remote sensing
Article
Evaluating Multi-Sensor Nighttime Earth
Observation Data for Identification of Mixed vs.
Residential Use in Urban Areas
Christoph Aubrecht 1, *,† and José Antonio León Torres 2,†
1
2
*
†
Social, Urban, Rural & Resilience (GSURR), The World Bank, Washington, DC 20006, USA
Institute of Engineering, National Autonomous University of Mexico (UNAM), Mexico City 04510, Mexico;
[email protected]
Correspondence: [email protected]
These authors contributed equally to this work.
Academic Editors: Ioannis Gitas and Prasad Thenkabail
Received: 28 November 2015; Accepted: 25 January 2016; Published: 4 February 2016
Abstract: This paper introduces a novel top-down approach to geospatially identify and distinguish
areas of mixed use from predominantly residential areas within urban agglomerations. Under the
framework of the World Bank’s Central American Country Disaster Risk Profiles (CDRP) initiative,
a disaggregated property stock exposure model has been developed as one of the key elements for
disaster risk and loss estimation. Global spatial datasets are therefore used consistently to ensure
wide-scale applicability and transferability. Residential and mixed use areas need to be identified
in order to spatially link accordingly compiled property stock information. In the presented study,
multi-sensor nighttime Earth Observation data and derivative products are evaluated as proxies to
identify areas of peak human activity. Intense artificial night lighting in that context is associated with
a high likelihood of commercial and/or industrial presence. Areas of low light intensity, in turn, can be
considered more likely residential. Iterative intensity thresholding is tested for Cuenca City, Ecuador,
in order to best match a given reference situation based on cadastral land use data. The results and
findings are considered highly relevant for the CDRP initiative, but more generally underline the
relevance of remote sensing data for top-down modeling approaches at a wide spatial scale.
Keywords: top-down modeling; urban areas; nighttime lights; DMSP; VIIRS; human activity;
residential use; mixed use; global spatial data; CDRP
1. Introduction
Issues of urban development are increasingly being addressed at the global scale, with
international non-governmental organizations (NGOs) and development institutions often setting
the path and moving the public agenda forward. Regularly published reports such as the United
Nations’ World Urbanization Prospects [1] or the World Bank’s World Development [2] and Global
Monitoring Reports [3] address fundamental issues and define key research questions to be tackled by
the scientific and international development community. In that context it has become more and more
evident that spatial data is playing a crucial role for consistent cross-regional analyses and unbiased
evaluation of locally implemented actions. Remote sensing data in particular provide a rich and
globally consistent source for analyses at multiple levels. At the global scale, different aspects have to
be considered than for local-level spatial analyses, including consistency, scalability, retraceability etc.
Several global project initiatives address these issues in various thematic domains. The World Bank’s
Global Urban Growth Data Initiative, for example, addresses pending issues of regional definition
and data incompatibilities and supports the international collaborative setup and development of
Remote Sens. 2016, 8, 114; doi:10.3390/rs8020114
www.mdpi.com/journal/remotesensing
Remote Sens. 2016, 8, 114
2 of 19
a consistent data set of global urban extents and associated population distribution patterns. In
the same context, the Global Human Settlement Working Group, established under the umbrella of
the Group on Earth Observations (GEO) (www.earthobservations.org/ghs), aims at establishing a
new generation of global settlement measurements and products based on consistent high-resolution
satellite imagery analysis.
The presented study has been carried out within the framework of the World Bank’s Country
Disaster Risk Profiles (CDRP) project initiative which has been successfully implemented at the
continental scale for Central America [4] and is currently being expanded to the Caribbean Region.
With the clear aim at extending to other regions, global applicability and easy transferability are
considered crucial for the model setup. Global spatial datasets are therefore used throughout the
CDRP project, with the presented approach specifically developed to support implementation of a
disaggregated property stock exposure model, one of the key elements for subsequent disaster risk
and loss estimation. While focusing primarily on natural hazards and risks, urban-rural identification
and intra-urban classification aspects are highly relevant for setting the basic spatial framework for
analysis [5].
This paper introduces a novel approach to geospatially identify and distinguish areas of mixed
use from predominantly residential areas within urban agglomerations. After initial urban-rural
classification at a 1 km grid level, that urban mask needs to be classified in residential and mixed use
areas in order to spatially link accordingly compiled property stock information (e.g., from global
tabular databases such as PAGER-STR [6]). The distinct identification of urban residential and mixed
use areas serves as crucial input to define inventory regions for subsequent exposure assessment.
Impervious Surface Area (ISA) data [7] based on remotely sensed nighttime lights from the Defense
Meteorological Satellite Program’s (DMSP) OLS sensor (Operational Linescan System) are used as
proxy to identify areas of peak human activity, often associated with a high likelihood of commercial
and/or industrial presence. ISA is chosen due to its inherent correlation with built-up area (providing
an indication on the percentage of built-up per grid cell) and thus good suitability for building stock
related land use classification. Several ISA thresholds are tested for a case study in Cuenca City,
Ecuador, in order to best match a given reference situation on the ground, where local-level cadastral
land use data is used to identify the actual distribution ratio of residential vs. mixed use areas.
Furthermore, unaltered nightlight intensity data as provided by the VIIRS sensor (Visible Infrared
Imaging Radiometer Suite) [8] are evaluated as alternative to the ISA data. With the DMSP program
fading out VIIRS provides the option for successive nighttime Earth Observation analyses due to its
low light imaging capability. We apply the same methodological steps as for the ISA data in order
to determine best-matching thresholds for binary land use classification and subsequently perform
a comparative analysis of the results. Also scale effects are accounted for in that regard with VIIRS
featuring a higher spatial resolution than OLS-based data products.
Preliminary results of this study were presented at the ECRS-1 conference [9]. Extensive further
research and integration of alternative data sources then lead to the multi-sensor approach illustrated in
this paper, highlighting the relevance of global remote sensing data for top-down modeling approaches
at wide spatial scale. Outcome is considered relevant for global urban spatial modeling in a variety of
topical domains including urban monitoring, disaster risk management, and regional development.
2. Study Area and Data
Due to availability of detailed in situ reference data for comparative analysis and evaluation of the
proposed methodology, the city of Cuenca, Ecuador, was chosen as study area. Cuenca City is located
in the mountainous southern region of Ecuador at an elevation of around 2500 m above sea level and
is the capital of the Azuay province (Figure 1). The city stretches across an area of roughly 70 km2
and had an urban population of 329,928 inhabitants according to the census 2010. Latest figures of
the Ecuador National Statistical Office (INEC) estimate an urban population of approximately 400,000
in 2015.
Remote Sens. 2016, 8, 114
Remote Sens. 2016, 8, 114
3 of 19
3 of 19
Study area Cuenca City, Ecuador.
Figure 1. Study
Ecuador.
The use
use of
of satellite-observed
satellite-observed nighttime
nighttime lights
lights has
has aa long
long tradition
tradition in
in research
research dealing
dealing with
with
The
monitoring
urban
areas
and
patterns
of
human
and
economic
activity
[10–16]
as
well
as
its
impact
on
monitoring urban areas and patterns of human and economic activity [10–16] as well as its impact on
the
environment
[17–19].
As
opposed
to
attempts
of
using
nighttime
lights
for
basic
delineation
of
the environment [17–19]. As opposed to attempts of using nighttime lights for basic delineation of
urban areas
areas or
or as
as weights
weights for
for population
population disaggregation,
disaggregation, in
in this
this paper
paper we
we rather
rather aim
aim at
at exploring
exploring the
the
urban
use
and
value
in
determining
intra-urban
characteristics.
use and value in determining intra-urban characteristics.
Public-domain applications
been restricted
restricted to
to one
one
Public-domain
applications of
of nighttime
nighttime Earth
Earth Observation
Observation have
have long
long been
satellite
sensor,
namely
the
Operational
Linescan
system
(OLS)
onboard
the
Defense
Meteorological
satellite sensor, namely the Operational Linescan system (OLS) onboard the Defense Meteorological
Satellite Program
Program (DMSP)
(DMSP) platform
platform [20].
[20]. More
Visible Infrared
Infrared Imaging
Imaging
Satellite
More recently,
recently, data
data from
from the
the Visible
Radiometer Suite
Suite (VIIRS)
(VIIRS) sensor
sensor onboard
onboard the
the Suomi
Suomi NPP
NPP satellite
satellite platform
platform have
have become
become available,
Radiometer
available,
providing
both
higher
spatial
and
radiometric
resolution
and
being
considered
as
natural successor
successor
providing both higher spatial and radiometric resolution and being considered as natural
to
the
fading-out
DMSP-OLS
series
[21].
The
commercial
satellite
EROS-B
also
offers
nighttime
to the fading-out DMSP-OLS series [21]. The commercial satellite EROS-B also offers nighttime
acquisition capability,
capability,even
even
at very
high spatial
resolution
[22]. However,
global-scale
and
acquisition
at very
high spatial
resolution
[22]. However,
global-scale
and temporally
temporally
continuous
open
data
availability
remains
restricted
to
DMSP-OLS
and
NPP-VIIRS,
continuous open data availability remains restricted to DMSP-OLS and NPP-VIIRS, therefore the only
therefore the
onlygiven
reasonable
choice
given
the scope ofCDRP
the above-outlined
CDRP
initiative.
In the
reasonable
choice
the scope
of the
above-outlined
initiative. In the
following,
we briefly
following,the
wetwo
briefly
the
two
sourcesCity
we use
the (1)
Cuenca
City case study,
the
introduce
dataintroduce
sources we
use
fordata
the Cuenca
casefor
study,
the Impervious
Surface(1)Area
Impervious
Surface
Area
(ISA)
product
derived
from
DMSP-OLS;
and
(2)
VIIRS
Day/Night
band
(ISA) product derived from DMSP-OLS; and (2) VIIRS Day/Night band light intensity data.
light intensity data.
2.1. Impervious Surface Area (ISA) Data, Derived from DMSP-OLS
2.1. Impervious Surface Area (ISA) Data, Derived from DMSP-OLS
The OLS sensor onboard the DMSP satellite series is able to detect faint light on the Earth’s
TheatOLS
the DMSPinsatellite
series
is able to
detect
faint designed
light on the
Earth’s
surface
nightsensor
due toonboard
its high sensitivity
the visible
spectrum.
While
initially
to monitor
surface
at
night
due
to
its
high
sensitivity
in
the
visible
spectrum.
While
initially
designed
to
monitor
cloud coverage, that low light imaging capacity allows identification of various light emitting sources
cloud coverage,
low light
capacity
of various
light
emitting
including
human that
settlements
andimaging
associated
human allows
activityidentification
patterns [20]. The
National
Geophysical
sources
including
human
settlements
and
associated
human
activity
patterns
[20].
The
National
Data Center (NGDC) of the National Oceanic and Atmospheric Administration (NOAA) is processing
Geophysical
Center (NGDC)
of the
National
Oceanic
Atmospheric
Administration
and
archivingData
OLS imagery,
thereby also
making
certain
derivedand
products
accessible
to the public.
(NOAA) is data
processing
and
archiving
OLS imagery,
thereby
also area
making
certain
products
DMSP-OLS
was first
used
to approximate
impervious
surface
(ISA)
in the derived
early 2000s
in the
accessible
to
the
public.
DMSP-OLS
data
was
first
used
to
approximate
impervious
surface
area
development of a national-scale model for the conterminous United States [23]. The ISA approach
(ISA)
in
the
early
2000s
in
the
development
of
a
national-scale
model
for
the
conterminous
United
was then consequently adjusted to global scale whereby a radiance-calibrated annual composite of
States [23].
Theis ISA
approach
was then
consequently
adjusted
to global scale
nighttime
lights
analyzed
in conjunction
with
ancillary data
such as population
counts.whereby
Output isa
radiance-calibrated
annual
composite
of
nighttime
lights
is
analyzed
in
conjunction
with
ancillary
consistently provided at 30 arc-sec spatial resolution giving indication on the distribution of manmade
data such
as population
counts.
is consistently
provided at 30 arc-sec spatial resolution
surfaces
including
buildings,
roadsOutput
and related
elements [7].
3
Remote Sens. 2016, 8, 114
4 of 19
giving indication on the distribution of manmade surfaces including buildings, roads and related
4 of 19
elements [7].
Due to its relevance for a broad set of applications, global impervious surface or general
Duearea
to itsmapping
relevancehas
forbeen
a broad
setfocus
of applications,
impervious
surface
ordifferent
general built-up
built-up
in the
of attentionglobal
for a while
with data
from
satellite
area
mapping
has
been
in
the
focus
of
attention
for
a
while
with
data
from
different
satellite
sensors used and various approaches implemented (overview provided by [24]), recentsensors
efforts
used
and various
approaches
implemented
provided
by [24]), recent
including
high
including
high resolution
products
such as(overview
the Global
Urban Footprint
(GUF)efforts
[25] and
the Global
resolution
products such
the Global
Footprint (GUF)
[25]data
andset
theisGlobal
Human
Settlement
Human Settlement
Layeras(GHSL)
[26].Urban
The DMSP-derived
ISA
unique
in a sense
that it
Layer
(GHSL)
[26].extract
The DMSP-derived
ISA
data set
is unique
a sense
that it does
directly
does not
directly
built-up from
satellite
imagery
butinuses
artificial
nightnot
lighting
asextract
proxy
built-up
satellite
imagerygeneral
but uses
artificial night
lighting
as proxy
While a detected
measure.from
While
a detected
correlation
between
night
lights measure.
and impervious
surfaces
general
correlation
night ISA
lightsproduct,
and impervious
basis for
the global
ISA
provides
the basis between
for the global
inherent surfaces
patternsprovides
point to the
different
human
activities
product,
inherent patterns
point
to different
human
(e.g.,Given
commercial,
industrial)
ratherofthan
(e.g., commercial,
industrial)
rather
than mere
builtactivities
structures.
the scope
and purpose
the
mere
built study
structures.
Given therelation
scope and
purposewith
of the
presented
study
two-sided relation
presented
this two-sided
to built-up
a specific
weight
on this
non-residential
human
to
built-up
with aisspecific
weightrelevant.
on non-residential
human
activity
is set
particularly
relevant.
activity
patterns
particularly
Figure 2 shows
the
globalpatterns
ISA data
for the year
2010
Figure
2 shows
globalCity
ISA study
data set
for the year 2010 extracted for the Cuenca City study area.
extracted
for thethe
Cuenca
area.
Remote Sens. 2016, 8, 114
Figure
2. Impervious
(ISA)—30 arc-sec
arc-sec grid
grid for
for Cuenca
Cuenca City.
City.
Figure 2.
Impervious Surface
Surface Area
Area (ISA)—30
2.2. VIIRS Day/Night Band Data
2.2. VIIRS Day/Night Band Data
Since 2011 the VIIRS sensor onboard the Suomi NPP satellite platform provides a natural
Since 2011 the VIIRS sensor onboard the Suomi NPP satellite platform provides a natural successor
successor to DMSP-OLS with its panchromatic Day/Night Band (DNB) detecting dim nighttime
to DMSP-OLS with its panchromatic Day/Night Band (DNB) detecting dim nighttime scenes in similar
scenes in similar manner. Using advanced processing schemes (e.g., excluding/correcting data
manner. Using advanced processing schemes (e.g., excluding/correcting data impacted by stray light),
impacted by stray light), NOAA-NGDC is producing global monthly composite products featuring
NOAA-NGDC is producing global monthly composite products featuring average radiance values at
average radiance values at 15 arc-sec spatial resolution [27]. As supplementary product the number
15 arc-sec spatial resolution [27]. As supplementary product the number of cloud-free observations
of cloud-free observations that was used to create the average composite is reported for each cell.
that was used to create the average composite is reported for each cell.
In addition to its superior spatial resolution, VIIRS offers a set of improvements over
In addition to its superior spatial resolution, VIIRS offers a set of improvements over OLS-derived
OLS-derived nighttime lights. These include lower light detection limits, improved dynamic range
nighttime lights. These include lower light detection limits, improved dynamic range as well as
as well as quantification and calibration options previously unavailable [21]. One of the few
quantification and calibration options previously unavailable [21]. One of the few disadvantages of
disadvantages of VIIRS, on the other hand, refers to the later overpass time (after midnight), when
VIIRS, on the other hand, refers to the later overpass time (after midnight), when outdoor lighting is at
outdoor lighting is at a significantly lower level as compared to early evening when OLS
a significantly lower level as compared to early evening when OLS acquisitions are made. Figure 3
acquisitions are made. Figure 3 illustrates the VIIRS DNB data for the June 2015 monthly composite
illustrates the VIIRS DNB data for the June 2015 monthly composite extracted for the Cuenca City
extracted for the Cuenca City study area (left). For comparative purposes, we also aggregate that
study area (left). For comparative purposes, we also aggregate that data to a 30 arc-sec grid (right),
data to a 30 arc-sec grid (right), thus matching the resolution of the ISA product.
thus matching the resolution of the ISA product.
2.3. Cadastral
Cadastral Data
Data
2.3.
Since 2010
2010 the
theMunicipality
MunicipalityofofCuenca
Cuencahas
hasintensified
intensified
efforts
collect
specific
information
Since
thethe
efforts
to to
collect
specific
information
on
on
all
buildings
located
in
the
urban
area
of
Cuenca
City
enabling
the
construction
of
a complete
all buildings located in the urban area of Cuenca City enabling the construction of a complete
cadastral
cadastral including
database including
geo-localization
and information
oncharacteristics.
building characteristics.
database
geo-localization
and information
on building
4
Remote
2016, 8,
Remote Sens.
Sens. 2016,
8, 114
Remote Sens. 2016, 8, 114
55of
of19
19
5 of 19
Figure 3. VIIRS Day/Night Band (DNB) data for Cuenca City. 15 arc-sec grid (left). Aggregated 30
Figure
(DNB)
data
forfor
Cuenca
City.
15 15
arc-sec
grid
(left).
Aggregated
30
Figure 3.
3. VIIRS
VIIRSDay/Night
Day/NightBand
Band
(DNB)
data
Cuenca
City.
arc-sec
grid
(left).
Aggregated
arc-sec grid (right).
arc-sec
gridgrid
(right).
30 arc-sec
(right).
More specifically, this cadastral database contains detailed information on the use of each
More specifically, this cadastral database contains detailed information on the use of each
Moreallowing
specifically,
cadastral
contains
detailed information
on the useoccupancy
of each building,
building,
forthis
example
thedatabase
distinction
of residential
and nonresidential
types.
building, allowing for example the distinction of residential and nonresidential occupancy types.
allowing
for example
theas
distinction
and nonresidential
types.
Data sources
Data sources
that served
input for of
theresidential
cadastral data
base originate occupancy
from different
national
entities
Data sources that served as input for the cadastral data base originate from different national entities
that
input for theofcadastral
originate
fromof
different
national
entities(INEC),
such as the
suchserved
as theasMunicipality
Cuenca,data
the base
National
Institute
Statistics
and Census
the
such as the Municipality of Cuenca, the National Institute of Statistics and Census (INEC), the
Municipality
of Cuenca,Water
the National
InstituteService
of Statistics
and Census
(INEC), (ETAPA),
the Telecommunications,
Telecommunications,
and Sewage
Company
of Cuenca
the National
Telecommunications, Water and Sewage Service Company of Cuenca (ETAPA), the National
Water
and Sewage
Service
Company(SNGR)
of Cuenca
(ETAPA),
the National
Secretariat
for Risk
Management
Secretariat
for Risk
Management
and
the University
of Cuenca.
After
a validation
and
Secretariat for Risk Management (SNGR) and the University of Cuenca. After a validation and
(SNGR)
and
the
University
of
Cuenca.
After
a
validation
and
filtering
process,
the
cadastral
building
filtering process, the cadastral building data base eventually comprises 65,436 records [28].
Each
filtering process, the cadastral building data base eventually comprises 65,436 records [28]. 2 Each
data
basefootprint
eventually
comprises
65,436 records
Each building
footprint
record area
is georeferenced
building
record
is georeferenced
and[28].
includes
information
on built-up
(in m ) and
building footprint record is georeferenced and2 includes information on built-up area
(in m2) and
2
and
includes
information
on
built-up
area
(in
m
)
and
occupancy
type
(Figure
4).
Residential
buildings
occupancy type (Figure 4). Residential buildings thereby cover an area of 12.9 km2 , complemented
occupancy type
(Figure 4). Residential
buildings thereby cover
an area of 12.9 km , complemented
2
2
2
thereby
an area of 12.9built-up
km , complemented
by a 4.3 km non-residential built-up area.
by a 4.3 cover
km non-residential
area.
by a 4.3 km2 non-residential built-up area.
Figure 4.
4. Cadastral
Cadastral building
footprints
of
Cuenca
City,
classified in
in residential
residential and
and non-residential
non-residential
Figure
buildingfootprints
footprintsof
of Cuenca
CuencaCity,
City, classified
classified
Figure
4. Cadastral
building
in residential
and non-residential
occupancy
types.
occupancy types.
types.
occupancy
For comparative purposes, we aggregate building footprint data to a 15 arc-sec and a 30 arc-sec
For
For comparative
comparative purposes,
purposes,we
weaggregate
aggregatebuilding
buildingfootprint
footprintdata
datato
toaa15
15arc-sec
arc-secand
andaa 30
30 arc-sec
arc-sec
grid respectively, thus matching the resolutions of the two analyzed nighttime lights data sets.
grid
respectively,thus
thusmatching
matching
resolutions
of two
the analyzed
two analyzed
nighttime
lights
data
sets.5
grid respectively,
thethe
resolutions
of the
nighttime
lights data
sets.
Figure
Figure 5 illustrates the aggregated grids for the non-residential share of the built-up area. On top,
Figure
5 illustrates
the aggregated
grids
for the non-residential
share ofarea.
the built-up
area. On top,
illustrates
the aggregated
grids for the
non-residential
share of the built-up
On top, non-residential
non-residential built-up percentage is shown. At the bottom, each cell’s contribution to the total
non-residential built-up percentage is shown. At the bottom, each cell’s contribution to the total
built-up area is visualized, whereby a main cluster in the center of the city is clearly depicted.
built-up area is visualized, whereby a main cluster in the center of the city is clearly depicted.
5
5
Remote Sens. 2016, 8, 114
6 of 19
built-up percentage is shown. At the bottom, each cell’s contribution to the total built-up area is
visualized,
whereby
Remote Sens. 2016,
8, 114 a main cluster in the center of the city is clearly depicted.
6 of 19
Figure 5. Cadastral data for Cuenca City aggregated to grids at 15 arc-sec (left) and 30 arc-sec (right)
Figure 5. Cadastral data for Cuenca City aggregated to grids at 15 arc-sec (left) and 30 arc-sec (right)
resolution. Non-residential built-up percentage (top). Percent-contribution of each cell to the total
resolution. Non-residential built-up percentage (top). Percent-contribution of each cell to the total
non-residential built-up
built-up area
area (bottom).
(bottom).
non-residential
3. Methods
3. Methods
As outlined in the introduction, the presented study was carried out under the framework of
As outlined in the introduction, the presented study was carried out under the framework of
the World Bank’s Central American Country Disaster Risk Profiles (CDRP) Initiative [4]. With that
the World Bank’s Central American Country Disaster Risk Profiles (CDRP) Initiative [4]. With that
kind of continental and global models, the implemented scale level plays an important role in
kind of continental and global models, the implemented scale level plays an important role in defining
defining the basic spatial units of analysis. Working on a 30 arc-sec resolution grid level (i.e.,
the basic spatial units of analysis. Working on a 30 arc-sec resolution grid level (i.e., approximately
approximately 1 km at the equator)—frequently used for global models—the spatial identification
1 km at the equator)—frequently used for global models—the spatial identification and distinction
and distinction of unique inventory regions is often not unambiguously possible at the grid cell level
of unique inventory regions is often not unambiguously possible at the grid cell level due to the
due to the well-studied mixed pixel issue [29,30]. While large urban residential areas as well as
well-studied mixed pixel issue [29,30]. While large urban residential areas as well as certain dedicated
certain dedicated industrial zones are still often built in rather compact manner and can thus indeed
industrial zones are still often built in rather compact manner and can thus indeed cover entire grid
cover entire grid cells, particularly commercial areas are commonly intertwined with residences
cells, particularly commercial areas are commonly intertwined with residences forming wider areas of
forming wider areas of mixed use. In order to appropriately identify urban non-residential areas in a
mixed use. In order to appropriately identify urban non-residential areas in a spatial top-down model
spatial top-down model it is therefore considered reasonable to assume a certain share of residential
it is therefore considered reasonable to assume a certain share of residential occupancy throughout
occupancy throughout and consider grid cells that also include a non-residential share as areas of
and consider grid cells that also include a non-residential share as areas of mixed use.
mixed use.
For identification of those built-up urban areas that also feature a share of non-residential use,
For identification of those built-up urban areas that also feature a share of non-residential use,
we refer to the above-outlined nighttime Earth Observation data and derivative products as proxy
we refer to the above-outlined nighttime Earth Observation data and derivative products as proxy
measure. The assumption hereby is that intense lighting in that context is associated with a high
measure. The assumption hereby is that intense lighting in that context is associated with a high
likelihood of commercial and/or industrial presence, commonly clustered in certain parts of a city
likelihood of commercial and/or industrial presence, commonly clustered in certain parts of a city
(such as central business districts and/or peripheral commercial zones). Areas of low light intensity, in
(such as central business districts and/or peripheral commercial zones). Areas of low light intensity,
turn, can be considered more likely residential.
in turn, can be considered more likely residential.
The main objective of this study is thus to identify the light intensity thresholds that match best
The main objective of this study is thus to identify the light intensity thresholds that match best
the separated distribution of residential vs. mixed use areas on the ground. DMSP-OLS derived
the separated distribution of residential vs. mixed use areas on the ground. DMSP-OLS derived ISA
data and VIIRS-DNB data are both evaluated and comparatively analyzed for that purpose. It
should be noted that the presented approach is proposed only for pre-identified urban areas [31], as
for rural regions coarse-scale lighting intensity has reduced spatial correlation with built-up and
other additional aspects come into play.
Remote Sens. 2016, 8, 114
7 of 19
ISA data and VIIRS-DNB data are both evaluated and comparatively analyzed for that purpose. It
should be noted that the presented approach is proposed only for pre-identified urban areas [31], as
for rural regions coarse-scale lighting intensity has reduced spatial correlation with built-up and other
additional aspects come into play.
Referring to the Cuenca City cadaster data we distinguish purely residential areas from areas
of mixed use. Using the building footprint area data we obtain that 75% of the total built-up area of
Cuenca City features residential occupancy, complemented by 25% non-residential occupancy. At the
aggregated 15 arc-sec level, the top 25% mixed use cells (covering 25 km2 out of the total 98.75 km2 )
account for 92% of the city’s total non-residential built-up area. We then use this bottom-up-determined
distribution ratio to identify the appropriate lighting intensity thresholds in the top-down model.
In order to define the relevant data value histograms for the threshold identification, we select
all cells of the respective ISA and VIIRS data sets that fall within the pre-defined urban test case area
of Cuenca City. In the case of the ISA data, the min-max value range is thereby identified as 5.7–77.8.
For the VIIRS data, the min-max value range is identified as 3.3–73.8. In order to factor out potential
effects generated by the mere difference in spatial resolution between ISA and VIIRS data we aggregate
the original 15 arc-sec VIIRS data to a 30 arc-sec grid, thus enabling direct spatial comparability to
the ISA grid. We iteratively apply several threshold cut-off points in the identified value ranges and
compare the resulting areas of relatively low and relatively high ISA and VIIRS values respectively to
the aggregated cadastral data. The eventually selected final cut-off point is that threshold value that
produces the best-matching output with regard to the 75:25 cadaster-based residential vs. mixed use
area distribution ratio.
4. Results
4.1. Identification of Residential vs. Mixed Use Areas Using ISA Data
Table 1 illustrates the various tested ISA threshold values and the corresponding building use
distribution ratios as derived from comparative spatial overlay with the aggregated cadastral data
at the 30 arc-sec grid level. ISA min-max range and respective threshold values are shown in the left
part of the table, with the percentile column indicating the relative value distribution. In mathematical
terms the percentile value is derived as follows: (Threshold–Min)/(Max–Min). Specifically, that means
that for the highlighted ID 4 the ISA value of 42 indicates the median value (50th percentile) in the
distribution histogram. Half of the values in the study area under consideration thus feature an ISA
value lower than 42 and the other half feature a higher value.
Table 1. ISA distribution thresholds and corresponding building use distribution ratios (grey indicating
selected best-matching threshold).
ISA Distribution
ID
1
2
3
4
5
Built-Up Area Distribution
Min
Threshold
Max
Percentile
Residential Use
Mixed Use
5.7
5.7
5.7
5.7
5.7
15
22
39
42
59
77.8
77.8
77.8
77.8
77.8
13%
23%
46%
50%
74%
32.00%
52.00%
70.00%
74.00%
96.00%
68.00%
48.00%
30.00%
26.00%
4.00%
Spatially overlaid on the aggregated cadastral building use density grid (at the 30 arc-sec level,
comparable to the ISA grid), that results in a 74% residential ratio and a 26% mixed use share, thus
best-matching the bottom-up-derived 75% residential share (ISA data is provided in integer numbers,
thus making it impossible to exactly match that 75% target value). Figure 6 maps the binary land use
classification (residential use vs. mixed use) for the 5 tested ISA thresholds respectively. The share of
mixed use area decreases thereby corresponding to the higher ISA cut-off points.
Remote Sens. 2016, 8, 114
Remote Sens. 2016, 8, 114
8 of 19
8 of 19
Figure 6.
6. Binary
Binary land
land use
use classification
classification of
of Cuenca
Cuenca City
City based
based on
on ISA
ISA thresholds
thresholds from
from Table
Table 1.
1. Blue
Blue
Figure
indicates
residential
and
orange
mixed
use.
Table
record
IDs
are
indicated
in
the
figure
as
1–5.
indicates residential and orange mixed use. Table record IDs are indicated in the figure as 1–5.
Having the building-level cadastral data at hand enables not only determination of the binary
Having the building-level cadastral data at hand enables not only determination of the binary
land use distribution ratio, but furthermore allows us to consequently evaluate the degree of spatial
land use distribution ratio, but furthermore allows us to consequently evaluate the degree of spatial
overlap as a measure of model output accuracy. Using the above-identified ISA threshold, thus best
overlap as a measure of model output accuracy. Using the above-identified ISA threshold, thus best
matching the relative distribution of the two occupancy types (residential and mixed use), 82.8% of
matching the relative distribution of the two occupancy types (residential
and mixed use), 82.8% of
the total non-residential building stock of Cuenca City (3.6 of 4.3 km22) is indeed captured within the
the total non-residential building stock of Cuenca City (3.6 of 4.3 km ) is indeed captured within the
selected top-down-derived binary mixed use mask.
selected top-down-derived binary mixed use mask.
4.2. Identification of Residential vs. Mixed Use Areas Using VIIRS Data in Original Spatial Resolution
4.2. Identification of Residential vs. Mixed Use Areas Using VIIRS Data in Original Spatial Resolution
Applying the same approach of iterative thresholding as illustrated for the ISA data we use
Applying the same approach of iterative thresholding as illustrated for the ISA data we use VIIRS
VIIRS data to perform comparative analysis. As outlined above, VIIRS data is evaluated in that
data to perform comparative analysis. As outlined above, VIIRS data is evaluated in that context first
context first at its original resolution level and furthermore at aggregated level matching the ISA
at its original resolution level and furthermore at aggregated level matching the ISA resolution in order
resolution in order to guarantee direct spatial comparability and factor out potentially biased scale
to guarantee direct spatial comparability and factor out potentially biased scale effects.
effects.
4.2.1. Application of VIIRS Data in Original Spatial Resolution
4.2.1. Application of VIIRS Data in Original Spatial Resolution
As shown above with the ISA data, Table 2 relates the various tested VIIRS threshold values to
As shown above
with the
ISA data,
2 relatesratios.
the various
tested
threshold
valuesare
to
the corresponding
cadastral
building
useTable
distribution
Several
lightVIIRS
intensity
thresholds
the corresponding
cadastral building
use distribution ratios.built-up
Several area
lightdistribution
intensity thresholds
are
applied
iteratively, approximating
the occupancy-type-specific
shares on the
applied
iteratively,
approximating
the
occupancy-type-specific
built-up
area
distribution
shares
on
ground. Spatial overlay of the VIIRS data and the corresponding aggregated cadastral building use
the ground.
overlay
the VIIRS
datathat
and the
the53rd
corresponding
aggregated
cadastral
building
density
grid Spatial
(at the 15
arc-secoflevel)
indicates
percentile threshold
exactly
matches
the
use
density
grid
(at
the
15
arc-sec
level)
indicates
that
the
53rd
percentile
threshold
exactly
matches
land use distribution as derived from the cadastral information, i.e., 75% residential and 25% mixed
the share.
land use
distribution
asbinary
derived
from
cadastral (residential
information,
i.e.,
residential
25%
use
Figure
7 maps the
land
use the
classification
use
vs.75%
mixed
use) for 5and
selected
mixed use share.
Figure The
7 maps
theofbinary
(residential
use vs. mixed
use)
for 5
thresholds
respectively.
share
mixedland
use use
areaclassification
decreases thereby
corresponding
to the
higher
selected
thresholds
respectively.
The
share
of
mixed
use
area
decreases
thereby
corresponding
to
the
VIIRS cut-off points.
higher VIIRS cut-off points.
Table 2. VIIRS distribution thresholds (original 15 arc-sec grid) and corresponding building use
Table 2. VIIRS
distribution
thresholds
(original
15 arc-sec
grid) and
corresponding
building
use
distribution
ratios
(grey indicating
selected
best-matching
threshold,
orange
indicating 50%
threshold
distribution
ratios (grey indicating selected best-matching threshold, orange indicating 50%
for
comparison).
threshold for comparison).
VIIRS Distribution
Built-up Area Distribution
ID
VIIRS Distribution
Built-up Area Distribution
Min
Threshold
Max
Percentile
Residential Use
Mixed Use
ID
Min
Threshold
Max
Percentile
Residential Use
Mixed Use
1
3.2
16.0
73.8
18%
32.00%
68.00%
3.23.2
16.025.0
73.8 73.8 18% 31%
32.00%52.00%
68.00%
2 1
48.00%
3 2
29.00%
3.23.2
25.038.5
73.8 73.8 31% 50%
52.00%71.00%
48.00%
4
3.2
40.6
73.8
53%
75.00%
25.00%
3
3.2
38.5
73.8
50%
71.00%
29.00%
5
3.2
57.0
73.8
76%
96.00%
4.00%
4
3.2
40.6
73.8
53%
75.00%
25.00%
5
3.2
57.0
73.8
76%
8
96.00%
4.00%
Remote Sens. 2016, 8, 114
Remote Sens. 2016, 8, 114
9 of 19
9 of 19
Figure
Figure 7.
7. Binary land use classification of Cuenca City based on thresholds from original VIIRS data
(Table
and
orange
mixed
use.use.
Table
record
IDs are
in the in
figure
(Table 2).
2).Blue
Blueindicates
indicatesresidential
residential
and
orange
mixed
Table
record
IDsindicated
are indicated
the
as
1–5. as 1–5.
figure
degree of spatial overlap between the binary classified VIIRS data and the
Results regarding the degree
correspondinglyaggregated
aggregatedcadastral
cadastralgrid
grid
indicate
of total
the total
non-residential
building
correspondingly
indicate
thatthat
76%76%
of the
non-residential
building
stock
2
2
stock
of Cuenca
Cityof
(3.27
of 4.3
kmcaptured
) are captured
the selected
top-down-derived
use
of
Cuenca
City (3.27
4.3 km
) are
withinwithin
the selected
top-down-derived
mixedmixed
use mask
mask (using
the identified
best-matching
53rd percentile
threshold).
(using
the identified
best-matching
53rd percentile
threshold).
4.2.2. Application of VIIRS Data Aggregated to a 30 arc-sec
arc-sec Grid
Grid
In order to perform comparative analysis at identical scale levels, VIIRS data is aggregated to a
30 arc-sec
iterative
threshold
determination.
Table Table
3 illustrates
the various
threshold
arc-secgrid
gridbefore
before
iterative
threshold
determination.
3 illustrates
the tested
various
tested
values
from
the
aggregated
VIIRS
data
and
the
corresponding
cadastral
building
use
distribution
ratios.
threshold values from the aggregated VIIRS data and the corresponding cadastral building
use
To
guarantee ratios.
direct comparability
thecomparability
ISA-based analysis,
the ISA-based
thresholds for
the aggregated
VIIRS
distribution
To guarantee with
direct
with the
analysis,
the thresholds
data
areaggregated
applied in VIIRS
such a data
way are
thatapplied
the building
usea distribution
(shown
the right part
of
for the
in such
way that theratios
building
use in
distribution
ratios
Table
3) in
arethe
identical
to the
tests carried
out before
thecarried
ISA data.
(shown
right part
of Table
3) are identical
to using
the tests
out before using the ISA data.
Table
Table 3.3. VIIRS
VIIRS distribution
distribution thresholds
thresholds (aggregated
(aggregated 30
30 arc-sec
arc-sec grid)
grid) and
and corresponding
corresponding building
building use
use
distribution
ratios
(grey
indicating
selected
best-matching
threshold,
orange
indicating
50%
threshold
distribution ratios (grey indicating selected best-matching threshold, orange indicating 50%
for
comparison).
threshold
for comparison).
ID
ID
VIIRS
Distribution
VIIRS
Distribution
Built-up
Distribution
Built-up
AreaArea
Distribution
ThresholdMax MaxPercentile
PercentileResidential
Residential
Mixed
Min Min Threshold
Use Use Mixed
UseUse
1
22
3
43
54
2.7 2.7
2.7 2.7
2.7
2.7 2.7
2.7 2.7
12.2 12.2
19.3 19.3
34.1
34.1 36.9
36.9 57.1
65.1 65.1
65.1 65.1
65.1
65.1 65.1
65.1 65.1
15%
55%
5
2.7
57.1
65.1
90%
27%
50%
15%
27%
50%
55%
90%
32.00%
32.00%
52.00%
52.00%
70.00%
70.00%
75.00%
96.00%
75.00%
68.00%
68.00%
48.00%
48.00%
30.00%
30.00%
25.00%
4.00%
25.00%
96.00%
4.00%
Figure 8 maps the binary land use classification (residential use vs. mixed use) for the 5 tested
Figurein
8 the
maps
the binaryVIIRS
land data
use classification
vs. mixed
use)
the of
5 tested
thresholds
aggregated
respectively. (residential
As with theuse
previous
tests,
thefor
share
areas
thresholds
the aggregated
VIIRS
data corresponding
respectively. Astowith
the previous
the share of areas
with
mixedin
occupancy
decreases
thereby
the higher
cut-off tests,
points.
with Spatial
mixed overlay
occupancy
decreases
thereby
corresponding
to the higher cut-off
points.
of the
aggregated
VIIRS
data and the corresponding
cadastral
building use density
Spatial overlay
thepercentile
aggregated
VIIRS data
and the the
corresponding
building and
use
grid indicates
that theof55th
threshold
best matches
target valuecadastral
of 75% residential
density
griduse
indicates
that
the 55th
percentile
threshold
the target
25%
mixed
shares (as
derived
from
the in situ
cadaster best
data),matches
thus slightly
highervalue
than of
for75%
the
residential
and
25%
mixed
use
shares
(as
derived
from
the
in
situ
cadaster
data),
thus
slightly
higher
original-resolution VIIRS data. Evaluating the degree of spatial overlap between the aggregated VIIRS
than and
for the
VIIRSgrid,
data.we
Evaluating
the
degree
spatial
overlap between
the
data
the original-resolution
corresponding cadastral
detect that
79%
of theoftotal
non-residential
building
2 ) is captured within
aggregated
VIIRS
cadastral
grid, we
detect that 79%binary
of themixed
total
stock
of Cuenca
Citydata
(3.4 and
of 4.3the
kmcorresponding
the selected
top-down-derived
non-residential
building stock of Cuenca City (3.4 of 4.3 km2) is captured within the selected
use
mask.
top-down-derived binary mixed use mask.
9
Remote Sens. 2016, 8, 114
Remote Sens. 2016, 8, 114
10 of 19
10 of 19
Figure 8.
8. Binary
Binary land
land use
use classification
classification of
of Cuenca
Cuenca City
City based
based on
on thresholds
thresholds from
from aggregated
aggregated VIIRS
VIIRS
Figure
data (Table
(Table 3).
3). Blue
Blue indicates
indicates residential
residential and
and orange
orange mixed use. Table
Table record
record IDs
IDs are
are indicated
indicated in
in the
the
data
figure as
as 1–5.
1–5.
figure
5. Discussion
5.
Discussion
The application
application of
of ISA
ISA and
and alternatively
alternatively VIIRS
VIIRS data
data to
to identify
identify intra-urban
intra-urban occupancy
occupancy type
type
The
distribution patterns
patterns that
that we
we outline
outline in
in this
this paper
paper and
and the
the corresponding
corresponding findings
findings include
include several
several
distribution
interesting
aspects
for
further
discussion.
In
the
following
we
highlight
three
relevant
points
at
interesting aspects for further discussion. In the following we highlight three relevant points at
different
stages
in
the
model
setup.
First,
we
provide
some
background
information
on
selection
different stages in the model setup. First, we provide some background information on selection
criteria of
of VIIRS
VIIRS data.
data. Then,
Then, we
we discuss
discuss the
the actual
actual differences
differences in
in the
the outcome
outcome of
of the
the proposed
proposed binary
binary
criteria
land
use
classification
approach
when
implemented
using
ISA
vs.
VIIRS
data.
Finally,
we
highlight
land use classification approach when implemented using ISA vs. VIIRS data. Finally, we highlight the
the impact
proper
spatial
delineation
the model
outcome
by applying
a spatially
impact
that that
proper
urbanurban
spatial
delineation
has onhas
theon
model
outcome
by applying
a spatially
shrunk
shrunkmask
urban
for the Cuenca
urban
formask
the Cuenca
City testCity
case.test case.
5.1.
5.1. VIIRS Data Selection
Initial
Initial nightlights
nightlights data selection has a big
big influence
influence on
on the
the model
model outcome
outcome and
and is
is particularly
particularly
important
important in aa sense
sense that
that VIIRS
VIIRS data
data is
is provided
provided by
by NOAA-NGDC
NOAA-NGDC as
as basic
basic monthly
monthly average
average light
light
intensity
intensity composites
composites whereas ISA data
data comes
comes as
as aa fully-processed
fully-processed product
product derived
derived from
from annual
annual
DMSP-OLS
and calibrated
calibratedwith
withancillary
ancillarybuilt-up
built-upreference
referenceinformation.
information.The
The
number
DMSP-OLS composites and
number
of
of
cloud-free
observations
a crucial
factor
for producing
average
composites
as excessive
cloud-free
observations
is aiscrucial
factor
for producing
average
composites
as excessive
cloudcloud
cover
cover
can obscure
light-emitting
sources
on the
ground.
monthlyproducts
productsfewer
fewer observations
observations are
can obscure
light-emitting
sources
on the
ground.
In In
monthly
are
potentially
composite
grids
as compared
to yearly
products
and average
values
potentially available
availabletotocompute
compute
composite
grids
as compared
to yearly
products
and average
values
can therefore
moreturn
easily
turn
outskewed
to be skewed
and non-representative
of extended
can
therefore
more easily
out
to be
and non-representative
in casein
ofcase
extended
cloud
cloud incover
in the respective
month.
are obviously
other influencing
thatlight
can
cover
the respective
month. There
areThere
obviously
other influencing
parameters parameters
that can impair
impair light identification
suchfactors
as obscuring
factors
like
smoke
or fog reflections
and misleading
reflections
identification
such as obscuring
like smoke
or fog
and
misleading
from snow
cover,
from snow
cover,
lightning
or the
aurora.
cloud cover
clearly
considered
theinmost
lightning
or the
aurora.
However,
cloud
coverHowever,
is clearly considered
theismost
relevant
parameter
the
relevantofparameter
in the context
the compositing
process,
particularly
equatorial
regions
such
context
the compositing
process,of
particularly
in equatorial
regions
such asin
the
study area
of Ecuador.
as the
study we
area
of Ecuador.
For recent
our study
we readily-processed
evaluated the 6 monthly
most recent
available
For
our study
evaluated
the 6 most
available
composites
(at
readily-processed
(at the
of writing)
January–June
2015 (other
the
time of writing)monthly
coveringcomposites
January–June
2015time
(other
monthly covering
composites
were only available
as
monthly composites
were having
only available
as preliminary
versions
havingarea
lower
the
preliminary
beta versions
lower quality).
For thebeta
Cuenca
City study
thequality).
monthlyFor
VIIRS
Cuenca City
areaJune
the 2015
monthly
VIIRS
composites
of May
and of
June
2015 feature
the highest
composites
ofstudy
May and
feature
the highest
average
number
cloud-free
observations
(see
average
of cloud-free
observations
Table
4), number
thus providing
best data
reliability. (see Table 4), thus providing best data reliability.
Figure 99illustrates
illustratesthe
thelight
lightintensity
intensityvalues
values
6 analyzed
monthly
composites
as well
as
Figure
forfor
thethe
6 analyzed
monthly
composites
as well
as the
the corresponding
number
of cloud-free
observations
at a pixel-by-pixel
For
theintensity
light intensity
corresponding
number
of cloud-free
observations
at a pixel-by-pixel
basis. basis.
For the
light
grids,
grids, darker
blueindicate
tones indicate
intensity.
the cloud-free
observation
grids,
dark
blue
darker
blue tones
higher higher
intensity.
For theFor
cloud-free
observation
grids, dark
blue
would
would represent
best situation
cloud
cover
any day
month)
whereas
green,
represent
the best the
situation
(no cloud(no
cover
at any
dayatduring
the during
month)the
whereas
green,
yellow
and
yellow
andindicate
red colors
indicatedata
decreasing
data
reliability
to fewerobservations).
cloud-free observations).
red
colors
decreasing
reliability
(due
to fewer(due
cloud-free
10
Remote Sens. 2016, 8, 114
Remote Sens. 2016, 8, 114
11 of 19
11 of 19
Table
4. Average
number
of cloud-free
observations
in VIIRS
2015 2015
monthly
composites
for the
Table
4. Average
number
of cloud-free
observations
in VIIRS
monthly
composites
forCuenca
the
City
studyCity
areastudy
(greyarea
indicating
eventually
selected selected
monthlymonthly
composite).
Cuenca
(grey indicating
eventually
composite).
MonthMonth
January
January
February
February
March March
April April
May
June May
June
Number
Observation
Band (DNB)
NumberofofCloud-Free
Cloud-Free
Observation Day/Night
Day/Night
BandValue
(DNB) Value
Min
Max
Average
Min
Max
Min
Max
Average
Min
Max
2
9
5.47
3.8
86.9
2
9
5.47
3.8
86.9
6
4.52
3.52
129.1
22
6
4.52
3.52
129.1
00
4
1.98
0.0
92.3
4
1.98
0.0
92.3
33
8
5.37
3.9
55.8
8
5.37
3.9
55.8
6
11
8.97
3.5
55.7
11 11
8.97
3.5
55.7
66
8.81
3.3
73.8
6
11
8.81
3.3
73.8
Theoretically,
justjust
a couple
of high-quality
observations
can be can
sufficient
to produce
appropriate
Theoretically,
a couple
of high-quality
observations
be sufficient
to an
produce
an
composite
product.
While
the
monthly
composites
of
May
and
June
have
the
highest
appropriate composite product. While the monthly composites of May and June have thenumber
highest of
cloud-free
other months’other
composites
thus feature
value
number observations,
of cloud-free observations,
months’can
composites
can very
thus similar
feature light
very intensity
similar light
distributions
(as itdistributions
is the case for(as
example
thefor
April
composite).
therefore
explicitly
to data
intensity value
it is thefor
case
example
for theWe
April
composite).
We refer
therefore
reliability
as
an
indicator
as
opposed
to
general
data
quality.
explicitly refer to data reliability as an indicator as opposed to general data quality.
Figure
VIIRS
dataofofthe
thefirst
firstsix
sixmonths
months of
of 2015
2015 for
for the
ofof
average
Figure
9. 9.
VIIRS
data
the Cuenca
CuencaCity
Citystudy
studyarea.
area.Grids
Grids
average
light
intensity
(top).
Grids
showing
the
number
of
cloud-free
observations
at
the
cell
level
used
to to
light intensity (top). Grids showing the number of cloud-free observations at the cell level
used
produce the average light intensity composites (bottom).
produce the average light intensity composites (bottom).
11
Remote Sens.
Sens. 2016,
2016, 8,
8, 114
114
Remote
12of
of19
19
12
Although the May composite has the highest number of cloud-free observations on average,
the Mayidentical
composite
highest
number(see
of cloud-free
average,
that
that Although
value is almost
tohas
thethe
June
composite
Table 4). observations
In that case,onan
additional
value
is
almost
identical
to
the
June
composite
(see
Table
4).
In
that
case,
an
additional
parameter
parameter should be identified to justify selection. On the one hand visual inspection of the cell-level
should
be identified
to justify
selection. could
On thegive
one an
hand
visual
inspection
of the cell-level
distribution
distribution
of cloud-free
observations
extra
indication
on potential
data quality.
If, for
of
cloud-free
observations
could give an
indication
oncenter
potential
datanon-residential
quality. If, foractivity
example,
example,
more
cloud-free observations
areextra
found
in the city
(where
is
more
cloud-free
observations
are
found
in
the
city
center
(where
non-residential
activity
is
expected),
expected), that could be beneficial given the context of the presented study. Another parameter
that
be beneficial
contextrange,
of the presented
study.ofAnother
the
couldcould
be the
detected given
light the
intensity
with detection
higher parameter
intensitiescould
(i.e., be
likely
detected
light
intensity
range,
with
detection
of
higher
intensities
(i.e.,
likely
non-obscured)
being
non-obscured) being potentially favorable. Following the latter criterion, higher intensity levels are
potentially
favorable.
criterion,
higher
are identified
the June
identified in
the June Following
compositethe
as latter
compared
to the
May intensity
data set levels
(see Table
4). Other in
secondary
composite
as compared
theinto
Mayaccount
data setinfluencing
(see Table 4).
Other secondary
selection
could take
selection criteria
could to
take
parameters
that impair
lightcriteria
identification
(as
into
account
influencing
parameters
that
impair
light
identification
(as
mentioned
above).
Data
on
mentioned above). Data on those parameters is usually not publicly available though. As intra-urban
those
parameters
is usually
not publicly
available distributed
though. As for
intra-urban
observationsthe
at
cloud-free
observations
at cell-level
are similarly
the May cloud-free
and June composites,
cell-level
are similarly
distributed
for was
the May
and June
thefactor
higher
intensity
higher detected
light intensity
range
eventually
thecomposites,
determining
in detected
selectinglight
the June
data
range
eventually
set forwas
the test
study. the determining factor in selecting the June data set for the test study.
To
the the
differences
in spatial
patternspatterns
betweenbetween
the 6 available
composites,
To further
furtherhighlight
highlight
differences
in spatial
the 6monthly
available
monthly
cell-by-cell
light
intensity
deviations
of
every
grid
to
the
eventually
selected
June
composite
are
composites, cell-by-cell light intensity deviations of every grid to the eventually selected June
computed
illustrated
in as
Figure
10. In line
with the
above, the
May composite
compositeas
are
computed
illustrated
in Figure
10.observations
In line withdescribed
the observations
described
above,
matches
June dataset
most
also inmost
that closely
regard.also
Besides,
overall
patterns
those
the May the
composite
matches
theclosely
June dataset
in thatthe
regard.
Besides,
theofoverall
cell-by-cell
deviations
align
adequately
with
the
corresponding
grids
showing
the
number
of
cloud-free
patterns of those cell-by-cell deviations align adequately with the corresponding grids showing the
observations
(see Figure
9). The February
and
composites,
for example,
showfor
the
largest
number of cloud-free
observations
(see Figure
9).March
The February
and March
composites,
example,
deviations
to the June
grid ontoa the
cell-by-cell
basis,
assumingly
confirming
the poorer
reliabilitythe
of
show the largest
deviations
June grid
on athus
cell-by-cell
basis,
thus assumingly
confirming
those
when referring
the low
number
of cloud-free
observations.
March grid
poorergrids
reliability
of thosetogrids
when
referring
to the low
number ofSpecifically
cloud-freethe
observations.
can
be considered
unusable,
while
may be
additional
reasons
thebe
extreme
lightreasons
intensities
Specifically
the March
grid can
be there
considered
unusable,
while
therefor
may
additional
for
observed
during
the
month
of
February
(e.g.,
night
parades
and
other
associated
carnival
celebration
the extreme light intensities observed during the month of February (e.g., night parades and other
activities
the middle
of the month).
associatedincarnival
celebration
activities in the middle of the month).
Figure 10. Cell-by-cell deviations to the selected June grid.
Table 55illustrates
illustratesaaset
setofof
tested
VIIRS
threshold
values
using
the May
as alternative
in
Table
tested
VIIRS
threshold
values
using
the May
data data
as alternative
in order
order
to
demonstrate
potential
model
output
variation
in
case
a
different
monthly
composite
was
to demonstrate potential model output variation in case a different monthly composite was selected.
selected.
When the
applying
the 53rdthreshold
percentile
thresholdin(highlighted
in 5)green
in Table the
5) best
that
When
applying
53rd percentile
(highlighted
green in Table
that delivered
delivered
the
best
match
to
the
cadastral
data
in
the
June
composite,
a
65:35
building
use
match to the cadastral data in the June composite, a 65:35 building use distribution split was obtained
distribution
split
was
obtained
for
the
May
composite,
thus
significantly
overestimating
the
for the May composite, thus significantly overestimating the non-residential share. In the May data the
non-residential
In as
thefitting
Maythreshold
data the(highlighted
64th percentile
as approximating
fitting threshold
64th
percentile isshare.
identified
in greyisinidentified
Table 5) best
the
(highlighted
in
grey
in
Table
5)
best
approximating
the
aggregated
cadastral
grid.
aggregated cadastral grid.
12
Remote Sens. 2016, 8, 114
13 of 19
Table 5. VIIRS distribution thresholds (original 15 arc-sec grid) and corresponding building use
distribution ratios using the monthly composite for May ratios (grey indicating selected best-matching
threshold, green indicating previously identified June threshold for comparison).
VIIRS Distribution
ID
1
2
3
6
9
Built-up Area Distribution
Min
Threshold
Max
Percentile
Residential Use
Mixed Use
3.5
3.5
3.5
3.5
3.5
16.0
29.5
31.1
37.0
50.6
55.6
55.6
55.6
55.6
55.6
24%
50%
53%
64%
90%
32.00%
62.00%
65.00%
75.00%
96.00%
68.00%
37.00%
35.00%
25.00%
4.00%
5.2. Comparative Analysis of ISA- and VIIRS-Based Results of the Binary Land Use Classification
The second aspect to be discussed is a comparison of the model output when using ISA and
VIIRS-DNB data respectively. This is relevant in several aspects, most specifically (1) in terms of
evaluating feasibility of continued applicability of the presented approach with the DMSP program
fading out as well as (2) to assess the impact and examine expected multisided improvements due to
VIIRS’ improved spatial and radiometric resolution as compared to OLS.
To factor out potential influences caused by the higher spatial resolution we first compare the
findings of the ISA-based analysis to those using a correspondingly aggregated 30 arc-sec VIIRS
grid. Results prove to be similar in fact, with a 55th percentile threshold identified as best fit to
distinguish residential and mixed occupancy areas in the VIIRS data as compared to the 50% threshold
in the ISA data. In case of applying the same 50% threshold to the aggregated VIIRS composite, the
obtained occupancy distribution in the correspondingly aggregated cadastral data would show a
70:30 residential-mixed split as compared to the targeted 75:25 ratio. When evaluating the degree of
spatial overlap as a measure of model output accuracy, applying the respectively identified best-fitting
thresholds to both data sets results in a slightly better capturing of non-residential built-up area in the
binary mixed use mask that is derived from the ISA data (83%) as compared to the aggregated VIIRS
data based mask (79%). If again the 50% threshold was applied to the VIIRS composite instead of the
identified 55th percentile threshold, approximately 84% of the non-residential built-up area would be
captured. While thereby a marginally better result is achieved in terms of capturing non-residential
built-up area, the residential-mixed distribution ratio would be skewed and mixed use areas would
actually be overrepresented spatially.
Applying the VIIRS data in its original spatial resolution (15 arc-sec), the best-fitting threshold
value to approximate the targeted 75:25 residential-mixed distribution pattern is identified at the 53rd
percentile. This is slightly below the threshold value identified for the aggregated VIIRS composite
(55th percentile). 76% of the total non-residential building stock of Cuenca City (3.27 of 4.3 km2 ) is
captured within the derived mixed use mask. That value is below both the 79% value when using the
aggregated VIIRS data and the 83% value when using the ISA data. In case of applying the initial 50%
threshold for the binary classification, 79% of the non-residential building stock would be captured at
a 71:29 occupancy type ratio distribution.
Checking those numbers it therefore appears that using ISA data renders a better model
performance than using VIIRS data both in original and aggregated form, inasmuch as more
non-residential built-up area is detected in the binary masks that were derived using optimized
thresholding to match residential-mixed occupancy distribution ratios. However, while a higher
percentage of the non-residential built-up area is captured, ISA-derived mixed use areas are slightly
more scattered. Taking VIIRS as input data source clusters the detection more in a sense that the
average cell-level non-residential built-up density is higher in those binary occupancy type mask
derivatives. Using the original-resolution composite, 76% of the total non-residential building stock
(3.27 km2 ) is captured within 25.75 km2 , thus featuring an average non-residential built-up density
Remote Sens. 2016, 8, 114
14 of 19
Remote Sens. 2016, 8, 114
14 of 19
km2 .
of 12.7% per
When using ISA data, 83% of the total non-residential building stock (3.57 km2 ) is
With
the
threshold
values
and associated
are2 . obviously rather similar for the
captured within 32 km2 , thus
an average
density ofparameters
11.1% per km
VIIRSand
another interesting
evaluative
perspective
is to derive
With
theISA-based
threshold approaches,
values and associated
parameters
are obviously
rather similar
for thea
corresponding
binary
mask
from
the
aggregated
cadastral
data
and
then
check
spatial
pattern
VIIRS- and ISA-based approaches, another interesting evaluative perspective is to derive
a
concurrence
to
the
nightlights
products.
Figure
11
shows
the
binary
classification
of
the
aggregated
corresponding binary mask from the aggregated cadastral data and then check spatial pattern
cadastral data
both for products.
the 15 arc-sec
(left)
and thethe
30binary
arc-secclassification
(right) aggregate.
The binary
concurrence
to (top),
the nightlights
Figure
11 shows
of the aggregated
mask
separates
cells
that
contribute
strongly
to
the
total
non-residential
area
from
cells
thatmask
only
cadastral data (top), both for the 15 arc-sec (left) and the 30 arc-sec (right) aggregate. The binary
have a marginal
share.
This approach
is congruent
to the nightlights
approach
in aa
separates
cells that
contribute
strongly to
the total non-residential
area thresholding
from cells that
only have
sense
that
it
aims
at
separating
high-intensity
from
low-intensity
cells
(referring
marginal share. This approach is congruent to the nightlights thresholding approach in a sense that to
it
“non-residential”
observed parameter).
The thresholds
are determined
in a way that,
for the
aims
at separating as
high-intensity
from low-intensity
cells (referring
to “non-residential”
as as
observed
nightlights data
thresholding,
75:25 residential-mixed
occupancy
type reference
ratio split is
parameter).
The thresholds
arethe
determined
in a way that, as
for the nightlights
data thresholding,
matched
best-possible.
For
the
15
arc-sec
grid
the
threshold
is
identified
at
0.25%
(i.e.,
cells
the 75:25 residential-mixed occupancy type reference ratio split is matched best-possible.that
Forhave
the
a
percent-contribution
to
the
total
non-residential
area
of
less
or
equal
than
0.25%),
while
for
30
15 arc-sec grid the threshold is identified at 0.25% (i.e., cells that have a percent-contribution to thethe
total
arc-sec grid thearea
derived
threshold
is 1%. Interestingly,
thearc-sec
difference
exactly
reflects
non-residential
of less
or equalvalue
than 0.25%),
while for the 30
grid in
thefact
derived
threshold
the
scale
difference
between
the
two
datasets
(i.e.,
factor
of
4).
The
binary
15
arc-sec
classification
value is 1%. Interestingly, the difference in fact exactly reflects the scale difference between the two
results in(i.e.,
a non-residential
mask
(in 15
dark
blue)classification
that captures
88.3%in of
the total non-residential
datasets
factor of 4). The
binary
arc-sec
results
a non-residential
mask (in
2, thus an average density of 14.9% per km2
building
area
of
Cuenca
City
on
an
area
of
25.5
km
dark blue) that captures 88.3% of the total non-residential building area of Cuenca City on an area of
2 average density in the VIIRS-derived 15 arc-sec binary mask). The 30
(compared
to the
12.7%/km
25.5
km2 , thus
an average
density
of 14.9% per km2 (compared to the 12.7%/km2 average density in
arc-sec
mask on the
hand
captures
86%.
the
VIIRS-derived
15other
arc-sec
binary
mask).
The 30 arc-sec mask on the other hand captures 86%.
Figure 11.
11. Binary
Binary classification
classification of
of the
the non-residential
non-residential cadastral
cadastral built-up
built-up area
area aggregated
aggregated to
to 15
15 arc-sec
arc-sec
Figure
(top-left) and
and30 arc-sec
30 arc-sec
(top-right)
on the percent-contribution
to the total
(top-left)
grids grids
(top-right)
based onbased
the percent-contribution
to the total non-residential
non-residential
areaBest-matching
of Cuenca City.
Best-matching
binary
classifications
as derived
from
15 arc-sec
area
of Cuenca City.
binary
classifications
as derived
from 15 arc-sec
VIIRS
(bottom-left)
VIIRS
and 30 arc-secdata.
ISA (bottom-right) data.
and
30 (bottom-left)
arc-sec ISA (bottom-right)
For comparative
comparative purposes,
purposes, the
the bottom
bottom two
two illustrations
illustrations in
in Figure
Figure 11
11 show
show the
the above-presented
above-presented
For
best-matching
binary
masks
derived
from
the
15
arc-sec
VIIRS
and
the
30
arc-sec
ISA
data. Visually
Visually
best-matching binary masks derived from the 15 arc-sec VIIRS and the 30 arc-sec ISA data.
evaluatingspatial
spatialdistribution
distribution
extent
the non-residential
in maps
the two
maps
reveals
evaluating
andand
extent
of theofnon-residential
class in class
the two
reveals
interesting
interesting
patterns.
The
VIIRS-derived
mask
covers
the
south-western
corner
of
the
corresponding
patterns. The VIIRS-derived mask covers the south-western corner of the corresponding cadaster-based
cadaster-based non-residential mask well and misses out on the north-eastern corner whereas it is
the other way around with the ISA-derived mask. VIIRS in that context seems not to detect
14
Remote
Remote Sens.
Sens. 2016,
2016, 8,
8, 114
114
15
15of
of19
19
above-average light intensities from the Cuenca City Airport (Aeropuerto Mariscal La Mar),
non-residential mask well and misses out on the north-eastern corner whereas it is the other way around
whereas it is a major contributing factor in the ISA data. The latter could be explained with the
with the ISA-derived mask. VIIRS in that context seems not to detect above-average light intensities
inherent data configuration of ISA, which per se is more correlated with built-up area rather than
from the Cuenca City Airport (Aeropuerto Mariscal La Mar), whereas it is a major contributing factor
pure light intensity.
in the ISA data. The latter could be explained with the inherent data configuration of ISA, which per
se
more correlated
built-up
rather than
pure light
intensity.
5.3.isEvaluating
Model with
Sensitivity
via area
Application
of Different
Spatial
Urban Delineation
5.3. Evaluating
Model
Sensitivity
via Application
of Different
Delineation approach with a
For further
evaluation
of the
model sensitivity
we Spatial
re-runUrban
the implemented
geospatially shrunk urban mask. While in the above-outlined implementations all the ISA and VIIRS
For further evaluation of the model sensitivity we re-run the implemented approach with a
grid cells were considered that fall within a pre-defined urban area of Cuenca City, now a more
geospatially shrunk urban mask. While in the above-outlined implementations all the ISA and VIIRS
central part of the urban agglomeration is selected. Two tests are carried out in that context. For the
grid cells were considered that fall within a pre-defined urban area of Cuenca City, now a more central
first one, we keep the same built-up area occupancy type distributions (75% residential vs. 25%
part of the urban agglomeration is selected. Two tests are carried out in that context. For the first
mixed use for VRIIS original resolution and 74% residential vs. 26% mixed use for the aggregated
one, we keep the same built-up area occupancy type distributions (75% residential vs. 25% mixed use
grid). In the second test, we keep the same threshold values identified above as the best match,
for VRIIS original resolution and 74% residential vs. 26% mixed use for the aggregated grid). In the
respectively, for each dataset (50th percentile for the ISA data and 53rd percentile for the VIIRS
second test, we keep the same threshold values identified above as the best match, respectively, for
data).
each dataset (50th percentile for the ISA data and 53rd percentile for the VIIRS data).
For the first test, the derived best-matching thresholds are now higher for both datasets. For the
For the first test, the derived best-matching thresholds are now higher for both datasets. For
ISA data the 55th percentile and for the VIIRS-original and aggregated grids the 63rd and 65th
the ISA data the 55th percentile and for the VIIRS-original and aggregated grids the 63rd and 65th
percentile are identified respectively. This was expected as predominantly residential areas in the
percentile are identified respectively. This was expected as predominantly residential areas in the
periphery of the city are now not included in the newly-defined urban mask and those cells
periphery of the city are now not included in the newly-defined urban mask and those cells (featuring
(featuring lower ISA and light intensity values) are thus missing in the histograms. The threshold
lower ISA and light intensity values) are thus missing in the histograms. The threshold increment
increment is higher for the VIIRS data application (roughly 10%–12%-increase) as compared to the
is higher for the VIIRS data application (roughly 10%–12%-increase) as compared to the ISA data
ISA data application (5%-increase). This aspect can be associated with different sensitivity of the
application (5%-increase). This aspect can be associated with different sensitivity of the identified
identified VIIRS and ISA thresholds due to varying histogram distributions (see Figure 12). Given
VIIRS and ISA thresholds due to varying histogram distributions (see Figure 12). Given the purpose
the purpose of the presented modeling, a more even histogram distribution could imply less
of the presented modeling, a more even histogram distribution could imply less sensitivity in the
sensitivity in the threshold determination.
threshold determination.
Figure
study.
Figure 12.
12. VIIRS-DNB (aggregated) and ISA histograms for the Cuenca city case study.
In fact, using the ISA composite, it only takes an increment of 24% (raising the threshold from
In fact, using the ISA composite, it only takes an increment of 24% (raising the threshold from
50% to 74%) to change the building occupancy type distribution ratio from 74:26 to 96:4. Regarding
50% to 74%) to change the building occupancy type distribution ratio from 74:26 to 96:4. Regarding
the aggregated DNB-VIIRS it would require a 35% increment (raising the threshold from 55% to
the aggregated DNB-VIIRS it would require a 35% increment (raising the threshold from 55% to
90%) to achieve the same theoretic change of the built-up area distribution. Small threshold shifts
90%) to achieve the same theoretic change of the built-up area distribution. Small threshold shifts
therefore have a bigger impact when using ISA as compared to VIIRS. To illustrate and emphasize
therefore have a bigger impact when using ISA as compared to VIIRS. To illustrate and emphasize this
this statistically, we use a sample of 10 value pairs each for ISA, original-resolution VIIRS, and
statistically, we use a sample of 10 value pairs each for ISA, original-resolution VIIRS, and aggregated
aggregated VIIRS as compared to the cadastral building use distribution, and run a linear
VIIRS as compared to the cadastral building use distribution, and run a linear regression (see Figure 13).
regression (see Figure 13). Considering all the value pairs, in fact the original-resolution VIIRS data
Considering all the value pairs, in fact the original-resolution VIIRS data shows the steepest slope in
shows the steepest slope in the regression (1.2468) whereas the aggregated VIIRS data indeed show
the regression (1.2468) whereas the aggregated VIIRS data indeed show the flattest slope (1.0341) with
the flattest slope (1.0341) with ISA in between (1.1613). The aggregated VIIRS would thus be the
ISA in between (1.1613). The aggregated VIIRS would thus be the least sensitive to threshold shifting
least sensitive to threshold shifting in a sense that the building use distribution ratios would
in a sense that the building use distribution ratios would accordingly deviate less from the target value
accordingly deviate less from the target value (see dashed line in Figure 13). While steepest when
(see dashed line in Figure 13). While steepest when considering all value pairs, the slope of the original
considering all value pairs, the slope of the original VIIRS graph matches ISA almost identically
15
Remote Sens. 2016, 8, 114
16 of 19
Remote Sens. 2016, 8, 114
16 of 19
VIIRS graph matches ISA almost identically around the relevant target value (dashed line). Threshold
around
target
valuewould
(dashed
line). Threshold
shiftsimpact
in theon
nightlights
products
would
shifts
in the
the relevant
nightlights
products
therefore
have a similar
the resulting
building
use
therefore
have
a
similar
impact
on
the
resulting
building
use
distribution
ratios.
distribution ratios.
Figure 13.
13. Plot
Plotof
ofresidential
residentialbuilding
buildinguse
use
ratio
data
percentile
from
histogram
distribution
for
ratio
vs.vs.
data
percentile
from
histogram
distribution
for ISA
ISA (extended
sample
of Table
1), original-resolution
VIIRS
(extended
sample
of Table
2),
(extended
datadata
sample
of Table
1), original-resolution
VIIRS
(extended
datadata
sample
of Table
2), and
and aggregated
(extended
data sample
3). The
lineresidential
shows thebuilding
residential
aggregated
VIIRSVIIRS
(extended
data sample
of Table of
3). Table
The dashed
linedashed
shows the
use
building
use ratio
for(75%)
Cuenca
City (75%)
deriveddata.
fromRegression
cadastral equations
data. Regression
equations
are
ratio
for Cuenca
City
as derived
fromas
cadastral
are colored
according
colored
according
to the respective graphs.
to
the respective
graphs.
In the
thesecond
secondtest
testusing
using
best-matching
thresholds
identified
the initial
In
thethe
best-matching
thresholds
identified
withwith
the initial
urbanurban
mask mask
(50th
(50th
percentile
for
ISA
and
53rd
and
55th
percentile
respectively
for
VIIRS)
the
newly
obtained
percentile for ISA and 53rd and 55th percentile respectively for VIIRS) the newly obtained built-up
area
built-up
area
occupancy
type
distribution
for
the
ISA
data
now
corresponds
to
a
50%
residential
occupancy type distribution for the ISA data now corresponds to a 50% residential and 50% mixed and
use
50% mixed
usetheshare
fordistribution
the VIIRS now
data shows
the distribution
shows a pattern
56%
share
while for
VIIRSwhile
data the
a pattern of now
56% residential
and 44%ofmixed
residential
and the
44%original
mixed resolution
use considering
the original
resolution
andthe
a aggregated
48:52 ratio 30
taking
in
use
considering
and a 48:52
ratio taking
in account
arc-sec
account
thenewly
aggregated
arc-secarea
grid.
These newly
derived built-up
area
type
grid.
These
derived30built-up
occupancy
type distribution
patterns
are occupancy
similar for both
distribution
patterns
are
similar
for
both
data
sources
(ISA
and
VIIRS)
and
clearly
overestimate
the
data sources (ISA and VIIRS) and clearly overestimate the share of mixed use area. This, again, was
share
of
mixed
use
area.
This,
again,
was
expected
in
the
same
way
than
the
first
test
result
expected in the same way than the first test result inasmuch as in the selected central part of the urban
inasmuch
as ina the
selected
centralof
part
of the urban
area there
are a decreased
number ofperiphery.
residential
area
there are
decreased
number
residential
buildings
as compared
to the sub-urban
buildings
as
compared
to
the
sub-urban
periphery.
Both tests are correlated in a sense that they give indication on higher light intensity values being
Both in
tests
are correlated
a sense
that they
indication
on higher
light intensity
clustered
central
core urban in
areas
of Cuenca
Citygive
whereas
sub-urban
areas feature
dimmervalues
lights
beingconsequently
clustered in central
coreISA
urban
areason
of average
Cuenca as
City
whereas
sub-urban
areas feature
dimmer
(and
also lower
values)
a result
of higher
residential
densities.
This
lights (and
consequently
alsothe
lower
ISA values)
on average
aspre-identification
a result of higher of
residential
exercise
basically
highlights
importance
of correct
spatial
the urbandensities.
area for
This exercise
basically analysis.
highlights
the urban
importance
correct overspatialorpre-identification
the urban
subsequent
intra-urban
If the
mask isofspatially
under-defined, theofappropriate
area
for
subsequent
intra-urban
analysis.
If
the
urban
mask
is
spatially
overor
under-defined,
the
nightlights threshold values would de- or increase respectively.
appropriate nightlights threshold values would de- or increase respectively.
16
Remote Sens. 2016, 8, 114
17 of 19
6. Conclusions and Outlook
The presented result of the ISA data application is very interesting as it in fact backs up the prior
non-evaluated assumption implemented in the Central American CDRP model to use ISA median
values as a threshold for the binary land use classification of residential and mixed use areas. At the
continental scale, without ground reference data as are available for the presented Cuenca City test
case study, the use of the median value seemed most appropriate as it introduces the least possible
subjectivity and merely separates a certain data set in high and low according to its histogram without
additionally induced statistical skew.
With that initially assumed median value (50%) threshold for the binary ISA classification
confirmed through comparative in situ data analysis for an accurately defined urban agglomeration,
the presented case study is considered very beneficial for the overall implementation process of the
CDRP initiative. Also, the second re-run of the model with a geospatially shrunken, more central
urban mask that showcased the correspondingly expected threshold upward shifts provides another
back-up for the model validity as well as underlining the importance of accurate urban delineation in
the first place.
It has to be noted that with the presented Cuenca City test case, those findings have to date just
been evaluated for that one particular city and caution is advised when it comes to directly transferring
those conclusions to other cities. With the CDRP exposure and subsequent risk and loss models already
implemented for all of Central America, further test studies can be carried out to increase the sample
size of the model evaluation and also test the approach in different regional settings. Cuenca City is
considered a rather typical Latin American city with regular patterns of clustered land use within
the urban agglomeration. Though, in Central America, basically no major deviations are expected
with regard to model applicability, it will be interesting to see testing results when extending to the
Caribbean and across as well as to cities of much larger spatial urban extent. Analysis of areas further
from the equator may furthermore be influenced by varying seasonal day duration as well as different
cloud cover patterns, two parameters which directly affect the nighttime lights compositing.
Testing VIIRS-DNB data as alternative to the DMSP-OLS-based ISA data is considered a crucial
step towards a continued applicability of the model. With the DMSP program fading out, VIIRS-DNB
is considered the natural successor to the OLS-based nightlights products. With certain visual
improvements expected due to VIIRS’s higher spatial and radiometric resolution as compared to
OLS, it is still highly valuable to get a clear idea about how these improvements eventually transfer
to the binary land use classification output. One specific finding in that context refers to the stronger
clustering and thus higher non-residential built-up density in VIIRS-derived binary classification
as compared to the ISA-based approach. A major and often-stated benefit of VIIRS with regard
to intra-urban pattern analysis is the much-improved radiometric resolution which eliminates the
restricting light intensity saturation issues in urban centers in OLS data [21]. For the purpose of the
presented study, however, this is not of major relevance as the OLS-derived ISA data refer to a specific
radiance-calibrated nightlights product where such saturation issues have already been addressed [32].
Only two ISA datasets are publicly available, however, for the years 2000 and 2010, thus limiting
potential direct applicability of the proposed approach in continuous time series analyses. Anyway,
even using the annually produced and publicly available OLS stable lights-product would likely not
result in a major deterioration of the binary classification as the high intensity values would still
be correctly identified irrespective of their relative lower displacement in the histogram due to the
saturation issue.
For future studies as well as potential longer-term time series analyses, the finding is very positive
in indicating that the DMSP-OLS-based ISA data and the more recent VIIRS data seem to be applicable
in very similar fashion as input data sources for the residential-mixed identification model. Two
main differences found using VIIRS data concern the varying threshold sensitivity and the amount of
built-up detected per square kilometer of land use. The procedure of binary land use classification
Remote Sens. 2016, 8, 114
18 of 19
using VIIRS is considered more flexible than using ISA, and has the potential to give a finer-scale
classification of residential and mixed used in urban areas.
Acknowledgments: The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and
Development/World Bank and its affiliated organizations, or those of its Executive Directors or the governments
they represent. Support from the World Bank’s Probabilistic Risk Assessment Program (CAPRA) and the
associated Country Disaster Risk Profiles (CDRP) initiative is acknowledged. An earlier and less comprehensive
version of this article showing preliminary results was presented as contribution to the 1st International Electronic
Conference on Remote Sensing (ECRS-1). The concepts discussed are grounded within recent work on setting up
a global-scale exposure model within the framework of the CDRP initiative, implemented in Central America and
the Caribbean. The authors appreciate initial conceptual discussions within the CDRP team, specifically with
Rashmin Gunasekera and Oscar Ishizawa, as well as comments from four anonymous reviewers that helped
improve the quality of the paper.
Author Contributions: Christoph Aubrecht developed the conceptual framework for the presented top-down
method and tested geospatial implementation in various Central American countries in the context of the CDRP
project initiative. José Antonio León Torres then implemented the developed approach for the Cuenca City
(Ecuador) case study and performed the comparative analysis with the bottom-up cadastral data. Following up on
the successful presentation of first results at the ECRS-1 conference, further research was carried out regarding the
additional integration of alternative nighttime EO data sources, leading to the multi-sensor approach illustrated in
this paper. Christoph Aubrecht developed that idea conceptually and José Antonio León Torres again implemented
for the Cuenca City test site. The paper was initiated by Christoph Aubrecht, with major contributions from both
authors in all sections.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
United Nations. World Urbanization Prospects: The 2014 Revision; United Nations: New York, NY, USA, 2014.
World Bank. World Development Report 2014: Risk and Opportunity—Managing Risk for Development; World
Bank: Washington, DC, USA, 2014.
World Bank & International Monetary Fund. Global Monitoring Report 2015/2016: Development Goals in an Era
of Demographic Change; World Bank: Washington, DC, USA, 2016.
Gunasekera, R.; Ishizawa, O.; Aubrecht, C.; Murray, S.; Blankespoor, B.; Pomonis, A.; Daniell, J. Developing
an Adaptive Exposure Model to Support the Generation of Country Disaster Risk Profiles. Earth-Sci. Rev.
2015, 150, 594–608. [CrossRef]
Aubrecht, C.; Gunasekera, R.; Ishizawa, O. Urban characterization aspects and associated spatial scale
considerations in the context of global-level exposure modeling: Comparative data evaluation. In
Proceedings of the Second European Conference on Earthquake Engineering and Seismology (2ECEES),
Istanbul, Turkey, 24–29 August 2014.
Jaiswal, K.; Wald, D.; Porter, K. A Global Building Inventory for Earthquake Loss Estimation and Risk
Management. Earthq. Spectra 2010, 26, 731–748. [CrossRef]
Elvidge, C.D.; Tuttle, B.T.; Sutton, P.C.; Baugh, K.E.; Howard, A.T.; Milesi, C.; Bhaduri, B.; Nemani, R. Global
Distribution and Density of Constructed Impervious Surfaces. Sensors 2007, 7, 1962–1979. [CrossRef]
Miller, S.D.; Mills, S.P.; Elvidge, C.D.; Lindsey, D.T.; Lee, T.F.; Hawkins, J.D. Suomi satellite brings to light
a unique frontier of nighttime environmental sensing capabilities. Proc. Natl. Acad. Sci. USA 2012, 109,
15706–15711. [CrossRef] [PubMed]
Aubrecht, C.; León Torres, J.A. Top-down identification of mixed vs. residential use in urban areas: Evaluation
of remotely sensed nighttime lights for a case study in Cuenca City, Ecuador. In Proceedings of the 1st
International Electronic Conference on Remote Sensing (ECRS-1), 22 June–5 July 2015; 2015; Volume 1,
p. d005. [CrossRef]
Welch, R. Monitoring urban population and energy utilization patterns from satellite data. Remote Sens.
Environ. 1980, 9, 1–9. [CrossRef]
Elvidge, C.D.; Baugh, K.E.; Kihn, E.A.; Kroehl, H.W.; Davis, E.R.; Davis, C.W. Relation between satellite
observed visible-near infrared emissions, population, economic activity and electric power consumption.
Int. J. Remote Sens. 1997, 18, 1373–1379. [CrossRef]
Sutton, P.; Roberts, D.; Elvidge, C.; Baugh, K. Census from Heaven: An estimate of the global human
population using night-time satellite imagery. Int. J. Remote Sens. 2001, 22, 3061–3076. [CrossRef]
Remote Sens. 2016, 8, 114
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
19 of 19
Doll, C.N.H.; Muller, J.-P.; Morley, J.G. Mapping regional economic activity from night-time light satellite
imagery. Ecol. Econ. 2006, 57, 75–92. [CrossRef]
Sutton, P.C.; Cova, T.J.; Elvidge, C.D. Mapping “Exurbia” in the conterminous United States using nighttime
satellite imagery. Geocarto Int. 2006, 21, 39–45. [CrossRef]
Elvidge, C.; Sutton, P.C.; Tuttle, B.T.; Ghosh, T.; Baugh, K.E. Global urban mapping based on nighttime lights.
In Global Mapping of Human Settlement; Gamba, P., Herold, M., Eds.; CRC Press: Boca Raton, FL, USA, 2009;
pp. 129–144.
Zhou, Y.; Smith, S.J.; Elvidge, C.D.; Zhao, K.; Thomson, A.; Imhoff, M. A cluster-based method to map urban
area from DMSP/OLS nightlights. Remote Sens. Environ. 2014, 147, 173–185. [CrossRef]
Rich, C.; Longcore, T. Ecological Consequences of Artificial Night Lighting, 1st ed.; Island Press: Washington, DC,
USA, 2006.
Aubrecht, C.; Elvidge, C.D.; Longcore, T.; Rich, C.; Safran, J.; Strong, A.E.; Eakin, C.M.; Baugh, K.E.;
Tuttle, B.T.; Howard, A.T.; et al. A global inventory of coral reef stressors based on satellite observed
nighttime lights. Geocarto Int. 2008, 23, 467–479. [CrossRef]
Rodrigues, P.; Aubrecht, C.; Gil, A.; Longcore, T.; Elvidge, C. Remote sensing to map influence of light
pollution on Cory’s shearwater in São Miguel Island, Azores Archipelago. Eur. J. Wildlife Res. 2012, 58,
147–155. [CrossRef]
Elvidge, C.D.; Imhoff, M.L.; Baugh, K.E.; Hobson, V.R.; Nelson, I.; Safran, J.; Dietz, J.B.; Tuttle, B.T. Night-time
lights of the world: 1994–1995. ISPRS J. Photogramm. Remote Sens. 2001, 56, 81–99. [CrossRef]
Elvidge, C.D.; Baugh, K.E.; Zhizhin, M.; Hsu, F.-C. Why VIIRS data are superior to DMSP for mapping
nighttime lights. Proc. Asia-Pac. Adv. Netw. 2013, 35, 62–69. [CrossRef]
Levin, N.; Johansen, K.; Hacker, J.M.; Phinn, S. A new source for high spatial resolution night time
images—The EROS-B commercial satellite. Remote Sens. Environ. 2014, 149, 1–12. [CrossRef]
Elvidge, C.D.; Milesi, C.; Dietz, J.B.; Tuttle, B.T.; Sutton, P.C.; Nemani, R.; Vogelmann, J.E. U.S. constructed
area approaches the size of Ohio. EOS Trans. Am. Geophys. Union 2004, 85, 233–240. [CrossRef]
Potere, D.; Schneider, A.; Angel, S.; Civco, D. Mapping urban areas on a global scale: Which of the eight
maps now available is more accurate? Int. J. Remote Sens. 2009, 30, 6531–6558. [CrossRef]
Esch, T.; Taubenböck, H.; Roth, A.; Heldens, W.; Felbier, A.; Thiel, M.; Schmidt, M.; Müller, A.; Dech, S.
TanDEM-X mission—New perspectives for the inventory and monitoring of global settlement patterns.
J. Appl. Remote Sens. 2012, 6. [CrossRef]
Pesaresi, M.; Huadong, G.; Blaes, X.; Ehrlich, D.; Ferri, S.; Gueguen, L.; Halkia, M.; Kauffmann, M.; Kemper, T.;
Lu, L.; et al. A global human settlement layer from optical HR/VHR RS data: Concept and first results.
IEEE J. Sel. Top. Appl. Earth Observ. Remote Sens. 2013, 6, 2102–2131. [CrossRef]
Baugh, K.; Hsu, F.-C.; Elvidge, C.; Zhizhin, M. Night lights compositing using the VIIRS day-night band:
Preliminary results. Proc. Asia-Pac. Adv. Netw. 2013, 35, 70–86. [CrossRef]
Municipalidad de Cuenca (GAD). Catastro Urbana de la Ciudad de Cuenca; Departamento de Avalúos y
Catastros: Cuenca, Ecuador, 2015.
Foody, G.M.; Cox, D.P. Sub-pixel land cover composition estimation using a linear mixture model and fuzzy
membership functions. Int. J. Remote Sens. 1994, 15, 619–631. [CrossRef]
Fisher, P. The pixel: A snare and a delusion. Int. J. Remote Sens. 1997, 18, 679–685. [CrossRef]
Aubrecht, C.; Gunasekera, R.; Ishizawa, O.; Pita, G. The flipside of “urban”—A novel model for
built-up-adjusted rural-urban pattern identification and population reallocation. In Internal CDRP Working
Paper 2015; World Bank: Washington, DC, USA, 2015.
Elvidge, C.D.; Baugh, K.E.; Dietz, J.B.; Bland, T.; Sutton, P.C.; Kroehl, H.W. Radiance calibration of DMSP-OLS
low-light imaging data of human settlements. Remote Sens. Environ. 1999, 68, 77–88. [CrossRef]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons by Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).