falls largely within the
domain of ‘experts’. Scientists can apply ground-based tools for measuring
the functioning of the soil in terms of infiltration and runoff and thus provide a
strong indicator of the health of the ecosystem. However, these
methodologies are imperfect and may be data intensive. Local ecological
knowledge of the functioning of the ecosystem, rainfall patterns and
indicators of soil moisture, and traditional practices of water conservation, for
example, act as valuable sources of information in valuing these ecosystem
services. Knowledge of this type can be shared amongst stakeholders and
scientists through a deliberative approach and contribute to an informed
valuation of the service.
29
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Key literature
Description
Stakeholders
Reference
Explains the criticality of water
conservation within drylands – a
constraint on ecosystem resilience and
a key element of trade-off in dryland
management
Ecological system
Bautista, S., B.J. Orr, et al.
(2010). Evaluating the Restoration
of Dryland Ecosystems in the
Northern Mediterranean. Water
and Sustainability in Arid Regions.
G. Schneier-Madanes and M.-F.
Courel, Springer Netherlands:
295-310.
Describes the importance of the
knowledge of rangeland pastoralists in
the Kalahari of vegetation and how it
informs them about the functioning of
the ecosystem
Local land users
Thomas and Twyman (2004).
Good or bad rangeland? Hybrid
knowledge, science, and local
understandings of vegetation
dynamics in the Kalahari. Land
Degradation and Development
15(3): 215-231
Landscape functional analysis approach
to monitoring the properties of the
ecosystem – infiltration and runoff of
rainwater
Ecological system
Tongway, D.J. and Hindley, N.
(2004) Landscape Function
Analysis: Procedures for
Monitoring and Assessing
Landscapes. CSIRO, Brisbane,
Australia.
4.5
Cultural Landscape and Heritage
It is inappropriate to think in monetary terms for monitoring this ecosystem
service, instead an approach that looks retrospectively and prospectively at
trends, considering recent changes in the extent to which the socialecological system allows users to continue to exercise choice in respecting
cultural practices and self-defined tradition, is more appropriate.
Stakeholder-led discussions aimed at achieving an assessment of the cultural
heritage of the landscape may aim to identify cultural ideals and aspirations
and identify those factors that limit the freedom of stakeholders to live-out
these ideals. Equally, asking stakeholders to give preferences (and
explanations) for alternative environments may offer useful information about
the relative values that they assign to cultural landscapes. Using such factors
as a platform for discussion might help to generate targeted discussion about
an otherwise intangible subject. Once again the skill of the researcher is in
finding the most locally-appropriate methods for describing and monitoring
the value of this ecosystem service.
30
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Key literature:
Description
Stakeholders
Reference
Describes the landscape of SouthWestern Niger as the product of a
history of interactions between social
and biophysical processes and the
contemporary culture of economic
diversification
Local land
users
Batterbury, S. (2001) Landscapes of
diversity: a local political ecology of
livelihood diversification in SouthWestern Niger. Ecumene 8 (4): 437-464
Explains how the aims of landscape
restoration differ from those of dryland
ecosystem restoration, and discusses
the challenges of balancing traditional
techniques, nature conservation and the
well-being of local groups
Local land
users
Moreira et al (2006) Restoration
principles applied to cultural landscapes.
Journal for Nature Conservation 14(3-4):
217-224
Case studies of World Heritage cultural
landscapes from around the world
Local
indigenous
populations
Rössler, M. (2006) World Heritage
cultural landscapes: A UNESCO flagship
programme 1992-2006. Landscape
Research 31(4): 333-353
5 Conclusion
A coupled human-environment approach to monitoring dryland restoration
requires a close connection with the social-ecological system being studied.
Complex and multi-scale system dynamics mean that it is difficult to paint a
generic picture of sustainable land management and it is similarly unlikely that
a set of indicators, devised away from the context of study, will represent the
most appropriate metric for dryland restoration success.
This paper has explored the idea of using the value of ecosystem services as
overarching, but locally adaptable concept for monitoring dryland
restoration and has highlighted some of the key ecosystem services, the
values of which combined might be considered a good indicator for this
purpose. Throughout, evidence from key literature and local case studies
have been used to show how this generic indicator might be translated into a
practical research focus in the field (and with the use of remotely sensed
data).
The value of ecosystem services fits well within a coupled humanenvironmental approach because it focuses not solely on the changing
biophysical properties of the ecosystem, but rather on the relationship that
these changing properties have on interacting social, cultural and economic
spheres, considering more precisely how the changing ecosystem affects the
well-being of the ecosystem’s many stakeholders.
It is strongly suggested that the particular basket of ecosystem used as an
indicator in any given study is adapted to be contextually-appropriate and
31
Department of Geography and the Environment
S.Whitfield and H. J.Geist
closely compatible with pre-defined aims of restoration. Stakeholder-led and
participatory research methodologies can help to ensure that this achieved.
Similarly the ways in which value is measured should also be relevant to the
aims of the restoration and appropriate to the situation of the stakeholders.
Close collaboration between experts in social sciences and ecology and
dryland stakeholders will allow for the sharing of knowledge and deliberation
over the value of ecosystem services. The real strength of this approach to
monitoring is in its holistic approach to understanding the many values of the
ecosystem and in its emphasis on contextual system intricacies. It is about
monitoring the ability of the ecosystem to meet the needs and contribute to
the well-being of its stakeholders as well as sharing knowledge and
incorporating consideration about how it might continue to meet stakeholder
needs in the future.
6 References
Adger W.N.; Benjaminsen T.A.; Brown, K.; and Svarstad, H. (2001) ‘Advancing a political
ecology of global environmental discourses’ Development and Change 32:681–715.
Adger WN, Hughes TP, Folke Carpenter SR, Rockstrom J (2005) Social ecological resilience to
coastal disasters. Science 309:1036–1039.
Alverson,H. (1984) The wisdom of tradition in the development of dry-land farming: Botswana.
Human Organisation 43(1): 1-8.
Archer et al (2009) Climate change, groundwater and intensive commercial farming in the
semi-arid northern Sandveld, South Africa. Journal of Integrative Environmental
Sciences 6(2): 139-155.
Asner, G.P.; Archer, S.; Hughes, R.F.; Ansley, R.J.; and Wessman, C.A. (2003) Net changes in
regional woody vegetation cover and carbon storage in Texas Drylands, 1937–1999.
Global Change Biology 9(3): 316-335.
Bainbridge, D.A. (2007) A guide for desert and dryland restoration: new hope for arid lands.
Island Press: Washington.
Balmford, A. and Bond, W. (2005). Trends in the state of nature and implications for human
well-being. Ecology Letters. 8: 1218-1234.
Barbier, B. (1996) Impact of market and population pressure on production, incomes and
natural resources in the dryland savannas of West Africa: bioeconomic modelling at
the village level. International Food Policy Research Institute.
Batterbury, S. (2001) Landscapes of diversity: a local political ecology of livelihood
diversification in South-Western Niger. Ecumene 8 (4): 437-464.
Bautista, S.; Orr, B.J.; Alloza, J.A. and Vallejo, R.V. (2010) Evaluating the Restoration of Dryland
Ecosystems in the Northern Mediterranean. In: G. Schneider-Madanes and M.F. Courel
(eds.) Water and Sustainability in Arid Regions Bridging the Gap Between Physical and
Social Sciences. Heidelberg: Springer.
Benayas, J.M.R; Bullock, J.M. and Newton, A.C. (2008) Creating woodland islets to reconcile
ecological restoration, conservation, and agricultural land use. Frontiers in Ecology
and the Environment. 6: 329—336.
32
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Besbes et al (2010) Changing Water Resources and Food Supply in Arid Zones: Tunisia.
Chapter 7 in Schneier-Madanes and Coure (eds.) Water and Sustainability in Arid
Regions Bridging the Gap Between Physical and Social Sciences. Heidelberg: Springer.
Berkes F, Folke C, Gadgil M. (1995) Traditional Ecological Knowledge, Biodiversity, Resilience
and Sustainability. Pages 281–299 in Perrings CA, Mahler KG, Folke C,Holling CS,
Jansson B, eds. Biodiversity Conservation: Problems and Policies. Dordrecht (The
Netherlands): Kluwer Academic.
Bennett, E.M.; Peterson, G.D.; Gordon, L.J. (2009) Understanding relationships among multiple
ecosystem services Ecology Letters 12(2): 1394-1404.
Butler,C.D. and Oluoch-Kosura, W. (2006) Linking Future Ecosystem Services and Future
Human Well-being Ecology and Society 11(1): 30.
Carpenter, S.R.; Bennett, E.M.; and Peterson, G.D. (2006) Scenarios for Ecosystem Services: An
Overview. Ecology and Society 11(1): 29.
Carpenter, S. R., and W. A. Brock. (2004). Spatial complexity, resilience and policy diversity:
fishing on lake-rich landscapes. Ecology and Society 9(1): 8.
Carretero,E.M.; Dalmasso, A.; and Trione, S. (2007) Carbon Storage in Larrea Divaricata andL.
cuneifolia (Zygophyllaceae) in Drylands of Central-Western Argentina. Arid Land
Research and Management 21(4): 273-285.
Chambers, R. 1994. Participatory Rural Appraisal (PRA): Analysis of Experience. World
Development 22(9):1253-1268.
Christanto, N.; Hadmoko, D.S.; Westen, C.J.; Lavigne, F.; Sartohadi, J.; and Setiawan, M.A.
(2009) Characteristic and Behavior of Rainfall Induced Landslides in Java Island,
Indonesia : an Overview. Geophysical Research Abstracts 11.
Clewell, A.F., and Aronson, I. (2006) ‘Motivations for the restoration of ecosystems’
Conservation Biology 20 (2): 420–428.
Cline-Cole, R. (1998) Knowledge claims and landscape: alternative views of the fuelwood -degradation nexus in northern Nigeria? Environment and Planning D: Society and
Space 16(3) 311 –346.
Costanza, R., R. d’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.
V. O’Neill, J. Paruelo, R. G. Raskin, P. Sutton and M. van den Belt. 1997. The value of
the world’s ecosystem services and natural capital. Nature 387:253–260.
Cotula, L. (ed), 2006, Land and Water Rights in the Sahel - Tenure Challenges of Improving
Access
to
Water
for
Agriculture,
London,
IIED,
Issue
Paper
139,
http://www.iied.org/pubs/pdf/full/12526IIED.pdf.
Daily, G.C.; Alexander, S.; Ehrlich, P.R.; Goulder, L.; Lubchenco, J.; Matson, P.A.; Mooney, H.A.;
Postel, S.; Schneider, S.H.; Tilman, D.; and Woodwell, G.M. (1997). Ecosystem services:
Benefits supplied to human societies by natural ecosystems. Issues in Ecology 2:1-16.
Daily, G.C.; Söderqvist , T.; Aniyar, S.; Arrow, K.; Dasgupta, P.; Erlich, P.R.; Folke, C.; Jansson, A;
Jansson, B-G.; Kautsky, N.; Levin, S.; Lubchenco, J.; Mäler, K-G.; Simpson, D.; Starrett,
D.; Tilman, D.; and Walker, B. (2000) The Value of Nature and the Nature of Value
Science 289(5478): 395-396.
Danielsen, F., H. Beukema, N. D. Burgess, F. Parish, C. A. Brühl, P. F. Donald, D. Murdiyarso, B.
Phalan, L. Reijnders, M. Struebig, and E. B. Fitzherbert. 2008. Biofuel plantations on
forested lands: double jeopardy for biodiversity and climate. Conservation Biology
23(2):348–358.
Defra (2007) An introductory guide to valuing ecosystem services. www.defra.co.uk
33
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Defries, R.S.; Foley, J.A.; and Asner, G.P. (2004) Land-use choices: balancing human needs
and ecosystem function Frontiers in Ecology and the Environment 2: 249–257.
De Groot, R.S.; Wilson, M.A.; and Boumas, R.M.J. (2002) A typology for the classification,
description and valuation of ecosystem functions, goods and services. Ecological
Economics 41(3): 393-408.
Dietz A.J.; Ruben, R.; and Verhagen, A. [eds](2004) The impact of climate change on
drylands: with a focus on West Africa. Kluwer Academic Publishers
Dipper, B. (1998) Monitoring and Post-auditing in Environmental Impact Assessment: A Review.
Journal of Environmental Planning and Management 41(6): 731-747
Dovie, D.B.K.; Witkowski, E.T.F.; Shackleton, C.M. (2004) The Fuelwood Crisis in Southern Africa –
Relating Fuelwood Use to Livelihoods in a Rural Village. GeoJournal 60(2): 123-133
Dregne, H.E. (2002) Land Degradation in the Drylands. Arid Land Research and Management
16(2): 99-132
Dryland Science for Development Consortium Working Group One (DSD WG1, 2009)
‘Integrated methods for monitoring and assessing desertification/land degradation
processes and drivers (land quality)’ White Paper of the DSD Working Group 1.
Dubroeucq, D. and Livenais, P. (2004) Land cover and land use changes in relation to social
evolution – a case study of Northern Chile. Journal of Arid Environments 56(2): 193-211.
Farrow, A. and Winograd, M. (2001) Land use modelling at the regional scale: an input to rural
sustainability indicators for Central America. Agriculture, Ecosystems and Environment
85: 249-268.
Fisher, B.; Turner, R.K.; and Morling, P. (2009) Defining and classifying ecosystem services for
decision making. Ecological Economics 68(3): 643-653.
Folke, C. (2006) ‘Resilience: The emergence of a perspective for social-ecological systems
analyses’ Global Environmental Change 16(3): 253-267.
Folke, C.; Carpenter, S.R.; Walker, B.; Scheffer, M.; Elmqvist, T.; Gunderson, L.; and Holling, C.S.
(2004) ‘Regime Shifts, Resilience and Biodiversity in Ecosystem Management’ Annual
Review of Ecology, Evolution, and Systematics 35: 557-581.
Gao, Y.; Qiu, G.Y.; Shimizu, H.; Tobe, K.; Sun, B.; and Wang, J. (2002) A 10-year study on
techniques for vegetation restoration in a desertified salt lake area. Journal of Arid
Environments 52(4): 483-497.
Geist, H.J. (2005) ‘The causes and progression of desertification (Ashgate Studies in
Environmental Policy and Practice)’. Ashgate: Aldershot.
Geist H.J.and Lambin, E.F. (2004) ‘Dynamic causal patterns of desertification’. BioScience
54(9): 817-829.
Glenn, E.; Squires, V.; Olsen, M.; and Frye, R. (1993) Potential for carbon sequestration in the
drylands. Water, Air and Soil Pollution 70(1-4): 341-355.
Gordon, I.J. (2007) Linking land to ocean: feedbacks in the management of socio-ecological
systems in the Great Barrier Reef catchments. Hydrobiologia 591(1): 25-33
Guaseelan, V.N. (2009) Biomass estimates, characteristics, biochemical methane potential,
kinetics and energy flow from Jatropha curcuson dry lands. Biomass and Bioenergy
33(4): 589-596.
Havstad et al (2007) Ecological services to and from rangelands of the United States.
Ecological Economics 64(2): 261-268.
34
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Hein, L.; van Koppen, K.; de Groot, R.S.; and van Ierland, E.C. (2006) Spatial scales,
stakeholders and the valuation of ecosystem services. Ecological Economics 57(2):
209-228.
Herrmann, S.M.; Anyamba, A.; and Tucker, C.J. (2005) Recent trends in vegetation dynamics
in the African Sahel and their relationship to climate. Global Environmental Change
Part A 15(4): 394-404.
Holling, C.S. (1973) ‘Resilience and stability of ecological systems’. Annual Review of Ecology
and Systematics 4:1-23.
Holling, C.S. (2001) Understanding the complexity of economic, ecological, and social
systems, Ecosystems 4: 390–405.
Hope, R.A.; Jewitt, G.P.W.; Gowing, J.W. (2004)Linking the hydrological cycle and rural
livelihoods: a case study in the Luvuvhu catchment, South Africa. Physics and
Chemistry of the Earth, Parts A/B/C 29(15-18): 1209-1217.
Huber-Sannwald, E; Maestre, F.T; Herrick, J.E.; and Reynolds, J.F.(2006) Ecohydrological
feedbacks and linkages associated with land degradation: a case study from Mexico
Hydrological Processes 20: 3395-3411.
IPCC (2007) IPCC, 2007: Climate Change 2007: Impacts, Adaptation and Vulnerability.
Contribution of Working Group II to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P. Palutikof,
P.J. van der Linden and C.E. Hanson, Eds., Cambridge University Press, Cambridge, UK,
976 pp.
Illius, A.W. and O’Connor, T.G.O. (1999) On the relevance of nonequilibrium concepts to arid
and semiarid grazing systems. Ecological Applications 9(3): 798-813.
Janssen, MA, and SR Carpenter. 1999. Managing the resilience of lakes: a multi-agent
modelling approach. Conservation Ecology 3:1.
Jollands, N., Ruth, M., Golubiewski, N., and Bernier, C., 2007. The climate's long-term impact
on New Zealand infrastructure (CLINZI) project – A case study of Hamilton City, New
Zealand. Journal of Environmental Management 83(4):460–477.
Jiang, H (2003) Stories Remote Sensing Images Can Tell: Integrating Remote Sensing Analysis
with Ethnographic Research in the Study of Cultural Landscapes. Human Ecology
31(2): 215-232.
Jiang, L.; Yufen, T.; Zhijie, Z.; Tianhong, L. And Jiankua, L. (2005) Water Resources, Land
Exploration and Population Dynamics in Arid Areas - The Case of the Tarim River Basin
in Xinjiang of China Population and Environment 26(6): 471-503
Kakuru, W.; Okia; C.; Okorio, J. (2004) Strategy for Agroforestry Development in Uganda’s
Drylands. Paper presented at the Drylands Agroforestry Workshop 1st-3rd September.
ICRAF Headquarters, Nairobi-Kenya.
Kerven, C. [eds] (2007) The Real Jewels of the Kalahari: Drylands Ecosystem Goods and
Services in Kgalagadi South District, Botswana. IUCN Paper.
Kinzig, A. P., P. Ryan, M. Etienne, H. Allison, T. Elmqvist, and B. H. Walker. 2006. Resilience and
regime shifts: assessing cascading effects. Ecology and Society 11(1): 20.
Koch, EW; Barbler, EB; Silliman, BR; Reed, DJ; Perillo, GM; Hacker, SD; Granek, EF; Primavera,
JH; Muthlga, NI; Polasky, S; Halpern, BS; Kennedy, CJ; Kappel, CV; Wolanski, E (2009)
Non-linearity in ecosystem services: temporal and spatial variability in coastal
protection Frontiers in Ecology and the Environment 7(1): 29-37
35
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Komatsu et al (2005) Evaluation of agricultural sustainability based on human carrying
capacity in drylands - a case study in rural villages in Inner Mongolia, China.
Agriculture, Ecosystems and Environment 108(1): 29-43
Kumar, M. and Kumar, P. (2008) Valuation of the ecosystem services: A psycho-cultural
perspective. Ecological Economics 64(4): 808-819
Lal, R. (2002) Soil carbon sequestration in China through agricultural intensification, and
restoration of degraded and desertified ecosystems. Land Degradation and
Development 13(6): 469-478.
Lal, R. (2004) Carbon Sequestration in Dryland Ecosystems. Environmental Management 33(4):
528-544.
Lal, R. (2009) Sequestering carbon in soils of arid ecosystems. Land Degradation and
Development 20(4): 441-454.
Lambin EF, Geist H, Reynolds JF, Stafford Smith DM (2009) Coupled human-environment
system approaches to desertification: linking people to pixels In: Röder A, Hill J (eds).
Recent advances in remote sensing and geoinformation processing for land
degradation assessment (ISPRS Book Series). CRC Press/Balkema: Leiden, NL, 3-14 .
Lambin E.F. and Meyfroidt P. (2010). Land use transitions: Ecological feedback versus
exogenous socio-economic dynamics, Land Use Policy, 27: 108-118.
Liu, J.; Dietz, T.; Carpenter, S.R.; Alberti, M.; Folke, C.; Moran, E.; Pell, A.N.; Deadman, P.; Kratz,
T.; Lubchenco, J.; Ostrom, E.; Ouyang, Z.; Provencher, W.; Redman, C.L.; Schneider,
S.H.; and Taylor, W.W. (2007) ‘Complexity of Coupled Human and Natural Systems’.
Science 317: 1513 – 1516.
Luck, G.W.; Daily, G.C.; and Ehrlich, P.R. (2003) Population diversity and ecosystem services.
Trends in Ecology and Evolution 18(7): 331-336.
McCauley, D.J. (2006) Selling out on Nature. Nature 443: 27-28.
Masri, Z.; Zöbisch, M.; Bruggeman, A.; Hayek, P.; and Kardous, M. (2003) Wind erosion in a
marginal Mediterranean drylands area: a case study from the Khanasser Valley, Syria.
Earth Surface Processes and Landforms 28(11): 1211-1222.
Millennium Ecosystem Assessment, (MA, 2005a). Ecosystems and Human Well-being: Synthesis.
Island Press, Washington, DC.
Millennium Ecosystem Assessment (MA, 2005b) ‘Ecosystems and Human Well-being:
Desertification Synthesis’. World Resources Institute: Washington, DC.
Moreira et al (2006) Restoration principles applied to cultural landscapes. Journal for Nature
Conservation 14(3-4): 217-224
Moreira, F.; Queiroz, A.I.; and Aronson, J. (2006) Restoration principles applied to cultural
landscapes. Journal for Nature Conservation 14(3-4): 217-224.
Mortimore, M.J. and Adams, W.M. (2001) Farmer adaptation, change and ‘crisis’ in the Sahel.
Global Environmental Change 11(1): 49-57.
Mukhopadhyay, D. (2009) Indigenous Knowledge and Sustainable Natural Resource
Management in the Indian Desert. The Future of Drylands 3: 161-170.
Nelson, E.; Mendoza, G.; Regetz, J.; Polasky, S.; Tallis, H.; Cameron, R.D.; Chan, K.M.A.; Daily,
G.C.; Goldstein, J.; Kareiva, P.M.; Lonsdorf, E.; Naidoo, R.; Ricketts, T.H.; and Shaw, M.R.
(2009) Modelling multiple ecosystem services, biodiversity conservation, commodity
production, and tradeoffs at landscape scales. Frontiers in Ecology and Environment
7(1): 4-11.
36
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Nyariki,D.M.; Wiggins, S.L.; and Imungi, J.K. (2002) Levels and causes of household food and
nutrition insecurity in dryland Kenya Ecology of Food and Nutrition 41(2): 155-176.
Olsson and Ardö (2002) Soil carbon sequestration in degraded semiarid agro-ecosystems –
perils and potentials. Ambio 31: 471–477.
Owour, B.; Eriksen, S.; and Mauta, W. (2005) Adapting to Climate Change in a Dryland
Mountain Environment in Kenya Mountain Research and Development 25(4):310-315.
Pereira, H.M.; Domingos, T.;and Vicente, L. (2006) Assessing Ecosystem Services at Different
Scales in the Portugal Millennium Ecosystem Assessment. In Reid, W.V.; Berkes, F.;
Wilbanks, T.; Capistrano, D. [eds] Concepts and Applications in Ecosystem Assessment.
Island Press
Perez et al (2007) Can carbon sequestration markets benefit low-income producers in semiarid Africa? Potentials and challenges. Agricultural Systems 94(1): 2-12.
Portnov, B.A. and Safriel, U.N.( 2004) Combating desertification in the Negev: dryland
agriculture vs. dryland urbanization. Journal of Arid Environments 56(4): 659-680.
Proctor,W.; Cork, S.; Lamgridge, J.; Langston, A.; Abel, N.: Howden.; Anderies, m.; Parry, R.;
Shelton, D. (2002) Assessing Ecosystem Services in Australia. Ecosystem Services
Project.
Reed, M.S. and Dougill, A.J. (2010) Linking Degradation Assessment to Sustainable Land
Management: a decision support system for Kalahari pastoralists. Journal of Arid
Environments 74: 149–155.
Reed, M.S.; Fraser, E.D.G.; Dougill, A.J. (2006) ‘An adaptive learning process for developing
and applying sustainability indicators with local communities’. Ecological Economics
59: 406–418.
Reid, W.V. (2006) Nature: the many benefits of ecosystem services. Nature 443: 749.
Reyers, B.; O'Farrell, P.J.; Cowling, R.M.; Egoh, B.N.; Le Maitre, D.C.; Vlok, J.H.J. (2009)
Ecosystem services, land-cover change, and stakeholders: finding a sustainable
foothold for a semiarid biodiversity hotspot. Ecology and Society 14(1): 23.
Reynolds, J.F.; Mark Stafford Smith, D.; Lambin, E.F.; Turner, II, B.L.; Mortimore, M.; Batterbury,
S.P.J.; Downing, T.E.; Dowlatabadi, H.; Fernández, R.J.; Herrick, J.E.; Huber-Sannwald,
E.; Jiang, H.; Leemans, R.; Lynam, T.; Maestre, F.T.; Ayarza, M.; and Walker, B. (2007)
‘Global desertification: Building a Science for Dryland Development’ Science
316(5826): 847-851.
Reynolds, J.F. and Stafford Smith, D.M. [eds] (2002) Global Desertification: Do Humans Cause
Deserts? Dahlem Workshop Report 88, Dahlem University Press, Berlin.
Rössler, M. (2006) World Heritage cultural landscapes: A UNESCO flagship programme 19922006. Landscape Research 31(4): 333-353.
Russelle M.P.; Entz, M.H.; and Franzluebbers, A.J. (2007) Reconsidering Integrated CropLivestock Systems in North America. Agronomy Journal 99: 325-334.
Sá e Silva (2009) Use and knowledge of fuelwood in three rural caatinga (dryland)
communities in NE Brazil. Environment, Development and Sustainability 11(4): 833-851.
Safriel, U.N. (2006) Dryland Development, Desertification and Security in the Mediterranean.
Desertification in the Mediterranean Region. A Security Issues. NATO Security through
Science Series 3: 227-250.
Safriel, U. (2009) ‘Deserts and desertification: Challenges but also opportunities’ Land
degradation and development 20(4): 353-366.
37
Department of Geography and the Environment
S.Whitfield and H. J.Geist
Scanlon , B.R.; Reedy, R.C.; Stonestrom, D.A.; Prudic, D.E. and Dennehy, K.F. (2005) Impact of
land use and land cover change on groundwater recharge and quality in the
southwestern US. Global Change Biology 11(10): 1577-1593.
Scheffer, M.; Carpenter, S.; Foley, A.; Folke, C.; Walker, B. (2001) Catastrophic shifts in
ecosystems. Nature 413: 591-596.
Scholes, R.J. (2009) Syndromes of dryland degradation in southern Africa. African Journal of
Range and Forage Science 26(3): 113-125.
Schulz, J.J.; Cayuela, L.; Echeverria, C.; Salas, J.; and Benayas, J.M.R. (2010) Monitoring land
cover change of the dryland forest landscape of Central Chile (1975–2008) Applied
Geography 30(3): 436-447.
Schwilch, G., Bachmann, F. and Liniger, H. (2009), Appraising and selecting conservation
measures to mitigate desertification and land degradation based on stakeholder
participation and global best practices. Land Degradation and Development, 20:
308–326.
Sivanappan, R.K. (1995) Soil and water management in the dry lands of India. Land Use Policy
12(2): 165-175.
Spangenberg, J.H. (2002) Institutional sustainability indicators: an analysis of the institutions in
Agenda 21 and a draft set of indicators for monitoring their effectivity. Sustainable
Development 10(2): 103-115.
Standish,R.J.; Cramer, V.A.; Wild, S.L.; Hobbs, R.J. (2007) Seed dispersal and recruitment
limitation are barriers to native recolonization of old-fields in western Australia. Journal
of Applied Ecology 44(2): 435-445.
Stohlgren,T.J.; Chase, T.N.; Pielke, R.A.; Kittel. T.G.F.; Baron, J.S. (1998)Evidence that local land
use practices influence regional climate, vegetation, and stream flow patterns in
adjacent natural areas. Global Change Biology 4: 495-504.
Swinton et al (2007) Ecosystem services and agriculture: Cultivating agricultural ecosystems
for diverse benefits. Introduction to a special edition of Ecological Economics 64(2).
Thomas, D.S.G.; Sporton, D.; and Perkins, J. (2000) The environmental impact of livestock
ranches in the Kalahari, Botswana: natural resource use, ecological change and
human response in a dynamic dryland system. Land Degradation and Development
11(4): 327-341.
Toman, M. (1998) Why not to calculate the value of the world’s ecosystem services and
natural capital. Ecological Economics 25: 57-60.
Tomich, T.P.; Argumedo, A.; Baste, I.; Camac, E., Filer, C.; Garcia, K.; Garbach, K.; and Geist,
H. (2010) Conceptual frameworks for ecosystem assessment: their development,
ownership, and use In: Ash N [ed] Ecosystems and human well-being: a manual for
assessment practitioners. Island Press: Washington, DC, 71-113.
Tovey, J.P. (2008) Whose rights and who’s right? Valuing ecosystem services in Victoria,
Australia. Landscape Research 33(2): 197-209 .
Upton, C. (2010) Living off the land: Nature and nomadism in Mongolia . Geoforum (Article in
Press).
van Jaarsveld, A.S.; Biggs, R.; Scholes, R.J.; Bohensky, E.; Reyers, B.; Lynam, T.; Musvoto, C.;
and Fabricus, C. (2005) Measuring conditions and trends in ecosystem services at
multiple scales: the Southern African Millennium Ecosystem Assessment (SAfMA)
experience. Philosophical Transitions of the Royal Society of Biological Sciences
360(1454): 425-441.
Wack P. (1985). Scenarios: Uncharted waters ahead. Harvard Business Review 63 (5).
38
Department of Geography and the Environment
Coupled Human-Environment Indicators for the Assessment of Dryland Restoration
Walker, B.; Carpenter, S.; Anderies, J.; Abel, N.; Cumming, G.; Janssen, M.; Lebel, L.; Norberg,
J.; Peterson, G.D.; and Pritchard, R. (2002) ‘Resilience management in socialecological systems: a working hypothesis for a participatory approach’. Conservation
Ecology 6(1): 14.
Walker, B. and Meyers, J.A. (2004) Thresholds in Ecological and Social–Ecological Systems: a
Developing Database. Ecology and Society 9(2): 3.
Warren, A. (2002). Land degradation is contextual. Land Degradation and Development 13,
449-459.
Watts, M. (1987) Drought, environment and food security: some reflections on peasants,
pastoralists and commoditization in dryland West Africa. In Glantz, M.H. [eds] Drought
and hunger in Africa: denying famine a future. Cambridge University Press,
Cambridge.
White, R.P.; Tunstall, D.; Henninger, N. (2002) Ecosystem approach to drylands- building
support for new development. World Resources Institute.
Wenlan, F.; Qigang, Z.; Baolei, Z.; Wancun, Z.; Ainong, L.; Haizhen, Z.; and Wei, X. (2006) GISbased spatial analysis and modeling for landslide hazard assessment: A case study in
upper Minjiang River basin. Wuhan University Journal of Natural Sciences 11(4): 847852.
Wollenberg E., D. Edmunds & L. Buck. (2000). Anticipating change: Scenarios as a tool for
adaptive forest management. Bogor, Indonesia: Center for International Forestry
Research.
Worm, B.; Barbier, E.B.; Beaumont, N.; Duffy, J.E.; Folke, C.; Halpern, B.S.; Jackson, J.B.C.; Lotze,
H.K.; Micheli, F.; Palumbi, S.R.; Sala, E.; Selkoe, K.A.; Stachowicz, J.J.; and Watson, R.
(2006) Impacts of Biodiversity Loss on Ocean Ecosystem Services. Science 314(5800):
787.
Young, O.R.; Lambin, E.F.; Alcock, F.; Haberl, H.; Karlsson, S.I.; McConnall, W.J.; Myint, T.; PahlWostl, C.; Polsky, C.; Ramakrishnan, P.S.; Schroeder, H.; Scouvart, M.; and Verburg, P.H.
(2006) ‘A portfolio approach to analysing complex human-environment interactions:
Institutions and land change’ Ecology and Society 11(2): 31
Zaal, F. and Oostendorp, R.H. (2002) "Explaining a Miracle: Intensification and the Transition
Towards Sustainable Small-scale Agriculture in Dryland Machakos and Kitui Districts,
Kenya," World Development, 30(7): 1271-1287.
Zhen, L.; Ochribat, B.; Lv, Y.; Wei, Y.J.; Liu, X.L.; Chen, J.Q.; Yao, Z.J.; and Li, F.(2010) Comparing
patterns of ecosystem service consumption and perceptions of range management
between ethnic herders in Inner Mongolia and Mongolia. Environmental Research
Letters 5(1).
Zucca, C.; Bautista, S.; and Previtali, F. (2009) ‘Desertification: Prevention and Restoration’. In:
R. Lal (ed.) Encyclopedia of soil science (Second Edition). New York: Taylor & Francis
Group.
39
Department of Geography and the Environment
Ground-based Assessment Methods and Indicators
II. Ground-based Methods and Indicators for the
Biophysical Assessment of Desertification
Prevention and Restoration Actions
Working Paper for the PRACTICE Project
Susana Bautista
Ángeles G. Mayor
41
S. Bautista and A. G. Mayor
Susana Bautista1 and Ángeles G. Mayor2 (October 2010)
1 Department of Ecology
University of Alicante
Apdo. 99
03080 Alicante
Spain
[email protected]
2 CEAM Foundation
Charles R. Darwin 14
Parque Tecnológico
46980 Paterna (Valencia)
[email protected]
42
Ground-based Assessment Methods and Indicators
1 Introduction
There is an increasing demand for the development and implementation of
appropriate assessment methods to measure progress on combating
desertification (UNCCD 2009). The adoption of best practices in prevention
and restoration efforts requires a good use of the existing expertise and
information, as well as improved understanding on the impacts of the
strategies and techniques applied on the target socio-ecological systems.
Both requirements can be met through the systematic evaluation of the
actions applied and further dissemination of the results.
Evaluation is the key element linking practice and the advances in both
science and technology. Moreover, monitoring and evaluation are critical
components of an adaptive management approach. For any given
prevention or restoration action, evaluation provides feedback for the finetuning of the treatments and techniques applied, and thereby helps address
the uncertainty inherent to ecosystem dynamics. For the general practice of
dryland management, evaluation helps managers, farmers, and other land
users learn from past efforts and adapt management strategies and
techniques in response to spatial and temporal variation in environmental
and socio-economic conditions. Evaluation is also needed to establish costeffective thresholds for the various management alternatives, and to identify
priority areas where actions could be most effective.
Identifying the most appropriate conceptual framework, approach and
indicator suite is crucial for a more widespread, harmonized, scientificallysound, and effective monitoring and assessment of prevention and
restoration actions. This working paper addresses this challenge by reviewing
and discussing the state of the art on ground-based ecosystem assessment
methods and approaches, and presenting a strategy and an associated
suite of indicators and methods for the ground-based assessment of dryland
management.
Although most of the approaches discussed here are applicable to a wide
variety of management actions, this working paper focuses on the evaluation
of prevention and restoration actions to combat degradation of dryland
ecosystems.
2 Evaluation approaches
The approaches for evaluating prevention and restoration actions are many,
including, among others, measuring the achievement of specific goals and
stages (Zedler 1995, Zedler and Callaway 1999); comparisons between
managed and non-managed areas, between management alternatives, or
between restored and reference target areas (Brinson and Rheinhardt 1996,
Gaboury and Wong 1999, Rey Benayas et al. 2009); comparisons with natural
range of variability (Hobbs and Norton 1996, Parker and Pickett 1997, Allen et
43
S. Bautista and A. G. Mayor
al. 2002); evaluation of the degree of (self)sustainability of the managed or
restored ecosystem (SER 2004): and direct assessment of ecosystem quality
according to a set of accepted standards. Though the different approaches
partially overlap, there are particular pros and cons associated with the
implementation of each of them that are discussed below.
The most obvious evaluation approach is to measure the degree of
achievement of the proposed objectives. Ideally, based on the goals of the
prevention or restoration action and on the knowledge and understanding of
the ecology of the system, explicit predictions are made of expected
responses by biotic and abiotic ecosystem components that will then be
monitored for evaluation. In turn, designing the appropriate monitoring and
evaluation program helps refine and explicitly state the project specific
objectives. The fact that restored ecosystems are not static also points to the
need for establishing the suitable time frame in which to assess the
achievement of the various stages envisioned.
In some cases, expectations could be written as statements of testable
hypotheses (Thayer et al. 2003). However, our limited understanding of the
processes, factors, and biotic interactions that control ecosystem dynamics
often limits the definition of the specific outcomes that could be expected,
and therefore only broadly-defined objectives are established for most of the
actions implemented. Moreover, both the knowledge framework and the
social context are dynamic, and each influences management decisions
and objectives. For example, since the 1990s, mitigating climate change has
become a core objective of afforestation and reforestation programs
worldwide. In the past, the main objectives of reforestation projects were
wood production, soil protection from erosion, and flood control (Vallejo et al.
2006, Bautista et al. 2010); while in the last decades the objectives have
shifted to other ecosystems goods and services such as improvement of water
quality, recreation, improvement of wildlife habitats, fire prevention,
biodiversity conservation, etc. Many projects that could be considered as
highly successful in meeting originally established objectives, would meet
none or very few of the current social demands regarding biodiversity
conservation and ecosystem services.
Restoration ecologists usually advocate the use of target or model
communities as reference systems to set restoration goals and evaluate
restoration success (e.g., Aronson et al. 1993, Aronson and Le Floc’h 1996,
Brinson and Rheinhardt 1996, White and Walker 1997, SER 2004, Ruiz-Jaen and
Aide 2005). A reference system is any ecosystem or landscape showing the
structure and function that is expected for an area to be deemed
successfully restored or managed. A restoration process aimed at
reconstructing a prior ecosystem and re-establishing former communities is,
however, a very difficult task, particularly at the landscape level (Henry and
Amoros 1995, Bradshaw 1997, van Diggelen et al. 2001). Several authors have
called for an alternative approach based on evaluation criteria that focus on
the functional aspects of the reference system, using specific services or
certain functions as reference conditions (e.g., Brinson and Rheinhardt 1996,
44
Ground-based Assessment Methods and Indicators
Choi 2004). Falk (2006) proposed to replace the more static concept of
reference conditions by reference dynamics: a process-centred approach
that places emphasis on ecological functions and ecosystem processes.
Historical data on pre-disturbed conditions or remnants of historic natural
areas are common forms of target references (Holl and Cairns 2002, Hobbs
and Harris 2001). However, given the dynamic nature of communities in a
changing environment and socio-economic context, several studies point to
the usefulness of using historical data as reference information (Pickett and
Parker 1994, Hobbs and Norton 1996, Choi 2004). Some authors suggest that
rather than focus on restoring to some primeval state, a more profitable
approach would be to focus on repairing damaged systems to the extent
possible, considering both the ecological potential for restoration and
societal desires (Higgs 1997, Hobbs and Harris 2001). In this approach, the prerestored, degraded system can be considered as the reference with which to
evaluate restoration. Defining the degree of improvement that could be
considered a success is the particular challenge of this approach.
Evaluation can be centred on the comparison of management alternatives.
This approach does not rule out including either reference systems or
degraded systems within the set of compared cases. Methods for generic
functional analyses (e.g., Tongway and Hindley 2004, Herrick et al. 2005) are
of particular interest for comparative evaluation, as they provide indices that
can be directly comparable across sites differing in area or scale.
Finally, the outcomes of the actions implemented could be assessed through
a variety of quality indicators that reflect current social demands, yet they
may not be the original target outcomes considered by the management
project. When no real, existing reference is available, or when there is no
adequate information about previous degraded conditions, this approach
can be the appropriate framework for evaluation. This approach is related to
existing tools and methods for ecosystem monitoring and assessment that
focus on the evaluation of ecosystem status and integrity. For example, WWFWorld Wide Fund for Nature and IUCN-International Union for Conservation of
Nature have developed an approach to landscape assessment of forest
quality that can also be used to evaluate restored forests and woodlands. This
method is based on the following criteria: (1) Authenticity, (2) Forest health,
(3) Environmental benefits, and (4) Social and cultural values. The method
relies on the assessment of a large suite of indicators that provide information
on these four criteria (WWF 2002).
All the approaches above implicitly consider restoration evaluation as the
evaluation of success (Hobbs and Harris 2001). However, success is a
subjective and somehow unclear and elusive concept (Zedler 2007) that
does not fully recognize and accommodate the many potential sources of
variation and uncertainty concerning management and restoration
outcomes, including the diverse, even contrasting perspectives among the
various stakeholders. Rather than merely following a success vs failure
approach, evaluation may be viewed as a process of creating information
45
S. Bautista and A. G. Mayor
and knowledge on the restoration actions implemented, providing a more or
less comprehensive and multifaceted description of the restoration
outcomes. This approach considers evaluation as an information system that
collects and provides useful data on ecosystem and landscape responses to
restoration. It can therefore support any other approach to evaluation.
3 Selection of attributes and indicators for evaluation
Whatever the evaluation approach that is used, the selection of the actual
attributes and indicators to be assessed is key to the evaluation process. The
structural and functional attributes of ecosystems do not always linearly
covary, (Cortina et al. 2006, Rey Benayas et al. 2009). Therefore, results may
vary greatly depending on the variables considered.
In selecting the appropriate indicators, there are a number of key aspects to
consider. First of all, an appropriate, scientifically justifiable framework must
underlie the selection of the assessment indicators. Second, depending on
the assessment goals and conceptual framework the appropriate number of
indicators, their spatial and temporal scope of application, and the scale,
methods, and resolution of the measurements can greatly vary. Finally, the
selected indicators should be relevant and clearly linked to underlying
primary processes, based on well-understood conceptual models, reliable,
sensitive to important changes but robust against natural variability, and also
be simple and measurable with a reasonable level of effort and cost (NRC
2000. Dale and Beyeler 2001, Jorgersen et al. 2005). Because of the large
spatial and temporal variability of ecosystems, particularly in drylands, recent
studies suggest focusing ecosystem assessment on ‘slow’ variables (Carpenter
and Turner 2000), both biophysical and socio-economic (e.g., soil fertility,
market access), as high variability in ‘fast’ variables may mask fundamental
trends and long-term changes (Reynolds et al. 2007). Similarly, the variables
used should have low spatial variability – outside of recognised gradients.
3.1
Evaluation frameworks
Evaluation criteria have evolved parallel to changes in conceptual
frameworks and perspectives for restoration, which have in turn been
reflected in the type of attributes and indicators selected for monitoring and
assessment. Thus, traditional evaluation frameworks that commonly focused
on technical aspects of compliance success (e.g., seedling survival rates in
forest plantations; Alloza 2003), have given way to approaches that focus on
ecosystem health and integrity (Xu et al. 2001, Ruiz-Jaen and Aide 2005).
The SER Primer (SER 2004) proposed a list of nine attributes that can be used as
conceptual frameworks for designing ecological restoration projects and
evaluating restoration success. It includes diversity and other structural
properties (such as presence of indigenous species and presence of
46
Ground-based Assessment Methods and Indicators
functional groups necessary for long-term stability), and general ecosystem
functions (such as resilience to natural disturbances and self-sustainability).
As discussed by Whiffield and Geist (see Working paper I, this report), focus
has recently shifted towards socio-ecological assessments, which recognise
the coupled nature and dynamics between human and ecological systems.
The Millennium Ecosystem Assessment provides a mechanism for integrating
human and environmental systems through the concept of ecosystem
services, defined as “the benefits people obtain from ecosystem” (MA 2005).
The MA framework is commonly used in desertification research5, and has
recently been used to assess potential impacts of ecological restoration (Rey
Benayas 2009).
Finally, the need to base assessment of ecosystem health on stakeholder
perspectives, and the need to integrate local and scientific knowledge are
increasingly acknowledged (ICCD 2007, Reed et al. 2008). Addressing these
needs would involve using a conceptual framework that facilitates the
integration of scientifically developed indicators and locally identified
indicators, capturing the variety of local perspectives and expertise through
participatory approaches.
3.2
Number, domain and scale of application of the assessment indicators
There is no universal prescription for what to measure in order to describe the
ecosystem response to management actions. In general, the spectrum of
alternatives ranges from project-specific options to more broadly applicable
selection of indicators, and from simple indicators of ecosystem integrity to
indicator suites of a large variety of attributes (Fig. 1).
In making these choices of evaluation approaches and indicators, there are
several tradeoffs to consider. On the one hand, the use of single indicators of
ecosystem integrity can be a very cost-effective option that reduces
monitoring effort, but it requires a profound knowledge of how well the
indicator represents the structural and functional conditions expected for the
restored system. Furthermore, the loss of information when many variables are
integrated into a single index could mask real differences between
management and restoration options. On the other hand, the more site- or
project-specific the indicators, the more useful the resulting information for
local managers to adjust restoration practice within an adaptive
management framework. However, the evaluation results of such a tailored
approach would apply only to the site and conditions under study, hindering
the applicability to broad geographic areas and the comparison of
restoration strategies across a variety of sites and regions.
Evaluation approaches based on few site-specific indicators would fit
management actions with relatively straightforward, project-specific
5
See more on this topic in Chapter 1, this report.
47
S. Bautista and A. G. Mayor
objectives or actions that are applied to a particular and relatively small
piece of land over a limited period of time. For example, the local recovery of
the population of certain species is often the primary goal of restoration, due
either to ecological, economic and/or cultural reasons. The achievement of
such a specific goal can be evaluated through the monitoring of few
indicators related to the dynamics and sustainability of the target population
(Bash and Ryan 2002). However, the selection of appropriate indicators is not
so straightforward for projects with more general goals, such as those
implemented to prevent desertification and restore degraded drylands. The
ultimate goal of the actions implemented to combat desertification is to
improve human well-being through the recovery of ecosystem health and
integrity, and the conservation or enhancement of the provision of dryland
ecosystem goods and services. Which metrics should be used to evaluate
projects with such general objectives?
Figure 1. General range of alternatives for ecological evaluation of management
and restoration projects as defined by the number of indicators and their domain of
application (Adapted from Bautista and Alloza 2009).
4.2.1 Evaluation approaches based on few holistic indicators
Simplification towards essential indicators that could characterise ecosystem
recovery adequately is obviously a cost-effective approach to evaluation. For
example, measures concerning indicator species, umbrella species, guilds, or
assemblages of indicator species are often used as surrogates of ecosystem
48
Ground-based Assessment Methods and Indicators
function and integrity (e.g., Williams 1993, Patten 1997). The structural and
functional requirements of indicator species should reflect the conditions
expected in the healthy ecological system. This approach requires the
development of a conceptual model that outlines the structure of the
community, including interrelations among ecosystem components (Block et
al. 2001). Therefore, the accurate use of these indicators depends on a high
level of knowledge of the target system.
Although evaluation protocols based on indicator species are relatively sitespecific, when based on general taxa or guilds (e.g., bird populations,
biological crusts) they could be applied to a broad range of sites and project
types (Neckles et al. 2002, Bowker et al. 2006). General indicators of
community structure, such as species richness, diversity, and evenness, can
also be used for evaluating general ecosystem response to prevention and
restoration actions (e.g., Reay and Norton 1999, Passell 2000, Gómez-Aparicio
et al. 2009).
Vegetation cover is the most common single indicator used for evaluating
management projects and restoration success, as it is often assumed that the
recovery of fauna and ecological processes will follow the establishment of
vegetation (Ruiz-Jaen and Aide 2005). Since vegetation cover is relatively
easy to assess, it is commonly used as a surrogate of ecosystem functions and
habitat quality (Reay and Norton 1999, Wilkins et al. 2003, Wildham et al.
2004). However, vegetation cover alone cannot always reflect how well an
ecosystem is functioning, and, of course, do not take into account species
identities and their potential role as keystone species, noxious weeds, or any
other particular role played by single species or functional groups.
During the last decade, a variety of functional assessment approaches have
been proposed (Tongway and Hindley 2004, Herrick et al. 2005). These
approaches rely on few functional indices that commonly integrate a suite of
metrics. For example, the “Landscape Functional Analysis” (LFA)
methodology (Tongway and Hindley 1995, 2004) assesses ecosystem
functional status through a set of easily recognizable soil and landscape
features, from which indices of infiltration, stability and nutrient cycling are
derived. These indices are expected to reflect the status of water
conservation, soil conservation, and nutrient cycling processes in the target
ecosystems.
Much of the recent scientific evidence suggests that ecosystems do not
always undergo predictable and more or less gradual trajectories. Indeed,
ecosystems can exhibit threshold dynamics, change between alternative
metastable states that are then reinforced by internal feedbacks (Scheffer
and Carpenter 2003, Rietkerk et al. 2004, Suding et al. 2004, Bestelmeyer 2006,
Suding and Hobbs 2009). These facts have triggered recent and ongoing
research on indicators that have the capacity for early warning of sudden
changes, state shifts, and, in particular, of imminent dryland degradation
(e.g., Kéfi et al. 2007). The general assumption of these approaches is that
certain variables exhibit threshold values that can indicate ecosystem state
49
S. Bautista and A. G. Mayor
shifts. Similarly, degraded systems that have shifted to a new state that is
reinforced by internal feedbacks cannot be restored to the previous state
unless certain thresholds are passed (Whisenant 1999, Suding et al. 2004). Due
to hysteresis dynamics, thresholds for degradation and recovery paths are
very different. Knowledge about restoration thresholds is still very scarce
(Maestre et al. 2006), which limits the definition of reference target values for
evaluation and the development of state-change indicators for evaluating
restoration success. However, this is a hot topic in ecological indicator
research and new insights and advances can be expected in the near future.
4.2.2 Evaluation approaches based on large suites of indicators
Evaluation based on a relatively large suite of indicators (Fig. 1) aims to
present a more comprehensive diagnosis of the restoration and
management effects. An integrated suite of indicators may be specific for
certain sites, problems or project types (Keddy and Drummond 1996, Davis
and Muhlberg 2002, Palmer et al. 2005) or may mostly rely on general metrics
that can be used for assessing a wide range of cases (SER 2004). Several
authors have promoted approaches that combine both general and casespecific indicators (e.g., Neckles et al. 2002, Jorgersen et al. 2005).
The use of multiple indicators maximizes the amount and variety of
information provided on the management impacts and is the best approach
possible when there is not sufficient scientific knowledge to support the use of
single holistic indicators as proxies for the function and integrity of the target
dryland ecosystem. Numerous authors have proposed lists of attributes that
can be used for evaluating management actions and restoration success
(e.g., Aronson and Le Floc’h 1996, Hobbs and Norton 1996, SER 2004, Palmer
et al. 2005, Ruiz-Jaen and Aide 2005).
According to this view, evaluation should rely on the widest range possible of
attributes and perspectives, provided they are relevant. The challenge for this
approach is to organize and integrate the profuse information in a
harmonized way, allowing conclusions to be drawn, avoiding redundant
information, and keeping the number of attributes assessed manageable in
practical terms. For example, for evaluating forest restoration actions in
northern Mediterranean drylands, a protocol based on a wide variety of
indicators was recently developed by the REACTION project (Bautista et al.
2004). This protocol was not only conceived as an evaluation methodology,
but also as an information system designed to compile and disseminate the
information derived from an integrated assessment of the restoration projects
evaluated. The large suite of indicators used were integrated into a small set
of ecological, environmental, socio-economic, and cultural attributes that
focused on both human and natural aspects of the systems evaluated
(Bautista and Alloza 2009, Bautista et al. 2010).
50
Ground-based Assessment Methods and Indicators
4 Common indicators and methods for the evaluation of
prevention and restoration actions to combat desertification.
4.1
An evaluation approach based on the comparative assessment of the
provision of ecosystem services
Management actions in degraded drylands may range from “managing as
usual” approaches to adaptive management approaches that base
improvement on the feedback provided by monitoring and assessment
programs. Since the actions applied are choices from the many options
available, assessment and evaluation can be based on a comparative
analysis of the outcomes and impacts. There are cases in which objectives or
reference values for the attributes of interest, as well as their acceptable
range of variability, are precisely defined, and hence the evaluation
approach should largely rely on measuring the achievement of those welldefined objectives and target values for any given single action. However,
most prevention and restoration actions to combat dryland degradation
share similar goals, such as the general improvement of ecosystem health,
the enhancement of productivity and other ecosystem services, or the
general improvement of people well-being. Therefore, we advocate here a
comparative assessment of the various management actions considered in
any given target area, including comparison with degraded (managing as
usual) areas or with healthy reference areas, if appropriate.
We propose to base the comparative assessment of prevention and
restoration actions on (1) key common indicators that represent overall
functioning of dryland ecosystems, and (2) site-specific indicators identified
by local stakeholders that are relevant to the objectives and the particular
context conditions. By combining scientifically- and locally-derived indicators
this approach aims at promoting the integration of scientific and local
knowledge on dryland processes, and capturing both the specificities of the
particular dryland systems assessed, and therefore the variation among
regions, countries, and sites, and the impacts on key underlying processes
that are common across drylands.
Following the rationale presented by Whitfield and Geist (2010, Working paper
I, this report), the Millennium Assessment framework is adopted here for
selecting and framing both common indicators and site-specific indicators.
The ‘Ecosystem Services’ concept and categories provide a suitable
framework for analysing in a harmonised and consistent way the information
provided by the – a priori – wide variety of global and site-specific indicators
that could be identified by experts and local stakeholders (see, for example
Rey Benayas et al. 2009).
Next section will present and discuss a suite of common indicators for
assessing the impacts of prevention and restoration actions on dryland
ecosystems. The selected indicators focus on biophysical attributes and
51
S. Bautista and A. G. Mayor
processes. The socio-economic assessment of actions to
desertification is addressed in Working papers I and IV in this report.
4.2
combat
Common indicators for assessing dryland management and restoration
Regulating & Supporting
Service
Provisioning
Service
In selecting an appropriate common suite of criteria and indicators for the
biophysical assessment of actions to combat desertification, a well-balanced
basket of ecosystem services should be considered, covering the three broad
categories of provisioning, regulatory and supporting services (the assessment
of cultural services is addressed in Working papers I and IV, this report), and
focusing on key services in drylands (MA 2005). Figure 2 summarises the
criteria and suite of indicators proposed here, including potential metrics for
their assessment. The selected criteria and indicator also seek to maximize
possible synergies with global programmes pursuing ecosystem health and
associated human well-being, such as the United Nation Framework on
Climate Change (UNFCC) or the Convention on Biodiversity (CBD).
Metrics
Criteria
Indicator
Goods (food, fiber,
timber, fuel wood,
fresh water)
Site-specific
Primary
production
Vascular plant ANPP
Above-ground biomass /
Growth of key species
C sequestration
Soil Organic C (SOC)
Standard SOC measures
Water regulation
Plant cover & pattern
Indicator-specific
Plant cover
Bare-soil connectivity
Soil conservation
& Nutrient cycling
Soil surface condition
Infiltration & Nutrient
Cycling Indices (mLFA)
Biodiversity
Diversity of vascular
plants
Species composition and
abundance of vascular
plants. Diversity indices
ANPP: Above-ground Net Primary Productivity
SOC: Soil Organic Carbon
mLFA: modified Landscape Functional Analyses (Tongway and Hindley 2004)
Figure 2. Common criteria, indicators and metrics for assessing prevention and
restoration actions to combat dryland degradation
52
Ground-based Assessment Methods and Indicators
The assessment of goods derived from biological productivity, such as food,
fiber, fuel wood or timber, should be part of any common protocol to
evaluate the impacts of prevention and restoration actions, as these
provisions represent basic materials and quantifiable benefits for most dryland
people. However, in spite of its relevance across drylands, the key goods
provided and their relative importance vary between sites, countries and
regions, including countries and regions where provision of goods is of small
importance as compared with regulating, supporting or cultural services.
Therefore, the suitable indicators, as well as the associated metrics, depend
on the specific site context and should be selected by the local stakeholders
and experts. Some general approaches for evaluating land productivity, in
terms of provisioning services, are discussed by Onofri et al (Working paper IV,
this report).
For those cases where the actions applied could have a direct impact on the
provision of fresh water, this service should be assessed as part of the
evaluation protocol. For instance, reforestation or afforestation actions can
modify acquifer recharge or can prevent the rise of saline groundwater near
the surface (Walsh et al. 2003, Calder 2007), and therefore the use of fresh
water availability –measured, for instance, as the aquifer water level– as an
indicator of the restoration impacts on ecosystem services is very much
appropriate. However, as above, the suitability of this indicator is highly
dependent on the site and the type of actions assessed. For a majority of
management and restoration actions, the impacts on fresh water availability
are indirect off-site impacts, resulting from changes in water regulation and
conservation services (see below).
Primary Production is a key supporting service on which many other services
depend. In drylands, the best indicator for assessing this service would be the
Above-ground Net Primary Production (ANPP) of vascular plants. The groundbased assessement of total plant ANPP is however a challenging task. The
assessment of changes in above-ground biomass or growth of few species or
functional groups of particular interest could be a better approach for the
ground-based assessment of primary production (Figure 2). When the spatial
extent of the actions is large enough, remote sensing based methods
probably are the most cost-effective approaches for assessing changes and
long-term trends in primary production (see Working paper III, this report).
NPP in drylands is highly constrained by water availability, and large interannual climatic variations cause important fluctuations that could mask the
impacts of the actions applied. Both ground and remote-sensing based
approches must therefore adjust their metrics for rainfall variability. The ratio of
net primary production to precipitation – the rain-use efficiency (commonly
calculated from remotely sensed vegetation indices and rain gauge data) is
often used as indicator of land degradation or recovery. Given the high
variability and fast response of this service to changes in water inputs, it is
important to focus the assessment on the ecosystem resilience in response to
this variation: how much NPP is in step with rainfall? Does NPP recover rapidly
following drought? (Prince et al. 2004). It is also important to assess long-term
53
S. Bautista and A. G. Mayor
trends, and to consider the
management different actions.
same
time
periods
when
comparing
Appropriate land management practices can greatly increase both soil and
vegetation carbon stocks and thereby contribute to global climate
regulation. Soil organic carbon and above-ground biomass are proposed
here as carbon sequestration indicators. These two indicators have been
included within the Essential Climate Variables list, recently released by the
UNFCCC (GCOS 2010), and therefore their use could strength the linkages
among global environmental programmes.
The soil organic carbon pool is two times the size of the atmospheric pool and
to and a half times the size of the biotic pool. Strategies to increase the soil
carbon pool include soil restoration and woodland regeneration, no-till
farming, cover crops, nutrient management, improved grazing, water
conservation and harvesting, and agroforestry practices (Lal 2004). Soil
organic carbon (SOC) is considered to be a ‘slow’ variable, and it can be
asessed through well-known and widely applied available protocols. Ideally,
the assessment of carbon sequestration rates would require the long-term
monitoring of SOC. Furthermore, the rate of soil organic carbon sequestration
depends on soil texture and structure, rainfall and temperature. However, for
the comparative analysis of management strategies within a given site, total
top-soil organic carbon, relative to the implementation time of the actions
evaluated, can be a suitable metric for evaluation purposes.
Direct estimation of aboveground biomass through destructive harvesting
can be used to quantify above-ground carbon stocks. However, there are
available methods for converting simple ground and remote sensing based
measurement into plant carbon stocks (Brown et al. 2005, Gibbs et al. 2007).
For instance, in forest areas, measurements of diameter at breast height (DBH)
alone or in combination with tree height can be converted to estimates of
forest carbon stocks using allometric relationships. Actually, species-specific
allometric relationships are not needed to generate reliable estimates of
forest carbon stocks. Grouping all species together and using generalized
allometric relationships, stratified by broad forest types or ecological zones,
can be highly effective (Brown 2002).
Water regulation and conservation, frequently tied to the conservation of soil
and nutrients trough runoff, is probably the most essential function in drylands
(Whitford, 2002), and thus, its evaluation has a special interest for their
management. However, measuring water fluxes is highly time consuming and
costly, especially at larger scales. Vegetation cover is commonly used as a
surrogate of water conservation potential, as its relationships with general
hydrological functioning have been widely reported. However, a number of
studies in semiarid areas have shown that not only cover but also shape,
spatial orientation and arrangement of plant patches within a landscape can
greatly influence hydrological functioning (e.g., Ludwig et al. 1999,
Puigdefábregas 2005, Bautista et al. 2007). During the last decade, several
functional indices that assume a tied relationship between the spatial pattern
54
Ground-based Assessment Methods and Indicators
of the vegetation and the hydrological functioning of drylands have been
proposed (Bastin et al. 2002, Imeson and Prinsen 2004, Ludwig et al. 2007,
Mayor et al. 2008). The theoretical framework for these approaches considers
that landscapes occur along a continuum of functionality from systems with
low bare-soil connectivity that conserve all resources to those that have no
plant patches and leak all resources (Ludwig and Tongway 2000). The index
developed by Mayor et al. (2008), Flowlength, is a simple metric of bare-soil
connectivity that integrates both plant cover and pattern, and which has
shown strong positive relationships with hillslope runoff and sediment yields. It
can be easily estimated from vegetation maps, aerial photos, or highresolution satellite images, combined with elevation data.
Indicators and metrics for assessing water conservation are generally used to
assess soil conservation as well. In addition, a number of indicators that
specifically rely on soil and surface properties have been proposed for the
assessment of soil functioning, in particular soil conservation and nutrient
cycling (Tongway and Hindley, 2004, Herrick et al. 2005). For example, the
“Landscape Functional Analysis” (LFA) methodology (Tongway and Hindley,
1995, 2004) assesses ecosystem functional status through a set of easily
measured soil and landscape features, from which indices of infiltration,
stability and nutrient cycling are derived. These indices are expected to
reflect the status of water conservation, soil conservation, and nutrient cycling
processes in the target ecosystems. The LFA approach has been applied in
many dryland areas worldwide (e.g., Tongway and Hindley 2000, McR. Holm
et al. 2002, Maestre and Puche 2009), where LFA indices have shown positive
relationships with quantitative measures of (top)soil features such as
aggregate stability, water infiltration capacity and nutrient content.
Biodiversity is involved in most services provided by dryland ecosystems
(Blignaut and Aronson 2008). Recent results indicate that biodiversity is
positively related to the ecological functions that support the provision of
ecosystem services in restored areas (Rey Benayas et al. 2009). Biodiversity
assessments typically focus on particular biota, ranging from general groups
(e.g., plants, vertebrates, herbaceous species) to specific taxa or guilds (e.g.,
butterflies, resprouting shrubs), often resulting in a combined approach based
on biodiversity and indicator taxa (Kerr et al. 2000). Here we propose to focus
on the diversity of vascular plants as indicator of overall biodiversity. When
certain species or functional groups play a key functional role within the
drylands assesses, measuring the abundance of the these key species will
contribute to information on the biodiversity status of the system.
5 Conclusions
In summary, the approach advocated here for the ground-based biophysical
assessment of prevention and restoration actions combines the use of few
essential indicators that could characterise ecosystem recovery for a majority
of drylands worldwide with a site-specific suite of indicators selected by local
expert and stakeholders. Both common and site-specific indicators are
55
S. Bautista and A. G. Mayor
expected to jointly represent a well-balanced basket of ecosystem services,
with common indicators mostly focusing on water and soil conservation,
nutrient cycling, carbon sequestration, and biological diversity, while sitespecific indicators capture the provision of goods and fresh water, including
the net primary production of key species or functional groups.
6 References
Allen, C.D., Savage, M., Falk, D.A., Suckling, K.F., Swetnam, T.W., Schulke, T., Stacey, P.B.,
Morgan, P., Hoffman, M., and Klingel, J.T. 2002. Ecological restoration of southwestern
ponderosa pine ecosystems: a broad perspective. Ecological Applications 12(5):
1418-1433.
Alloza, J.A. 2003. Análisis de repoblaciones forestales en la Comunidad Valenciana.
Desarrollo de criterios y procedimientos de evaluación. Ph.D. Thesis, Universitat
Politècnica de València, Spain.
Aronson J. and Le Floc'h E. 1996. Vital landscape attributes: Missing tools for restoration
ecology. Restoration Ecology 4(4): 377-387.
Aronson J., Floret, C., Le Floc'h, E., Ovalle, C., and Pontanier R. 1993. Restoration and
rehabilitation of degraded ecosystems in arid and semi-arid lands. I. A view from the
south. Restoration Ecology 1: 8-17.
Bash, J.S. and Ryan, C.M. 2002. Stream restoration and enhancement projects: Is anyone
monitoring? Environmental Management 29(6): 877-885.
Bastin, G.N., Ludwig, J.A., Eager, R.W., Chewings, H., Liedloff, C., 2002. Indicators of landscape
function: comparing patchiness metrics using remotely-sensed data from rangelands.
Ecological Indicators 1: 247-260.
Bautista, S. and Alloza, J.A. 2009. Evaluation of forest restoration projects. In S. Bautista, J.
Aronson, and J. Vallejo (eds.), Land Restoration to Combat Desertification. Innovative
Approaches, Quality Control and Project Evaluation. CEAM Foundation, pp. 47-72.
Bautista, S., J.A. Alloza, and V.R. Vallejo. 2004. Conceptual framework, criteria, and
methodology for the evaluation of restoration projects. The REACTION approach.
Fundación CEAM, Valencia. (http://www.ceam.es/reaction/).
Bautista, S., Mayor, A. G., Bourakhouadar, J., Bellot, J., 2007. Plant spatial pattern predicts
hillslope runoff and erosion in a semiarid Mediterranean landscape. Ecosystems 10(6):
987-998.
Bautista, S., Orr, B.J., Alloza, J.A., and Vallejo, V.R. 2010. Evaluation of the restoration of
dryland ecosystems in the northern Mediterranean: Implications for practice. In M-F.
Courel and G. Schneier-Madanes (eds.), Water in Arid and Semi-arid Zones. Advances
in Global Change Research. Springer, Dordrecht, The Netherlands, pp. 295-310.
Bestelmeyer, B.T. 2006. Threshold concepts and their use in rangeland management and
restoration: The good, the bad, and the insidious. Restoration Ecology 14: 325-329.
Blignaut, J. and Aronson, J. 2008. Getting serious about maintaining biodiversity. Conservation
Letters 1: 12-17.
Block, W., Franklin, A., Ward, J., Ganey, J., and White, G., 2001. Design and implementation of
monitoring studies to evaluate the success of ecological restoration on wildlife.
Restoration Ecology 9(3): 293-303.
Bowker, M.A., Belnap, J., and Miller, M.E. 2006. Spatial modeling of biological soil crusts to
support rangeland assessment and monitoring. Rangeland Ecology and
Management 59: 519-529.
56
Ground-based Assessment Methods and Indicators
Bradshaw, A.D. 1997. What do we mean by restoration?. In K. Urbanska, N. Webb and P.
Edwards (eds.), Restoration Ecology and Sustainable Development. Cambridge
University Press, Cambridge, UK, pp. 8-14.
Brinson, M.M. and Rheinhardt, R. 1996. The role of reference wetlands in functional assessment
and mitigation. Ecological Applications 6: 69-76.
Brown, S. 2002 Measuring carbon in forests: current status and future challenges
Environmental Pollution 116: 363–72.
Brown, S., Pearson, T., Slaymaker, D., Ambagis, S., Moore, N., Novelo, D., and Sabido, W. 2005.
Creating a virtual tropical forest from three-dimensional aerial imagery: application for
estimating carbon stocks Ecological Application 15: 1083–95.
Calder, I. 2007. Forests and water—ensuring forest benefits outweigh water costs. Forest
Ecology and Management 251:110–120.
Carpenter, S.R. and Turner, M.G. 2000. Hares and tortoises: Interactions of fast and slow
variables in ecosystems. Ecosystems 3(6): 495-497.
Choi, Y.D. 2004. Theories for ecological restoration in changing environment: Toward 'futuristic'
restoration. Ecological Research 19: 75-81.
Cortina, J., Maestre, F.T., Vallejo, V.R., Baeza, M.J., Valdecantos, A., and Pérez-Devesa, M.
2006. Ecosystem structure, function, and restoration success: Are they related? Journal
for Nature Conservation 14: 152-160.
Dale, V.H. and Beyeler, S.C. 2001. Challenges in the development and use of ecological
indicators. Ecological Indicators 1: 3-10.
Davis, J.C. and Muhlberg, G.A. 2002. The evaluation of wetland and riparian restoration
projects. Anchorage, AK. Alaska Dep. of Fish and Game, Habitat and Restoration
Division (http://www.state.ak.us/adfg/habitat/geninfo/publictn/riparianproj.pdf).
Falk, D.A. 2006. Process-centred restoration in a fire-adapted ponderosa pine forest. Journal
for Nature Conservation 14: 140-151.
Gaboury, M. and Wong, R. 1999. A framework for conducting effectiveness evaluations of
watershed restoration projects. Watershed Restoration Technical Circular 12. Ministry
of Environment, Lands and Parks and Ministry of Forest. British Columbia.
GCOS (Global Climate Observing System) 2010. Implementation Plan for the Global
Observing System for Climate in Support of the UNFCCC (August 2010 Update) GCOS138 (GTOS-184, GTOS-76, WMO-TD/No.1523. Geneva: WMO GCOS. Available Online:
http://www.wmo.int/pages/prog/gcos/Publications/gcos-138.pdf
Gibbs H.K., Brown S., Niles J.O., and J.A. Foley. 2007. Monitoring and estimating tropical forest
carbon stocks: making REDD a reality. Environmental Research Letters 2(4), 045023.
doi:10.1088/1748-9326/2/4/045023.
Gómez-Aparicio, L., Zavala, M.A., Bonet, F.J., and Zamora, R. 2009. Are pine plantations valid
tools for restoring Mediterranean forests? An assessment along abiotic and biotic
gradients. Ecological Applications 19(8): 2124-2141.
Henry, C.P. and Amoros, C. 1995. Restoration ecology of riverine wetlands. I. A scientific base.
Environmental Management 19(6): 891-902.
Herrick, J.E., van Zee, J.W., Havstad, K.M., and Whitford, W.G. 2005. Monitoring Manual for
Grassland, Shrubland and Savanna Ecosystems. USDA-ARS Jornada Experimental
Range, Las Cruces, New Mexico. University of Arizona Press, Tucson.
Higgs, E.S. 1997. What is good ecological restoration? Conservation Biology 11: 338-348.
Hobbs, R.J. and Harris, J.A. 2001. Restoring ecology: Repairing the Earth's ecosystems in the
new millennium. Restoration Ecology 9(2): 239-246.
57
S. Bautista and A. G. Mayor
Hobbs, R.J. and Norton, D.A. 1996. Towards a conceptual framework for restoration ecology.
Restoration Ecology 4(2): 93-110.
Holl, K.D. and Cairns Jr, J. 2002. Monitoring and appraisal. In M.R. Perrow and A.J. Davy (eds.),
Handbook of Ecological Restoration, I: Principles of Restoration. University of Anglia,
Cambridge University Press, pp. 411-432.
ICCD/COP8/CST 2007. Report of the fifth meeting of the Group of Experts of the Committee
on Science and Technology. Addendum: Methodologies for the Assessment of
Desertification at Global, Regional and Local Levels. ICCD/COP(8)/CST/2/Add.6.
Imeson, A.C., Prinsen, H.A.M., 2004. Vegetation patterns as biological indicators for identifying
runoff and sediment source and sink areas for semi-arid landscapes in Spain.
Agriculture, Ecosystems and Environment 104 333-342.
Jorgersen, S.E., Xu, F-L., Salas, F., and Marques, J.C. 2005. Application of indicators for the
assessment of ecosystem health. In S.E. Jorgensen, R. Costanza, and F-L. Xu (eds.),
Handbook of Ecological Indicators for Assessment of Ecosystem Health. Lewis
Publishers, Inc. USA, pp. 5-66.
Keddy, P.A. and Drummond, C.G. 1996. Ecological properties for the evaluation,
management, and restoration of temperate deciduous forest ecosystems. Ecological
Applications 6(3): 748- 762.
Kéfi, S., Rietkerk, M., Alados, C.L., Pueyo, Y., Papanastasis, V.P., ElAich, A., and de Ruiter, P.C.
2007. Spatial vegetation patterns and imminent desertification in Mediterranean arid
ecosystems. Nature 449: 213-217.
Kerr, J.T., Sugar, A., and Packer, L. 2000. Indicator taxa, rapid biodiversity assessment, and
nestedness in an endangered ecosystem. Conservation Biology 14(6): 1726-1734.
Lal, R. 2004. Soil carbon sequestration impacts on global climacte change and food security.
Science 304: 1623-167.
Ludwig, J. A., Tongway, D., Freudenberger, D., Noble, J., Hodgkinson, K., 1997. Landscape
Ecology, Function and Management: Principles from Australia’s Rangelands. CSIRO,
vol. 158. CSIRO, Collingwood, Australia.
Ludwig, J.A., Bastin, G.N., Chewings, V.H., Eager, R.W., and Liedloff, A.C. 2007. Leakiness: A
new index for monitoring the health of arid and semiarid landscapes using remotely
sensed vegetation cover and elevation data. Ecological Indicators 7: 442-454.
Ludwig, J.A. and Tongway, D.J. 2000. Viewing rangelands as landscape systems. In O. Arnalds
and S. Archer (eds.), Rangeland Desertification. Kluwer Academic Publishers,
Dordrecht, The Netherlands, pp. 39-52.
Ludwig, J.A., Tongway, D.J., and Marsden, S.G. 1999. Stripes, strands or stipples: modelling the
influence of three landscape banding patterns on resource capture and productivity
in semiarid woodlands, Australia. Catena 37: 257-273.
MA
(Millennium Ecosystem Assessment). 2005. Ecosystems and Human
Desertification Synthesis. World Resources Institute: Washington, DC.
Well-being:
Maestre, F.T. and Puche, M.D. 2009. Indices based on surface indicators predict soil
functioning in Mediterranean semi-arid steppes. Applied Soil Ecology 41: 342–350.
Maestre, F.T., Cortina, J., and Vallejo, V.R. 2006. Are Ecosystem composition, structure, and
functional status related to restoration success? A test from semiarid Mediterranean
steppes. Restoration Ecology 14(2): 258-266.
Mayor, A.G., S. Bautista, E. E. Small, M. Dixon, and J. Bellot (2008), Measurement of the
connectivity of runoff source areas as determined by vegetation pattern and
topography: A tool for assessing potential water and soil losses in drylands, Water
Resour. Res., 44, W10423, doi:10.1029/2007WR006367.
58
Ground-based Assessment Methods and Indicators
McR. Holm, A., Bennet, L.T., Loneragan, W.A., Adams, M.A., 2002. Relationships between
empirical and nominal índices of landscape function in the arid shrubland of Western
Australia. Journal of Arid Environments 50: 1-21.
Neckles, H.A., Dionne, M.D., Burdick, D.M., Roman, C.T., Buchsbaum, R., and Hutchins, E. 2002.
A monitoring protocol to assess tidal restoration of salt marshes on local and regional
scales. Restoration Ecology, 10: 556-563.
NRC (U.S. National Research Council). 2000. Ecological Indicators for the Nation. Chapter 3: A
Framework for Indicator Selection. Washington D.C.: The National Academies Press.
180 pp. Online: http://www.nap.edu/openbook.php?record_id=9720&page=51
Palmer, M.A., Bernhardt, E.S., Allan, J.D., Lake, P.S., Alexander, G., Brooks, S., Carr, J., Clayton,
S., Dahm, C.N., Follstad-Shah, J., Galat, D.L., Loss, S.G., Goodwin, P., Hart, D.D., Hassett,
B., Jenkinson, R., Kondolf, G.M., Lave, R., Meyer, J.L., O'donnell, T.K., Pagano, L., and
Sudduth, E. 2005. Standards for ecologically successful river restoration. Journal of
Applied Ecology 42(2): 208-217.
Parker, V.T. and Pickett, S.T.A. 1997. Restoration as an ecosystem process: implications of the
modern ecological paradigm. In K. Urbanska, N.Webb, and P. Edwards (eds.),
Restoration Ecology and Sustainable Development. Cambridge University Press,
Cambridge, UK, pp. 17-32.
Passell, H.D. 2000. Recovery of bird species in minimally restored Indonesian thin strips mines.
Restoration Ecology 8: 112-118.
Patten, M.A. 1997. Reestablishment of a rodent community in restored desert scrub.
Restoration Ecology 5: 156-161.
Pickett, S.T.A. and Parker, V.T. 1994. Avoiding the old pitfalls: opportunities in a new discipline.
Restoration Ecology 2: 75-79.
Prince, S.D., Brown De Colstoun, E., and Kravitz, L.L. (2004). Evidence from rain-use efficiencies
does not indicate extensive Sahelian desertification. Global Change Biology 4(4): 359–
374.DOI: 10.1046/j.1365-2486.1998.00158.x
Puigdefábregas, J. 2005. The role of vegetation patterns in structuring runoff and sediment
fluxes in drylands. Earth Surface Processes and Landforms 30, 133-147.
Reay S.D. and Norton, D.A. 1999. Assessing the success of restoration plantings in a temperate
New Zealand forest. Restoration Ecology 7: 298-308.
Reed, M.S., Dougill, A.J., Baker, T.R. 2008. Participatory indicator development: What can
ecologists and local communities learn from each other? Ecological Applications
18(5): 1253–1269
Rey Benayas, J.M., Newton, A.C., Diaz, A., and Bullock, J.M. 2009. Enhancement of
biodiversity and ecosystem services by ecological restoration: A meta-analysis.
Science 325: 1121-1124. DOI: 10.1126/science.1172460.
Reynolds, J.F., Stafford Smith, D.M., Lambin, E.F., Turner II, B.L., Mortimore, M., Batterbury, S.P.J.,
Downing, T.E., Dowlatabadi, H., Fernández, R.J., Herrick, J.E., Huber-Sannwald, E.,
Jiang, H., Leemans, R., Lynam, T., Maestre, F.T., Ayarza, M., Walker, B. 2007. Global
desertification: Building a science for dryland development. Science 316: 847-551.
DOI: 10.1126/science.1131634.
Rietkerk, M., Dekker, S.C., de Ruiter, P.C., and van de Koppel, J. 2004. Self-organized
patchiness and catastrophic shifts in ecosystems. Science 305: 1926-1929.
Ruiz-Jaen, M.C. and Aide, T.M. 2005. Restoration success: How is it being measured?.
Restoration Ecology 13(3): 569-577.
Scheffer, M., and Carpenter, S.R. 2003. Catastrophic regime shifts in ecosystems: linking theory
to observation. Trends in Ecology and Evolution 18(12): 648-656.
59
S. Bautista and A. G. Mayor
SER (Society for Ecological Restoration, Science & Policy Working Group). 2004. The SER Primer
on Ecological Restoration (http://www.ser.org/).
Suding, K.N. and Hobbs, R.J. 2009. Threshold models in restoration and conservation: a
developing framework. Trends in Ecology and Evolution 24: 271-279.
Suding, K.N., Gross, K.L., and Houseman, G.R. 2004. Alternative states and positive feedbacks
in restoration ecology. Trends in Ecology and Evolution 19(1): 46-53.
Thayer, G.W., McTigue, T.A., Bellmer, R.J., Burrows, F.M., Merkey, D.H., Nickens, A.D., Lozano,
S.J., Gayaldo, P.F., Polmateer, P.J., and Pinit, P.T. 2003. Science-Based Restoration
Monitoring of Coastal Habitats, I: A Framework for Monitoring Plans Under the Estuaries
and Clean Waters Act of 2000 (Public Law 160-457). NOAA Coastal Ocean Program
Decision Analysis Series No. 23. NOAA National Centers for Coastal Ocean Science,
Silver Spring, MD.
Tongway, D.J. and Hindley, N. 2004. Landscape Function Analysis: Procedures for Monitoring
and Assessing Landscapes. CSIRO Publishing, Brisbane.
Tongway, D.J., Hindley, N., 2000. Assessing and monitoring desertification with soil indicators.
In: Arnalds, O., Archer, S. (Eds.), Rangeland Desertification, Advances in Vegetation
Sciences, vol. 19. Kluwer Academic Publishers, London, UK, pp. 89-98.
Tongway, D.J. and Hindley, N. 1995. Assessment of soil condition of tropical grasslands. CSIRO
Ecology and Wildlife, Canberra.
UNCCD 2009. UNCCD 1st Scientific Conference: Synthesis and recommendations.
ICCD/COP(9)/CST/INF.3. UN Convention to Combat Desertification. Available Online:
http://www.unccd.int/php/document2.php?ref=ICCD/COP%289%29/CST/INF.3
Vallejo, V.R., Aronson, J., Pausas, J.G., and Cortina, J. 2006. Restoration of Mediterranean
woodlands. In J. van Andel and J. Aronson (eds.), Restoration Ecology: The New
Frontier. Oxford, Blackwell Science, Oxford, pp. 193-209.
van Diggelen, R., Grootjans, Ab P., and Harris, J.A. 2001. Ecological restoration: State of the
art or state of the science?. Restoration Ecology 9(2):115-118.
Walsh, P.G., C.V. Barton, and K.D. Montagu. 2003. Macquarie catchment pilot project, New
South Wales, Australia: a cost-effective, market-based approach to reducing salinity
through tree planting. Unasylva 54 (212): 37–39.
Whisenant, S.G. 1999. Repairing Damaged Wildlands. Cambridge University Press, Cambridge,
UK.
White, P.S. and Walker, J.L. 1997. Approximating nature's variation: selecting and using
reference sites and reference information in restoration ecology. Restoration Ecology
5: 338-249.
Whitford, W.G. 2002. Ecology of desert systems. Elsevier Science Ltd., London, England.
Windham, L., Laska, M.S., and Wollenberg, J. 2004. Evaluating urban wetland restorations:
case studies for assessing connectivity and function. Urban Habitats 2(1): 130-146.
Wilkins, S., Keith, D.A., and Adam, P. 2003. Measuring success: evaluating the restoration of a
grassy eucalypt woodland on the Cumberland Plain, Sydney, Australia. Restoration
Ecology 11: 489-503.
Williams, K.S. 1993. Use of terrestrial arthropods to evaluate restored riparian woodlands.
Restoration Ecology 1: 107-116.
WWF, 2002. Indicators for measuring progress towards forest landscape restoration.
(http://www.panda.org/downloads/forests).
60
Ground-based Assessment Methods and Indicators
Xu, F-L., Tao, S., Dawson, R.W., Li, P-L., and Cao, L. 2001. Lake ecosystem health assessment:
indicators and methods. Water Research 13: 3157-3167.
Zedler, J.B. 2007. Success: An unclear, subjective descriptor of restoration outcomes.
Ecological Restoration 25(3): 162-168.
Zedler, J.B. 1995. Salt marsh restoration: lessons from California. In J. Cairns (ed.), Rehabilitating
Damaged Ecosystems. Lewis Publishers, Boca Raton. Florida, pp. 75-95.
Zedler, J.B. and Callaway, J.C. 1999. Tracking wetland restoration: Do mitigation sites follow
desired trajectories?. Restoration Ecology 7(1): 69-73.
61
Remote Sensing Based Assessment
III. Remote Sensing Based Options for
Assessment of Mitigation and Restoration
Actions to Combat Desertification
Working paper for the practice project
Achim Röder
Marion Stellmes
Joachim Hill
Thomas Udelhoven
62
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
© 2010 by A. Röder, M. Stellmes, J. Hill, Th. Udelhoven
Remote Sensing Department
Trier University
D-54286 Trier, Germany
[email protected],
trier.de
[email protected],
[email protected],
udelhove@uni-
www.feut.de
63
Remote Sensing Based Assessment
1 Introduction
The capability of ecosystems to supply goods and services that are essential
for human well-being are strongly impacted upon by many factors,
desertification and land degradation being among the most important ones
in dryland areas. This impact varies greatly, driven by the degree of aridity,
the variability of precipitation and the pressure exerted by people on the
resources.
A variety of measures are being devised that aim at a general recovery of
degraded ecosystems, or the mitigation of negative development. Evidently,
any such measure requires definition of an intended ecosystem state, of what
a “healthy” ecosystem looks like, or what services and functions it needs to
supply. In very general terms, restoration goals might be grouped according
to “species”, “ecosystem functions” and “ecosystem services” (Harris and van
Diggelen 2008).
In principle, there is no possibility to measure the success of restoration or
mitigation actions from remote sensing data directly. Rather, approaches
need to be designed to match the intervention scale of the action to be
addressed, and the development goal pursued by this action. In that sense,
remote sensing based options are closely linked to studies of land use / land
cover change. Yet, in this case locations under management or intervention
scenarios are known, and the task is to evaluate if the intended change (or
no-change if desired) can actually be detected in a spatially explicit manner.
As a direct consequence of these basic statements, first considerations of
which approach is best suited require a clear definition of
• The ecological problem being tackled
• The type of management intervention applied (including location and
spatial extension)
• The intended management or development goal
This information defines the type of remote sensing data to be utilized as well
as the indicators that might be employed to evaluate the degree of success
of management actions.
Therefore, the remote sensing component of a protocol to assess the success
of restoration and mitigation practices proposed by PRACTICE will draw from
concepts and techniques established in the context of land use / land cover
change assessments. Traditionally, these have been based on diachronic or
multi-date assessments of land use / land cover (LULC) maps, and an
evaluation of the changes between classes between two or more dates.
Such analyses may involve characterizing LULC for different dates and
interpreting change, or mapping change directly from diachronic data sets
and attributing the type of change to apparent patterns (Coppin et al.
2004a; Lu et al. 2004). Accordingly, great importance is attributed to the
selection of appropriate classification algorithms, as well as strategies of
defining training and validation data. In recent years, there has been growing
64
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
evidence that modifications of land cover, even within one broad LULC
category, may be of equal importance than changes between classes (in this
context termed transformations or conversions) (Lambin et al. 2006; Turner II et
al. 2007). Implications of gradual changes within vegetated areas relate for
instance to aspects as diverse as fire risk, soil erosion or carbon sequestration.
Assessing modifications, particularly if slow processes are considered,
necessitates a different approach that allows tracking gradual
developments, i.e. time series analysis using long-term datasets (Hill et al.
2004).
It follows from these considerations that the analysis of restoration measures
using remote sensing data relies on the derivation of suitable spatial
information for the required dates. This information can either relate to LULC
information or to the quantification of suitable indicators, such as proportional
cover of surfaces types (vegetation, sand, soil,…) or the derivation of more
specific biophysical indicators, i.e. leaf area index, leaf water content etc.
While LULC information is commonly employed to characterize change
between two or more dates, the temporal development of more specific
indicators may be quantified using time series analysis techniques. To move
towards a more comprehensive evaluation, information derived either way
subsequently needs to be integrated with auxiliary spatial and non-spatial
data, such as knowledge of applied management techniques, socioeconomic data, climate data etc. The outcome of this integration may be an
advanced evaluation, but could also serve to drive specific assessment
models at the landscape scale.
Today, a wide range of sensor systems placed on space-based and airborne
platforms are available. They are characterized by geometric resolution,
spatial coverage, repetition rate and history of data acquisition. These need
to be selected depending on target processes and areas, and on the type of
information product desired, which is in turn dependent on potential end
users. In addition, these systems are equipped with a largely varying number
of spectral bands determining their information content. Figure 1 provides an
overview of sensor systems grouped according to spectral and spatial
resolution.
Besides these technical specifications, the history and repetition rate of image
acquisition of the different missions are determining their potential for longterm analyses. Although various approaches for intercalibration across
different sensors exist (Jiang et al. 2008; Röder et al. 2005; Steven et al. 2003),
coherence of information is facilitated by the use of consistent datasets. In
such setups, spectral and geometric sensor setups as well as orbit parameters
allow for a direct integration of information derived from different sensors of
the same family.
Most prominent examples of long-term, globally available data archives are
the NOAA-AVHRR and Landsat systems, available since 1981 and 1972,
respectively. In recent years they have been complemented by further
relevant sensors, in particular the MODIS mission started in 1999.
65
Remote Sensing Based Assessment
Figure 1. Important operational and experimental remote sensing systems grouped
according to spectral and spatial resolution (Hill et al. 2004, modified)
These are the archetypical examples of the major categories of multitemporal data: coarse scale imagery with reduced spatial (and often
spectral) capabilities, but daily coverage allowing for the provision of
different stabilized, multi-day composites on the one hand; and finer scale
data with better spatial (and sometimes spectral) discrimination capabilities
at the expense of revisit intervals. While these data can be considered largely
operational and cost-efficient, commercial and experimental systems may
provide excellent information for particular questions, but these are
commonly covering only limited areas, and are either expensive (Very High
Resolution (VHR) data such as Ikonos, Quickbird, World-View) or experimental
in nature, such as airborne hyperspectral (Aviris, HyMap, Ares) or LiDAR
systems, which can – as of now – only be acquired in the frame of specific
campaigns (although different countries in Europe have already initiated
country-wide LiDAR campaigns).
Besides their technical characteristics, the level of detail resolved by these
systems, along with the area covered in total, largely determines their
potential use. Coarse-scale, hypertemporal systems such as NOAA-AVHRR,
Spot-VEGETATION or MODIS are highly relevant for policy makers and regional
to national decisions makers. On the other hand, systems with better spatial
66
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
resolution (e.g. Landsat, Aster, Spot) are more consistent with regional to local
land management, and finally VHR data may help in tracking processes at
the level of individual plots.
Given all these considerations, this document will focus on routinely available
remote sensing data sources, and aim at maintaining a high degree of
operational applicability in methods wherever possible. The techniques and
methods presented will be structured according to the major scales
described before, followed by some considerations on opportunities and
limitations associated with different scales.
2 Coarse scale
2.1
Data sources
For detection of large-scale effects from regional to country and global
levels, archives of satellite data with high temporal resolution provide a
temporally intermittent view across large areas (Bastin et al. 1995; Hill et al.
1998; Hostert et al. 2001b). With their high repetition rate, they allow to
distinguish long-term changes in surface reflectance from seasonal and/or
annual oscillations (Duchemin et al. 1999; Erbrecht 2003; Senay and Elliott
2002). Since temporal resolution of remote sensing systems is typically
enhanced at the expense of spatial resolution those systems operate on
coarser scales.
In particular data archives derived from the Advanced Very High Resolution
Radiometers (AVHRR-) onboard the National Oceanic and Atmospheric
Administration (NOAA-) satellites provide invaluable and irreplaceable
historical information about surface conditions (Tucker et al. 2005). Monitoring
relative biomasses and greenness indicators such as NDVI reveals information
about timing and distribution of phytophenological events such as greeningup, rate of biomass accumulation and onset of vegetation senescence. This is
of paramount importance in different contexts, among others to study the
influence of climatic conditions, the success of land management systems,
and impacts of overgrazing, exploitation of natural resources, pests and
diseases. Although new, global, land vegetation-sensing instruments such as
SeaWiFS land data (launch October 1997), SPOT4/VEGETATION (launch in
April 1998), MODIS from the Terra and Aqua platforms (launch in 2000 and
2002, respectively) or MERIS (onboard of Envisat; launch in March 2002) supply
higher quality data compared to the AVHRR, their relatively short operation
duration so far is the limiting factor for analyzing long-term surface conditions.
For instance, at the time of MODIS launch, there was already nearly a 20-year
NDVI global data set (1981 - 1999) available from the NOAA- AVHRR series.
Thus, AVHRR data will remain a fundamental component regarding the need
for the longest possible time series required for understanding long-term
variability in the earth terrestrial biosphere (Cihlar et al. 2004).
67
Remote Sensing Based Assessment
The existence of hypertemporal data archives enables the application of
well-established methods in time-series analysis to assess changes in surface
conditions and in addition to forecast trajectories to some degree, for
instance to evaluate the performance of national drought policies or
rangeland management programs in drylands. In addition, many time series
analysis techniques, such as autoregressive and moving average models,
Continuous Wavelet Transform or Empirical Mode Decomposition, require
temporal equally sampled data which are usually not available from local
scale systems.
By nature, results based on these data offer only explanations on the scale of
observation, thereby, local scale processes and management practices are
likely to remain undetected. At present, hypertemporal data archives with
global coverage derive from the AVHRR, Spot VEGETATION, MODIS and MERIS
sensors.
The most frequently used global AVHRR archives are compiled from the
"afternoon" NOAA operational meteorological satellites (NOAA7, 9, 11, 14, 16
and 18) and include the Pathfinder AVHRR Land (PAL) data and the data
from the global inventory monitoring and modeling studies (GIMMS).
The 8x8 km GIMMS set additionally includes data from the NOAA-16 and
NOAA-17 sensor and comprises bimonthly NDVI composites (days 1-15 and
16-end of the month) covering the period 1981 to 2006. The data are
available to the scientific community e.g. through the Global Land Cover
Facility
at
the
University
of
Maryland
(http://glcf.umiacs.umd.edu/data/gimms). The data pre-processing scheme
is described in Tucker et al. (2005).
The MEDOKADS poses an example for a Local Area Coverage (LAC) data set
with higher spatial (1x1 km) and temporal (daily) resolution compared to the
PAL and GIMMS data sets. The data are collected, processed and distributed
by the Institute of Meteorology, Free University of Berlin, Germany and cover
the period from 1989 until 2005. This data set comprises full resolution AVHRR
channel data and collateral data in latitude-longitude representation, with a
resolution of 0.01° in both directions (Friedrich and Koslowsky 2009).
The Spot VEGETATION Programme provides daily coverage of the entire earth
at a spatial resolution of 1 km. The instrument and associated ground services
for data archival, processing and distribution are operational since April 1998.
The first VEGETATION instrument is part of the SPOT 4 satellite and a second
payload, VEGETATION 2, is now operationally operated onboard SPOT 5. Data
and product information are available at http://www.spot-vegetation.com.
MODIS was launched by NASA aboard the Earth Observing System (EOS)
satellite Terra on December 18, 1999. A second MODIS instrument was
launched May 4, 2002 on the EOS-Aqua platform. With both instruments in
operation, MODIS provides a morning and afternoon view with global, neardaily repeat coverage, complementing the spectral, spatial, and temporal
coverage of the other research instruments aboard each platform. Its
detectors measure 36 spectral bands between 0.405 and 14.385 µm, and it
68
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
acquires data at three spatial resolutions -- 250m, 500m, and 1,000m. As part
of the MODIS mission, data provided range from surface reflectance and
albedo data sets to derived vegetation-related indices (NDVI, EVI, FAPAR,
LAI) and even LULC data provided at a yearly resolution. More detailed
information on MODIS data products and algorithms can be found at
http://modis.gsfc.nasa.gov/data/.
MERIS is a programmable imaging spectrometer on ESA’s ENVISAT platform
with fifteen spectral bands in the solar reflective spectral range that can be
selected by ground command. A Product Handbook that guides users to
select and to use MERIS data and that explains the way these data are
processed
and
organized
is
available
online
at
http://envisat.esa.int/handbooks/meris/.
2.2
Diachronic analyses
There have been many efforts to apply various classification approaches to
coarse-scale, hypertemporal data, commonly making use of the fact that
these data are able to trace phenological dynamics of different vegetation
units over the year. As a result, different LULC products derived from different
sensor systems exist (Bartholome and Belward 2005; Friedl and Brodley 1997;
Friedl et al. 2002; Friedl et al. 2010; Loveland et al. 2000). Provided their
availability for suitable time steps, they can be employed to map LULC
change either aggregated for larger units or at a per-pixel level (Coppin et al.
2004a). Alternatively, per-year differences in aggregated NDVI values (see
below) that exceed certain thresholds were shown to be related to LULC
change processes under specific assumptions (Lunetta et al. 2006).
Different limitations need be considered in this context. On a general level,
the coarse resolution of 500 m (best resolution for MODIS LULC products;
Globcover derived from MERIS provides 300 m but is only available for one
date) or less involves considerable problems given the heterogeneous
structure of many landscapes (Stellmes et al. 2010). In such cases, the
problem of mixed pixels may cause erroneous classification of fine-grained
landscape patterns. Globally applicable mapping approaches are frequently
reported to be of highly variable quality, with best results commonly being
achieved in regions where algorithms had been developed and calibrated
(Cohen et al. 2003; Price 2003). In addition, maps of the same region derived
from different sensor systems are not necessarily congruent, which is owed to
differences in classification techniques and LULC definitions (Herold et al.
2008). As a result, it seems advisable to base diachronic analyses of coarse
scale data on consistent data sets from the same sensor. This limits the use of
MODIS data, as these are only available since 2000.
69
Remote Sensing Based Assessment
2.3
Vegetation indices
Vegetation is considered a major indicator in remote sensing based
approaches, because it is spectrally distinct from many other surface types
and it can be characterized using different methods. Subsequently, such
information can thematically be linked to biophysical parameters, such as
leaf area index, biomass etc. and/or be interpreted in different thematic
contexts, such as grazing, fire (and post-fire recovery) etc.
Using remote sensing data, ‘vegetation’ can be characterized at different
scales and using a range of methodological approaches, depending on the
target information required.
Traditionally, the derivation of vegetation-related information has been based
on the calculation of simple spectral indices such has the Vegetation Index
(VI) or Normalized Difference Vegetation Index (NDVI), which is commonly
calculated as
NDVI =
NIR − R
NIR + R
where NIR and R denote reflectance of the near infrared and red portions of
the wavelength spectrum, respectively. The red part is sensitive toward
photosynthetically absorbed radiation, whereas surface reflectance
increases more sharply in the near-infrared for green vegetation than for soils
or senescent vegetation (Curran 1980). NDVI values vary between −1.0 to 1.0.
Larger values are associated with higher levels of green vegetation density
and abundance for a given plant, but spectral reflectance can generally be
different for different plants. Negative values indicate non-vegetated features
such as water, barren areas, ice, snow or clouds. It was shown that the NDVI
partially compensates multiplicative noise effects in multiple bands such as
changing illumination conditions, atmospheric attenuation, and viewing
aspect-factors that strongly affect observed radiances in the reflective
domain of the solar spectrum (Gutman 1991; Lyon et al. 1998). NDVI suffers
from a number of drawbacks, such as the influence of soil background
especially for bright soils and sparse vegetation canopies, or saturation effects
for very dense vegetation layers (Baret and Guyot 1991; Huete et al. 1985;
Verstraete 1994a). In addition, it does not eliminate differences that arise from
intrinsic differences between the different AVHRR instruments including
spectral response functions, sensor degradation and orbital drift effects
(Cracknell 1997). Despite these limitations, NDVI registers spatial information
about short- and long-term changes in their temporal contexts, even at the
coarse scale.
Alternative measures related to vegetation were developed to account for
the influence of background soil and atmosphere, including Tasselled Cap
(Kauth and Thomas 1976), Perpendicular Vegetation Index (PVI, (Richardson
and Wiegand 1977)), Weighted Difference Vegetation Index (WDVI, (Clevers
1988)), Soil Adjusted Vegetation Index (SAVI, (Major et al. 1990)), and nonlinear Global Evironmental Monitoring Index (GEMI, (Pinty and Verstraete
70
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
1992)). The computation of these indices requires knowledge of the spectral
properties of the background soil, to define the so-called soil line. The
distance of a vegetation pixel from the soil line is then interpreted as a
measure for the amount of vegetation. However, since the reflectance
properties of the soil are spatially variable, they cannot be generally defined
and the NDVI often produces a comparable quality in estimating vegetation
cover (Schmidt and Karnieli 2001).
Another promising measure for vegetation cover that overcomes some of the
limitations of the NDVI is based on spectral mixture analysis (SMA). The
classical SMA derives the fraction of vegetation cover on the basis of multispectral image data and also allows the characterization of different surface
materials (Hill et al. 1998; Smith et al. 1994). A modified SMA technique that
improves the vegetation cover estimation in low vegetated areas by
eliminating the effect of the background signal on the NDVI was introduced
by Sommer et al. (European Commission 1998) and further developed by
Stellmes et al. (2006) and Weissteiner (2008). This approach is based on the
assumption, that vegetation canopy should predominately control the
position on an AVHRR land surface pixel within the feature space spanned up
by NDVI and surface temperature. The amount of vegetation canopy is
inversely proportional to surface temperature, due to a variety of factors
including the lower heat capacity and thermal inertia of vegetation
compared to soils and the latent heat transfer through evapotranspiration
(Gates 1980; Goward and Hope 1989).
In the context of the MODIS mission, the Enhanced Vegetation index (EVI)
was developed and incorporated into the MODIS products (Huete et al.
2002). It is defined as follows:
EVI =
NIR − R
*G
L + NIR + c1 * R − c 2 * B
The EVI utilizes the ratio of the reflectance in the near-infrared band (pNIR,
MODIS 841-876 nm) and the red band (pRED, MODIS 620-670 nm). It also
employs the blue band (pBLUE, MODIS 459-479 nm) to reduce atmospheric
effects, the coefficients C1 and C2 of aerosol resistance, L for the
background stabilization and the gain factor G.
Vegetation indices have been correlated with various vegetation parameters
such as LAI, biomass, canopy cover, and fraction of absorbed
photosynthetically active radiation (fAPAR). The EVI was found to have a
more linear relationship with these parameters compared to the NDVI (Huete
et al. 2002). In addition, specific MODIS LAI and FAPAR products are available
based on a general reflectance transfer model (Myneni et al. 2002; Tian et al.
2002a, b). Yet, since image quality does often not support parameterization
of this model, it is often replaced by a scaled NDVI, thus questioning its
71
Remote Sensing Based Assessment
superiority over traditional vegetation indices (Yang et al. 2006a; Yang et al.
2006b).
Time-series analysis on daily raw hypertemporal images is problematic
because the data are strongly contaminated with clouds and cloudshadows. In order to produce nearly cloud free time-series with reduced
atmospheric contamination such data are regularly aggregated using
different compositing techniques, such as maximum-value compositing
(MVC) (Holben 1986) or Best Index Slope Extraction (Viovy et al. 1992).
2.4
Time-series analysis using hypertemporal satellite archives
Hypertemporal remote sensing data provide the opportunity to discriminate
between short-lived fluctuations in surface reflectance, several types of
phenologies and long-term transient signals. These patterns might be related
to degradation processes in drylands such as overgrazing. This chapter
provides an overview about relevant time-series analysis (TSA-) techniques
that are necessary to decouple transient, cyclic, and stochastic components
in time records (Stellmes et al. 2010; Verbesselt et al. 2010). Figure 2 shows an
example of how a (simulated) temporal NDVI profile can be decomposed
into these individual trend components.
Figure 2: Different components of hypertemporal time series: (a) trend (transient), (b)
cyclic, (c) noise (Udelhoven 2010)
TSA-methods are not explained in detail since they are subject of many
textbooks (Chatfield 2004; Shumway and Stoffer 2006; Wei 1990).
72
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
2.4.1 Preprocessing
As a first step the data archives should be cleaned from inconsistencies such
as missing values and outliers and extraneous noise. Missing values and
outliers can be replaced by (seasonal) averages of the affected time-series
or by interpolation from surrounding data. A reliable tool for outlier detection
is Chebyshev's Theorem, which states that a random variable, irrespective of
its distribution, will take on a value within k standard deviations s of the mean
with a probability (confidence level) of at least 1-(1/k2) (Lohninger 1999). The
factor k can be determined according to a user defined confidence level.
However, it is recommended to exercise care on outlier removal, since outliers
sometimes represent true extreme values that may be more informative than
the other values, for instance in land degradation questions. Extraneous noise
that introduces high frequent components in the data can be are removed
by Savitzky-Golay filtering (Savitzky and Golay 1964).
A widely used method to analyze climatic impacts on vegetation cover is to
compare anomalies from a climatic time-series (temperature and/or rainfall)
and the NDVI. There are several options to derive anomalies:
−
Anomalies from the total mean are differences d of each measured
values and the arithmetical mean of the respective series:
d = Xt − X
The anomalies between two series can be directly compared to each
other by calculating standardized anomalies, which are obtained by
dividing d by the respective standard deviation of the series.
−
Anomalies from seasonal mean
Similar to a., but here global means are replaced by the respective
seasonal parameters.
d = X t − X seas.
Anomalies from long-term seasonal averages can be considered as
differences of the observed quantities from the long-term seasonal
reference state. The anomalies can be standardized by dividing d by
the seasonal standard deviations. This removes seasonal components
(e.g. the annual vegetation growth cycle) from the data: The resulting
series has zero seasonal means and seasonal standard deviations are
one, respectively. The removal of seasonal effects from a record is a
precondition for significance assessment of trends, if ordinary-leastsquare regression is used as a trend model (see below). Figure 3 shows
an example how the seasonal component is removed from the series
by calculating (standardized) seasonal anomalies.
73
Remote Sensing Based Assessment
Figure 3: NDVI series (left), differences from the seasonal mean (middle) and
standardized seasonal differences (right). While in the original series the annual cycle
is clearly visible, seasonal differences are eliminated in the pre-treated series.
2.4.2 Assessment of seasonal components
Seasonal components in a time-series are undesirable elements for trenddetection, yet, they bear important information about the phenological
cycle. The most important methods to characterize the non-stationary
behavior of time-series seasonal components are discrete Fourier transform,
continuous wavelet transform (CWT), and empirical mode decomposition
(EMD).
Discrete Fourier transform (DFT), that is in practice computed by the Fast
Fourier transform (FFT), returns a two-sided spectrum F(λ) in complex form
which is scaled and converted to polar form to obtain the power spectrum or
periodogram I(λ).
real ( F (λ )) 2 + imaginary( F (λ )) 2
I (λ ) =
N
The power spectrum consists of the Fourier frequencies (Fλ) λ=0,…,N/2.
(Schlittgen and & Streitberg 1999). I(λi) represents the magnitude of a
frequency λi while the phase information is lost. Therefore, the original signal
cannot be fully reconstructed from I(λ).
74
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Figure 4: A MEDOKADS monthly NDVI series (left) and raw periodogram of the series
(right). Peaks in the periodogram correspond to the fundamental frequency (192
month), the annual cycle (12 months) and related harmonics (six and four months,
respectively) which in many cases represent no independent frequency compounds
if a cyclic phenomenon has non-sinusoidal shape.
Apart from the magnitude of a certain Fourier frequency λ the DFT as
complex measure also bears information about the phase. The phase term for
each λ is recovered from the real
and imaginary
part of F(λ) using the
relation
P (λi ) = tan −1 ⎡⎣ ℑ { F (λi )} / ℜ { F (λi )}⎤⎦
Magnitudes and phases in NDVI data are closely linked to agro-biological
phenomena, such as land cover conditions in response to seasonal pattern of
rainfall and temperatures (Andres et al., 1994). Figure 4 provides an example
where both measures were used to describe the spatial patterns of the
annual vegetation cycle in PAL-NDVI data (1982 – 1999).
To account for local-frequency information the DFT must be modified. The
windowed Fourier transform (WFT) uses a sliding rectangular window of length
T that moves along the time-series with the sampling rate of ∆t and the length
T = N∆t. For each window the FFT derives the Fourier frequencies T-1 to (2∆t)-1.
The length of the window is selected such that the basic cyclic components
of interest are represented by these frequencies. For instance, the
development of the annual cycle’s amplitude in monthly temperature data
can be assessed by selecting T as any integer multiple of 12. However, for
complex signals with several mixed frequency components the selection of a
proper window that balances time and frequency resolution is difficult or
even impossible (Stefanovska and Bracic 1999). For those cases the CWT has
been proven to be very useful in analyzing data with gradual frequency
changes (Torrence and Compo 1998). It produces a two-dimensional
75
Remote Sensing Based Assessment
scalogram for each time-series that keeps both time and frequency
information assessable.
Figure 5: CWT output from TimeStats’s interactive CWT function, using a GIMMS NDVI
series (1982-2003) from the western fringe of the Nile delta. The scale-averaged series
is related to the annual cycle (mother wavelet: Morlet). The trend map (right) derives
from a linear regression model applied to the corrected MEDOKADS and highlights
the estimated NDVI changes in the observation period.
As an example figure 5 demonstrates the CWT decomposition for a
detrended monthly GIMMS-NDVI series (1982-2003) from the western fringe of
the Nile delta. The image shows the original series (a), the scalogram
denoting the wavelet power spectrum at different scales (b), the scaleaveraged series, which is associated in this example with the wavelet power
of the annual growth cycle (c), and the global wavelet (d) that integrates the
wavelet power at each scale over time. Crosshatched regions in the
scalogram indicate the cone of influence (COI), where results are
questionable since wavelet power is lowered by edge effects (Torrence and
Compo 1998). Two striking features in the CWT results are a positive
development in the strength of the annual vegetation growth cycle outside
the COI and the onset of a semi-annual cycle (at the 0.5 year period in the
scalogram) starting in the middle of the observation period. In addition a
NDVI trend map (figure 5, right) suggests a positive development in
vegetation cover along the fringes of the delta. This points to land
reclamation activities by irrigation during the observation period.
76
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
2.4.3 Transitional components
The detection of spatial trend patterns is an important issue in long-term
environmental studies. A trend is characterized by its functional form,
direction and magnitude and should be interpreted with respect to its
statistical significance (Widmann and Schär 1997). There exist many
techniques for trend detection, including (non-) linear regression analysis and
non-parametric methods. In general, parametric tests are more powerful than
nonparametric versions, provided the conditions for parametric significance
testing are satisfied (Hirsch and Slack 1984). The distribution or density
estimation in nonparametric trend models is data-driven and requires no a
priori assumptions on the functional form. The relative robustness of
nonparametric methods against missing values and irregularly spaced values
is also higher than in case of the parametric models.
The simplest trend model is linear regression analysis:
y = β ⋅ time + a
where the coefficient β determines the strength and direction of a trend. The
linear model can easily be extended by the consideration of higher order
terms (e.g. time2) or Fourier polynomials to account for non-linear and cyclic
behaviour.
A powerful non-parametric methods is the modified seasonal Kendall (MSK-)
test, which is suited for monotonic trends detection independent from the
functional type (e.g. linear, quadratic) of the trend. In addition the MSK test
accounts for seasonality and temporal autocorrelation in the series. Thus,
there is no need to remove the annual cycle from the data to enable
significance tests. The test statistics are derived individually for each season.
The H0 assumes all observations in each season are randomly ordered,
whereas the H1 expects a monotone trend (increasing or decreasing) in at
least one season (Hirsch and Slack 1984).
2.5
Integrated interpretation concepts
Based on the time series components extracted from hypertemporal time
series, advanced interpretation approaches may be implemented if sufficient
background information and auxiliary data are available. Although these
approaches have been developed to map changes in species composition
or land use patterns, they might be adapted to characterize the long-term
result of mitigation and restoration measures across large areas.
2.5.1 Mapping functional vegetation types
In dryland ecosystems, plant communities are effective indicators of land
condition and are strongly related to land use intensities. For instance, plant
composition reflects succession patterns after disturbances (e.g. clearing,
77
Remote Sensing Based Assessment
fire), and has been shown to be linked to grazing intensities (Noy-Meir 1998;
Quinn 1986; Trabaud 1981). A typical example is savannah ecosystems, where
trees and greases interact in many ways (Jeltsch et al. 1996; Jeltsch et al.
2000; Scholes and Archer 2997), and where tree and shrub encroachment
resulting from high stocking rates is frequently causing problems in terms of
carbon and nitrogen cycles and wildfire dynamics (Gillson and Hoffman 2007;
Skarpe 1992). In accordance with information on sensor characteristics given
in section 1, different observation scales offer specific opportunities and
limitations for assessing plant communities. At the coarse scale, geometric
and spectral resolutions are limited, impeding identification of plant species.
However, the high temporal resolution provides information on phenological
cycles, which can in turn be used to map major groups of plant communities
(i.e. functional types), such as grasslands, shrublands, forests etc. Such
information is essential for large area management of grazing livestock or to
predict phenological cycles of trees and greases for ecological applications
(Archibald and Scholes 2007; Skarpe 1992). Various techniques exist to extract
the required characteristics from time series data, which include multivariate
exploratory methods such as principal components analysis (Eastman and
Fulk 1993; Lobo and Maisongrande 2006), fitting of mathematical functions
such as sigmoidal functions (Fischer 1994a, b; Zhang et al. 2003) or Gaussian
functions (Jönsson and Eklundh 2002) to derive models representing
vegetation phenological cycles, as well as methods of harmonic analysis and
wavelet analysis (Li and Kafatos 2000). In particular, harmonic analysis has
been shown by many studies to be useful in assessing vegetation dynamics
(Azzali and Menenti 2000; Moody and Johnson 2001). Yet, in cases where
subtle phenological differences exist in semi-arid rangeland vegetation, high
order polynomials may be required to completely represent all features. For
instance, (Geerken et al. 2005) employed five annual harmonics to
differentiate functional vegetation types in a grazed steppe in Syria. Since
incorporation of higher order Fourier coefficients tends to generate spurious
oscillations in the modelled time series (Chen et al. 2004), higher order terms
can be down-weighted or their amplitudes be tapered (Hermance 2007;
Hermance et al. 2007). Alternatively, vegetation discrimination may be based
on Fourier analysis which is used to quantify the similarity between the annual
cycles of an individual pixel with a reference cycles (Evans and Geerken
2006; Geerken et al. 2005).
2.5.2 Syndrome-based mapping approaches
Human population and natural renewable resources are main components of
human-environment systems, which can be affected by climatic or socioeconomic disturbances. For instance, large areas of the European
Mediterranean are believed to be affected by land degradation processes,
mainly emerging from conflicts between past and present land uses, or
between economic and ecological priorities (Brandt and Thornes 1996;
Puigdefabregas and Mendizabal 1998). Although the different definitions of
78
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
and perspectives on desertification illustrate the difficulty in identifying a
unifying concept or explanation (Geist 2005; Geist and Lambin 2004), the
concept of “syndromes”, developed in a wider global change concepts,
may serve to provide an interpretation framework for land degradation and
desertification. The basic idea behind it is the description of change
archetypical, co-evolutionary patters of human-nature interactions. Downing
and Lüdeke (2002) have supplied a first attempt at linking vulnerability
concepts to degradation and identifying relevant global change syndromes.
Time series components as described before contain information that may be
related to land degradation based on a clear, thematic definition of
syndromes relevant in a specific area. For instance, trends in vegetation cover
may be related to greening-up following land abandonment, often
accompanied by increased water uptake and fire risk, while degradation
due to overuse is characterized by negative trends. On the contrary, shifts in
peaking times and the magnitude of the vegetation signal point to changes
in land use, for instance to intensification of agriculture, often associated with
excessive water consumption, use of fertilizers etc. Hill et al. (2008) have
implemented an interpretation scheme based on the linear trend
component, phase and phase shift, and change in magnitude of time series
fluctuations to map major syndromes between 1989 to 2004 based on the
MEDOKADS archive. They used upper and lower percentiles for each factor
and developed a rule base to combine them into major groups of change
patterns for the Iberian Peninsula. For natural and semi-natural areas they
defined syndrome classes “rural exodus”, “disaster” and “overexploitation”
and mapped these on a spatially explicit level. In order to understand if
apparent changes are driven by climatic conditions (such as precipitation
variability, droughts etc.) or if they are human-induced, rainfall data need to
be integrated in a second step (Wessels 2007). As, depending on resilience
and elasticity, this influence may manifest at different times after precipitation
(or drought) events, specific models can be employed to quantify this effect
and thus differentiate between human- and climate-induced changes, e.g.
lag models (Udelhoven et al. 2009).
2.5.3 Integrated land condition assessment
Land condition is frequently assessed based on vegetation related
information, one of the reasons being the possibility to derive this information
in a spatially explicit manner from remote sensing data. In dryland
ecosystems, vegetation abundance is strongly linked to water availability,
and plants aim at maximizing the efficiency of water use through optimization
processes. Based on this relation, Boer (1999) has proposed a theoretical
framework for land degradation assessments relying on site water balance
modeling, and integrating vegetation-related temporal information with
rainfall to evapotranspiration ratios (Boer and Puigdefabregas 2003, 2005).
Central to such approaches is the concept of rain use efficiency (RUE), i.e.
the ratio of net primary productivity (NPP) to precipitation (Le Houérou 1984),
79
Remote Sensing Based Assessment
and the assumption that land degradation is resulting in decreasing RUE
values. Simplified approaches of land degradation mapping based on RUE
derived from remote sensing time series and interpolated rainfall data were
developed by Hountondjo et al. (2009) and Bai et al. (2008). Other authors
argue that in more humid areas, rainfall is not a limiting factor to NPP.
Accordingly, RUE gradients may merely reflect gradients in precipitation if
viewed across large areas (Wessels 2009). The problem of applicability of land
condition assessment across climatic zones has been addressed by the 2dRUE
concept developed by del Barrio et al. (2010). It draws from the concept that
any reduction of plant biomass and NPP below those of land which under
equivalent environmental conditions is not desertified. Major ecosystem
attributes required in the model are annual average biomass and seasonal or
inter-annual growth peaks, both of which can be derived from climatically
de-trended expressions of rain use efficiency, requiring time series of remotely
sensed vegetation estimates and continuous fields of precipitation and
temperature data. To mitigate the fact that in areas of varying climatic
conditions, the driest areas tend to exhibit largest scores because of their very
low precipitation values opposed to very high evpotranspiration estimates, a
climatic de-trending was introduced. It is based on a spatially explicit
calculation of aridity index using the expression of potential
evapotranspiration given by Hargreaves & Samani (1982). The potential of this
method has been demonstrated for the Iberian peninsula (Del Barrio et al.
2010) and for the Horqin Sandy Lands region in North Western China, which is
located in the agro-pastoral zone between the Inner Mongolian Plateau and
the Northeast Plains (Hill et al. 2010).
3 Fine scale
3.1
Data sources and time series setup
The analysis of landscape dynamics at fine scales may be based on different
sensor systems.The longest continuity of data is provided by the Landsat
mission, which was initiated in 1972 with the launch of Landsat-1 Multi Spectral
Scanner (MSS). It acquires data at a spatial resolution of 79*79 m2 and in four
spectral bands, i.e. the blue, green red bands in the visible (VIS) and in the
near-infrared (NIR) wavelength range of the electromagnetic spectrum. In
1984, the Thematic Mapper (TM) replaced the MSS system, with an
improvement in geometric, spectral and radiometric resolution (Richards and
Jia 2005). At a spatial resolution of 30*30 m2, bands are positioned in the VIS
(blue, green, red), the NIR, plus 2 bands in the mid-infrared section (MIR), thus
providing better spectral discrimination capabilities. In addition, a thermal
band is available at a 120*120 m2 spatial resolution. In 2001, the Enhanced
Thematic Mapper (ETM+) was introduced onboard Landsat-7 (Goward et al.
2001; Masek et al. 2001). It has essentially the same technical specifications as
Landsat-5 for optical bands, but a panchromatic band was added with 15*15
m2 geometric resolution, and two thermal bands (high gain and low gain)
80
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
acquire data at 60*60 m2 resolution. Data from this system have provided
excellent quality until 2003, when failure of the scan line corrector occurred,
causing line dropouts particularly around the edges of the data frame
(Markham et al. 2004). Although approaches for gap-filling have been
developed (Maxwell et al. 2007; Pringle et al. 2009; Roy et al. 2008), this has
caused a decline in the usability of Landsat-7 data, such that Landsat-5 was
switched back on and continues to deliver information. Meanwhile the
Landsat continuity mission has been approved and launch of Landsat-8 has
been scheduled for December 2012 (Wulder et al. 2008). All Landsat-systems
have a Nadir-viewing geometry with only limited scan-angles to the sides of
7.5°, which, together with the orbital characteristics, leads to revisit times of 16
days.
In addition to the Landsat data suite, a number of potential complementary
data sources exist, such as the SPOT and ASTER sensors launched in 1986 and
1999, respectively (Kramer 2002).
The SPOT-4 and -5 (since 2002) HRV systems provide information in four bands,
VIS (green, red), NIR and MIR, at 20*20 m2 (SPOT-4) and 10*10 m2 (SPOT-5),
respectively. In addition, a panchromatic band at 5*5 m2 is available
onboard SPOT-5. The SPOT system can be programmed to observe specific
locations, and high revisit frequencies can be achieved by pointing the
sensor to up to 23.5° off-nadir viewing positions. However, this acquisition
mode has significant implications on image geometries and complicates
geometric image registration (Richards and Jia 2005). SPOT data are
acquired on demand and can be bought off-the-shelve later on. Yet, there is
no regular revisit interval and acquisition depends on ordering status, such
that no global coverage comparable to the Landsat system exists.
Besides these operational and commercial (SPOT) systems with long-term
monitoring capabilities, other systems have been placed in orbit in recent
years, such as the ASTER and Rapideye systems. Aster provides information in
15 spectral bands, with varying geometric resolutions according to the
wavelength range covered (VIS/NIR: 15*15 m2, MIR: 30*30 m2, TIR: 90*90 m2).
The ASTER sensor acquires images upon previous registration of test sites and
ordering. Recorded images can be ordered off-the-shelve at low cost;
however, no complete coverage of the globe is available (Richards and Jia
2005). In 2008, the Rapideye system has been launched, providing images in
five bands (VIS Blue, Green, Red, VIS/NIR Red Edge, NIR) and a pixel size of 5*5
m2. Again, data need to be acquired for registered test sites for following
acceptance of a research proposal, or bought from the Rapideye Company
directly. Similar to other sensors, no global coverage is available.
In addition, further experimental sensors exist, such as multi-angle viewing
CHRIS-PROBA, MISR etc. While all of these systems offer potentially higher
information content due to specific acquisition techniques or improved
spectral, spatial and temporal characteristics, they do not provide global
coverage but only image registered test sites, so far only cover short periods
of time, and mission continuity is not safeguarded for any of them. This
81
Remote Sensing Based Assessment
accounts even more for experimental airborne systems, such as hyperspectral
or LiDAR systems. Despite the enhanced spectral and geometric capabilities
of these systems, they are largely scientific, such that proposals need to be
submitted to attain imagery of specific test sites or monitoring areas. Given
these considerations, monitoring or assessment approaches that need to
incorporate the spatial domain should ideally be based on routinely available
data that cover sufficient periods of time and are also collectable in a
retrospective manner, rendering the Landsat system still the prime choice. This
is particularly true given that a large amount of data is now made freely
available through the USGS6. Yet, this does not account for all of the data
covering western, central and southern Europe, which needs to be acquired
through commercial providers at higher cost. Given their availability, higher
level experimental data may still be incorporated to answer specific questions
or complement other information.
The setup of the data base needs to be based upon a sound knowledge of
the processes to be detected. In the case of assessing already implemented
restoration measures, exact knowledge of applied measures might hence
greatly facilitate this work, with implications for work load associated with
processing and analysis of data etc. Following traditional change analysis
concepts, data bases are frequently compiled of a set of multi-temporal
data, i.e. a number of data sets within on specific observation date (e.g. one
year). This ensures phenological information may be employed to map
different types of natural, semi-natural and agricultural vegetation types. If this
is repeated for different time steps, change in land use/land cover (LULC)
classes may be inferred. In recent times, this concept of reducing “change”
to alteration between broad land use categories has been questioned, since
structural changes within one stable class may be equally important in
different environmental contexts (e.g. increasing or decreasing vegetation
cover within a “grassland” land use category). To address both potential
types of change, the terms land use transformation (or conversion) and
modification were introduced (Lambin et al. 2006). While LULC conversion
can be assessed following the concept described before, modification issues
may require a more densely populated time series enabling to apply time
series analysis techniques such as those described in section 2. Although
possible in rare cases (Scarth et al. 2010), hypertemporal analysis is usually
impeded by the amount of processing involved, as well as the fact that
clouds frequently prevent establishment of an appropriate, equally spaced
time series. Therefore, time series are often set up in a way to acquire one
image per year at comparable phenological dates and allow assessing the
transient or linear trend component (Hostert et al. 2003b; Hostert et al. 2001b;
Röder et al. 2008c).
If transformation and modification processes need to be addressed for a
specific area, satellite image time series should incorporate (i) yearly or twoyearly data for a given seasonal window and (ii) multi-temporal data sets
6
www.glovis.usgs.gov
82
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
within a year that are placed in relation to time frames potentially relevant for
governing physical and socio-economic framework conditions, such as
accession to political institutions, legal events, droughts etc. In case the
required time series cannot be procured using a single sensor type, specific
intercalibration approaches might be required to ensure quantitative
consistency (compare section 3.2.2) (de Vries et al. 2007; Dinguirard and
Slater 1999; Röder et al. 2005; Teillet et al. 2007).
3.2
Standardization requirements
Satellite imagery exhibits a number of unique properties, which result from the
technical process of recording images of a part of the Earth’s surface while
this is rotating, in combination with technical characteristics of the sensor and
the satellite platform. This denies a direct comparison of images from different
dates and sensors, although this is a mandatory component of any
surveillance system.
3.2.1 Geometry
The interplay between sensor characteristics with movement of the satellite
platform and the rotation of the Earth surface results in geometric distortions
that prevent a direct integration with other spatial data bases.
As a result of the opto-mechanical image acquisition of the Landsat sensors
and the rotation of the Earth, different sources of error emerge, which are
generally accounted for by the system correction prior to delivery ([USGS]
1984; Richards and Jia 2005). In addition to these systematic or systeminherent errors, images acquired with opto-mechanical scanners show a
typical cartographic representation of the Earth’s surface, with a parallel
projection in along-track and a central projection in across-track viewing
direction. The latter significantly influences cartographic comparability with
the magnitude of distortion depending on the distance of the pixel from the
satellite nadir and the elevation of the surface element. Consequently,
especially in terrain with a strong relief gradient, it is crucial to adequately
account for this terrain-induced distortion to ensure the consistency of satellite
images with each other, as well as with other sources of spatial information
(Itten and Meyer 1993; Pala and Pons 1995; Welch et al. 1985).
Finally, combining the resulting data set with additional information layers and
producing maps requires transferring this planar space-oblique Mercator
projection into a pre-defined reference cartographic projection system
(Lillesand and Kiefer 2000). The type of reference system depends largely on
the geometric resolution of data employed, the scale to be addressed by the
analysis, and the spatial extension of target areas. For local to regional scale
applications (1:250,000 or larger) and target areas in mid-latitudes, the UTM
system is widely used. In addition, there is a range of locally adapted
spheroids and datums, such that the UTM system can be well adapted to
83
Remote Sensing Based Assessment
different areas. In areas at very high latitudes, other ortho-projected
references systems with narrow zones may be more convenient.
Depending on the data provider, images may already have been rectified to
a cartographic reference system. The USGS provides ortho-rectified Landsat
imagery (Level 1T product) at excellent quality, which has been processed
employing SRTM-based elevation information to account for non-systematic
distortions. Yet, data from other providers may still require this essential
processing step. Provided the availability of a suitable source of digital
elevation information (e.g. DEMs from the Shuttle Radar Topography Mission,
SRTM, or made available from the ASTER mission based on stereoscopic
image pairs), ground control points are required to link pixel to geographic
coordinates. The accuracy largely depends on the scale of reference maps
and whether the images provide sufficient suitable features that can be
identified with high confidence. More sophisticated methods are frequently
based on window-based cross-correlation approaches. Using statistical
correlations with search windows, they are able to exploit topographic
features, and instead of maps geo-referenced digital elevation sources may
be used as references to register one image as a master image, whereupon
other images may be co-registered (Hill and Mehl 2003).
3.2.2 Radiometry
The envisaged data processing and analysis strategy generally determines
the level of accuracy required from radiometric data pre-processing. For
simple classification approaches, radiometric processing may be negligible or
simple correction approaches may be sufficient (Song et al. 2001). However,
with the development of quantitative analysis techniques, changes
becoming apparent from multi-temporal analyses should only result from
intrinsic surface properties (Hall et al. 1991; Vogelmann et al. 2001). The signal
at the sensor is a result of solar irradiance received at the ground, which
interacts with surface elements, and is partially reflected back towards the
sensor. However, the direct analysis of multi-temporal data sets based on
spectral properties of these surface elements is impeded by radiometric
properties of satellite images, which cause them to deviate from “true”
surface reflectance values and which can be summarized under two major
components.
a) Sensor calibration: Incoming radiance is converted to a digital signal,
which can be described by a calibration function. The further radiometric
processing of satellite images requires physically meaningful values; hence it is
necessary to calibrate the digital numbers (DN) in satellite images back to
radiance values. Since the sensitivity of detectors is known to decay with
sensor age, the same incoming radiance signal will not result in the same
radiance value for different dates, and the corresponding error can be
expected to propagate through the subsequent processing and
interpretation chain (Teillet 1997; Thome et al. 1997). Calibration procedures
84
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
are further complicated when different sensor systems need to be integrated
within one multi-temporal data base.
b) Atmosphere- and topography-induced distortions: The signal recorded at
the sensor is a result of solar radiance propagating through the Earth’s
atmosphere, interacting with the surface, and being partially reflected back
through the atmosphere towards the sensor. Besides surface properties, which
are the target information required, there are a number of factors affecting
this signal, such as the distance between Earth and Sun; geometric viewing
constellation between sun, sensor and earth surface (including the slope and
orientation of the surface); and the constitution of the atmosphere, as
characterized by gaseous and aerosol components and their concentration.
(Schott 1997; Tanre et al. 1990; Vermote et al. 1997).
In order to meet the criteria for a general monitoring framework formulated
before, it is mandatory to remove these influences of topography,
illumination, shadows, viewing directions and atmospheric properties to attain
physical quantities (i.e. surface reflectance) that can then be considered to
be primary indicators for further interpretation, (Du et al. 2002; Schott 1997).
Different approaches to radiometric correction exist, with the most common
ones being empirical correction approaches, or a full modeling of processes
associated with transfer of radiation through the atmosphere.
Empirical line correction
The empirical line method is an atmospheric correction technique that
provides an alternative to radiative transfer modeling approaches. It offers a
relatively simple means of surface reflectance calibration, providing that a
series of calibration targets are available, that are invariant in time and for
which measurements are available. This technique has been applied with
variable success to both airborne data and coarser spatial resolution satellite
sensor data. Earth observation satellite data have a spatial resolution (4- 30
m) which permits to identify a number of homogeneous targets with a range
of contrasting reflectances that are suitable for the empirical calibration.
The approach is based on matching the image DNs and true reflectance of
two (or more) targets found in the respective scene. "Correction" is
accomplished by forcing the scene spectral data to match selected field
reflectance spectra by calculating a linear regression for each spectral band
to equate DN and reflectance, calculate as
ρ t = a0 + a1 DN
with ρ t being the target reflectance, a0 and a1 offset and gain coefficients,
and DN denoting the uncalibrated digital number as recorded by the sensor.
The empirical line method is an efficient method to produce reflectance
values for further interpretation at the expense of non-linear effects or
topography-induced illumination variations (de Vries et al. 2007; Karpouzli and
Malthus 2003).
85
Remote Sensing Based Assessment
Radiative transfer correction
Available radiative transfer codes make extensive use of analytical
expressions and pre-selected atmospheric models to correct for atmospheric
absorption, scattering and pixel adjacency effects, where diffusion and
absorption processes are assumed to be independent. Upward and
downward transmission coefficients are derived by introducing the auxiliary
quantity of optical thickness τ which measures the total extinction of a light
beam due to molecular and aerosol scattering when passing through an
airmass. Multiple scattering is accounted for, and the absorbing atmospheric
gases (H2O, O3, CO2, O2) are assumed to condense at the top of the
atmosphere and at the top of the layer between the earth surface and the
sensor altitude. In the absence of measurements of atmospheric properties,
the scattering optical depth may also be estimated from scene data itself (Hill
and Sturm 1991; Teillet and Fedosejevs 1995). The major problem in retrieving
reflectance factors from operational satellites lies in the sensitivity to the
absolute radiance calibration of the sensors. In-flight calibrations for LandsatTM and SPOT data are conducted at various high-reflectance sites which is
useful for determining at least an absolute radiometric calibration gain and
monitoring the sensor degradation with time (Chander et al. 2004; Chander
and Markham 2003; Helder et al. 2008; Teillet et al. 2001; Thome et al. 1997).
However, it must be also noted that the successful use of radiative transfer
calculations for high spectral resolution imagers depends on the availability of
key parameters that adequately characterize atmospheric conditions during
the flight (i.e. absorption optical depth, in particular that of water vapor).
Figure 6 illustrates the radiance paths that need to be accounted for in the
context of radiative transfer modeling, although the influence of topography
is not considered yet.
Figure 6: Simplified
representation of the
radiative transfer through
atmospheres to be
considered for correcting
satellite measurements in the
reflective domain.
86
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Rather than being expressed as radiance, the various quantities of the image
forming equation are commonly described in terms of equivalent at-satellite
reflectance which is defined as
ρ* =
µ0 is the cosine of the solar zenith angle θ0; E0 denotes the solar
irradiance at the top of the atmosphere [mW/cm2/µm], and L is
the target radiance registered at the satellite, [mW/cm2/sr/µm].
The correction coefficient d accounts for the daily variation in the
sun-to-earth distance and is obtained from d = 1/au2, where au
denotes the average distance from sun to earth (149.6 x 106 km) in
astronomical units (Markham and Barker, 1986).
π ⋅L
E0 ⋅ µ 0 ⋅ d
The signal ρ* in reflective bandpasses is directly related to the target
reflectance factor ρt which is affected from two major effects: absorption by
atmospheric gases (mainly water vapor and ozone) and scattering by
aerosols and molecules; emission at the normal atmospheric temperature is
very small and can be ignored. For a uniform Lambertian surface these
effects can be conveniently expressed as a series of radiation interactions in
the ground-atmosphere system (Tanre et al. 1990; Vermote et al. 1997):
⎧
ρ * =t gas ↓↑⋅⎨ρ at +
⎩
T ↓⋅[t d ↑⋅ρ t +t s ↑⋅〈 ρ 〉 ]⎫
⎬
1− 〈 ρ 〉⋅s
⎭
In this equation, ρt denotes the target reflectance
factor, <ρ> the background contribution to the
apparent reflectance of the target, and s is the
spherical albedo; ρat is the intrinsic atmospheric signal
component, T↓ is the total downward, td↑ the direct
and ts↑ the scattered (i.e., diffuse) upward
denotes
the
combined
transmittance;
tgas
transmittance of the atmospheric gases.
The target reflectance ρt depends on the photometric properties of the
surface and the distribution of the illumination which is a complex
combination of attenuated solar irradiance arriving at the ground Eg, the
diffuse sky light Es and the reflected ground radiance from the target's
surroundings. Upward and downward transmission coefficients are derived by
introducing the auxiliary quantity of optical thickness τ which measures the
total extinction of a vertical light beam passing through an air mass. Optical
depth largely depends on density and size distribution of particles (molecules
and aerosols) in the atmosphere and on their scattering and absorption
properties, which need to be taken into account in the modeling process
(Aranuvachapun 1985).
In mountainous terrain, further improvements of the radiative transfer
approach may be attained by considering topography-induced illumination
variations, requiring the calculation of the angle between the normal to the
terrain surface and incident irradiation (cos λ), a shade mask, a visible sky
mask, and the cosine of the slope. Integrating these parameters with the
properties derived from the atmosphere model allows separate treatise of
direct and diffuse irradiance components (Hay and McKay 1985; Hill et al.
1995).
Various approaches exist for radiometric correction of Landsat-type imagery
based on the 5S or 6S transfer codes. Hill & Mehl (2003) have proposed a
87
Remote Sensing Based Assessment
geometric and radiometric processing chain that was successfully
implemented in the context of different long-term monitoring studies (Hostert
et al. 2003a; Röder et al. 2008a; Röder et al. 2008c). It is based on the 5S
transfer code (Tanré 1990), while atmospheric transmission is characterized
using the Modtran code (Berk et al. 1999) along with aerosol phase functions
derived from Aranuvachapun (1985). Using the anisotropy index by Hay and
McKay (1985), diffuse and direct irradiance terms are separately treated and
differentiated into anisotropic and isotropic components to adequately
represent global irradiation and attenuation by shading effects. Commonly,
such models can be parameterized based on knowledge of atmospheric
water vapor concentrations and optical thickness being represented by the
Angstrøm coefficients (Angstrom 1964). If there are not atmospheric sounding
data concurrent with satellite overpasses, such information can be derived
from data bases such as AERONET (Holben et al. 1998; Sinyuk et al. 2007;
Smirnov et al. 2000), based on dark targets such as clear water bodies, forests
etc. (Teillet and Fedosejevs 1995), or be estimated using other sensors as
references (Hill et al. 2010; Roy et al. 2008). Although such approaches are
only an approximation of actual atmospheric conditions at the time of
overpass, they have been shown to produce consistent time series supporting
retrospective analyses.
3.3
Derivation of vegetation-related information
There are a number of indices that may be employed to characterize the
state and ground cover of vegetation, and which may be related to other
parameters such as LAI, biomass etc.
In this context, numerous studies have addressed the issue of sparse
vegetation mapping. Traditionally, the derivation of vegetation-related
information has been based on the calculation of simple spectral indices
such as the Simple Vegetation Index (VI) or the Normalized Difference
Vegetation Index (NDVI, compare section 2). While their application is simple
and efficient, they suffer from a number of drawbacks, such as the influence
of soil background especially for bright soils and sparse vegetation canopies,
or saturation effects for very dense vegetation layers (Baret and Guyot 1991;
Huete et al. 1985; Verstraete 1994b). Some authors suggested introducing a
soil line as a reference, resulting in the Soil Adjusted Vegetation Index (SAVI),
or the Perpendicular Soil Adjusted Vegetation Index (PSAVI) (Pinty and
Verstraete 1992). These helped resolve some of the ambiguities, but for
variable soils and lithological backgrounds the concept of a universal soil line
cannot be maintained (Gilabert et al. 2002; Huete 1988; Rondeaux 1995;
Verstraete 1994a). Other approaches include the Global Environmental
Monitoring Index (GEMI) (Pinty and Verstraete 1992) or transformations of the
feature space, such as the Tasseled Cap transformation (Crist and Kauth
1986; Kauth and Thomas 1976). Another problem common to these indices is
the comparability of index values derived from remote sensing systems with
different spectral characteristics. Nonetheless, vegetation indices are still
88
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
widely used in manifold applications and are still being further developed
(Gilabert et al. 2002; Turner et al. 1999), in particular in relation to
hyperspectral data (Schlerf et al. 2005).
However, all of these indices remain confined to the pixel as the reference
spatial entity, which poses different problems, in particular when
heterogeneous environments are regarded. In these cases, the average size
of homogeneous surface features frequently lies below the pixel size, such
that the recorded values have to be considered mixtures of different
materials (Cracknell 1998; Fisher 1997). Consequently, the pixel signal can be
assumed to be a combination of the reflectance of a limited number of
surface features. Provided the availability of information on these reference
materials, the signal can be decomposed into proportions of these materials,
which is known as ‘Spectral Mixture Analysis’ (SMA) (Adams et al. 1986; Smith
et al. 1990; Smith et al. 1994). Elmore et al. (1993) have demonstrated the
superiority of vegetation cover derived using SMA over NDVI for sparsely
vegetated areas, and various authors have successfully applied the
methodology to infer information on vegetation or soil cover in a quantitative
way in different environmental settings (GarciaHaro et al. 1996; Hostert et al.
2003a; Roberts et al. 1993; Smith et al. 1994). An extension of the approach is
represented by multi-date unmixing approaches (Kuemmerle et al. 2006;
Rogan et al. 2002; Shoshany and Svoray 2002) or the establishment of multiple
endmember setups (García-Haro et al. 2005; Okin et al. 1998). In many cases,
it is aimed at relating information on vegetation cover to other structural
parameters, such as LAI, above ground biomass etc. (Calvao and Palmeirim
2004; Fang et al. 2005; Lacaze 1996, 2005).
Beside vegetation, soil-related parameters have been successfully derived
from remote sensing data using a variety of approaches. For instance, Hill et
al. (1995) used SMA to relate soils and bedrock fractional cover to
degradation status for a test site in Southern France. Hill & Schütt (1993) used
an empirical function derived from soil samples and hyperspectral
measurements to infer a linear relation between reflectance and soil organic
matter, which was upscaled to Landsat-TM imagery using a forwardbackward spectral unmixing approach to restore soil-related information.
Spectral mixture analysis (SMA) is based on the assumption that the majority
of spectral variation encountered in multispectral imagery can be ascribed to
a mixture of a limited number of surface materials, which can be
characterized by reflectance spectra (Smith et al., 1990). Most frequently,
these mixtures occur at the sub-pixel level, resulting in mixed-pixel spectra.
SMA attempts to model the multispectral reflectance recorded in a pixel as a
mixture of representative ‘prototype’ spectra, which are termed ‘spectral
endmembers’
(EM).
The
approach
computationally
decomposes
multispectral measurements into a finite number of pure spectral components
by describing the recorded signal as a result of linear spectral mixing (Adams
et al. 1986; Smith et al. 1990) and yields the relative abundance of the
different endmembers in a pixel according to
89
Remote Sensing Based Assessment
n
Ri = ∑ F j • REij + ε i and
j =1
with
n
∑F
j =1
j
=1
Ri = reflectance of the mixed spectrum in band i
RE ij = reflectance of the endmember spectrum j in band i
F j = fraction of endmember j
n=
εi =
number of spectral endmembers
residual error in band i
This equation is solved through a least squares regression, while constraining
the sum of fractions to one. A unique solution is possible as long as the
number of spectral components (endmembers) does not exceed the number
of bands plus one. However, it has been emphasized that the intrinsic
dimensionality of the data is lower than the number of bands due to
autocorrelation between bands (Settle and Drake 1993; Small 2004). For
instance, Small (2004) gives a maximum of 4 EM for Landsat ETM+ data;
however, this number has also been shown to depend on the spectral
contrast of the selected endmembers in relation to the sensor sensitivity and
the spectral variance within the image (Drake and Settle 1989).
Composition of the endmember sets is the crucial step in spectral unmixing.
These may be derived from the image itself (Boardman et al. 1995; GarciaHaro et al. 1999), or - given limitations in finding homogeneous pixels in
moderate resolution image data (Bateson and Curtiss 1996) – from
hyperspectral reflectance measurements of different surface materials (Hill et
al. 1995; Lacaze 1996). In any case, establishment of an endmember set that
encompasses the spectral dimensionality of the data set is a complicated
task that largely determines the quality of results (Chang et al. 2006; Dennison
et al. 2004; Dennison and Roberts 2003; Franke et al. 2009; García-Haro et al.
2005; Numata et al. 2007; Rogge et al. 2006).
In addition, an artificial shade endmember may be included to account for
albedo differences between different pixels and to model shading effects
(Adams et al. 1986; Roberts et al. 1993). Once a suitable endmember model is
determined and the abundance estimates for the different components is
calculated, a normalization of the results with respect to the shade
component can be performed by proportionally assigning it to the other
fractions. The required normalization is calculated according to:
f =
with
1
and
(1 − FShade )
( n −1)
∑F
j =1
j
• f =1
f = shade normalisation factor
FShade = fraction estimate for the shade endmember
F j = fraction estimate for endmember j
90
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Fractions resulting from the unmixing process should range from 0 to 1. In
addition, the root-mean squared error may be employed as a measure of the
mathematical accuracy of the SMA model.
⎡
⎢
∑
k =1 ⎢
⎣
RMSE =
m
with
2
⎛ n ⎛
⎞ ⎤
⎜ ∑ ⎜ R jk − R jk ′ ⎞⎟ ⎟ / n ⎥
⎜ j =1 ⎝
⎠ ⎟⎠ ⎥
⎝
⎦
m
R jk = modelled reflectance in band j and pixel k
′
R jk = measured reflectance in band j and pixel k
n = number of spectral bands
m = number of pixels
The RMSE provides a measure of the spectral variability explained by the
model, and its calculation per pixel yields an RMSE image that allows a visual
assessment which surface materials may not have been correctly presented
in the model.
Once a mathematically sound result has been attained, the next crucial step
in any vegetation-related analysis is to validate results. Here, significant
scaling issues are major obstacles to direct comparison between groundbased and satellite-based results, since the way of acquiring data is often
entirely different from a conceptual point of view (Brogaard and Olafsdottir
1997; Röder et al. 2008a). However, given ground-sampling schemes are set
up to be consistent with remote sensing image acquisition, very good results
have been demonstrated (Armston et al. 2008). As an alternative some
authors employ high resolution imagery or aerial photography as an
intermediate step between field measurements and moderate scale image
data such as Landsat-TM/ETM+ (Hill et al. 2009; Kuemmerle et al. 2006).
Given the availability of hyperspectral imagery, enhanced vegetationrelated information may be derived through empirical relations to groundbased observation, e.g. using logistic regression models, support vector
regression models etc. Such approaches have been employed to map forest
biophysical parameters (Schlerf et al. 2005), nutritional content (Asner and
Martin 2009; Skidmore et al. 2010), or map native vs. invasive plant species
(Asner et al. 2008a; Asner et al. 2008b).
A more complex approach to deriving vegetation related biophysical
parameters at leaf-, tree- or canopy level is the application of reflectance
models that model the reflectance signal as a function of these parameters,
and utilizes model inversion or lookup strategies to conclude on target
variables such as LAI, water content, leaf angle etc. (Kimes et al. 2000). To run
such models at canopy level, models need to integrate leaf-, tree-, and
geometric components to account for geometric properties of canopies. A
proven combination is to use the PROSPECT model to account for leaf
reflectance as a function of biophysical properties (chlorophyll, water
concentration, leaf structure etc.) (Jacquemoud and Baret 1990) with a
model to represent the full plant reflectance taking into account
91
Remote Sensing Based Assessment
background, leaf angle, woody plant components etc. (e.g. SAIL)
(Jacquemoud 1995; Verhoef 1984), and link this to a turbid medium model for
the geometric configuration of the landscape in terms of trees/shrubs,
interspersed areas with no vegetation or a grass understorey, and
shading/illumination effects. One frequently employed model is the GeoSail
model (Huemmrich 2001), but models and algorithms are continuously being
further developed (Verhoef and Bach 2007) to accommodate new sensor
systems, such as multi-angular systems etc. (Chopping et al. 2008). Alternative
approaches are ray tracing models that represent the path of each individual
photon within vegetation canopies (Govaerts and Verstraete 1998), which is
becoming more attractive through the use of ground-based LiDAR systems
(Loudermilk et al. 2009; Strahler et al. 2008). Yet, such model-based
approaches are often very demanding in terms of input data (e.g.
hyperspectral data) , and were found to be limited in complex, semi-arid
environments (Hill et al. 2009; Röder et al. 2008b). Therefore, their application
to such environments is a research issue that would need to be specifically
addressed.
3.4
Fine-scale temporal analysis
There are different ways in which to assess temporal dynamics of vegetationrelated information as described in section 3.3. One is to assess temporal
trajectories for different positions based on the respective pixel values for
different dates. This has frequently been done to assess post-fire dynamics
(García-Haro et al. 2001; Riano et al. 2002; Viedma et al. 1997) or the impact
of fire severity (Díaz-Delgado et al. 2003), for instance by comparing pre- and
post-fire values and calculating the difference between recovered and prefire vegetation states. With the availability of long-term records with sufficient
points in time, full trend analysis is possible and was successfully employed to
map post-fire recovery rates (Röder et al. 2008a) or the impact of grazing
schemes (Hostert et al. 2003b; Röder et al. 2008c; Scarth et al. 2010). Also,
evenly spaced time series allow for the identification of singular disturbance
events, e.g. fires, clear-cuts etc. by threshholding (Röder et al. 2008a) or by
moving temporal window time series analysis (Verbesselt et al. 2010).
Fine-scale linear trend analysis of remotely sensed images is partially
congruent with the approaches presented in section 2 for coarse-scale
imagery. Therefore, principles and formula presented above are not
repeated here. However, and beside rare exceptions (Scarth et al. 2010),
limitations of data availability and processing efforts lead to fine scale time
series can usually not be compiled in a manner to support analysis of
frequency components. Rather, characterization of trends in rangelands that
might indicate grazing induced degradation of resources has to be based on
assessing the transient component by means of a linear trend analysis.
Hence, temporal analysis of fine scale imagery is confined to linear trend
analysis of the target indicator by minimizing the sum of least squares
92
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
between given and estimated values for each fix point in the series
(Udelhoven 2010). The offset of the resulting function characterizes the level of
vegetation cover at the start of the observation period, while the gain
describes the direction and magnitude of the development. The statistical
significance of the resulting function is assessed using a t-test, while the cyclic
and the white noise component have to be considered sources of ambiguity
in the resulting function. In order to reduce this ambiguity to a minimum,
careful procurement of the data base is required. Ideally, it should be
composed of one image per year, and image acquisition dates should be
selected to represent comparable states of phenological cycles, such as
maximum photosynthetic activity, or late summer conditions after
senescence of herbaceous vegetation.
Figure 7 illustrates possible temporal profiles for a series of Landsat-TM and –
EMT+ images, and indicates the corresponding statistical estimates (Röder
2005). While profiles a and b show significant negative and positive trends,
profiles c and d are illustrative of stable situations, where no major change in
vegetation cover can be noted except for minor fluctuations. Although these
do not pertain to ‘trends’ in the true sense of the meaning, there is a high
relevance to such profiles as they are indicative of stable ecological
situations, in particular in grazing areas. Hence, the root-mean squared is
another important factor of multi-temporal statistics, as it is independent of
the magnitude and direction of change, such that stable conditions may be
identified through low error (RMSE) estimates.
1
Vegetation cover
(a)
Gain = -9.697E-005
RMSE/mean = 0.163
R2 = 0.8123
sig.
0
Gain = 8.0134E-005
RMSE/mean = 0.0833
R2 = 0.8567
sig.
0.5
1984
1985
1998
1999
2000
1993
1994
1995
1996
1997
1988
1989
1990
1991
1992
1984
1985
1986
1987
0
Time
Time
Gain = 1.416E-006
RMSE/mean = 0.0297
R2 = 0.0108
not sig.
(c)
0
(d)
0.5
Gain = -3.3076E-006
RMSE/mean = 0.2026
R2 = 0.0105
not sig.
Time
1998
1999
2000
1993
1994
1995
1996
1997
1998
1999
2000
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
0
1984
1985
1986
1987
0.5
1
Vegetation cover
Vegetation cover
1
1988
1989
1990
1991
1992
0.5
(b)
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
Vegetation cover
1
Time
Figure 7: Sample temporal profiles of green vegetation cover in the Lagadas area, illustrating
negative (a), positive (b) and indifferent (c, d) temporal development as a result of different
factors and how these ‘translate’ into statistical parameters; more specific explanations in the
text.
93
Remote Sensing Based Assessment
3.5
Land use / land cover change analysis
Trend analysis of vegetation-related parameters is a viable means of
characterizing within-class changes, for instance caused by grazing of
shrubland and grassland areas. However, understanding the often complex
and interactive process regimes at the landscape level requires
complementing such trends by information on changes in surface types and
relating these to underlying socio-economic and political causes where
possible (Geist and Lambin 2004; Stafford-Smith and Reynolds 2002; Turner II et
al. 2007). Remote sensing data offer opportunities to provide spatially explicit
information on processes of land change. On a fine scale, there are different
strategies that can be employed (Coppin et al. 2004b; Lu et al. 2004), and
which can principally be assigned to two categories: (i) a so-called postclassification can be performed, where the satellite images are classified first
and next a change detection analysis is conducted based on the thematic
classes derived before, (ii) the change detection analysis can be performed
by spectral images itself since land use/land cover changes involve
alterations of the spectral properties of a surface, e.g. by multi-temporal
Principal Component Analysis (García-Haro et al. 2001), Change Vector
Analysis (Johnson and Kasischke 1998) or image differencing (Rogan and Yool
2001). Contrary to a post-classification change analysis, the thematic
assignment of these changes has to be done afterwards. This may be
complicated in the absence of precise knowledge of the changes to be
expected, and requires the availability of images covering comparable
dates. In addition, land cover maps of different periods can serve as useful
sources of information for a variety of assessment approaches, such that the
focus of this overview will be on classifying individual data sets.
For land use classifications based on Landsat data, it is important to
characterize individual time slices using a sequence of images, thus
incorporating phenology information relevant to differentiate different types
of (semi-)natural vegetation and crops (Dymond et al. 2002; Schriever and
Congalton 1995). Again, two major options for classification may be
considered: supervised and unsupervised classification. Supervised
classification methods are effective in identifying complex land cover classes
if detailed a-priori knowledge of the study area and good training data exist,
for instance derived from field surveys or aerial photographs (Cihlar et al.
1998). Then, different classification approaches can be employed, such as
minimum distance and parallel-epiped classifiers, artificial neural networks or
maximum likelihood classifiers (Benediktsson et al. 1990; Richards and Jia 2005;
Smits et al. 1999). Recently, support vector machine (SVM) classifiers have
been introduced, which discriminate classes by fitting separating hyperplanes
in the feature space based on training samples (Foody and Mathur 2006;
Huang et al. 2002; Sanchez-Hernandez et al. 2007). SVMs are trained in an
iterative and controlled manner, and represent an excellent alternative
where land cover classes are spectrally similar or where only small training
data sets are available. Where only little reference information is available,
unsupervised classification approaches may be preferable (Bauer and
94
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Coppin 1994; Lark 1995) and they have been rated more robust and
repeatable (Cihlar et al. 1998; Wulder et al. 2004). Unsupervised classification
is based on a statistical clustering approach to map pixels to spectral classes
without prior information on the number and types of classes. Using a
minimum distance approach, each pixel is compared to a number of clusters
in the spectral feature space, which are characterized by their class statistics.
In an automated, iterative way classes are rearranged until maximum
separability is achieved, and the user may interact by assigning thresholds
and termination criteria. Further approaches to classification include objectoriented classification or rule-based, decision tree classification of different
input layers (Blaschke et al. 2004; Friedl and Brodley 1997; San Miguel-Ayanz
and Biging 1996; Shandley et al. 1996). Some authors also suggest combining
the strengths of different classifiers by using decision fusion methods
(Udelhoven et al. 2010) or by applying random forest classification based on
various runs with different subsamples from the training sample, e.g. through
bagging and boosting methods (Breiman 2001; Gislason et al. 2006; Lawrence
et al. 2006; Pal 2007).
In order to exploit the strengths of the individual approaches in the best
possible manner, hybrid approaches have been suggested that combine
supervised and unsupervised classification strategies. Three strategies can be
distinguished: first, unsupervised clustering is useful to stratify input images prior
to subsequent supervised classifications (Hill et al. 2004; Tommervik et al. 2003);
second, unsupervised methods can reveal spectrally homogenous areas for
optimized training sample and ground truth collection (McCaffrey and
Franklin 1993; Rees and Williams 1997); and third, manually collected training
data can be clustered into spectrally homogeneous sub-classes for use in a
subsequent supervised classification, for instance using Support Vector
Machine classifiers (Bauer and Coppin 1994; Kuemmerle et al. 2009; Pal and
Mather 2005; Stuckens et al. 2000).
Once classification of land use and land cover at discrete dates has been
achieved, change may be characterized qualitatively in the form of a fromto matrix or through a spatial representation of major change classes (Coppin
et al. 2004b). In any case, exact geometric co-registration of images is a
prerequisite; also, it may prove necessary to reduce the number of identified
classes to a limited set of target classes, thus facilitating a better spatial
discrimination of change patterns.
A particular challenge is often encountered where large areas need to be
classified, but reference information is only available for a limited area, One
option to overcome such limitations was presented by Knorn et al. (2009),
who propose a chain classification approach to propagate training data
between adjacent images, thus enabling multi-scene classification across
image tiles and seasonal differences.
Regardless of the methodology employed, it is essential to provide accuracy
measures in order to enable users to evaluate the accuracy of results
attained (Congalton and Green 1999; Foody 2004). Different sources of
95
Remote Sensing Based Assessment
information can serve validation purposes, such as data collected in the field,
aerial photography; recently, Google Earth ™ images have presented an
alternative in the absence of other credible data sources (Knorn et al. 2009;
Kuemmerle et al. 2009)
3.6
Integrated interpretation concepts
Once information on land use and cover and on vegetation-related
attributes has been procured, it can be assimilated into assessment models in
different manners. It is beyond the scope of this document to review all of
these; therefore, some studies are briefly referenced for illustration purposes. In
particular, any such approach needs to be tailored to the problem
addressed.
Owing to their spatially explicit nature, remote sensing data serve as prime
input for various data interpretation concepts.
Categorical maps derived from land use land cover classification can be
employed to quantify landscape spatial patterns over time using aggregated
and non-aggregated spatial metrics (McGarigal 2002; Ostapowicz et al. 2008;
Soille and Vogt 2009; Vogt et al. 2009; Vogt et al. 2007). Such information may
be linked to a variety of questions, i.e. mapping of habitats (Kuemmerle et al.
2010), assessing landscape connectivity in terms of fire propagation
(Rodriguez Gonzalez et al. 2008), monitor protected areas (Townsend et al.
2009) etc.
Land use change models of varying nature, e.g. based on cellular automata
or dynamic agent-based simulations etc. A review of methods and
approaches is provided by Verburg et al. (2004) and Koomen et al. (2007).
While remote sensing derived data are able to reveal different types of land
use change, it is equally important to understand why stakeholders take their
decisions on how to use resources, either to inform land management or
policy making, or to employ such data in model based approaches.
Therefore, integrating spatially explicit data with socio-economic information
is an important interpretation issue. Various authors stress the importance of
investigating coupled human-environment systems (Lambin et al. 2006;
Rindfuss et al. 2004a; Rindfuss et al. 2004b) and advocate the integration of
physical and socio-economic data at a spatially explicit level. Lorent et al.
(2008) surveyed the income situation of individual households dwelling on
livestock or agricultural production and related this information to land use
change and vegetation trend information derived from remote sensing and
GIS analyses. They could show the dependence of land use decisions on
subsidies in a marginal area in Northern Greece, which is essential to establish
enhanced land management and ultimately policy schemes.
In the context of grazing management, the integration of remotely sensed
data with auxiliary data has been shown to be an efficient tool for modeling
livestock dynamics and suggesting management interventions. Different
studies have used time series of Landsat data to estimate the long-term
96
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
impact of grazing based on an integrated evaluation of different time series
parameters, e.g. by calculating degradation indices (Hostert et al. 2003b;
Hostert et al. 2001a; Röder et al. 2008c). Harris & Asner (2003) et al. employed
airborne spectrometry data to map grazing impact, while Röder et al. (2007)
have adapted the grazing gradient approach introduced by Pickup &
Chewings (1994) using cost surface modeling and identified spatially explicit
gradients in grazing pressure in a heterogeneous Mediterranean rangeland in
Northern Greece. Svoray et al. (2009) have coupled a grassland production
model with climate and remote sensing data to model forage availability,
and relate this to livestock grazing behavior acquired based on GPS collars.
One frequently used option is to calibrate and validate models using remote
sensing based information from different time steps, i.e. by providing input
data for early time periods, and employing remote sensing data acquired at
a later stage for validation purposes. Duguy et al. (2007) have employed fuel
maps derived from a time series of satellite data to parameterize a fire
propagation model and model a particular fire event. Once the model had
been established, they simulated the effect of different fuel treatments and
suggested landscape interventions for the prevention of fire spread.
As stated in section 2.5, land condition is frequently assessed based on
vegetation related information, since this may be efficiently derived using
remote sensing data. Opposed to hypertemporal analyses, however,
selection of appropriate dates is crucial in representing yearly phenological
cycles. Following ecological optimization theories, vegetation abundance in
dryland ecosystems is strongly linked to water availability. Boer (1999) has
proposed a theoretical framework for land degradation assessments relying
on site water balance modeling, and integrating vegetation-related
temporal information with rainfall to evapotranspiration ratios (Boer and
Puigdefabregas 2003, 2005). Central to such approaches is the concept of
rain use efficiency (RUE), i.e. the ratio of net primary productivity to
precipitation (Le Houérou 1984), and the assumption that land degradation is
resulting in decreasing RUE values. Integrating different geospatial data
layers, he employed Specht’s evaporative coefficient to predict vegetation
index values under different local conditions, and used the difference
between the predicted and actual values to map land degradation based
on undegraded reference sites (Boer 1999).
4 Plot scale
As shown in figure 1, different sensors exist that allow addressing specific
question at the plot scale with a very high geometric resolution (VHR), such as
the Quickbird-, Ikonos- or WorldView-systems, with geometric resolution of
0.5*0.5 to 2.4*2.4 m2. Alternatively, aerial imagery may be utilized. The high
spatial resolving capacity of these systems is commonly achieved at the
expense of spectral resolution, such that these systems commonly are
characterized by two to three bands in the visible portion of the wavelength
97
Remote Sensing Based Assessment
spectrum, and one band in the near infra-red. In principle, this allows for
application of all of the concepts presented in section 3. Yet, high acquisition
costs most frequently impede acquisition of extensive VHR data sets. This is
added to by high demands in processing capacities imposed by large data
volumes and by the often complicated image geometry resulting from the
combination of very high resolution and side-looking viewing properties. A
review of image processing concepts for VHR imagery may be found in Toutin
et al. (Toutin 2004) with a focus on satellite based VHR imagery, or Mikhail et
al. (2001) for general photogrammetric principles and approaches, e.g.
bundle block adjustment of stereo image pairs etc.
VHR image data are an excellent source of information at a high level of
spatial detail. They are frequently employed for visual interpretation and to
map specific surface features. Automatic and semi-automatic approaches
include for instance visual delineation of surface features, threshold-based
mapping of target features, or simple classification approaches.
Visual mapping is commonly employed to quickly produce emergencyrelated maps for disaster relief (rapid mapping, e.g. www.zki.dlr.de).
A combination of thresholds applied to the panchromatic band and to
vegetation index values derived from the multi-spectral bands was employed
to map woody cover in a rangeland ecosystem in Northern Greece (Hill et al.
2009; Röder et al. 2008b) and separate tree and grass vegetation in an
African Savannah ecosystem (Archibald and Scholes 2007). Yet, some authors
have successfully employed maximum likelihood classifiers to map invasive
species (Laba et al. 2008) or mangrove stands (Wang et al. 2004),
characterize biomass volumes (Leboeuf et al. 2007), or inventory forest
resources (Song et al. 2010). In general, VHR data have been shown to be
consistent with medium resolution imagery (e.g. Landsat, Spot) (Goward et al.
2007; Soudani et al. 2006a; Soudani et al. 2006b) They are therefore an
excellent source of information to complement data covering larger areas,
and may be used for calibration/validation purposes or to provide for
intermediate explanation levels when upscaling from ground-based
information to higher aggregation levels (Kuemmerle et al. 2006).
5 References
Adams, J.B., Smith, M.O., & Johnson, P.E. (1986). Spectral mixture modelling: a new analysis of
rock and soil types at Viking Lander. Journal of Geophysical Research, 91, 8113-8125
Angstrom, A. (1964). The parameters of atmospheric turbidity. Tellus, 16, 64-75
Aranuvachapun, S. (1985). Satellite remote sensing of atmospherical optical depth spectrum.
In E.M. Rollin, M.D. Steven & P.M. Mather (Eds.), Atmospheric Corrections for Remote
Sensing. Proceedings of a Remote Sensing Workshop on Atmospheric Corrections,
22nd May 1985. Nottingham, UK: University of Nottingham
Archibald, S., & Scholes, R.J. (2007). Leaf green-up in a semi-arid African savanna - separating
tree and grass response to environmental cues. Journal of Vegetation Science, 18,
583-594
98
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Armston, J.D., Denham, R.J., Danaher, T.J., Scarth, P.F., & Moffiet, T.N. (2008). Prediction and
validation of foliage projective cover from Landsat-5 TM and Landsat-7 ETM+ imagery.
In, 14th Australasian Remote Sensing and Photogrammetry Conference (pp. 1-28).
Darwin, Northern Territory, Australia
Asner, G.P., Jones, M.I., Martin, R.E., Knapp, D.E., & Hughes, R.F. (2008a). Remote sensing of
native and invasive species in Hawaiian forests. Remote Sensing of Environment, 112,
1912-1926
Asner, G.P., Knapp, D.E., Kennedy-Bowdoin, T., Jones, M.O., Martin, R.E., Boardman, J., &
Hughes, R.F. (2008b). Invasive species detection in hawaiian raninforests using
airborne imaging spectroscopy and LiDAR. Remote Sensing of Environment, 112, 19421955
Asner, G.P., & Martin, R.E. (2009). Airborne spectranomics: mapping canopy chemical and
taxonomic diversity in tropical forests. Frontiers in Ecology and the Environment, 7, 269276
Azzali, S., & Menenti, M. (2000). Mapping vegetation-soil-climate complexes in southern Africa
using temporal Fourier analysis of NOAA-AVHAA NDVI data. International Journal of
Remote Sensing, 21, 973-996
Bai, Z.G., Dent, L.G., Olsson, L., & Schaepman, M.R. (2008). Proxy global assessment of land
degradation. Soil Use and Management, 24, 223-234
Baret, F., & Guyot, G. (1991). Potentials and limits of vegetation indices for LAI and APAR
assessment. Remote Sensing of Environment, 35, 161-173
Bartholome, E., & Belward, A. (2005). GLC2000: A new approach to global land cover
mapping from earth observation data. International Journal of Remote Sensing, 26,
1959-1977
Bastin, G.N., Pickup, G., & Pearce, G. (1995). Utility of AVHRR data for land degradation
assessment: a case study. International Journal of Remote Sensing, 16, 651-672
Bateson, A., & Curtiss, B. (1996). A method for manual endmember selection and spectral
unmixing. Remote Sensing of Environment, 55, 229-243
Bauer, M.E., & Coppin, P.R. (1994). Processing of Multitemporal Landsat Tm Imagery to
Optimize Extraction of Forest Cover Change Features. IEEE Transactions on
Geoscience and Remote Sensing, 32, 918-927
Benediktsson, J.A., Swain, P.H., & Ersoy, O.K. (1990). Neural Network Approaches Versus
Statistical-Methods in Classification of Multisource Remote-Sensing Data. IEEE
Transactions on Geoscience and Remote Sensing, 28, 540-552
Berk, A., Anderson, G.P., Acharya, P.K., Chetwynd, J.H., Bernstein, L.S., Shettle, E.P., Matthew,
M.W., & Adler-Golden, S.M. (1999). MODTRAN4 user’s manual. In. Hanscom AFB, MA
01731-3010: Air Force Research Laboratory, Space Vehicles Directorate, Air Force
Material Command
Blaschke, T., Burnett, C., & Pekkarinen, A. (2004). Image Segmentation Methods for Objectbased Analysis and Classification. In S.M.d. Jong & F.D.v.d. Meer (Eds.), Remote
Sensing Image Analysis. Including the Spatial Domain. (pp. 211-236): Kluwer Academic
Publishers
Boardman, J.W., Kruse, F.A., & Green, R.O. (1995). Mapping target signatures via partial
unmixing of AVIRIS data. In, 5th annual JPL Airborne Earth Science Workshop (pp. 2326)
Boer, M.M. (1999). Assessment of dryland degradation: linking theory and practice through
site water balance modelling. Utrecht: Department of Physical Geography, Utrecht
University
99
Remote Sensing Based Assessment
Boer, M.M., & Puigdefabregas, J. (2003). Predicting potential vegetation index values as a
reference for the assessment and monitoring of dryland condition. International
Journal of Remote Sensing, 24
Boer, M.M., & Puigdefabregas, J. (2005). Assessment of dryland condition using spatial
anomalies of vegetation index values. International Journal of Remote Sensing, 26,
4045-4065
Brandt, J., & Thornes, J.B. (1996). Mediterranean desertification and land use. London: Wiley &
Sons
Breiman, L. (2001). Random forests. Machine Learning, 45, 5-32
Brogaard, S., & Olafsdottir, R. (1997). Ground-truths of Ground-lies? In, Lund Electronic Reports
in Physical Geography
Calvao, T., & Palmeirim, J.M. (2004). Mapping Mediterranean scrub with satellite imagery:
biomass estimation and spectral behaviour. International Journal of Remote Sensing,
25, 3113-3126
Chander, G., Helder, D.L., Markham, B.L., Dewald, J.D., Kaita, E., Thome, K.J., Micijevic, E., &
Ruggles, T.A. (2004). Landsat-5 TM reflective-band absolute radiornetric calibration.
IEEE Transactions on Geoscience and Remote Sensing, 42, 2747-2760
Chander, G., & Markham, B. (2003). Revised Landsat-5 TM radiometric calibration procedures
and postcalibration dynamic ranges. IEEE Transactions on Geoscience and Remote
Sensing, 41, 2674-2677
Chang, C.-I., Wu, C.-C., Liu, W.-M., & Ouyang, Y.-C. (2006). A new growing method for
simplex-based endmember extraction algorithm. IEEE Transactions on Geoscience
and Remote Sensing, 44, 2804-2819
Chatfield, C. (2004). The analysis of time-series. An introduction. Boca Ration: CRC Press
Chen, J., Jönsson, P., Tamura, M., Gu, Z., Matsushita, B., & Eklundh, L. (2004). A simple method
for reconstructing a high-quality NDVI time-series data set based on the SavitzkyGolay filter. Remote Sensing of Environment, 91, 332-344
Chopping, M.J., Su, L., Rango, A., Martonchik, J.V., Peters, D.P.C., & Laliberte, A. (2008).
Remote sensing of woody shrub cover in desert grasslands using MISR with a
geometric optical canopy reflectance model. Remote Sensing of Environment, 112,
19-34
Cihlar, J., Latifovice, R., Chen, J., Trishchenko, A., Du, Y., Fedosojevs, G., & Guindon, G. (2004).
Systematic corrections of AVHRR image composites for temporal studies. Remote
Sensing of Environment, 89, 217-233
Cihlar, J., Xia, Q.H., Chen, J., Beaubien, J., Fung, K., & Latifovic, R. (1998). Classification by
progressive generalization: a new automated methodology for remote sensing
multichannel data. International Journal of Remote Sensing, 19, 2685-2704
Clevers, J.G.P.W. (1988). The derivation of a simplified reflectance model for te estimation of
leaf area index. Remote Sensing of Environment, 35, 53-70
Cohen, W.B., Maiersperger, T.K., Yang, Z., Gower, S.T., Turner, D.P., Ritts, W.D., Berterretche, M.,
& Running, S.W. (2003). Comparisons of land cover and LAI estimates derived from
ETM+ and MODIS for four sites in North America: a quality assessment of 2000/2001
provisional MODIS products. Remote Sensing of Environment, 88, 233-255
Congalton, R.G., & Green, K. (1999). Assessing the Accuracy of Remotely Sensed
Data:Principles and Practices. Boca Raton: CRC Press
Coppin, P., Jonckheere, I., Nackaerts, K., & Muys, B. (2004a). change detection methods in
ecosystem monitoring: a review. International Journal of Remote Sensing, 25, 15651596
100
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Coppin, P., Jonckheere, I., Nackaerts, K., Muys, B., & Lambin, E. (2004b). Digital change
detection methods in ecosystem monitoring: a review. International Journal of
Remote Sensing, 25, 1565-1596
Cracknell, A.P. (1997). The Advanced Very High Resolution Radiometer (AVHRR). London:
Taylor & Francis
Cracknell, A.P. (1998). Synergy in remote sensing - what's in a pixel? International Journal of
Remote Sensing, 19, 2025-2047
Crist, E.P., & Kauth, R.T. (1986). The Tasseled Cap De-Mystified. Photogrammetric Engineering
& Remote Sensing, 52, 81-86
Curran, P. (1980). Multispectral remote sensing of vegetation amount. Progress in Physical
Geography, 4, 315-341
de Vries, C., Danaher, T., Denham, R., Scarth, P., & Phinn, S. (2007). An operational
radiometric calibration procedure for the Landsat sensors based on pseudo-invariant
target sites. Remote Sensing of Environment, 107, 414-429
Del Barrio, G., Puigdefabregas, J., Sanjuan, M.E., Stellmes, M., & Ruiz, A. (2010). Assessment
and monitoring of land condition in the Iberian Peninsula, 1989-2000. Remote Sensing
of Environment, 114, 1817-1832
Dennison, P.E., Halligan, K.Q., & Roberts, D.A. (2004). A comparison of error metrics and
constraints for multiple endmember spectral mixture analysis and spectral angle
mapper. Remote Sensing of Environment, 93, 359-367
Dennison, P.E., & Roberts, D.A. (2003). Endmember selection for multiple endmember spectral
mixture analysis using endmember average RMSE. Remote Sensing of Environment, 87,
123-135
Díaz-Delgado, R., Lloret, F., & Pons, X. (2003). Influence of fire severity on plant regeneration
through remote sensing imagery. International Journal of Remote Sensing, 24, 17511763
Dinguirard, M., & Slater, P.N. (1999). Calibration of Space-Multispectral Imaging Sensors: A
Review. Remote Sensing of Environment, 68, 194-205
Downing, T.E., & Lüdeke, M.K.B. (2002). Social geographies of vulnerability and adaptation. In
J.F. Reynolds & D.M. Stafford-Smith (Eds.), Global desertification. Do humans cause
deserts? (pp. 233-252). Berlin: Dahlem University Press
Drake, N.A., & Settle, J.J. (1989). Linear mixture modelling of Thematic Mapper data of teh
Peruvian Andes. Proceedings of the 9th EARSeL Symposium, Helsinki, Finland, 27th
June - 1st July 1989, 490-495
Du, Y., Teillet, P.M., & Cihlar, J. (2002). Radiometric normalization of multitemporal high
resolution satellite images with quality control for land cover change detection.
Remote Sensing of Environment, 82, 123-134
Duchemin, B., Goubier, J., & Courrier, G. (1999). Monitoring phenological key stages and
cycle duration of temperate deciduous forest ecosystems with NOAA/AVHRR data.
Remote Sensing of Environment, 76, 68-82
Duguy, B., Alloza, J.A., Röder, A., Vallejo, R., & Pastor, F. (2007). Modelling the effects of
landscape fuel treatments on fire growth and behaviour in a Mediterranean
landscape (eastern Spain). International Journal of Wildland Fire, 16, 619-632
Dymond, C.C., Mladenoff, D.J., & Radeloff, V.C. (2002). Phenological differences in Tasseled
Cap indices improve deciduous forest classification. Remote Sensing of Environment,
80, 460-472
101
Remote Sensing Based Assessment
Eastman, R.J., & Fulk, M. (1993). Long sequence time series evaluation using standardized
principla components. Photogrammetric Engineering & Remote Sensing, 59, 13071312
Erbrecht, T. (2003). Monitoring of land degradation and land use systems with high temporal
resolution remote sensing data. Trier
European Commission (1998). An Integrated Approach to Assess and Monitor Desertification
Processes in the Mediterranean Basin (DeMon-II). Final Report. In. Ispra, Italy: European
Commission, Directorate General Joint Research Centre
Evans, J.P., & Geerken, R. (2006). Classifying rangeland vegetation type and coverage using
a Fourier component based similarity measure. Remote Sensing of Environment, 105,
1-8
Fang, H., Lian, S., McClaran, M.P., van Leeuwen, W.J.D., Drake, S., Marsh, S.E., Thomson, A.M.,
Izaurralde, R.C., & Rosenberg, N. (2005). Biophysical characterization and
management effects on semiarid rangeland observed from Landsat ETM+ data. IEEE
Transactions on Geoscience and Remote Sensing, 43, 125-134
Fischer, A. (1994a). A model for the seasonal variations of vegetation indices in coarse
resolution data and its inversion to extract cop parameters. Remote Sensing of
Environment, 48, 220-230
Fischer, A. (1994b). A simple model for the temporal variations of NDVI at regional scale over
agricultural countries. Validation with ground radiometric measurements. International
Journal of Remote Sensing, 15, 1421-1446
Fisher, P. (1997). The pixel: A snare and a delusion. International Journal of Remote Sensing,
18, 679-685
Foody, G.M. (2004). Thematic map comparison: evluating the statistical significance of
differences in classification accuracy. Photogrammetric Engineering & Remote
Sensing, 70, 627-633
Foody, G.M., & Mathur, A. (2006). The use of small training sets containing mixed pixels for
accurate hard image classification: Training on mixed spectral responses for
classification by a SVM. Remote Sensing of Environment, 103, 179-189
Franke, J., Roberts, D.A., Halligan, K., & Menz, G. (2009). Hierarchical Multiple Endmember
Spectral Mixture Analysis (MESMA) of hyperspectral imagery for urban environments.
Remote Sensing of Environment, 113, 1712-1723
Friedl, M.A., & Brodley, C.E. (1997). Decision tree classification of land cover from remotely
sensed data. Remote Sensing of Environment, 61, 399-409
Friedl, M.A., McIver, D.K., Hodges, J.C.F., Zhang, X.Y., Muchoeny, D., Strahler, A.H., Woodcock,
C.E., Gopal, S., Schneider, A., Cooper, A., Baccini, A., Gao, F., & Schaaf, C. (2002).
Global land cover mappign from MODIS: Algorithms and early results. Remote Sensing
of Environment, 83, 287-302
Friedl, M.a., Sulla-Menashe, D., Tan, B., Schneider, A., Ramankutty, N., Sibley, A., & Huang, X.
(2010). MODIS Collection 5 global land cover: Algorithm refinements and
characterization of new datasets. Remote Sensing of Environment, 114, 168-182
Friedrich, K., & Koslowsky, D. (2009). Inter-comparison of MEDOKADS and NOAA/NASA
pathfinder AVHRR land NDVI time series. In A. Röder & J. Hill (Eds.), Recent Advancesin
Remote Sensing and Geoinformation Processing for Land Degradation Assessment
(pp. 103-116). Boca Raton, New York, Abingdon: CRC Press/Balkema
Garcia-Haro, F.J., Gilabert, M.A., & Melia, J. (1999). Extraction of endmembers from spectral
mixtures. Remote Sensing of Environment, 68, 237-253
García-Haro, F.J., Gilabert, M.A., & Melìa, J. (2001). Monitoring fire-affected areas using
Thematic Mapper data. International Journal of Remote Sensing, 22, 533-549
102
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
García-Haro, F.J., Sommer, S., & Kemper, T. (2005). A new tool for variable multiple
endmember spectral mixture analysis (VMESMA). International Journal of Remote
Sensing, 26, 2135-2162
GarciaHaro, F.J., Gilabert, M.A., & Melia, J. (1996). Linear spectral mixture modelling to
estimate vegetation amount from optical spectral data. International Journal of
Remote Sensing, 17, 3373-3400
Gates, D.M. (1980). Biophysical Ecology. New York: Springer
Geerken, R., Zaitchicki, B., & Evans, J.P. (2005). Classifying rangeland vegetation type and
coverage from NDVI time series using Fourier Filtered Cycle Similarity. International
Journal of Remote Sensing, 26, 5535-5554
Geist, H. (2005). The causes and progression of desertification. Aldershot: Ashgate Publishing
Limited
Geist, H.J., & Lambin, E.F. (2004). Dynamic causal patterns of desertification. Bioscience, 54,
817-829
Gilabert, M.A., Gonzàles-Piqueras, J., GarciaHaro, F.J., & Melia, J. (2002). A generalized soiladjusted vegetation index. Remote Sensing of Environment, 82
Gillson, L., & Hoffman, M.T. (2007). Rangeland ecology in a changing world. Science, 315, 5354
Gislason, P.O., Benediktsson, J.A., & Sveinsson, J.R. (2006). Random forests for land cover
classification. Pattern Recognition Letters, 27, 294-300
Govaerts, Y.M., & Verstraete, M.M. (1998). Raytran: A Monte Carlo ray-tracing model to
compute light scattering in three-dimensional heterogeneous media. IEEE Transactions
on Geoscience and Remote Sensing, 36, 493-505
Goward, S.N., Davis, P.E., Fleming, D., Miller, L., & Townshend, J.R.G. (2007). Empirical
comparison of Landsat 7 and IKONOS multispectral measurements for selected Earth
Observation Systems (EOS) validation sites. Remote Sensing of Environment, 88, 80-99
Goward, S.N., & Hope, A.S. (1989). Evapotranspiration from combined reflected solar and
emitted terrestrial radiations: preliminary FIFE results from ACHRR data. Advances in
Space Research, 9
Goward, S.N., Masek, J.G., Williams, D.J., Irons, J.R., & Thompson, R.J. (2001). The Landsat 7
mission. Terrestrial research and applications for the 21st century. Remote Sensing of
Environment, 78, 3-12
Gutman, G. (1991). Vegetation indices from AVHRR: an update and future prospects. Remote
Sensing of Environment, 35, 121-136
Hall, F.G., Strebel, D.E., Nickeson, J.E., & Goetz, S.J. (1991). Radiometric rectification - toward a
common radiometric response among multi-date, multisensor images. Remote
Sensing of Environment, 35, 11-27
Hargreaves, G.H., & Samani, Z.A. (1982). Estimating potential evapotransporation. Journal of
the Irrigation and Drainage Division - ASCE, 108, 225-230
Harris, J.A., & Asner, G.P. (2003). Grazing gradient detection with airborne imaging
spectroscopy on a semi-arid rangeland. Journal of Arid Environments, 55, 391-404
Harris, J.A., & van Diggelen, R. (2008). Ecological restoration as a project for global society. In
J. van Andel & J. Aronson (Eds.), Restoration Ecology (pp. 3-15). Malden, Oxford,
Carlton: Blackwell Publishing
Hay, J.E., & McKay, D.C. (1985). Estimating solar irradiance on inclined surfaces: A review and
assessment of methodologies. International Journal of Solar Energy, 3, 203-240
103
Remote Sensing Based Assessment
Helder, D.L., Markham, B., Thome, K.J., Brarsi, J.A., Chander, G., & Malla, R. (2008). Updated
radiometric calibration for the Landsat 5 Thematic Mapper reflective bands. IEEE
Transactions on Geoscience and Remote Sensing, 46, 3309-3325
Hermance, J.F. (2007). Stabilizing high-order, non-classical harmonic analysis of NDVI data for
average annual models by damping model roughness. International Journal of
Remote Sensing, 28, 2801-2819
Hermance, J.F., Jacob, R.W., Bradley, B.A., & Mustard, J.F. (2007). Extracting phenological
signals from multiyear AVHRR NDVI time series: ramework for applying high-order
annual splines with roughness damping. IEEE Transactions on Geoscience and Remote
Sensing, 45, 3264-3276
Herold, M., Mayaux, P., Woodcock, C.E., Baccini, A., & Schmullius, C. (2008). Some challenges
in global land cover mapping: an assessment of agreement and accuracy in existing
1 km datasets. Remote Sensing of Environment, 112, 2538-2556
Hill, J., Dach, C., Del Barrio, G., Stellmes, M., Helldén, U., & Changyao, W. (2010). Integrating
MODIS-EVI and Gridded Rainfall/Temperatur Fields for Assessing Land Degradation
Dynamics in Horqin Sandy Lands (Inner Mongolia, China). In EARSeL (Ed.), 30th EARSeL
Symposium 2010. Remote Sensing for Science, Education and Natural and Cultural
Heritage. 31 May - 3 June, 2010, Paris, France. Paris
Hill, J., Hostert, P., & Röder, A. (2004). Long-term Observation of Mediterranean Ecosystems
with Satellite Remote Sensing. In S. Mazzoleni, G. di Pasquale, M. Mulligan, P. di
Martino & F. Rego (Eds.), Recent Dynamics of the Mediterranean Vegetation and
Landscape (pp. 33-43). Chichester: John Wiley & Sons Ltd.
Hill, J., Hostert, P., Tsiourlis, G., Kasapidis, P., Udelhoven, T., & Diemer, C. (1998). Monitoring 20
years of increased grazing impact on the Greek island of Crete with earth observation
satellites. Journal of Arid Environments, 39, 165-178
Hill, J., & Mehl, W. (2003). Geo- und radiometrische Aufbereitung multi- und hyperspektraler
Daten zur Erzeugung langjähriger kalibrierter Zeitreihen. Photogrammetrie –
Fernerkundung – Geoinformation, 1, 7-14.
Hill, J., Mehl, W., & Radeloff, V. (1995). Improved forest mapping by combining corrections of
atmospheric and topographic effects. In J. Askne (Ed.) (pp. 143-151). Rotterdam,
Brookfield: A.A. Balkema
Hill, J., Röder, A., Mehl, W., & Tsiourlis, G.M. (2009). Retrieving rangeland vegation
characteristics through constrained inverse reflectance modelling of earth
observation imagery. In A. Röder & J. Hill (Eds.), Recent Advances in Remote Sensing
and Geoinformation Prcessing for Land Degradation Assessment (pp. 211-226).
London: Taylor&Francis
Hill, J., Stellmes, M., Udelhoven, T., Röder, A., & Sommer, S. (2008). Mediterranean
desertification and land degradation. Mapping related land use change syndromes
based on satellite observations. Global and Planetary Change, 64, 146-157
Hill, J., & Sturm, B. (1991). Radiometric correction of multitemporal Thematic Mapper data for
use in agricultural land-cover classification and vegetation monitoring. International
Journal of Remote Sensing, 12, 1471-1491
Hirsch, R., & Slack, J. (1984). A nonparametric trend test for seasonal data with serial
dependance. Water Resources Research, 20, 727-732
Holben, B.N. (1986). Characteristics of maximum value composite images for temporal AVHRR
data. International Journal of Remote Sensing, 7, 1417-1437
Holben, B.N., Eck, T.F., Slutsker, I., Tanré, D., Buis, J.P., Setzer, A., Vermote, E., Reagan, J.A.,
Kaufman, Y.J., Nakajima, T., Lavenu, F., Jankowiak, I., & Smirnov, A. (1998). AERONET-A Federated Instrument Network and Data Archive for Aerosol Characterization.
Remote Sensing of Environment, 66, 1-16
104
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Hostert, P., Roder, A., & Hill, J. (2003a). Coupling spectral unmixing and trend analysis for
monitoring of long-term vegetation dynamics in Mediterranean rangelands. Remote
Sensing of Environment, 87, 183-197
Hostert, P., Roder, A., Hill, J., Udelhoven, T., & Tsiourlis, G. (2003b). Retrospective studies of
grazing-induced land degradation: a case study in central Crete, Greece.
International Journal of Remote Sensing, 24, 4019-4034
Hostert, P., Roder, A., Jarmer, T., Udelhoven, T., & Hill, J. (2001a). The potential of remote
sensing and GIS for desertification monitoring and assessment. Annals of Arid Zone, 40,
103-140
Hostert, P., Röder, A., Jarmer, T., Udelhoven, T., & Hill, J. (2001b). The potential of remote
sensing and GIS for desertification monitoring and assessment. Annals of Arid Zone, 40,
103-140
Hountondji, Y.C., Sokpon, N., Nicolas, J., & Ozer, P. (2009). Ongoing desertification processes
in the sahelian belt of West Africa: an evidence from the rain-use efficiency. In A.
Röder & J. Hill (Eds.), Recent Advances in Remote Sensing and Geoinformation
Processing for Land Degradation Assessment (pp. 173-186). London: Taylor & Francis
Huang, C., Davis, L.S., & Townshend, J.R.G. (2002). An assessment of support vector machines
for land cover classification. International Journal of Remote Sensing, 23, 725-749
Huemmrich, K.F. (2001). The GeoSail model: a simple addition to the SAIL model to describe
discontinuous canopy reflectance. Remote Sensing of Environment, 75, 423-431
Huete, A.R. (1988). A Soil-Adjusted Vegetation Index (Savi). Remote Sensing of Environment,
25, 295-309
Huete, A.R., Didan, K., Mirua, T., Rodriguez, E.P., Gao, X., & Ferreira, L.G. (2002). Overview of
the radiometric and biophysical performance of the MODIS vegetation indices.
Remote Sensing of Environment, 83, 195-213
Huete, A.R., Jackson, R.D., & Post, D.F. (1985). Spectral response of a plant canopy with
different soil backgrounds. Remote Sensing of Environment, 17, 37-53
Itten, K.I., & Meyer, P. (1993). Geometric and radiometric correction of TM data of
mountainous forested areas. IEEE Transactions on Geoscience and Remote Sensing,
31, 764-770
Jacquemoud, G., & Baret, F. (1990). PROSPECT: a model of leaf optical properties spectra.
Remote Sensing of Environment, 34, 75-91
Jacquemoud, S. (1995). Extraction of vegetation biophysical parameters by inversion of the
PROSPECT+SAIL models on sugar beet canopy reflectance data. Application to TM
and AVIRIS sensors. Remote Sensing of Environment, 52, 163-172
Jeltsch, F., Milton, S.J., Dean, W.R.J., & van Rooyen, N. (1996). Tree spacing and coexistence
in semiarid savannas. Journal of Ecology, 84
Jeltsch, F., Weber, G.E., & Grimm, V. (2000). Ecological buffering mechanisms in savannas: A
unifying theory of long-term tree-grass coexistence. Plant Ecology, 161
Jiang, L., Tarpley, J.D., Mitchell, K.E., Zhou, S., Kogan, F.N., & Guo, W. (2008). Adjusting for longterm anomalous trends in NOAA's global vegetation index data sets. IEEE Transactions
on Geoscience and Remote Sensing, 46, 409-422
Johnson, R.D., & Kasischke, E.S. (1998). Change vector analysis: a technique for the
multispectral monitoring of land cover and condition. International Journal of Remote
Sensing, 19, 411-426
Jönsson, P., & Eklundh, L. (2002). Seasonality extraction by function fitting to time-series of
satellite sensor data. IEEE Transactions on Geoscience and Remote Sensing, 40, 18241832
105
Remote Sensing Based Assessment
Karpouzli, E., & Malthus, T. (2003). The empirical method for the atmospheric correction of
IKONOS imagery. International Journal of Remote Sensing, 24, 1143-1150
Kauth, R.J., & Thomas, G.S. (1976). The Tasselled Cap - A Graphic Description of the SpectralTemporal Development of Agricultural Crops as seen by Landsat. In, Symposium on
Machine Processing of Remotely Sensed Data. West Lafayette, Indiana, USA: The
Laboratory for Applications of Remote Sensing, Purdue University
Kimes, D.S., Knyazikhin, Y., Privette, J.L., Abuelgasim, A.A., & Gao, F. (2000). Inversion methods
for physically-based models. Remote Sensing Reviews, 18, 381-439
Knorn, J., Rabe, A., Radeloff, V.C., Kuemmerle, T., Kozak, J., & Hostert, P. (2009). Land cover
mapping of large areas using chain classification of neighboring Landsat satellite
images. Remote Sensing of Environment, 113, 957-964
Koomen, E., Stillwell, J., Bakema, A., & Scholten, H.J. (2007). Modelling land-use change.
Progress and Applications. Dordrecht, The Netherlands: Springer
Kramer, H.J. (2002). Observation of the Earth and its environment. Survey of missions and
sensors. Berlin, Heidelberg, New York: Springer
Kuemmerle, T., Chaskovskyy, O., Knorn, J., Radeloff, V.C., Keeton, W.S., & Hostert, P. (2009).
Forest cover change and illegal logging in the Ukrainian Carpathians in the transition
period from 1988 to 2007. Remote Sensing of Environment, 113, 1194-1207
Kuemmerle, T., Perzanowski, K., Chaskovskyy, O., Ostapowicz, K., Halada, L., Bashta, A.-T.,
Kruhlov, I., Hostert, P., Waller, D., & Radeloff, V.C. (2010). European Bison habitat in the
Carpathian Mountains. Biological Conservation, in print
Kuemmerle, T., Röder, A., & Hill, J. (2006). Separating grassland and shrub vegetation by
multidate pixel-adaptive spectral mixture analysis. International Journal of Remote
Sensing, 27, 3251-3271
Laba, M., Downs, R., Smith, S., Welsh, S., Neider, C., White, S., Richmond, M., Philpot, W., &
Baveye, P. (2008). Mapping invasive wetland plants in the Hudson River National
Estuarine Research Reserve using quickbird satellite imagery. Remote Sensing of
Environment, 112, 286-300
Lacaze, B. (1996). Spectral characterisation of vegetation communities and practical
approaches to vegetation cover changes monitoring. In J. Hill & D. Peter (Eds.), The
use of remote sensing for land degradation and desertification monitoring in the
Mediterranean basin - state of the art and future research. Proceedings of a
workshop, jointly organized by JRC/IRSA and DGXII/D-2/D-4, Valencia, 13-15 June
1994 (pp. 149-166). Valencia: Office for Official Publications of the European
Communities
Lacaze, B. (2005). Remotely-sensed optical and thermal indicators of land degradation. In M.
Oluic (Ed.), New strategies for European Remote Sensing. Proceedings of the 24th
Symposium of the European Association of Remote Sensing Laboratories, Dubrovnik,
Croatia, 25-27 May 2004 (pp. 211-217)
Lambin, E.F., & Geist, H.J. (2006). Land Use and Land Cover Change. Local Processes and
Global Impacts. Berlin, Heidelberg, New York: Springer Verlag
Lambin, E.F., Geist, H.J., & Rindfuss, R.R. (2006). Introduction: Local Processes with Global
Impacts. In E.F. Lambin & H.J. Geist (Eds.), Land-Use and Land-Cover Change (pp. 18). Berlin, Heidelberg, New York: Springer
Lark, R.M. (1995). A reappraisal of unsupervised classification 1. Correspondence between
spectral and conceptual classes. International Journal of Remote Sensing, 16, 14251443
106
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Lawrence, R.L., Wood, S.D., & Sheley, R.L. (2006). Mapping invasive plants using hyperspectral
imagery and Breiman Cutler classifications (Random Forests). Remote Sensing of
Environment, 100, 356-362
Le Houérou, N.H. (1984). Rain use efficiency: a unifying concept in arid land ecology. Journal
of Arid Environments, 7, 1-15
Leboeuf, A., Beaudoin, A., Fournier, R.A., Guindon, L., Luther, J.E., & Lambert, M.C. (2007). A
shadow fraction method for mapping biomass of northern boreal black spruce forests
using QuickBird imagery. Remote Sensing of Environment, 110, 488-500
Li, Z., & Kafatos, M. (2000). Interannual variability of vegetation in the United States and its
relation to El Nino/southern oscillation. Remote Sensing of Environment, 71, 239-247
Lillesand, T.M., & Kiefer, R.W. (2000). Remote Sensing and Image Interpretation. New York,
USA: John Wiley & Sons
Lobo, A., & Maisongrande, P. (2006). Searching for trends of change through exploratory
data analysis of time series of remotely sensed images of SW Europe. In J.A. Sobrino
(Ed.), Proceedings of the Second Conference on Recent Advances in Quantitative
Remote Sensing, Valencia (pp. 534-537). Valencia
Lohninger, H. (1999). Teach/Me Data Analysis. Berlin: Springer
Lorent, H., Evangelou, C., Stellmes, M., Hill, J., Papanastasis, V.P., Tsiourlis, G., Röder, A., &
Lambin, E.F.-. (2008). Land degradation and economic conditions of agricultural
households in a marginal region of northern Greece. Global and Planetary Change,
64, 198-209
Loudermilk, E.L., Hiers, J.K., O'Brien, J.J., Mitchell, R.J., Singhania, A., Fernandez, J.C., Cropper,
W.P.J., & Slattion, K.C. (2009). Ground-based LiDAR: a novel approach to quantify
fine-scale fuelbed characteristics. International Journal of Wildland Fire, 18, 676-685
Loveland, T.R., Reed, B.C., Brown, J.F., Ohlen, D.O., Zhu, Z., Yang, L., & Merchant, J.W. (2000).
Development of a global land cover characteristics database and IGBP DISCover
from 1 km AVHRR data. International Journal of Remote Sensing, 21, 1303-1330
Lu, D., Mausel, P., Brondizio, E., & Moran, E. (2004). Change detection techniques.
International Journal of Remote Sensing, 25, 2365-2407
Lunetta, R.S., Knight, J.F., Ediriwickrema, J., Lyon, J.G., & Worthy, L.D. (2006). Land-cover
change detectio nusing multi-temporal MODIS NDVI data. Remote Sensing of
Environment, 105, 142-154
Lyon, J.G., Yuan, D., Lunetta, R.S., & Elvidge, C.D. (1998). A change detection experiment
using vegetation indices. Photogrammetric Engineering & Remote Sensing, 64, 143150
Major, D., Baret, F., & Guyot, G. (1990). A ratio vegetation index adjusted for soil brightness.
International Journal of Remote Sensing, 11, 727-740
Markham, B.L., Storey, J.C., Williams, D.J., & Irons, J.R. (2004). Landsat sensor performance:
history and current status. IEEE Transactions on Geoscience and Remote Sensing, 42,
2691-2694
Masek, J.G., Honzak, M., Goward, S.N., Liu, P., & Pak, E. (2001). Landsat-7 ETM+ as an
observatory for land cover. Initial radiometric and geometric comparisons with
Landsat-5 Thematic Mapper. Remote Sensing of Environment, 78, 118-130
Maxwell, S.K., Schmidt, G.L., & Storey, J.C. (2007). A multi-scale segmentation approach to
filling gaps in Landsat ETM+ SLC-off images. International Journal of Remote Sensing,
28, 5339-5356
McCaffrey, T.M., & Franklin, S.E. (1993). Automated training site selection for large-area
remote-sensing image-analysis. Computers & Geosciences, 19, 1413-1428
107
Remote Sensing Based Assessment
McGarigal, K. (2002). Landscape pattern metrics. In A.H. El-Shaarawi & W.W. Piegorsch (Eds.),
Encyclopedia of environmetrics (pp. 1135-1142). Sussex: Wiley & Sons
Mikhail, E.M., Bethel, J.S., & McGlone, J.C. (2001). Introduction to Modern Photogrammetry.
New York: John Wiley & Sons
Moody, A., & Johnson, D.M. (2001). Land surface phenologies from AVHRR using the discrete
fourier transform. Remote Sensing of Environment, 73, 305-323
Myneni, R.B., Hoffman, S., Knyazikhin, Y., Privette, J.L., Glassy, J., Tian, Y., Wang, Y., Song, X.,
Zhang, Y., Smith, G.R., Lotsch, A., Friedl, M., Morisette, J.T., Votava, P., Nemani, R.R., &
Running, S.W. (2002). Global products of vegetation leaf area and fraction absorbed
PAR from year one of MODIS data. Remote Sensing of Environment, 83, 214-231
Noy-Meir, I. (1998). Effects of grazing on Mediterranean grasslands: the community level. In
V.P. Papanastasis & D. Peter (Eds.), Ecological basis of livestock grazing in
Mediterranean ecosystems. Proceedings of the International Workshop held in
Thessaloniki (Greece) on October 23-25, 1997 (pp. 27-39). Luxemburg: Office for
Official Publications of the European Communities
Numata, I., Roberts, D.A., Chadwick, O.A., Schimel, J., Sampaio, F.R., Leonidas, F.C., & Soares,
J.V. (2007). Characterization of pasture biophysical properties and the impact of
grazing intensity using remotely sensed data. Remote Sensing of Environment, 109,
314-327
Okin, G.S., Okin, W.J., Roberts., D.A., & Murray, B. (1998). Multiple endmember spectral mixture
analysis: application to an arid/semi-arid landscape. In R.O. Green (Ed.), 10th Airborne
Visible/Infrared Imaging Spectrometer (AVIRIS) Workshop. Pasadena
Ostapowicz, K., Vogt, P., Riitters, K.H., Kozak, J., & Estreguil, C. (2008). Impact of scale on
morphological spatial pattern of forest. Landscape Ecology, 23, 1107-1117
Pal, M. (2007). Ensemble learning with decision tree for remote sensing classification. World
Academy of Science, Engineering and Technology, 36, 258-260
Pal, M., & Mather, P.M. (2005). Support vector machines for classification in remote sensing.
International Journal of Remote Sensing, 26
Pala, V., & Pons, X. (1995). Incorporation of relief in polynom-based geometric correction.
Photogrammetric Engineering & Remote Sensing, 61, 935-944
Pickup, G., & Chewings, V.H. (1994). A grazing gradient approach to land degradation
assessment in arid areas from remotely sensed data. International Journal of Remote
Sensing, 15, 597-617
Pinty, B., & Verstraete, M.M. (1992). GEMI: A non-linear index to monitr global vegetation from
satellites. Vegetatio, 1335-1372
Price, J.C. (2003). Comparing MODIS and ETM+ data for regional and global land
classification. Remote Sensing of Environment, 86, 491-499
Pringle, M.J., Schmidt, M., & Muir, J.S. (2009). Geostatistical interpolation of SLC-off Landsat
ETM+ images. ISPRS Journal of Photogrammetry and Remote Sensing, 1-11
Puigdefabregas, J., & Mendizabal, T. (1998). Perspectives on desertification: western
Mediterranean. Journal of Arid Environments, 39, 209-224
Quinn, R.D. (1986). Mammalian herbivory and resilience in mediterranean-climate
ecosystems. In B. Dell, A.J.M. Hopkins & B.B. Lamont (Eds.), Resilience in
mediterranean-type ecosystems (pp. 113-128). Dordrecht, Boston, Lancaster: Kluwer
Academic Publishers
Rees, W.G., & Williams, M. (1997). Monitoring Changes in land cover induced by atmospheric
pollution in the Kola Peninsula, Russia, using Landsat-MSS data. International Journal of
Remote Sensing, 18, 1703-1723
108
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Riano, D., Chuvieco, E., Usting, Zomer, R., Dennison, P.E., & Roberts, D.A. (2002). Assessment of
vegetation regeneration after fire through multi-temporal analysis of AVIRIS images in
the Santa Monica Mountains. Remote Sensing of Environment, 79, 60-71
Richards, J.A., & Jia, X. (2005). Remote Sensing Digital Image Analysis. Berlin, Germany:
Springer
Richardson, A.J., & Wiegand, C.L. (1977). Distinguishing vegetation from soil background
information. Photogrammetric Engineering & Remote Sensing, 1541-1552
Rindfuss, R.R., Walsh, S.J., Turner, B.L., Fox, J., & Mishra, V. (2004a). Developing a science of
land change: Challenges and methodological issues. Proceedings of the National
Academy of Sciences of the United States of America, 101, 13976-13981
Rindfuss, R.R., Walsh, S.J., Turner II, B.L., Moran, E.F., & Entwisle, B. (2004b). Linking people and
pixels. In G. Gutman, A.C. Janetos, C.O. Justice, E.F. Moran, J.F. Mustard, R.R. Rindfuss,
D. Skole, B.L. Turner Ii & M.A. Cochraine (Eds.), Land Change Science (pp. 379-394).
Dordrecht: Kluwer Academic Publishers
Roberts, D.A., Smith, M.O., & Adams, J.B. (1993). Green Vegetation, Nonphotosynthetic
Vegetation, and Soils in Aviris Data. Remote Sensing of Environment, 44, 255-269
Röder, A. (2005). A remote sensing based framework for monitoring and assessing
Mediterranean rangelands. Case studies from two test sites in Spain and Greece. PhD
Thesis, University of Trier, Germany
Röder, A., Duguy, B., Alloza, J.A., Vallejo, R., & Hill, J. (2008a). Using long time series of Landsat
data to monitor fire events and post-fire dynamics and identify driving factors.
Remote Sensing of Environment, 112 (1), 259-273
Röder, A., Hill, J., & Tsiourlis, G. (2008b). Assessing scale transitions in leaf area index estimation
in a heterogeneous shrubland ecosystem. In, 14th Australasian Remote Sensing and
Photogrammetry Conference (pp. 1-10). Darwin, Northern Territory, Australia
Röder, A., Kuemmerle, T., & Hill, J. (2005). Extension of retrospective datasets using multiple
sensors. An approach to radiometric intercalibration of Landsat TM and MSS data.
Remote Sensing of Environment, 95, 195-210
Röder, A., Kuemmerle, T., Hill, J., Papanastasis, V.P., & Tsiourlis, G.M. (2007). Adaptation of a
grazing gradient concept to heterogeneous Mediterranean rangelands using cost
surface modelling. Ecological Modelling, 204, 387-398
Röder, A., Udelhoven, T., Hill, J., Del Barrio, G., & Tsiourlis, G.M. (2008c). Trend analysis of
Landsat-TM and -ETM+ imagery to monitor grazing impact in a rangeland ecosystem
in Northern Greece. Remote Sensing of Environment, (in print)
Rodriguez Gonzalez, J., Del Barrio, G., & Duguy, B. (2008). Assessing functional landscape
connectivity for disturbance propagaion on regional scales - A cost surface model
approach applied to surface fire spread. Ecological Modelling, 211, 121-141
Rogan, J., Franklin, J., & Roberts, D.A. (2002). A comparison of methods for monitoring
multitemporal vegetation change using Thematic Mapper imagery. Remote Sensing
of Environment, 80, 143-156
Rogan, J., & Yool, S.R. (2001). Mapping fire-induced vegetation depletion in the Peloncillo
Mountains, Arizona and New Mexico. International Journal of Remote Sensing, 22,
3101-3121
Rogge, D.M., Rivard, B., Zhang, J., & Feng, J. (2006). Iterative spectral unmixing for optimizing
per-pixel endmember sets. IEEE Transactions on Geoscience and Remote Sensing, 44,
3725-3736
Rondeaux, G. (1995). Vegetation monitoring by remote sensing: a review of biophysical
indices. Photo-Interprétation, 95, 197-213
109
Remote Sensing Based Assessment
Roy, D.P., Ju, J., Lewis, P., Schaaf, C., Gao, F., Hansen, M., & Lindquist, E. (2008). Multi-temporal
MODIS-Landsat data fusion for relative radiometric normalization, gap filling, and
prediction of Landsat data. Remote Sensing of Environment, 112, 3112-3130
San Miguel-Ayanz, J., & Biging, G.S. (1996). An iterative classification approach for mapping
natural resources from satellite imagery. International Journal of Remote Sensing, 17,
957-981
Sanchez-Hernandez, C., S., D., & Foody, G.M. (2007). Mapping specific habitats from remotely
sensed imagery: Support vector machine and support vector data description based
classification of coastal saltmarsh habitats. Ecological Informatics, 2, 83-88
Savitzky, A., & Golay, M.J.E. (1964). Smoothing and differentiating of data by simplified least
square procedure. Analytical Chemistry, 36, 1627-1638
Scarth, P., Schmidt, M., Röder, A., & Gill, T. (2010). Tracking grazing pressure and climate
interaction - the role of Landsat fractional cover in time series analysis. In, 15th
Australasian Remote Sensing and Photogrammetry Conference. Alice Springs,
Australia
Schlerf, M., Atzberger, C., & Hill, J. (2005). Remote sensing of forest biophysical variabiles using
HyMap imaging spectrometer data. Remote Sensing of Environment, 95, 177.194
Schlittgen, R., & & Streitberg, B.H.J. (1999). Zeitreihenanalyse. München, Wien: Oldenburg
Verlag
Schmidt, H., & Karnieli, A. (2001). Sensitivity of vegetation indices to substrate brightness in
hyper-arid environments: the Makhtesh Ramon Crater (Israel) case study. International
Journal of Remote Sensing, 22, 3503-3520
Scholes, R.J., & Archer, S.R. (2997). Tree-grass interactions in savannas. Annual Review of
Ecology and Systematics, 28
Schott, J.R. (1997). Remote Sensing. The image chain approach. Oxford, New York: Oxford
University Press
Schriever, J.R., & Congalton, R.G. (1995). Evaluating seasonal variability as an aid to covertype mapping from Landsat Thematic Mapper data in the Northeast.
Photogrammetric Engineering and Remote Sensing, 61, 321-327
Senay, G.B., & Elliott, R.L. (2002). Capability of AVHRR data in discriminating rangeland cover
mixtures. International Journal of Remote Sensing, 23, 299-312
Settle, J.J., & Drake, N.A. (1993). Linear Mixing and the Estimation of Ground Cover
Proportions. International Journal of Remote Sensing, 14, 1159-1177
Shandley, J., Franklin, J., & White, T. (1996). Testing the Woodcock Harward image
segmentation algorithm in an area of southern Californian chaparral and woodland
vegetation. International Journal of Remote Sensing, 17, 983-1004
Shoshany, M., & Svoray, T. (2002). Multidate adaptive spectral unmixing and its application to
analysis of ecosystem transition along a climatic gradient. In, Remote Sensing of
Environment (pp. 5-20)
Shumway, R.H., & Stoffer, D.S. (2006). Time Series Analysis and Its Applications: With R
Examples. New York: Springer
Sinyuk, A., Dubovik, O., Holben, B., Eck, T.F., Breon, F.-M., Martonchik, J., Kahn, R., Diner, D.J.,
Vermote, E.F., Roger, J.-C., Lapyonok, T., & Slutsker, I. (2007). Simultaneous retrieval of
aerosol and surface properties from a combination of AERONET and satellite data.
Remote Sensing of Environment, 107, 90-108
Skarpe, C. (1992). Dynamics of savanna ecosystems. Journal of Vegetation Science, 3, 293300
110
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Skidmore, A., Ferwerda, J.G., Mutanga, O., Van Wieren, S., Peel, M., Grant, R.C., Prins, H.H.T.,
Bektas Balick, F., & Venus, V. (2010). Forage quality of savannas - Simultaneously
mapping foliar protein and polyphenols for trees and grass using hyperspectral
imagery. Remote Sensing of Environment, 114, 64-72
Small, C. (2004). The Landsat ETM plus spectral mixing space. Remote Sensing of Environment,
93, 1-17
Smirnov, A., Holben, B.N., Eck, T.F., Dubovik, O., & Slutsker, I. (2000). Cloud-Screening and
Quality Control Algorithms for the AERONET Database. Remote Sensing of
Environment, 73, 337-349
Smith, M.O., Adams, J.B., & Gillespie, A.R. (1990). Reference Endmembers for Spectral Mixture
Analysis. In, 5th Australasian Remote Sensing Conference. Perth, Australia
Smith, M.O., Adams, J.B., & Sabol, D.E. (1994). Spectral Mixture Analysis - New Strategies for
the Analysis of Multispectral Data. In J. Hill & J. Mégier (Eds.), Imaging Spectrometry -a Tool for Environmental Observations (pp. 125-143): Springer Netherland
Smits, P.C., Dellepiane, S.G., & Schowengerdt, R.A. (1999). Quality assessment of image
classification algorithms for land-cover mapping: a review and a proposal for a costbased approach. International Journal of Remote Sensing, 20, 1461-1486
Soille, P., & Vogt, P. (2009). Morphological segmentation of binary patterns. Pattern
Recognition Letters, 30, 456-459
Song, C., Dickinson, M.B., Su, L., Zhang, S., & Yaussey, D. (2010). Estimating average tree
crown size using spatial information from Ikonos and QuickBird images: Across-sensor
and across-site comparisons. Remote Sensing of Environment, 114, 1099-1107
Song, C., Woodcock, C.E., Seto, K.C., Lenney, M.P., & Macomber, S.A. (2001). Classification
and change detection using Landsat TM data: when and how to correct atmospheric
effects? Remote Sensing of Environment, 75, 230-244
Soudani, K., François, C., le Maire, G., Le Dantec, V., & Dufrêne, E. (2006a). Comparative
analysis of IKONOS, SPOT, and ETM+ data for leaf area index estimation in temperate
coniferous and deciduous forest stands. Remote Sensing of Environment, 102, 161-175
Soudani, K., François, C., le Maire, G., Le Dantec, V., & Dufrêne, E. (2006b). Erratum to
"Comparative analysis of IKONOS, SPOT, and ETM+ data leaf area index estimation in
temperature coniferous and deciduous forest stands" [Remote Sensing of Environment
102 (2006) 161-175]. Remote Sensing of Environment, 103, 497-497
Stafford-Smith, D.M., & Reynolds, J.F. (2002). The Dahlem desertification paradigm: a new
approach to an old problem. In J.F. Reynolds & M.D. Stafford-Smith (Eds.), Global
desertification. Do humans cause deserts? (pp. 403-424). Berlin: Dahlem University
Press
Stefanovska, A., & Bracic, M. (1999). Physics of the human cardiovascular system.
Contemporary Physics, 40, 31-55
Stellmes, M., Udelhoven, T., Röder, A., Sonnenschein, R., & Hill, J. (2010). Dryland observation
at local and regional scale - Comparison of Landsat TM/ETM+ and NOAA AVHRR time
series. Remote Sensing of Environment, 114
Steven, M.D., Malthus, T.J., Baret, F., Xu, H., & Chopping, M.J. (2003). Intercalibration of
vegetation indices from different sensor systems. Remote Sensing of Environment, 88,
412-422
Strahler, A.H., Jupp, A.L.B., Woodcock, C.E., Schaaf, C., Yao, T., Zhao, F., Yang, X., Lovell, J.,
Culvenor, D., Newnham, G., Ni-Miester, W., & Boykin-Morris, W. (2008). Retrieval of
forest structural parameters using a ground-based lidar instrument (Echidna).
Canadian Journal of Remote Sensing, 34, 426-440
111
Remote Sensing Based Assessment
Stuckens, J., Coppin, P.R., & Bauer, M.E. (2000). Integrating contextual information with perpixel classification for improved land cover classification. Remote Sensing of
Environment, 71, 282-296
Svoray, T., Shafran-Nathan, R., Ungar, E.F., Arnon, A., & Perevolotsky, A. (2009). Integrating GPS
technologies in dynamic spatio-temporal models to monitor grazing habitats in dry
rangelands. In A. Röder & J. Hill (Eds.), Recent Advances in Remote Sensing and
Geoinformation Processing for Land Degradation Assessment (pp. 301-322). London:
Taylor&Francis
Tanré, D. (1990). Description of a computer code to simulate the satellite signal in the solar
spectrum: 5S code. International Journal of Remote Sensing, 11, 659-668.
Tanre, D., Deroo, C., Duhaut, P., Herman, M., Morcrette, J.J., Perbos, J., & Deschamps, P.Y.
(1990). Description of a computer code to simulate the satellite signal in the solar
spectrum - the 5S Code. International Journal of Remote Sensing, 11, 659-668
Teillet, P.M. (1997). A status overview of Earth observation calibration/validation for terrestrial
applications. Canadian Journal of Remote Sensing, 23, 291-298
Teillet, P.M., Barker, J.L., Markham, B.L., Irish, R.R., Fedosejevs, G., & Storey, J.C. (2001).
Radiometric cross-calibration of the Landsat-7 ETM+ and Landsat-5 TM sensors based
on tandem data sets. Remote Sensing of Environment, 78, 39-54
Teillet, P.M., & Fedosejevs, G. (1995). On the dark target approach to atmospheric correction
of remotely sensed data. Canadian Journal of Remote Sensing, 21, 374-387
Teillet, P.M., Fedosejevs, G., Thome, K.J., & Barker, J.L. (2007). Impacts of spectral band
difference effects on radiometric cross-calibration between satellite sensors in the
solar-reflective spectral domain. Remote Sensing of Environment, 110, 393-409
Thome, K., Markham, B., Barker, J., Slater, P., & Biggar, S. (1997). Radiometric calibration of
Landsat. Photogrammetric Engineering and Remote Sensing, 63, 853-858
Tian, Y., Woodcock, C.E., Wang, Y., Privette, J.L., Shabanov, N.V., Zhou, L., Zhang, Y.,
Buermann, W., Dong, J., Veikkanen, B., Häme, T., Andersson, K., Ozdogan, M.,
Knyazikhin, Y., & Myneni, R.B. (2002a). Multiscale analysis and validation of the MODIS
LAI product: I. Uncertainty assessment. Remote Sensing of Environment, 83, 414-430
Tian, Y., Woodcock, C.E., Wang, Y., Privette, J.L., Shabanov, N.V., Zhou, L., Zhang, Y.,
Buermann, W., Dong, J., Veikkanen, B., Häme, T., Andersson, K., Ozdogan, M.,
Knyazikhin, Y., & Myneni, R.B. (2002b). Multiscale analysis and validation of the MODIS
LAI product: II. Sampling strategy. Remote Sensing of Environment, 83, 431-441
Tommervik, H., Hogda, K.A., & Solheim, L. (2003). Monitoring vegetation changes in Pasvik
(Norway) and Pechenga in Kola Peninsula (Russia) using multitemporal Landsat
MSS/TM data. Remote Sensing of Environment, 85, 370-388
Torrence, C., & Compo, G. (1998). A practical guide to wavelet analysis. Bulletin of the
American Meteorological Society, 79, 61-78
Toutin, T. (2004). Review article: Geometric processing of remote sensing images: models,
algorithms and methods. International Journal of Remote Sensing, 25, 1893-1924
Townsend, P.A., Lookingbill, T.R., Kindon, C., & Gardner, R.H. (2009). Spatial pattern analysis for
monitoring protected areas. Remote Sensing of Environment, 113, 1410-1420
Trabaud, L. (1981). Postfire plant community dynamics in the Mediterranean basin. In J.M.
Moreno & W.C. Oechel (Eds.), The role of fire in Mediterranean-type ecosystems (pp.
1-15). New York, Berlin, Heidelberg: Springer
Tucker, C.J., Pinzon, J.E., Brown, M.E., Slayback, D., Pak, E.W., Mahoney, R., Vermote, E., & ElSaleous, N. (2005). An extended AVHRR 8-km NDVI data set compatible with MODIS
and SPOT Vegetation NDVI data. International Journal of Remote Sensing, 26, 44854498
112
A. Röder, M. Stellmes, J. Hill, and Th. Udelhoven
Turner, D.P., Cohen, W.B., Kennedy, R.E., Fassnacht, K.S., & Briggs, J.M. (1999). Relationships
between leaf area index and Landsat TM spectral indices across three tempreate
zone sites. Remote Sensing of Environment, 70, 52-68
Turner II, B.L., Lambin, E., & Reenberg, A. (2007). The emergence of land change science for
global environmental change and sustainability. Proceedings of the National
Academy of Sciences of the United States of America, 104, 20666-20671
Udelhoven, T. (2010). TimeStats: A Software Tool for the Retrieval of Temporal Patterns from
Global Satellite Archives. IEEE Journal of Selected Topics in Applied Earth Observations
and Remote Sensing, 99, 1-8
Udelhoven, T., Stellmes, M., Del Barrio, G., & Hill, J. (2009). Assessment of rainfall and NDVI
anomalies in Spain (1989-1999) using distributed lag models. International Journal of
Remote Sensing, 30, 1961-1976
Udelhoven, T., van der Linden, S., Waske, B., & Stellmes, M. (2010). Hypertemporal
classification of large areas using decision fusion. IEEE Geoscience and Remote
Sensing Letters, 6, 592-596
U.S.G.S. (1984). Landsat data user handbook. Alexandria
Verbesselt, J., Hyndman, R., Newnham, G., & Culvenor, D. (2010). Detecting trend and
seasonal changes in satellite image time series. Remote Sensing of Environment, 114,
106-115
Verburg, P.H., Schot, P.P., Dijist, M.J., & Veldkamp, A. (2004). Land use change modelling:
current practice and research priorities. GeoJournal, 61, 309-324
Verhoef, W. (1984). Light scattering by leaf layers with application to canopy reflectance
modeling: the SAIL model. Remote Sensing of Environment, 16, 125-141
Verhoef, W., & Bach, H. (2007). Coupled soil-leaf-canopy and atmosphere radiative transfer
modeling to simulate hyperspectral multi-angular surface reflectance and TOA
radiance data. Remote Sensing of Environment, 109, 166-182
Vermote, E.F., Tanre, D., Deuze, J.L., Herman, M., & Morcrette, J.J. (1997). Second simulation of
the satellite signal in the solar spectrum, 6S: an overview. IEEE Transactions on
Geoscience and Remote Sensing, 35, 675-686
Verstraete, M. (1994a). The contribution of remote sensing to montior vegetation and to
evaluate its dynamic aspects. In F. Veroustrate & R. Ceulemans (Eds.), Vegetation,
modeling and climage change effects (pp. 207-212). The Hague: SPB Academic
Publishing
Verstraete, M. (1994b). Retrieving canopy properties from remote sensing measurements. In J.
Hill & J. Mégier (Eds.), Imaging spectrometry - a tool for environmental observations
(pp. 109-125). Dordrecht, Bostong, London: Kluwer Academic Publishers
Viedma, O., Melia, J., Segarra, D., & GarciaHaro, F.J. (1997). Modeling rates of ecosystem
recovery after fires by using Landsat TM data. Remote Sensing of Environment, 61, 383398
Viovy, N., Arino, O., & Belwards, A.S. (1992). The best index slope extraction (BISE): a method
for reducing noise in NDVI time series. International Journal of Remote Sensing, 13
Vogelmann, J.E., Helder, D., Morfitt, R., Choate, M.J., Merchant, J.W., & Bulley, H. (2001).
Effects of landsat 5 thematic mapper and Landsat 7 enhanced thematic mapper plus
radiometric and geometric calibrations and corrections on landscape
characterization. Remote Sensing of Environment, 78, 55-70
Vogt, P., Ferrari, J.R., Lookingbill, T.R., Gardner, R.H., Riitters, K.H., & Ostapowicz, K. (2009).
Mapping functional connectivity. Ecological Indicators, 9, 64-71
113
Remote Sensing Based Assessment
Vogt, P., Riitters, K.H., Iwanowski, M., Estreguil, C., Kozak, J., & Soille, P. (2007). Mapping
landscape corridors. Ecological Indicators, 7, 481-488
Wang, L., Sousa, W.P., Gong, P., & Biging, G.S. (2004). Comparison of IKONOS and QuickBird
images for mapping mangrove species on the Caribbean coast of Panama. Remote
Sensing of Environment, 91, 432-440
Wei, W. (1990). Time series analysis. Univariate and multivariate methods. : Addison-Wesley
Weissteiner, C., Böttcher, K., Mehl, W., Sommer, S., & Stellmes, M. (2008). Mediterranean-wide
Green Vegetation Abundances for Land Degradation Asessment Derived from AVHRR
NDVI and Surface Temperature 1989 - 2005. Luxembourg: European Commission
Welch, R., Jordan, T.R., & Ehlers, M. (1985). Comparative evaluations of the geodetic
accuracy and cartographic potential of Landsat-4 and Landsat-5 Thematic Mapper
Image Data. Photogrammetric Engineering & Remote Sensing, 51, 1249-1262
Wessels, K.J. (2007). Can human-induced land degradation be distinguished from the effects
of rainfall variability? A case study in South Africa. Journal of Arid Environments, 68
Wessels, K.J. (2009). Letter to the Editor: Comments on Proxy global assessment of land
degradation by Bai et al. (2008). Soil Use and Management, 25
Widmann, M., & Schär, C. (1997). A principal component and long-term trend analysis of
daily precipitation in Switzerland. International Journal of Climatology, 17, 1333-1356
Wulder, M.A., Franklin, S.E., White, J.C., Cranny, M.M., & Dechka, J.A. (2004). Inclusion of
topographic variables in an unsupervised classification of satellite imagery. Canadian
Journal of Remote Sensing, 30, 137-149
Wulder, M.A., White, J.C., Goward, S.N., Masek, J.G., Irons, J.R., Herold, M., Cohen, J.,
Loveland, T.R., & Woodcock, C.E. (2008). Landsat continuity: Issues and opportunities
for land cover monitoring. Remote Sensing of Environment, 112, 955-969
Yang, W., Huang, D., Tan, B., Stroeve, J.C., Shabano, N.V., Knyazikhin, Y., & Nemani, R.R.
(2006a). Analysis of leaf area index and fraction of PAR absorbed by vegetation
products from the Terra MODIS sensor: 2000-2005. IEEE Transactions on Geoscience
and Remote Sensing, 44, 471-475
Yang, W., Tan, B., Huang, D., Rautiainen, M., Shabanov, N.V., Wang, Y., Privette, J.L.,
Huemmrich, K.F., Fensholt, R., Sandholt, I., Weiss, M., Ahl, D.E., Gower, S.T., Nemani,
R.R., Knyazikhin, Y., & Myneni, R.B. (2006b). MODIS leaf area index products: from
validation to algorithm improvement. IEEE Transactions on Geoscience and Remote
Sensing, 44, 1885-1898
Zhang, X., Friedl, M.A., Schaaf, C.B., Strahler, A.H., Hodges, J.C.F., Gao, F., Reed, B.C., &
Huete, A.R. (2003). Monitoring vegetation phenology using MODIS. Remote Sensing of
Environment, 84
114
Assessing and Valuating Ecosystem Services and Mitigation Strategies
IV. A Methodological Framework for
Assessing and Valuating Ecosystem Services
and Mitigation Strategies against
Desertification
Working Paper for the PRACTICE Project
Laura Onofri
Andrea Ghermandi
Paulo A.L.D. Nunes
115
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
Laura Onofri, Andrea Ghermandi and Paulo A.L.D. Nunes (November 2010)
Department of Economics
University of Venice Cà Foscari
San Giobbe 823
30121 Venice
Italy
[email protected]
116
Assessing and Valuating Ecosystem Services and Mitigation Strategies
1 Introduction
Drylands are territories where ecosystems, human economic activities and
biodiversity coexist. Human activities and economic use of natural resources
often imply a deterioration of natural, ecological and geo-physical resources.
There is, in fact, an evident trade-off expressed in term of objectives.
Economic productive activities generate income, profits or guarantee survival
At the same time, economic productive activities can negatively affect
delicate ecological equilibria. It is recently recognized that sustainable
development requires the preservation and respect of ecological equilibria,
when performing economic activities. It is also recently recognized that
natural ecosystems provide themselves services that might be of support to
human activities. These services are called ecosystem services. Therefore, the
trade off does not imply anymore a corner solution, bur should be interpreted
as an equilibrium condition. The choice does not focus anymore on the tradeoff between economic growth vs. natural resources preservation, but it
becomes a choice on how much natural resources should be used, so that
economic growth costs (loss of natural resources) equal economic growth
benefits at the margin Economists, in fact, study human behavior in order to
valuate, explain and predict. In the economist’s framework, economic
agents are rational and pursue a rational objective/perform choices in a
constrained setting. Constrained choice implies that every choice has
benefits and costs. Economists assess and measure, when feasible, the
monetary value of benefits and costs. The measurement of a monetary value
is straightforward when there is an institution (market), where demand and
supply meet. The market value is defined by several indicators, the most
important of which is the trade price. In addition, economists are able to
compute consumers’ surplus, producer surplus, total surplus. Non-market
values are more difficult to compute, since there is not an institution where
trade occurs and no price exists. Therefore, in order to assess non-market
values (benefits), it is important to adopt valuation methodologies that mimic
market behavior, in order to elicit stakeholders’ preferences for the good,
expressed in monetary terms (willingness-to-pay). Ecosystem services produce
both market and non-market values. In order to assess and measure
ecosystem services we need some classifications and a lot of data and
information.
The paper is organized as follows. Section 2 provides some definitions and
classifications. Section 3 surveys the most important valuation techniques
adopted in economic analysis, by highlighting costs and benefits of each
methodology Section 4 introduces, surveys and critically discusses socioeconomic indicators for assessing ecosystem services and anti-desertification
strategies. In addition, the section presents a meta-indicator that conveys
synthesized information with some causality content. Section 5 concludes.
117
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
2. Some taxonomy.
The paragraph introduces some basic taxonomy for defining and explaining
technical economic and valuation concepts. It is thus effectively the
application of ‘ecosystem valuation’ (7) by analysts to recognize and manage
the dependence and impact of human activities upon ‘ecosystem services’.
This is achieved through better understanding the economic value of
ecosystem services, and determining which stakeholders derive those values.
Such information can then be incorporated within the decision-making to
help the stakeholders to enhance their objectives and/or implement
sustainable development solutions. In this section we provide some taxonomy.
‘Ecosystem Services’ are ‘the benefits people obtain from ecosystems’. The
benefits can be broken down into four categories that include:
–
Provisioning services. The benefits that ecosystems provide in the form of
‘products’ or ‘goods’ that are consumed by humans or used in the
production of other goods. They include things such as fish, nuts, timber,
water and genetic resources;
–
Regulating services. The benefits obtained from an ecosystem’s control of
natural processes such as climate, disease, erosion, water quality and
flows, and pollination, as well as protection from natural hazards such as
storm and wave damage. “Regulating” in this context is a natural
phenomenon and is not to be confused with government policies or
regulations. They are ecosystem ‘functions’ and ‘regulatory processes’
that includes vegetation storing carbon, wetlands slowing down water
flows and cleansing water, and coral reefs and mangroves protecting
coastal infrastructure from erosion and storm damage;
–
Cultural services: The non-material benefits people obtain from
ecosystems such as recreation, spiritual values, and aesthetic enjoyment.
–
Supporting services: The natural processes such as nutrient cycling and
primary production that maintain the other services.
The value of supporting services is captured within the value of the above
three services and so should NOT be valued separately. ‘Ecosystem Valuation’
is defined in this context as ‘the placing of monetary values on ecosystem
services and associated changes in their quantity and or quality’. These
changes may be positive or negative. It is important to recognize that this
does not include the valuation of all ‘environmental externalities’8 only those
(7) Often referred to as 'environmental valuation'.
8 ‘Environmental externalities’ are ‘environmental impacts caused by one party that affect another, but which are
not accounted for financially by the party that caused the impact due to a lack of market price’. For example, a
company emitting pollution will typically not take into account the costs that its pollution imposes on others in terms
of damaged health or climate change costs. Environmental externalities include impacts to ecosystems, but also to
other receptors such as humans, buildings and structures (including cultural assets) and economic activities. In
addition, externalities can be either positive or negative. The impacts typically relate to adverse implications
associated with air emissions, discharges, spills, land-take, noise, sedimentation and waste disposal etc. However,
positive externalities occur when a company makes an environmental improvement (eg protecting a habitat) which
they do not benefit from directly.
118
Assessing and Valuating Ecosystem Services and Mitigation Strategies
specifically related to ecosystems. For example, this would include assessing
the value that carbon emissions have on ecosystems (eg agriculture and
environmental related tourism), but not the value of impacts to other
economic activities, buildings or people directly.
3 Total economic value and ecosystem services economic tools
Ecosystem services are crucial to human well-being (MEA, 2000) and are
associated to a wide range of benefits to human society. The value of
ecosystem services can be assessed in terms of their impacts on the provision
of inputs to production processes; in terms of their direct impact to human
welfare and in terms of their impact on the regulation of the nature
ecological functions relationship. Usually, market mechanisms that price the
total value of ecosystem services, given their complexity, are lacking.
Therefore, the economic valuation of ecosystem services require the use of
special valuation tools, in particular it requires the use of a holistic, integrated
assessment.
The paragraph surveys the most important valuation techniques adopted in
economic analysis, by highlighting costs and benefits of each methodology.
A well recognized and used framework for putting monetary values on
ecosystem services is that of ‘Total Economic Value’. As illustrated in Figure 1,
this categorizes the different ‘ecosystem services’ into the following type of
economic value:
•
Direct use values: These include raw materials and physical products
that are used directly for production, consumption and sale such as
those providing energy, shelter, foods, agricultural production, water
supply, transport and recreational facilities. These are effectively the
same as ‘provisioning services’ and recreation related ‘cultural
services’.
•
Indirect use values: These include the ecological functions that
maintain and protect natural and human systems through services
such as maintenance of water quality and flow, flood control and
storm protection, nutrient retention and micro-climate stabilization,
and the production and consumption activities they support. They are
effectively equivalent to ‘regulatory services’.
•
Option values: This is the ‘premium’ placed on maintaining a pool of
habitats, species and genetic resources for future possible uses, some
of which may not be known now, such as leisure, commercial,
industrial, agricultural and pharmaceutical applications. This type of
value potentially applies to each of the three main services.
•
Non-use values: This is the value of ecosystems regardless of their
current or future use, for cultural, spiritual, aesthetic, heritage and
119
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
biodiversity reasons. They comprise ‘existence’, ‘altruistic’ and
‘bequest’ values. For example, people are willing to pay to protect
whales and rainforests even though they may never use or see them in
the wild themselves. Non-use values can be significant, particularly for
maintaining unique ecosystems where large populations may be
willing to pay to protect them.
Figure.1.
Ecosystem Services and Total Economic Value
There are four main applications (or approaches) for ecosystem valuation (9)
each trying to answer different types of questions.
The most important and commonly used application is to evaluate ‘tradeoffs’ between different options.
This approach typically uses cost-benefit
analysis to compare the full range of costs and benefits over time associated
with different options.
For example, one design option for a new
development may result in a loss of 2 ha of wetlands, 1 ha of seagrass and 1
ha of sand dune.
Another option may result in the loss of 5ha of wetland. If all other things are
equal, which is the best option? From a purely economic perspective, the
answer depends on the difference in aggregated ‘marginal’ cost of losing
each ha of habitat. Ecosystem valuation can help evaluate that trade-off
through converting the impacts to a single unit of currency (money), although
other factors may also influence the ultimate decision.
The other applications include: attempting to estimate the total value of an
environmental asset; determining the distribution of benefits amongst
stakeholders, and identifying potential sources of surplus based on who
(9) Pagiola, S,von Ritter, K, and Bishop, J (2004) Assessing the Economic Value of Ecosystem Conservation. The World Bank Environment
Department. In collaboration with The nature Conservancy and IUCN.
120
Assessing and Valuating Ecosystem Services and Mitigation Strategies
benefits (or loses). Assessing values for the latter tends to involve combining
outputs from i) or ii), plus iii) below.
It is important to stress that there exist a hierarchy of valuation approaches.
Ecosystem ‘valuations’ may be in qualitative, quantitative or monetary terms.
Qualitative valuation typically involves describing the value as well as
indicating whether the value is likely to be of high, medium or low economic
value (eg a fishery value of a wetland may be medium whilst the non use
value may be high). Quantitative valuation involves describing the nature of
the value in terms of relevant quantitative information (eg the wetland is used
by 10 fishermen who catch 5 tons of fish per year, and its existence in good
quality is of significance to 50,000 local people).Monetary valuation actually
involves placing a ‘monetary’ or ‘dollar value’ on the impact (eg the wetland
generates US$ 2,000/year from its fish and the locals are willing to pay US$
500,000/year for its continued protection).
3.1 A survey of economic valuation techniques
As a result of most ecosystem services and environmental impacts not having
a monetary value expressed in a market-place/market-price, a diverse range
of environmental valuation techniques have been developed over the past
twenty to thirty years. This has historically been so that monetary values of
environmental assets and impacts can be assessed alongside other
economic and financial values for public decision-making purposes, to
encourage more sustainable outcomes.
There are three main categories of valuation techniques: Revealed
preference approaches; Cost-based approaches and Stated preference
approaches. These are briefly summarized in the table below.
Table 2. Monetary Based Ecosystem Valuation Techniques. Source: adapted from
‘WBCSD Corporate Ecosystem Valuation Scoping Study’
Category
Revealed
preference
approaches:
Technique
Description
Example
Market
prices
Market prices
How much it costs to buy an
ecosystem good or service, or
what it is worth to sell.
The market price of timber,
fish or water.
Production
function
approaches
Effect on
production
Relates changes in the output of a
marketed good or service to a
measurable change in ecosystem
goods.
The reduction in fishery
output as a result of
clearing a mangrove or
saltmarsh.
Travel costs
Using information on the amount
of time and money people spend
visiting an ecosystem for
recreation or leisure purposes to
elicit a value per visit.
The transport and
accommodation costs,
entry fees and time spent
to visit a National Park.
Hedonic
pricing
The difference in property prices or
wage rates that can be ascribed
to the different ecosystem qualities
or values.
The difference in prices
between houses
overlooking an area of
natural beauty or water
features compared to
similar houses that do not.
Look at the way in
which people reveal
their preferences for
ecosystem services
through market
Surrogate
production and
market
consumption
approaches
121
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
Category
Technique
Description
Example
Replacement
costs
The cost of replacing an
ecosystem good or service with
artificial or man-made products,
infrastructure or technologies, in
terms of expenditures saved.
The costs of flood
protection infrastructure
after the loss of catchment
protection forest
Mitigative or
avertive
expenditures
Additional purification
Expenditures required to mitigate
infrastructure required to
or avert the negative effects of the
maintain water quality
loss of ecosystem services (similar
standards after the loss of
to replacement costs).
natural wetlands
Cost-based approaches:
Look at the market trade-offs or
costs avoided of maintaining
ecosystems for their goods and
services
The costs incurred to property,
infrastructure and production
Damage costs when ecosystem services which
avoided
protect economically valuable
assets are lost, in terms of
expenditures saved.
Contingent
valuation
Infer ecosystem values by asking
people directly what is their
willingness to pay (WTP) for them or
their willingness to accept (WTA)
compensation for their loss saved.
How much would you be
willing to contribute
towards a fund to clean up
and conserve a river?
Choice
experiments
Presents a series of alternative
resource or ecosystem use options,
each defined by various attributes
set at different levels (including
price), and asks respondents to
select which option (ie sets of
attributes at different levels) they
prefer.
Respondents’ preferences
for conservation,
recreational facilities and
educational attributes of
natural woodlands.
Stated preference
approaches:
Ask consumers to state their
preference directly
The damage to roads,
bridges, farms and property
resulting from increased
flooding after the loss of
catchment protection
forest
In addition, a number of other ‘valuation’ techniques can be used, as
detailed in Table 3. These include ‘benefit transfer’ (or value transfer), which is
becoming an increasingly common means of valuation due to its relative
ease and cost-effectiveness. Benefit transfer is simply taking an economic
value determined for a good in one context (e.g . improvement in
environmental quality) and applying it to another site and context. Another
technique, which is commonly used in determining biodiversity offsets, is
‘Habitat equivalency analysis’. This involves determining the quantity of new
habitat required (ie restored or created) to offset the loss of goods and
services from a damaged area of similar habitat.
Table 3. Other Ecosystem Valuation Techniques. Source: adapted from ‘WBCSD
Corporate Ecosystem Valuation Scoping Study’
Category/
approach
Benefit (or
Value) Transfer
Technique
Description
Unit value (eg average This technique applies average
willingness to pay
WTP values taken from other
values)
studies. Ideally these values
are adjusted to account for
key differences in context such
as income levels and degree of
impact.
Example
The average willingness to
pay of recreational visitors to
one woodland in UK applied
to another similar woodland
in UK. If applied to a wood in
Poland, converted using
difference in GDP/capita
factor.
122
Assessing and Valuating Ecosystem Services and Mitigation Strategies
Category/
approach
Technique
Description
Example
WTP functions
This technique uses the benefit
function (or ‘bid’ function)
which is the formula that
explains the WTP value in terms
of key characteristics (eg
environmental and socioeconomic such as incomes).
Insert specific site related
variables (eg average
income, level of education)
into the WTP bid function for
visitor s to a woodland
determined for a similar
woodland elsewhere.
Meta analysis
This technique takes the results
from a number of studies and
analyses them in such a way
that the variations in WTP in
those studies can be
explained.
Analysis of many WTP studies
for woodlands to derive
trends in the key variables
affecting visitor WTP values
for woodlands, to establish a
suitable value or adjustments
for the site to be assessed.
Habitat equivalency
analysis
This techniques uses a scaling
process to determine the
amount of environmental
compensation required based
on units of habitat damaged
and created.
The approach may
determine that 3.5m2 of coral
reef needs to be replaced for
every 1m2 of coral reef
damaged.
The amount of environmental
compensation required is
based on units of resource
damaged and created.
500 trout need to be
replaced in a river damaged
by a pollution incident.
The amount of environmental
compensation required is
based on either ‘value to cost’
(where the cost of remediation
measures is set to the value of
damages) or ‘value to value’
(where the value of
remediation measures is set to
the value of damages).
The value of damages to
corals based on a stated
preference survey reveals a
loss of US$ 1 million, which is
then invested in coral
remedial and compensatory
measures that cost US$ 1
million or that provide US$ 1
million of benefits.
Resource equivalency
Natural
analysis
resource
damage
assessments
and biodiversity Value equivalency
analysis
offsetting
A number of other valuation techniques may be relevant for valuing
environmental externality impacts to human health. These include for
example, the cost of illness, loss of earnings, Quality Adjusted Life Years, but
are outside the scope of these guidelines.
Finally, a variety of new ‘tools’ are being developed to assist with ecosystem
valuation studies. These include various innovative web and Geographic
Information System (GIS) based approaches. Some have been developed
initially with public decision-making in mind (eg for governments assessing
economic trade-offs comparing net benefits to society as a whole from
different alternatives), and others with specific business applications. GIS can
be extremely useful to map out the economic aspect and environmental
features. It can help illustrate the assessment from a visual perspective and
assist with calculations such as determining the precise area of habitat
affected and the number of household or people within different distance
123
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
bands from an impact. The latter can be especially useful when assessing
non-use values.
3.2. Advantages, disadvantages and applicability of ecosystem services
valuation techniques
This section provides guidance on selecting ecosystem valuation techniques
and highlights some typical steps and issues associated with a number of the
more common valuation techniques likely to be used in valuation exercises.
The Table below indicates which valuation technique is more typically used to
value which category of ecosystem service.
Table 4. Application of Valuation Techniques for Different Types of Ecosystem
Services. Source: adapted from ‘WBCSD Corporate Ecosystem
Revealed preference
Total
Economic
Value
Ecosystem Services
Direct use
Provisioning
Indirect use
Regulating
Direct use
Non use
Cultural
Market
prices
Effect on
prod-uction
Travel Hedonic
costs pricing
Cost- Stated
Benefit
based preference transfer
Recreation
Aesthetic
However, in many situations there is a choice of technique, and it then
depends on the study context and circumstances as to the preferred
technique. For example, aspects such as accuracy required, and budget,
data and time availability may play a role. Note that the budgets (including
data costs), timelines, and skills required can vary significantly depending on
data already available, nature of the issue and who is involved in undertaking
the study.
Finally, Table 5 summarises some of the key features of the main ecosystem
valuation techniques.
124
Assessing and Valuating Ecosystem Services and Mitigation Strategies
Table 5. Comparison of Ecosystem Valuation Techniques. Source: adapted
from ‘WBCSD Corporate Ecosystem Valuation Scoping Study’
Technique
Advantages
Disadvantages
Data required
Market prices
+ A readily transparent
and defensible
method since based
on market data.
- Only applicable where a
market exists for the ecosystem
service and data is readily
available.
Market price to buy an
ecosystem product.
- Necessary to recognise and
understand the relationship
between the ecosystem
service and output of product.
Data on changes in the
output of a product.
+ It reflects an
individual’s willingness
to pay (WTP).
Effect on
production
+ If data is available, it
is a relatively
straightforward
technique to apply.
- Can be difficult to obtain
data on both change in the
ecosystem service and effect
on production.
Travel costs
+ Based on actual
behaviour (what
people do) rather than
a hypothetically stated
WTP.
+ The results are
relatively easy to
interpret and explain.
- Approach is limited to direct
use recreational benefits.
- Difficulties in apportioning
costs when trips are to multiple
destinations or are for more
than one purpose.
The costs involved to
process and bring the
ecosystem product to
market (e.g. processed
timber).
Data on cause and effect
relationship (eg loss of
fisheries due to loss of
seagrass or coral habitat).
The amount of time and
money that people
spend visiting an
ecosystem for recreation
or leisure purposes.
Motivations for travel.
- Considering travel costs alone
ignores the opportunity cost of
time whilst travelling (e.g.
should a proportion of the
individual’s hourly salary be
added?)
- Some individuals may have
moved house to live closer to
the site, reducing their travel
costs and therefore under
estimating their valuation of the
site.
Hedonic pricing
+ Readily transparent
and defensible
method since based
on market data and
WTP.
+ Property markets are
generally very
responsive so are good
indicators of values.
Replace-ment
costs
- Approach is largely limited to
benefits related to property.
- The property market is
affected by a number of
factors in addition to
environmental attributes, so
these need to be identified
and discounted.
+ Provides surrogate
measures of value for
regulatory services
(which are difficult to
value by other means).
- Can overestimate values.
+ A readily transparent
and defensible
method when based
on market data.
- The replacement service
probably only represents a
proportion of the full range of
services provided by the
natural resource.
- The technique does not
consider social preferences for
services or behaviour in the
absence of the services.
Usually data relating to
differences in property
prices or wage rates that
can be ascribed to the
different ecosystem
qualities (e.g. a
landscape view, distance
to water feature).
The replacement costs of
a proxy for the ecosystem
service. This needs to be
at least as effective as
the ecosystem service
and be the least cost
alternative. Society must
also have demonstrated
a willingness to pay for
the project.
125
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
Mitigative or
avertive
expenditures
+ Provides surrogate
measures of value for
regulatory services.
- Can overestimate values.
+ When goods are
non-marketed, it can
be easier to value the
costs of producing
benefits than the
benefits themselves.
Damaged costs
avoided
Contingent
valuation
+ Provides surrogate
measures of value for
regulatory services that
are difficult to value by
other means (eg storm,
flood and erosion
control).
+ Enables values to be
captured for both use
and non-use values.
+ Extremely flexible
since it can be used to
estimate the economic
value of virtually
anything.
+ Will give a much
more accurate
outcome than benefit
transfers.
- The approach is largely
limited to services related to
properties, assets and
economic activities.
- Can overestimate values.
- The results are subject to
numerous different bias from
respondents and hypothetical
in nature.
- E.g. respondents may express
a positive willingness to pay to
promote a ‘warm glow’ effect,
overestimating valuations.
- E.g. if the cost is perceived as
a government tax, respondents
may express a negative
willingness to pay,
underestimating valuations.
Data on the expenditures
required to mitigate or
avert the negative effects
of the loss of ecosystem
services.
Data on costs incurred to
property, infrastructure or
production as a result of
loss of ecosystem services.
Damages under different
scenarios including ‘with’
and ‘without’ regulatory
service.
Data on the amounts that
people would be willing
to pay for an ecosystem
service, or conversely,
what they would be
willing to accept as
compensation for loss of
an ecosystem service.
Obtained by asking
individuals to state their
preference directly.
- Non-use values derived can
still be questionable
- It is resource intensive.
Choice
experiments
+ Enables values to be
captured for both use
and non-use values.
+ Provides theoretically
more accurate values
for marginal changes
(eg values per percent
increase in coral
cover).
+ Will give a much
more accurate
outcome than benefit
transfers.
Benefits transfer
+ Low cost and rapid
method for estimating
recreational and nonuse values.
- The results are subject to bias
from respondents, is resource
intensive and hypothetical in
nature.
- It is resource intensive.
- Can be mentally challenging
for respondents to truly weigh
up the alternative choices
given to them in the time
available.
- The results can be
questionable unless carefully
applied.
Data on the individual
preferences of people
when presented with a
series of alternative
resource or ecosystem
use options.
Valuations from similar
studies elsewhere.
Data on key variables
from different studies (eg
GDP per person).
126
Assessing and Valuating Ecosystem Services and Mitigation Strategies
4 Valuing drylands ecosystem services and anti-desertification
policies
Within the context of the mounting global concern caused by land
degradation in drylands (defined as desertification in the text of the United
Nations Convention to Combat Desertification), desertification generates a
persistent reduction in the services provided by drylands ecosystems, leading
to unsustainable use of the drylands and their impaired development.
‘‘Desertification’’ means land degradation in arid, semiarid, and dry
subhumid areas resulting from various factors, including climatic variations
and human activities. Land degradation means reduction of or loss in the
biological or economic productivity and complexity of rain-fed cropland,
irrigated cropland, range, pasture, forest, or woodlands resulting from land
uses or from processes arising from human activities and habitation patterns
(UNCCD 1992). Drylands Ecosystem services might contribute (together with
other exogenous factors, including political marginality, slow growth of health
and education infrastructure, facilities and services) to directly affect humanwell-being. The Millennium Assesment conceptual framework (MA 2003)
broadly identifies five key components of human well-being.
• Basic materials for a good life. The low biological production of drylands,
constrained by water, is the ecosystem condition that limits the provision of
basic materials for a good standard of living. This also limits the livelihood
opportunities in drylands and often leads to practices, such as intensified
cultivation, that cannot be serviced due to low and further impaired nutrient
cycling and water regulation and provision, requiring adjustments. in
management practices or the import of nutrients and water provided by
services of other ecosystems.
• Health. The two key factors contributing to poor health in drylands are
malnutrition and limited access to clean drinking water, again reflecting their
low biological production and water provision. In Asia, for example, the
fraction of children under the age of five facing hunger is 36% in drylands
compared with 15% in the forest and woodland system (averaged for the
different subtypes of each of these systems) However, poor health is also
exacerbated by poor health-related infrastructures.
• Good Social Relations. The quality of social relations can be gauged in
terms of social strife (wars and political upheavals) and refugees.
Environmental refugees leave their homes due to environmental degradation
and lack of viable livelihoods. And the sale of stock, wage labor, borrowing of
cash for food, and the sale of valuables all precede their migration. Other
categories include people displaced for political reasons that may affect the
availability of services in drylands to which they have been relocated. Thus
demography and sociopolitical drivers, more than the direct condition of
ecosystem services, contribute to the quality of social relations.
• Security. Food security is an essential element of human wellbeing in
drylands and is related to socioeconomic marginalization, lack of proper
127
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
infrastructure and social amenities, and often the lack of societal resilience.
Climatic events like prolonged droughts and excessive floods also drive
insecurity in drylands. But sociopolitical drivers like land tenure practices that
relate to sharing and conservation of natural resources or that generate land
cover change that may limit traditional pastoral livelihood opportunities can
greatly affect food security.
• Freedom and Choice. Dryland people are not affected just by the unique
condition of the ecosystem but are also further restricted by local and global
political and economic factors, which can exert major limitations on their
freedom and life choices. With the exception of OECD countries, dryland
peoples are mostly politically marginalized; that is, their role in political
decision-making processes is perceived as being insignificant. Consequently,
market factors that determine dryland farmers’ decision-making and the
effects on their well-being are often critical.
What needs to be assessed, in this case, is whether or not there are thresholds
in the provision of services that, either independently or together with the
exogenous factors, cause human well-being to drop to a level that crosses
the poverty threshold or dramatically decrease welfare conditions.
In addition, the recognized importance of ecosystem services on human
activities has spurred to take action and starts implementing measures that
limit desertification in drylands. This part of socio-economic analysis, valuation
approach and technique choice should consider the impact of mitigation
and adaptation strategies in dryland selected areas, in particular, those
prevention and restoration actions that combat desertification in those areas.
The analysis can vary according to available data and valuation approach
and technique. However, the focus should be on comparing different
situations before and prevention/restoration action and valuate costs and
benefits of those actions. It is, therefore, important to valuate the socioeconomic impacts of several dimensions:
1) The (total) economic benefit of a dryland systems with respect to the
services they provide and the drivers that determine trends in their
provision;
2) The economic loss/costs that desertification is causing to delicate
dryland systems;
3) The economic impacts of mitigation and adaptation strategies to
desertification.
In order to perform an economic-valuation exercise, it is important to gather
data and select the proper valuation approach (qualitative, quantitative
and/or monetary) and technique (as surveyed in the previous paragraph).
Often the choice of the valuation approach and technique depends on
data availability and on the possibility to “build” data (like WTP)10. Integrated
assessment of ecosystem services in drylands before, during and after the
10
See Appendix 1.
128
Assessing and Valuating Ecosystem Services and Mitigation Strategies
policy should be the preferred valuation methods. However, for the sake of
simplicity, for the difficulty to gather information and for the costly
management of economic valuation techniques, socio-economic indicators
may represent an interesting tool for assessing anti-desertification policies
performance.
4.1 Evaluation of anti-desertification policies through socio-economic
indicators
In general, a policy evaluation process should evaluate the “success” (or
failure) of a given policy (in a certain sense it is the performance
measurement of public policies). However it is always important to find a solid
causal nexus/relationship between policy/measures and observed
facts/phenomena. For instance, suppose that in the period after the
reforestation policies in a selected area, tourism (measured in terms of
number of presences) has increased by 20%. Shall we conclude that the
mitigation/anti desertification policy has been effective (on a socioeconomic perspective) because a better environment has increased
tourism? Is this statement correct? The answer is: it depends!
Suppose that a policy generates a (potential) Outcome A and an observed
Outcome B. An evaluation comparing (potential) Outcome B and (observed)
Outcome A is a counterfactual evaluation. An evaluation considering only
Outcome B is a sort of “qualitative” evaluation; it is very common in the
evaluation of European Cohesion Policy. It suffers from problems of causality.
Describing observations via summaries, as in the previous example, is a critical
part of most evaluation research. But it is often not the primary goal. The goal
rather, is inference – the process using the facts we know to learn about facts
we do not know. There are two types of inference: (1) descriptive and (2)
causal11. To identify the impact of the policy on Y we need to consider also X
and Z (e.g. in the tourism example, we should consider the general economic
conditions). We need a counterfactual (i.e. observations without the policy).
When adopting socio-economic indicators for assessing the impact of policies
it is also very important to stress that:
Correlation ≠ causality
(in general, we need a theory or argument to link one variable to another! Or
we need a deep evaluation of inter-correlated phenomena!).
11
(1) Descriptive inference is different than data summary. The aim is still to find out
“something we do not know” (i.e. a particular function, the impact of a given
policy). By using case studies and data summary, researchers should prove the
existence of “links” (see e.g. meta-analysis).
(2) In causal inference, researcher wants to know whether one factor or a set of
factors leads to (or causes) some outcome. In general, causal inference is the
difference between two descriptive inferences.
129
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
Another point of policy economic evaluation refers to the timing of the
evaluation exercise. We can distinguish three types of valuation linked to the
timing:
a. Ex ante valuation (before the policy): cost benefit analysis
b. Ex post valuation (after the policy, success of the policy):
effectiveness
c. In fieri valuation (during the policy): monitoring.
It is crucial to emphasise the importance of framing the socio-economic
analysis within a holistic perspective, well represented by ecosystem services
valuation (the before and after the implementation of anti-desertification
practices). This approach (shortly summarized in Section 3) allows embracing
all the multidimensional implications of the problem at issue, without giving up
on its complexity or unduly broadening the scope of the analysis. However,
adopting complex valuation techniques is a costly research activity.
Therefore, another method to evaluate phenomena (desertification in the
case at issue) is the adoption of socio-economic indicators.
Broadly speaking, “an indicator can be defined as the measurement of an
objective to be met […]. An indicator produces quantified information with a
view to helping actors concerned with public interventions to communicate,
negotiate or make decisions. Within the framework of evaluation, the most
important indicators are linked to the success criteria of public interventions”
(EVALSED, p. 111).
In the various lists of desertification indicators that have been developed over
the years, the characterization of the socio-economic context against which
to evaluate the interventions to combat desertification is invariably
acknowledged as a crucial component next to environmental indicators. This
is natural if one considers that desertification is the result of both natural and
anthropogenic pressures, which lead to environmental degradation and loss
of biological and economic productivity (Rubio and Bochet, 1996).
According to those authors, the main purposes of desertification indicators
are the following:
(1) assess the current environmental and economic state;
(2) determine trends over time;
(3) detect detrimental stages in early stage to prevent irreversible
damage; and
(4) identify causes in order to specify proper management actions.
The UN Convention to Combat Desertification (UNCCD) has identified
desertification indicators as a necessary tool for both management and
monitoring the implementation of strategies to combat desertification.
Among the desirable characteristics of desertification indicator systems
according to UNCCD is simplicity. Lists with a limited number of indicators that
broadly apply to many stressors and sites are generally preferable, since they
tend to be simpler and more ready to use by the parties concerned (UN,
130
Assessing and Valuating Ecosystem Services and Mitigation Strategies
1997). At the same time, the indicators must be sensitive enough to detect
important changes without being by natural variability (MEA, 2000). When
extended lists are produced, a list of key indicators for common purposes
should be indicated and it should be possible for final users to select among
the remaining indicators only those that are most relevant for their purposes.
Reliability and objectivity of indicators are a crucial concern. According to
Reed et al. (2006), indicators should be accurate and bias-free, reliable and
consistent over space and time, scientifically robust and credible, verifiable
and replicable. They should provide timely information and be sensitive to
system stresses or the changes they are meant to indicate.
In the selection of the indicators, cost-effectiveness in the collection and
analysis of the data is an important concern since they may require
substantial resources in terms of finance, manpower, and time. Preferably,
already available information concerning measures undertaken and their link
with the rehabilitation of productive dryland ecosystems should be used.
Indicators characterized by measurability and ease to be operationalized
should be clearly preferred over purely conceptual ones, and they should be
related as much as possible to the existing onsite practices (Brandt et al.,
2003). They should be able to reflect spatial change in land function and over
time, i.e., to assess the present status as well as to identify time trends. A target
level, baseline or threshold against which to measure them should be defined
(Reed et al., 2006).
It is important to keep in mind that indicators are an aid to decision-making
and not an end in themselves. Therefore their understandability by policymakers is an important concern. Indicators should be clear and
unambiguous, simplify complex phenomena and facilitate communication of
information. They should measure what is important to stakeholders and
possibly have social appeal and resonance (Reed et al., 2006).
Several international bodies have worked on the definition of environmental
indicator systems, which bear importance in evaluating the extent, severity
and socio-economic aspects of desertification. Such institutions include:
The UN Convention to Combat Desertification (UNCCD,
http://www.eea.europa.eu/data-and-maps/indicators);
The Organization for Economic Cooperation and Development (OECD,
www.oecd.org/dataoecd/32/20/31558547.pdf);
The United Nations Development Programme (UNDP,
http://www.undp.org/fssd/crosscutting/sustdevmdg.html);
The United Nations Environment Programme (UNEP,
http://www.fao.org/docrep/w4745e/w4745e07.htm);
The World Bank (http://web.worldbank.org);
The European Environment Agency (EEA, http://www.eea.europa.eu/dataand-maps/indicators);
The World Meteorological Organization (WMO, http://www.wmo.int);
131
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
The UN Food and Agriculture Organization (FAO,
http://www.fao.org/ag/agl/agll/lada).
The investigation of the indicator systems put forward by such organizations or
proposed
by
other
EU
projects
such
as
Desertlinks
(http://www.kcl.ac.uk/projects/desertlinks)
and
Medalus
(http://www.kcl.ac.uk/projects/desertlinks), one can identify the common
trends in socio-economic desertification indicators.
Although the actual indicator may differ according to the spatial scale of
interest (e.g., global, national or project level), we can in general distinguish
four main categories of socio-economic desertification indicators:
(1) Resource indicators provide information on the means used to implement
programmes. For instance:
a. information about monetary investments for anti-desertification
policies in €;
b. or information about other resources (e.g. number of workers);
c. time frame;
d. costs of inaction.
e. Opportunity cost of the policy
(2) Output indicators represent the product of the programme activity.
Everything obtained in exchange for public expenditure. For instance:
a. hectares of reforested area or dryland;
b. information about productivity and profitability of the forest
area/land;
N
7
N
provisioningservice
Pr oductivity dryland
= ∑∑ TotalOutput in / ∑ Forest − LandArea n
n =1 i =1
n =1
Where i and n represent the type of forest/agricultural product and the
number of countries/regions located in each dryland area12.
In particular, current productivity values of forests in terms of provisioning of
the 7 types of Wood Forest Products (WFPs), i.e. industrial round wood, wood
pulp and recover paper, other processed wood products, sawn wood, wood-
12
Forest/Land productivity refers to productivity calculation when the land/forest is used to produce
marketable products (i.e agriculture, breeding). The same index can be calculated by considering
data on Harvesting/Silvicultural practices, when the land/forest is used to produce non-marketable
products, in order to highlight different resource productivity when human activities intervene or do not
interve;
132
Assessing and Valuating Ecosystem Services and Mitigation Strategies
based panels, paper and paperboard and wood fuel can be synthesized by
the following:
N
7
N
provisioningservice
Pr oductivityValue geo
− c lim aticregion = ∑∑ ExportValuein / ∑ ForestArea n
n =1 i =1
n =1
Where i and n represent the type of forest product and the number of
countries located in each geo-climatic area. This index is used by Fao, for the
valuation of forests provisioning services. Another way to calculate the (land)
productivity value is:
Pr oductivityValue = Pr oductivity * marketpriceoftheproduct
Where the land productivity (per hectare or squared meter) is multiplied by
the market price of a selected product cultivated in the land.
Data on Forest/Land profitability (how much profit/income is produced per
ha/squared meter, when the land is used for productive activities);
N
7
N
provisioningservice
Pr ofitability dryland
= ∑∑ ExportValue − Pr oftis in / ∑ Forest − LandArea n
n =1 i =1
n =1
Where i and n represent the type of forest/agricultural product and the
number of countries/regions located in each dryland area
c. soil water conservation measures;
d. hectares of land converted to agricultural use;
e. diversification of land-use and land-use intensity.
(3) Result indicators are the immediate advantages of the programme for
the direct beneficiaries. For instance:
a. decrease in child mortality or people migration in sub/Saharian areas
at risk of desertification;
b. unemployment rate;
c. net farm income;
d. water availability per capita;
e. population living above poverty level.
133
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
(4) Impact indicators are the consequences of the programme beyond its
direct and immediate interaction with the beneficiaries. For instance:
a. increase of the overall economy performance of the area (impact
on GDP/GNP per capita; total and social investment as a percentage
of GDP; development of new economic activities; tourism change and
number of tourists);
b. impact on social context of the local population (changes in
migration patterns and intensity; changes in population ageing patterns
and age structure; family size; population density).
4.2 Desertification policy assessment: A cost-effective meta indicator
For the Practice Project purposes, given the heterogeneity of the many case
studies and data sources and given the difficulties to gather the same kind of
information, we propose a cost-effective indicator:
CEI =
monetary cos tsofthepolicy − monetary cos tsofinaction
effectsofthepolicy − effectsofinaction
Let us analyse this “meta-indicator. On the numerator there is the difference
between the monetary costs of the policy (i.e. Euros transferred for planting
new trees, building irrigation systems/pipelines and so on) and the monetary
costs of inaction, that (obviously assume to be zero. No policy, no money
transfer, no public expenditure). It is very important to highlight that in this
simplified ratio we do not account for opportunity costs13. The denominator is
the difference between the (potential, expected) gain on the socioeconomic milieu from a selected measure (GDP of the area; child mortality
and many others) and the gains/losses, interpreted as the started benchmark.
Suppose that the costs of the mitigation policy are X units of money. So the
numerator is positive. Suppose that we want to evaluate the effects of the
policy in terms of impact on agriculture (the choice of the denominator is
discretional to the researcher). Anti-desertification measures are assumed to
improve agriculture. This might be done in terms of increased (if any) hectares
of cultivable land or in terms of increased income (if any) from the agricultural
13 James Buchanan has defined opportunity costs as expressing "the basic relationship between scarcity
and choice." The notion of opportunity cost plays a crucial part in ensuring that scarce resources are
used efficiently. Thus, opportunity costs are not restricted to monetary or financial costs: the real cost of
output forgone, lost time, pleasure or any other benefit that provides utility should also be considered
opportunity costs. Opportunity cost is, therefore, the cost related to the next-best choice available to
someone who has picked between several mutually exclusive choices. It is a key concept in economics
134
Assessing and Valuating Ecosystem Services and Mitigation Strategies
sector in the selected area.; or in terms of increased productivity of land or
farmers. Suppose that the numerator is always positive (money for the policy >
0 ⇒ inaction implies no money transfer for the policy). Suppose that our data
(gathered information) show that the GDP per capita in the agricultural
sector has increased with respect to the moment when no action was taken.
We can/cannot infer causality, but this is not the purpose of our indicators,
whose primary aim is “comparing” in a simple and synthetic informational
vehicle (the indicator) phenomena occurring before and after the policy.
Suppose that the denominator is positive because the gains from the policy
are larger than the effects of inaction on the selected informative indicator
(the agricultural sector). Since the numerator is always positive, this would
generate a positive CEI. Suppose now that the denominator is positive
because the gains from the policy are smaller than the effects of inaction on
the selected informative indicator (the agricultural sector). For instance, after
the policy action is taken the farmers’ GDP decreases. Since the numerator is
always positive, this would generate a negative CEI. Finally, suppose that the
denominator is zero because the gains from the policy equal the effects of
inaction. In this case, the CEI is zero.
This kind of simple indicator is very flexible and allows for direct and
straightforward comparability. In addition, if the different currencies are
expressed by unique single, selected currency (i.e. Euros or U.S. dollars), the
comparability can be possible also across different case studies.
5 Conclusion
Sustainable development requires the preservation and respect of ecological
equilibria, when performing economic activities. Natural ecosystems provide
themselves services that might be of support to human activities. These
services are called ecosystem services. Desertification is a threat to the
provision of ecosystem services and human well-being in drylands. For this
reason, anti-desertification policies are designed and implemented in order
to mitigate and adapt to climate changes. It is a challenging task, for
economists, to value ecosystem services and impacts of anti-desertification
policies. The valuation exercise can be performed through different valuation
techniques that aim at targeting and measuring the value of ecosystem
services (before and after the anti-desertification policy) in monetary terms.
Alternatively the assessment of the anti-desertification policy can be
performed through the adoption of indicators that convey synthesized socioeconomic information about ecosystem services in selected areas and time
periods. The first approach allows embracing all the multidimensional
implications of the problem at issue, without giving up on its complexity or
unduly broadening the scope of the analysis. However, adopting complex
valuation techniques is a costly research activity. Therefore, another method
to evaluate phenomena (desertification in the case at issue) is the adoption
of socio-economic indicators. This paper critically surveys economic valuation
techniques and socio-economic indicators for ecosystem services (and
135
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
policies), highlighting limits and advantages of such important informational
tools. The focus of the paper is on the survey, discussion and definition of key
socio-economic indicators for assessing anti-desertification practices. In
particular, the paper suggests a cost-effective meta-indicator, a flexible
measure for anti-desertification policy assessment. Such measure is easy to
compute; is not only a correlation measure, but synthesizes some causality
among selected variables and can be used for comparability in different
scenarios and selected areas. Further research should focus on the
calculation of such indicators in order to assess its applicability.
6 References
Brandt, J., N. Geeson, and A. Imeson (2003) A desertification indicator system for
Mediterranean Europe. Report of the DESERTLINKS project, available at:
http://www.kcl.ac.uk/projects/desertlinks/ (accessed September 2010).
Chiabai A., Travisi C. , A. Markandya, H. Ding, P. A.L.D. Nunes (2010) Economic Assessment of
Forest Ecosystem Services Losses: Cost of Policy Inaction," Working Papers 2010-13,
BC3.
Costanza, R., R. d’Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.
V. O’Neill, J. Paruelo, R. G. Raskin, P. Sutton and M. van den Belt. 1997. The Value of
the World’s Ecosystem Services and Natural Capital. Nature 387:253–260.
De Groot, R.S.; Wilson, M.A.; and Boumas, R.M.J. (2002) A Typology for the Classification,
Description and Valuation of Ecosystem Functions, Goods and Services. Ecological
Economics 41(3): 393-408
Fisher B. and K. Turner (2008) Ecosystem Services: Classification for Valuation. Biological
Conservation. 141. 1167-1169.
Fisher, B.; Turner, R.K.; and Morling, P. (2009) Defining and Classifying Ecosystem Services for
Decision Making. Ecological Economics 68(3): 643-653
Garrod G. and K. Willis (2010) Economic Valuation of the Environment: Methods and Case
Studies. Edward Elgar.
Geist, H.J. (2005) The Causes and Progression of Desertification. Ashgate Studies in
Environmental Policy and Practice. Ashgate: Aldershot
Ghermandi A., J.C.J.M. van den Bergh, L.M. Brander, H.L.F. de Groot, P.A.L.D. Nunes, (2008).
The Economic Value of Wetland Conservation and Creation: A Meta-Analysis,
Working Papers 2008.79, Fondazione Eni Enrico Mattei.
Kumar, M. and Kumar, P. (2008) Valuation of the Ecosystem Services: A Psycho-Cultural
Perspective. Ecological Economics 64(4): 808-819
MEA, Millennium Ecosystem Assessment (2000) Analytical Approaches for Assessing Ecosystem
Condition and Human Wellbeing. World Resource Institute, Washington DC. p.50.
MEA, Millennium Ecosystem Assessment, (2005a). Ecosystems and Human Well-being:
Synthesis.
Island Press, Washington, DC.
MEA, Millennium Ecosystem Assessment (2005b) ‘Ecosystems and Human Well-being:
Desertification Synthesis’. World Resources Institute: Washington, DC.
136
Assessing and Valuating Ecosystem Services and Mitigation Strategies
Nunes P.A.L.D. and J. C.J.M. van den Bergh (2001) Economic Valuation of Biodiversity:
Sense or Nonsense? Ecological Economics 39: 203-222.
Nunes
P. A.L.D., H. Ding, A. Markandya, (2009) The Economic Valuation of Marine
Ecosystems, Working Papers 2009.68, Fondazione Eni Enrico Mattei.
Nunes
P. A.L.D., R.Portela, N. Rao and S. Teelucksingh (2009). Recreational, Cultural and
Aesthetic Services from Estuarine and Coastal Ecosystems, Working Papers 2009.121,
Fondazione Eni Enrico Mattei
Onofri L, P.A.L.D. Nunes, H.Ding (2007). An Economic Model for Bioprospecting Contracts,"
Working Papers 2007.102, Fondazione Eni Enrico Mattei]
Onofri L. and P.A.L.D. Nunes (2004). The Profile of a “Warm-Glower”: A Note on Consumer’s
Behavior and Public Policy Implications Working Papers 2004.113, Fondazione Eni
Enrico Mattei.
Reed, M.S., E.D.G. Fraser, A.J. Dougill (2006) An Adaptive Learning Process for Developing
and Applying Sustainability Indicators with Local Communities. Ecological Economics
59: 406-418.
Rubio, J.L., and E. Bochet (1998) Desertification indicators as diagnosis criteria for
desertification risk assessment in Europe. Journal of Arid Environments 39: 113-120.
Turner K., I. Bateman,, W. N. Adger (2001) Economics of Coastal and Water Resources:
Valuing Environmental Functions,. Kluwer
Turner K. , S. Morse-Jones and B. Fisher ( 2010) Ecosystem Valuation: A Sequential Decision
Support System and Quality Assessment Issues. Annals of the New York Academy of
Science. 79: 101-123.
UN (1997) Report on ongoing work being done on benchmarks and indicators.
Intergovernmental negotiating committee for the elaboration of an international
convention to combat desertification in those countries experiencing serious drought
and/or desertification, particularly in Africa, 10th session, New York, 6-17 January 1997.
137
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
APPENDIX 1: Required Information for Performing an Economic Valuation
Study
In order to perform an economic-valuation exercise, it is important to gather
data and select the proper valuation approach (qualitative, quantitative
and/or monetary) and technique (as surveyed in the previous paragraph).
Often the choice of the valuation approach and technique depends on
data availability and on the possibility to “build” data (like WTP).
The paragraph, therefore, is listing required information (minimum threshold)
for economic valuation of drylands ecosystem services.
1) Drylands Supporting Services:
a) soil formation and soil conservation
b) Nutrient cycling supporting the service of soil development
c) Primary production
The information/data required for Supporting Services valuation are mostly
geo-physical. We can highlight:
•
Data and Information on above ground biomass:
•
Data on dryland temperature in a selected time period:
•
Data on dryland precipitation in a selected time period:
•
Information about the composition in the land cover/land cover
pattern in the area at study;
•
Information/data about soil death and soil fertility:
•
…
2) Drylands Regulating Services
a) water regulation;
b) climate regulation though carbon sequestration and/or surface
reflectance and evaporation;
c) Pollination and seed dispersal
The information/data required for Regulating Services valuation are mostly
geo-physical. We can highlight:
•
Data
on
number
of
natural
water
basins
and/or
constructed/engineering based streams/water pipelines or other
water infrastructure possibilities;
138
Assessing and Valuating Ecosystem Services and Mitigation Strategies
•
Data on Irrigation Costs:
•
Data on Carbon Reserves
dryland/forest area);
•
…
(forestry
map; extension
of
the
3) Drylands Provisioning Services:
a) Provisions derived from Biological Production (foods and fiber, wood
fuel, biochemicals; freshwater provisioning)
The information/data required for Regulating Services valuation are mostly
economic. We can highlight:
Data on land productivity per squared meter/ha for the selected
products cultivated in the land at desertification risk (e.g. x
kg/corn/ha);
Data on domestic prices of products produced in drylands (for
market purposes);
Information on the type and intensity of (market) economic activity
in the selected dryland;
Information (qualitative) on the type and intensity of bioprospecting
in the selected dryland;
Data on the number of water springs and/or water river basins and
reservoir in all or selected locations in drylands
Data on domestic prices of selected products cultivated in the
land at desertification risk (e.g. y euro/kg corn);
Data on land use patterns/composition in the land area at
desertification risk, including forestry, grassland and the intensity
of grazing activities, when applicable (e.g. for the case study Y
we have a total of land area of Z ha, distributed in z1 for crop1,
z2 crop2 and so on, or preferably a GIS map);
Data on production input (labour, water and land) costs/prices
for the production of selected products cultivated in the land at
desertification risk (e.g. irrigation costs);
Data on employment regarding the areas of study, including
the contribution of the cultivated land at desertification risk on
the creation of jobs/employment
….
4) Drylands Cultural Services:
a) Cultural Identity and diversity;
139
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
b) Cultural landscape and heritage value;
c) Services Knowledge Systems;
d) Spiritual Services;
e) Aesthetic and Inspiration Services;
f) Recreation and Tourism
The information/data required for Cultural Services valuation are mostly
economic. We can highlight:
• Data on Human Capital: (real GDP per capita/income of relevant
population; number of inhabitants; relevant population in the
nearby/and study site; level of human development (education level;
employment/unemployment ratios; Innovation level; Public and private
research and development investment; poverty indexes; time spent by
household members collecting water and fuel wood; percent of rural
population below the poverty line¸ children mortality and percent of
children underweight. These indicators vary across selected area and
depend on the level of economic and human development of
selected areas. Indicators about poverty or R&D investments are
calculated in different ways. Therefore, we just indicate the required
information and leave the selection of data/specific indicator ta a
casa-by-case selection ).
This kind of data is useful in order to define the profile of the relevant
population living in the drylands and for defining proper policies and to
perform at least a qualitative valuation.
• Data/Information on Cultural, Environmental and Aesthetic Values
I. number of Unesco sites, monuments, archeological sites; historic
sites, cultural identity sites; natural protected areas;
II. data constructed through valuation techniques (see above
paragraph 3);
III. Data/Information on in situ Education/Research Activities
IV. Data/information on particular
developed through the site
economic,
production
skills
• Data/Information on Governance Structures (information about the
way the land/forest is organized and managed; information about
property rights allocation (commons or privately owned land; for
developing countries, information on common land, quantity of annual
household consumption that is derived from common lands;
information about the management of natural resources (associations,
private companies; public authorities; information about trade habits
(contracts; exchange; franchising)
140
Assessing and Valuating Ecosystem Services and Mitigation Strategies
This kind of data is useful in order to define the governance structure,
through which the resources that provide ecosystem services in drylands
are managed.
Central or local government investments in infrastructure and accessibility
to credit can also influence sustainability or vulnerability of dryland
livelihoods as well as determine human well-being in these areas. Largescale government-driven projects can facilitate the sustainable
development of drylands, as seen in developed dryland areas in Israel,
California, and Australia. But if inappropriately designed, implemented or
managed, they can lead to desertification, as in the Aral Sea region. The
failure of African governments, for instance, to devolve power to affected
people and to link environmental degradation to economic policy has
been seen by some as a significant drawback in combating desertification
and drought As a result of these failings, many programs lack local support
or are undermined by conflicting trade and agricultural policies pursued
by governments.
• Data on Tourism (annual flows of visitors, characterization of the supply
side, including the number and type of hotels and restaurants););
• anthropogenic pressure (e.g., some index of pollution);
• climatic data (e.g., temperature, precipitation);
• type and intensity of economic activity taking place in the area
under consideration, including tourism;
• …
Biodiversity Services
Species richness declines with decreasing primary productivity and vice versa,
hence drylands species richness should decline with aridity. Indeed, the low
number of flowering plant species in the hyperarid
subtype rapidly increases with reduced aridity (in agreement with an increase
of between-‘‘broad biome’’ diversity) and peaks in the dry sub-humid
subtype. However, contrary to expectation, species diversity declines in nondryland temperate humid areas. But as might be expected, it is nearly half
that of tropical areas. It is important to assess and measure the dimensions,
structure and composition of dryland biodiversity, since biodiversity
conservation plays a role in the provision of drylands ecosystem services.
There are many dryland species that are directly involved in the provision of a
range of ecosystem services. One such example is African acacia (Ashkenazi
1995), which provides for soil development and conservation (roots, canopy,
and litter), forage (leaves and pods eaten by livestock), fuelwood (dead
twigs), and food (edible gums). It is also involved in nutrient cycling (symbiosis
with nitrogen-fixing bacteria), generates cultural services and it supports
other biodiversity: a large number of animal species depend on it for shelter,
141
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
shade, nest sites, and food. (Often this is of mutual benefit: wild and domestic
mammals disperse the seeds, thus determining the spatial distribution of the
species.)
The numerous dryland plant species of different growth forms jointly provide a
package of services through their ground cover and structure, which provide
the drylands’ most important services of water regulation and soil
conservation as well as forage and fuel wood provision and climate
regulation. In arid and semiarid areas, clumps of bushes and annuals
embedded in the matrix of a biological soil crust—which consists of an
assemblage of several species of cyano bacteria (that provide the added
benefit of nitrogen fixation), microalgae, lichens, and mosses—jointly
generate soil conservation and water regulation. In many arid and semiarid
areas, this biodiversity of ‘‘vegetation cover’’ and biological soil crusts is
linked to a diversity of arthropod species that process most of the living plant
biomass, constituting the first link of nutrient cycling.
The loss of biodiversity from drylands is not likely to affect all services uniformly.
Rather, primary production and the provisioning services derived from it, as
well as water provision, will be more resilient than recreation and ecotourism.
This is based on the observation that ecosystem services performed by top
predators will be lost before those performed by decomposers. The service of
supporting biodiversity (by generating and maintaining habitats of required
value and sample size) is expected to be degraded faster than the service of
provisioning biological products.
The direct threats to the service of supporting biodiversity include not only
land degradation but also habitat loss and fragmentation, competition from
invasive alien species, poaching, and the illegal trade in biodiversity products.
Indirect threats include the losses of the drylands-specific ‘‘keystone’’ species
and ‘‘ecosystem engineer’’ species (which modify the dryland environment
for the benefit of other species) . Finally, not only losses but also addition of
species may impair service provision. For example, Eucalyptus tree species
introduced to southern Africa have invaded entire catchments of natural
vegetation, causing large-scale changes in water balance and depriving
water from lower catchments
Required data/information is both ecological and economic:
• Data/information on local Biodiversity (census of mammals, birds,
plants…)
• Data/Information on Cultural, Environmental and Estethic Values of
Biodiversity:
V. number of existing and protected species in dryland;
VI. data constructed through valuation techniques (see above
paragraph 3);
142
Assessing and Valuating Ecosystem Services and Mitigation Strategies
Adaptation and Mitigation Strategies
The recognized importance of ecosystem services on human activities has
spurred to take action and starts implementing measures that limit
desertification in drylands. This part of socio-economic analysis, valuation
approach and technique choice should consider the impact of mitigation
and adaptation strategies in selected areas, in particular, those prevention
and restoration actions that combat desertification in selected area. The
analysis can vary according to available data and valuation approach and
technique. However, the focus should be on comparing different situations
before and prevention/restoration action and valuate costs and benefits of
those actions.
Required data/information is both ecological and economic:
• Information about restoration practices adopted in selected areas;
• Information/data about mitigation and adaptation strategies costs and
investments (i.e. management costs and restoration/prevention
technologies in drylands;
• Data/Information about (selected) stakeholders awareness and opinion
on adaptation/mitigation strategies;
• data constructed through valuation techniques (see above paragraph
3);
• ….
Finally, Box 1 highlights an important point about data gathering: time
frame and discounting rate.
Box 1: Selecting a study time-frame and a Discount Rate
An appropriate time-scale for the analysis is needed, for example relating to
the expected life of the product, project or asset, or perhaps more arbitrarily
set at a reasonable duration between say 25 to 100 years. This needs to
account for longer term implications, but bear in mind that going too far
into the future leads to i) considerable uncertainties and ii) future money
flows becoming significantly reduced depending on the discount rate used.
A suitable discount rate should be adopted when considering different
values over time. Depending on the nature of the assessment, this should
be either a company commercial rate or an appropriate social rate (if using
it for an economic analysis). It may be that several rates could be applied
to show the sensitivity of the outcome.
The UK government, for instance, adopts declining rate of discount 3.5% for first 30
years, 3% for years 31 to 75, 2.5% for 76 to 125, 2% for 126 to 200, 1.5% for 201 to 300
and 1% for over 300 years.
143
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
APPENDIX 2: A Pilot Case Study of Valuation to combat desertification; the
Piximanna Forest at Pula, Sardinia.
“Forests in drylands” is not an oxymoron, because forests do occur in drylands.
Eighteen percent of the area of the dryland system is occupied by the forest
and woodland system, though the probability of encountering forests in
drylands decreases with their aridity. In general, aridity increases inland, and
the forest and woodland systems prevail along the coasts. The dry sub-humid
dryland subtype that is adjacent to the forest and woodland system has the
greatest amount of overlap between the two systems, as compared with
other drylands subtypes. But the distribution of forest in dry sub-humid areas is
patchier than it is within the forest and woodland system outside the drylands.
Forests also occur in the much wider semiarid zone of Australia but are there
mostly confined to the less dry seaward direction. Forests occur in the dry subhumid subtype in Africa but are very scattered and rare in the semiarid zone.
In China and India, with dry sub-humid areas wider than in Australia, forests
penetrate deep into dry sub-humid areas. In Europe, where many dry subhumid areas are surrounded by non-drylands, forests are scattered all over the
dryland areas. In the Americas, forests are patchily distributed in dry subhumid and semiarid regions.
The forest of Piximanna14, at Pula represents an ideal case-study for valuating
measures to combat desertification, in the Practice framework. The forest has
drylands characteristics and, since the 1950’s, represents was subject to
mitigation actions, aiming at reforestation (from the 1950’s to the ‘70s) and
implementing silvicultural practices, that improve structural complexity and
ecosystem functions (from the ‘70s)15.
The valuation of (selected) ecosystem services and mitigation strategies
at Pula will be designed in such a way, that the adopted methodology
can be applied and/or extended to other Practice sites/partners, and will
follow the following organizational guidelines:
14
15
For details about the forest and mitigation strategies we remand to the Practice official documents.
Anti desertification practice is mostly reforestation (30 years, in two stages)
1) MITIGATION ACTIONS
> 50s, 60s, 70s: Reforestation actions (Pinus pinea L.,Pinus halepensis Mill.,Quercus suber L.) for watersheds protection;
> Last three decades: Silvicultural practices (selective thinning on conifer stands) to facilitate the natural
introduction of late-successional hardwood (improvement of structural complexity and ecosystem
functions)
2) MAIN DEGRADATION DRIVERS
> Overgrazing by domestic animals (mainly goats untill‚‘60s)
> Overexploitation of fuel - wood (untill ’50s)
> Forest fires
(Source: D’Angelo et. al. Practice Kick Off Meeting (2009)
144
Assessing and Valuating Ecosystem Services and Mitigation Strategies
1) Define the Scope
This step helps to scope out and define what to value and to define the
boundaries of the valuation exercise.
2) Plan the Study
This step goes into specific details as to how the valuation will actually be
undertaken in terms of which stakeholders will be involved, to undertake what
aspects, when, how long it may take and what it will cost.
3) Undertake the Valuation
This step is the actual valuation itself. It comprises a process of other sub-steps
that should be followed to undertake the valuation aspect of the ecosystem
services.
This involves aspects such as establishing the environmental
baseline, determining the stakeholders impacts and dependencies, assessing
which ecosystem services are affected, and valuing the changes. The
valuation methodology can be developed along the main points in Box 2.
For the Piximanna study the scope of the valuation is defined and
summarized by Table 3. We have selected 3 different application areas.
1) A “very restricted area characterised by the reforestation intervention;
2) A “restricted area” characterised by economic, ecological and geophysical similarities with the “non reforested area”. That is
we look for
situations that would depict us the status quo of the dryland if no mitigation
strategy had occurred
3) A “broad area” that covers the all forest district.
The valuation of different ecosystem services under different scenarios (three
different areas) and the related required information is summarized in Table 6.
Finally a list of potential stakeholders was defined. Next step is planning thee
study scope and undertaking the valuation.
145
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
Box 2. Ecosystem Services Valuation Methodology
1)
2)
3)
4)
5)
6)
7)
Finalise agreed objectives and initial methodology. This step may be
necessary if an external organisation has submitted a Proposal to
conduct the ecosystem valuation.
Undertake initial data collection. This could include a literature and
database review, stakeholder consultation (internal and external),
site visits and various surveys (eg biological, hydrological or physicochemical). The aim is to determine what information gaps need
filling and what information can be used for the valuation
techniques.
Select the valuation approach. This depends on the data availability
and on the objectives and methodology.
Valuation can be
qualitative, quantitative or monetary.
Design the valuation approach. This step could include designing a
draft stated preference questionnaire, undertaking a pilot
questionnaire survey for the stated preference surveys, designing
sampling strategies and draft questionnaire surveys et cetera.
Implement the valuation. This could involve a further literature and
database review to identify specific studies for benefit transfers,
additional stakeholder consultation, and implementing questionnaire
surveys.
Analyse the results. This should include assessing monetised and nonmonetised costs and benefits, considering the time-scales, discount
rate and undertaking econometric analysis.
Produce a report. This should set out the methodology, level of
consultation, results, assumptions, sensitivity analysis, conclusions and
potential applications etc.
146
Assessing and Valuating Ecosystem Services and Mitigation Strategies
Table 6. Piximanna Case Study Scoping. Valuation of reforestation and Silviculture
(anti-desertification mitigation policies) actions. The mitigation policies have been
effective because contributed to improve ecosystem functions?
Vey Restricted Area
Restricted Area
Broad Area
Supporting +
Biomass
Land cover
Simplified Land cover;
Biodiversity
Biodiversity (ecologicalbiophysical analysis)
Valuation of
biodiversity
inter-temporal
comparison (1955-19802010)
Transects in plot area
Biodiversity (literature
survey)
Regulating
Provisioning
Carbon reserves
Irrigation Costs
Irrigation Costs?
Number of water
springs
Number of water
reservoirs
Type and intensity of
economic activity
Type and intensity of
economic activity
Type and intensity of
economic activity
Productivity
Productivity
Profitability
Profitability
Prices (wood, honey,
mushrooms, breeding)
Prices
Opportunity Cost
Indicator
Policy Valuation
Meta-Indicator
Cultural
Environmental and
cultural sites…(census)
Census of Knowledge
Systems
(relevant activities)
Governance structure
of forestal resource
Valuation of cultural,
environmental and
aesthetic value/WTP
Valuation of
Aesthetic and
inspiration services
Valution/Census of
Knowledge Systems
Valuation/census of
recreationists and
recreation-cultural
system
Data on tourism (flows,
returns)
GDP, Number of
Inhabitants,
employment/unemploy
ment rate …..
Governance structures
(resources governance
structure)
Environmental and
cultural
sites…(mapping)
147
L. Onofri, A. Ghermandi, and P.A.L.D. Nunes
148