ENVIRONMENTAL ASSESSMENT OF AGRICULTURE AT A
REGIONAL SCALE IN THE PAMPAS OF ARGENTINA
E. F. VIGLIZZO∗, A. J. PORDOMINGO, M. G. CASTRO and F. A. LERTORA
INTA/CONICET, Centro Regional La Pampa, La Pampa, Argentina
(∗ author for correspondence, e-mail: [email protected])
(Received 23 May 2002; accepted 25 September 2002)
Abstract. Governments need good information to design policies. However, in the Argentine Pampas there are neither sufficient knowledge on environmental issues, nor clear perception of environmental alterations across space and time. The general objective of this work was to provide decision
makers with a scientifically sound set of indicators aiming at the assessment of current status and
future trends in the rural environment of this sensitive region. As driving criteria to select indicators,
we assumed that they had to be sound, simple to calculate, easy to understand, and easily applicable by decision makers. They are related closely to significant ecological structures and functions.
Twelve basic indicators were identified: (1) land use, (2) fossil energy use, (3) fossil energy use
efficiency, (4) nitrogen (N) balance, (5) phosphorus (P) balance, (6) nitrogen contamination risk,
(7) phosphorus contamination risk, (8) pesticide contamination, (9) soil erosion risk, (10) habitat
intervention, (11) changes in soil carbon stock, and (12) balance of greenhouse gases. Indicators were
geographically referenced using a geographic information system (GIS). The strength of this study is
not in the absolute value of environmental indicators, but rather in the conceptualization of indicator
and the identification of changing patterns, gradients and trends in space and time. According to our
results, we can not definitely say that agriculture in the Pampas, as a whole, tends to be sustainable
or not. While some indicators tend to improve, others keep stable, and the rest worsen. The relative
importance among indicators must also be considered. The indicators that showed a negative net
change are key to the identification of critical problems that will require special attention in the close
future.
Keywords: agro-environmental assessment, Argentine Pampas, regional scale, sustainability
indicators
1. Introduction
Environmental decision- and policy-makers require approximate quantification of
environmental properties such as vulnerability, potential risk, conservation status
or ability of ecosystems to recover after a perturbation (Villa and McLeod, 2002).
Knowledge of such properties is essential, but its precise characterization requires
investment in data collection, research and modeling that is not always possible
in developing countries. As a result, measures and indicators are often calculated
with poor scientific justification, arousing a controversy about their utility for sound
decision making. Thus, a scientific-based guidance on environmental assessment
suitable for decision makers seems to be quite urgent and necessary, especially
Environmental Monitoring and Assessment 87: 169–195, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
170
E. F. VIGLIZZO ET AL.
in agricultural ecosystems that suffer major changes in relatively short periods of
time. This is a demand that needs to be matched in the Pampas of Argentina.
Like any other economic activity involving nature, agriculture affects and is
affected by the environment. Mutual effects cut across different levels and scales
(Allen and Starr, 1982). In the Pampas region of Argentina, there are neither sufficient knowledge on environmental issues, nor clear perception of environmental
alterations across space and time (Viglizzo et al., 1997). We have learnt that land
use and technology were both major drivers of environmental alteration (Viglizzo
et al., 2001). However, some critical questions are still unanswered: How have
land use change and technology incorporation impacted on the rural environment?
What critical environmental trends are detectable? What impacts can we project
to the short- and mid-term? Certainly, the right answers are not simple, but this
information is necessary to address the problem.
Although indicators to measure social and economic changes are abundant in
Argentina, proper indicators for assessing environmental changes are scarce. However, it is increasingly recognized that they are essential to: (a) assess changes in
the rural environment at different geographic scales (farm, ecosystem, region and
country), (b) guide preventive strategies and corrective tactics, and (c) deliver practical recommendations for users that operate at those different scales (producers,
scientists, technicians, consultants, and policy makers that operate in national and
international organizations).
Governments need good information to design policies, and science and technology organizations are appropriate to provide it. At present, the conservation of
environmental goods and services (e.g., water cleaning, air purification, soil erosion
prevention, natural decontamination, mineral cycling, pollination, recreation, etc.)
is not still considered a cost in public accounting. However, an appropriate valuation of affected goods and services that aims at internalizing environmental costs
will be unavoidable in the near future. Pressures to do so increase in developed
countries, and will grow in developing countries as well.
The general objective of this work was to provide decision makers with a scientifically sound set of indicators that aim at assessing changes in the rural environment of the Pampas in Argentina. Specific objectives were: (a) to propose a
standard monitoring system to describe and quantify changes (progress, stability
or regress) in the rural environment; (b) to inform policy makers on changes in
the rural environment, in particular in areas of higher risk, and (c) to facilitate the
development of environmental policies, programs and projects.
2. The Pampas Ecoregion in Brief
The Pampas ecoregion is a vast, flat region of Argentina that comprises more that
50 million ha of arable lands for crop and cattle production (Hall et al., 1992).
Agriculture in the pampas has a short history (a little more than 100 yr) and shares
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
171
common features with the agricultural history of the American Great Plains. Both
ecoregions were mostly native rangelands until the end of the 19th century and
the beginning of the 20th, and lands were then allocated to crop (cereal crops
and oil seeds) and cattle production under dryland conditions. Provided that land
cultivation was accomplished with unsuited tillage systems and machinery, they
both were affected by heavy erosion episodes (dust bowls) especially on the fragile
lands during the first half of the 20th century (Cole et al., 1989; Covas, 1989; Lal,
1994).
The Pampas plain is not homogeneous in soils (Satorre, 2001). Using soil and
rainfall patterns, the region can be divided into five homogeneous agroecological
areas (Figure 1) as follows: (1) Rolling pampas, (2) Central pampas (which can
be subdivided in Subhumid on the East and Semiarid on the West), (3) Southern
pampas, (4) Flooding pampas, and (5) Mesopotamian pampas.
Rainfall and soil organic matter, granular structure, and nutrient contents decline from East to West. According to FAO (1989) criteria, deep and well-drained
soils predominate on the Rolling pampas, which provides conditions for continuous farming. In the Central pampas, despite increased wind erosion towards the
West, most of the lands are suited for cultivation. The Flooding and Mesopotamian
pampas are mostly devoted to cattle farming on native and introduced perennial
pastures. Limitations for crop production in these areas are associated with shallow
soil depth, frequent flooding, soil salinity, poor drainage, and water erosion.
The mixed grain crop-cattle production systems have extended nowadays over
most of the Pampas. Grain crops rotate among them and cattle pastures are integrated in mixed production programs under different land use schemes, all of
which depend on environmental limitations. Cattle production activities, on the
other hand, vary from cow-calf to cattle finishing. Rainfall regimes vary in time
and space, causing occasional droughts and flood episodes that transitorily affect
both crop and cattle production (Viglizzo et al., 1997).
3. Materials and Methods
Different sources of information have been utilised in this study: (1) two national
general censuses (years 1960 and 1988) that comprised the totality of farms scattered in 147 political districts, (2) one national survey for 1996, that comprised a
sample of farms in different areas, (3) a variety of production and yield statistics
regularly published by the Secretary of Agriculture of Argentina, and ((4) energy,
nitrogen (N), and phosphorus (P) concentration of inputs and outputs as determined
by various authors. Data on land use and crop yields were analysed for all districts.
Land use was expressed in terms of the relative area (%) of crops, pastures and
natural grasslands with respect to the total area devoted to farming activities. The
analysis was based only on predominant (crop and beef production) activities such
as wheat (Triticum aestivum L.), maize (Zea mays L.), soybean (Glycine max L.
172
E. F. VIGLIZZO ET AL.
Figure 1. Location of the Pampas ecoregion and different ecological areas in the Argentine territory.
Merr.), and sunflower (Helianthus annuus L.). Because of the lack of long term
data, beef production was estimated from equations for each ecologically homogeneous area (Viglizzo, 1982) that relate stocking rate (available data) to meat
production per hectare.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
173
3.1. S USTAINABILITY INDICATORS
As driving criteria to select indicators, we assumed that they have to be sound,
simple to calculate, easy to understand, and easily applicable by decision-makers.
Efforts were put in avoiding the generation of multiple indicators that would not
provide specific information, or that could become hard to interpret or use. Indicators here referred to significant ecological structures and functions, avoiding those
that had low environmental significance.
We faced the lack of sufficient data, variable quality of data and heterogeneity
of data sources. Emphasis were put on the homogenization of data basis for all
study areas. Thus, twelve basic indicators were defined: (1) land use, (2) fossil
energy (FE) use, (3) fossil energy use efficiency, (4) nitrogen (N) balance, (5) phosphorus (P) balance, (6) nitrogen contamination risk, (7) phosphorus contamination
risk, (8) pesticide contamination, (9) soil erosion risk, (10) habitat intervention,
(11) changes in soil carbon (C) stock, and (12) balance of greenhouse gases (GHG).
The databases were geographically referenced using a geographic information
system (GIS). Dot-density maps were built for most indicators. The procedure
allowed a graphic comparison of attributes on different areas. The detection of
spatial patterns and density gradients for each study environmental parameter was
the main outcome of this procedure.
3.2. L AND USE (I NDICATOR NO . 1)
Land use is an indicator of high relevance in agro-ecology. Because they affect
the functionality of agro-ecosystems, both land use and technology application are
determinants of sustainability in the rural environment (Viglizzo et al., 2001). Land
use refers to the purposes (Van Latesteijn, 1993; Rabbinge et al., 1994) by which
lands are allocated in agriculture. Changes in land use generate spatial patterns and
temporal trends of environmental alteration (International Geosphere-Biosphere
Program/Human Dimensions of Global Environmental Change Program, 1995).
This indicator was basic for calculating the rest of the indicators that were calculated here. In our case, different land use patterns in time and space were estimated
in terms of the proportional allocation (%) of the land to: (a) native rangeland,
(b) introduced perennial pastures, and (c) annual crops.
3.3. F OSSIL ENERGY USE (I NDICATOR NO . 2) AND FOSSIL ENERGY USE
EFFICIENCY (I NDICATOR NO . 3)
The use of fossil energy (FE) highly correlates with the intensification of agriculture. Being usually linked to contamination episodes and to greenhouse gases
emission, the increasing use of FE is frequently associated with environmental
degradation (Agriculture and Agri-Food Canada, 2000).
This indicator was calculated by considering the annual energy costs per hectare
of predominant inputs (fertilizers, seeds, concentrates, pesticides) and practices
174
E. F. VIGLIZZO ET AL.
(tillage, planting, weeding, harvesting, etc.) expressed in joules per hecatare. The
fossil energy cost of inputs and practices were obtained from different sources of
estimations (Reed et al., 1986; Stout, 1991; Conforti and Giampietro, 1997; Pimentel, 1999). Although it was not possible to check in detail the original procedures
to estimate such figures, we assume they provide an acceptable estimation of all
involved fossil energy costs.
The fossil energy use efficiency was calculated by considering the amount of
megajoules (Mj) of FE used to get one Mj of product. Calculations were made on
annual basis taking into account the proportional participation of each analyzed
activity. Under this scheme, the larger the amount of FE used to produce one unit
of energy, the less efficient the production process was. When the input/output ratio
decreases over time, the FE use efficiency increases and the relative environmental
impact decreases in equivalent proportion.
3.4. BALANCE OF NITROGEN (I NDICATOR NO . 4) AND PHOSPHORUS
(I NDICATOR NO . 5)
Among other nutrients, the adequate supply of nitrogen (N) and phosphorus (P) is
essential to the plant growth and development. The consecutive cultivation of land
through many years, alters the original nutrient endowment of arable soils. The
accumulation of effects generates imbalances that worsen over time. If extraction
exceeds supply over the years, the accumulation of negative balances may finally
provoke a severe nutrient depletion, reduce crop yields and lower economic returns.
Conversely, if supply exceeds extraction, the accumulation of positive balances
can overload the soil with nutrients, causing potential for both soil and water contamination. Ideally, a well-balanced input/output ratio of nutrients is essential for
achieving and maintaining soil sustainability.
A simplistic procedure was followed here. The annual average of available
N and P in soil was estimated as the difference between inputs and outputs per
hectare and per year, in each study area. A simplistic assumption was adopted:
(a) if extraction exceeds supply, soil fertility declines, and (b) if supply exceeds
extraction, the excess is cause of potential contamination. The export of N an P by
agricultural products was the only way of loss considered in this work. On the other
hand, the considered ways of nitrogen gain were (a) precipitation, (b) fertilizers,
(c) biologic fixation by legumes, and (d) purchased feed, later excreted as cattle
urine and manure. The predominant phosphorus gains are fertilizers and purchased
feed.
3.5. C ONTAMINATION RISK BY NITROGEN (I NDICATOR NO . 6) AND
PHOSPHORUS (I NDICATOR NO . 7)
The assessment of contamination risks by N and P is key for assessing the sustainability of intensive agriculture. For example, nitrate leaching into ground water can
be risky for human and animal health. On the other hand, the runoff of water with
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
175
nitrates and P can increase the eutrophication risk on ponds and lakes. The rapid
increase of algae and aquatic plants depletes the oxygen in water and alters the biota
composition in surface water. Given that the balance of P was always considered
negative in the Argentine Pampas, it was empirically assumed that contamination
risk was low. But considering that phosphorus fertilization has rapidly increased in
the 1990’s, the risk of contamination has probably increased, particularly in certain
areas.
The N contamination risk was calculated considering only the residual N when
the N balance was positive. Dividing the amount of residual N by the amount of
water available for nitrogen dilution (water excess), the N concentration in water
can be estimated. The excess of water was calculated on annual basis from a water
balance estimation, which takes into account the water gain by rainfall (mm yr−1 )
less the real evapotranspiration in the same period. Besides, the contamination risk
calculation proceeded only in areas where the excess of water exceeded the water
holding capacity of soils. We have utilized default values for water holding capacity
of average soils cited by McDonald, 2000. They were: (a) 100 mm for a sandy or
a sandy loam soil, (b) 150 mm for a loam soil, (c) 200 mm for a loam clay soil,
and (d) 250 mm for a clay soil. Therefore, if the excess of water (rainfall – real
evapotranspiration) was less than the water holding capacity of the soil, saturation
did not happen, leaching was absent and the water contamination risk was low. A
similar procedure was used to calculate the P contamination risk.
3.6. P ESTICIDES CONTAMINATION RISK (I NDICATOR NO . 8)
The pesticides contamination risk was described here by a relative index. This indicator seems to be useful for doing temporal and geographic comparisons. Strictly
speaking, absolute values are not meaningful. Calculation included the most common insecticides, herbicides and fungicides that predominated in different decades
for the main agricultural activities. The actual toxicity values (LD-50) used were
provided by manufacturers, and obtained from a well-known current pesticides
guide (CASAFE, 1997).The pesticide contamination risk arises concern on: (a) water and soil degradation by pesticides residues, (b) air quality degradation by the
volatile fraction of pesticides, and (c) the negative impact on biodiversity.
This indicator was based on the assessment of the relative toxicity of predominant pesticides packages used in different decades for various farming activities in
different areas of the Argentine Pampas. The land use allocation was the base layer
of information on which pesticide packages, and their corresponding toxicity values, were superimposed. Commercially recommended application rates of active
products were used for each area. The relative index for pesticide contamination
was the result of summing-up each pesticide contribution per hectare, after multiplying the proportion (%) of land allocated to each analyzed crop, by the toxicity
of each product. Using a GIS to superimpose various information layers, maps
176
E. F. VIGLIZZO ET AL.
of relative toxicity were produced for different ecological areas in the Pampas, in
different decades.
3.7. S OIL EROSION RISK (I NDICATOR NO . 9)
Soil quality was defined here as the soil capacity to sustain agricultural activities
over time without affecting its productivity and the quality of the environment.
Erosion has negative effects at the field level, and also can impact larger areas
outside the field. Normally, erosion is expressed in terms of lower chemical fertility,
yield falls, reduced soil infiltration and water retention, increased soil compaction
and runoff, shift in soil pH, and ditch formation. Outside the field boundaries,
the principal consequences are the sediment deposition on adjacent fields, sand
dunes formation, sedimentation and embankment of drainage ditches and natural
surface waters (ponds, lakes, creaks and rivers), water quality degradation, and
water contamination.
Unsuitable land use schemes and tillage practices can cause severe erosion in
fragile lands. In general, the high quality soils tolerate some degree of erosion
without affecting their productivity. In those cases, a natural soil forming process
seems to compensate erosion losses, at least in the short-term. However, temporal
trends in different areas need to be assessed and monitored to prevent undesirable
consequences.
In this work, the erosion risk was assessed by means of a relative index. The
method was based on the relative quantification of main responsible factors that
generate water or wind erosion. Four factors were multiplied: (a) land use (crop
and pasture allocation in lands), (b) soil fragility (valued through the organic carbon content), (c) tillage practices (e.g., conventional, minimum tillage, or no-till),
and (d) slope. Although digital elevation models could be very useful for getting
reliable slope data, this factor was considered irrelevant for our analysis because
the Argentine Pampas is very large flat plain with no steep slopes. Accepted criteria
(Agriculture and Agri-Food Canada, 2000) classify water erosion risk into five
categories: tolerable (less than 6 metric tons of sediment loss per hectare and year),
low (6 to 11 tons), moderate (11 to 12 tons), high (22 to 33 tons), and heavy (more
than 33 tons). The method used here allows only for estimation of a relative erosion
risk. A family of equations for getting a quantitative estimation of absolute losses
(expressed in ton ha−1 yr−1 ) due to erosion would have been ideal, but the lack of
such tool did not allow us to get quantitative comparisons.
3.8. H UMAN PRESSURE ON HABITATS (I NDICATOR NO . 10)
Over the centuries, agriculture has transformed natural habitats into human-designed habitats. Despite the fact that agriculture has historically benefited from
biodiversity, habitat intervention by agriculture greatly reduces natural biodiversity.
Biodiversity has provided agriculture (a) germplasm and gene variability, (b) pollinators, (c) beneficial insects and control agents, and (d) waste disposal and recyc-
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
177
ling pools. On the other hand, diversified agriculture could also benefit biodiversity,
through integrated farming schemes that favor the recovery of environments that
were impoverished by monoculture.
Land use change seems to be the most relevant impact factor on biodiversity
(Sala et al., 2000). But tillage practices and pesticides are vehicles of habitat aggression as well. The aggressive agronomic practices can reinforce the negative
impact of intensive land use on habitat and biodiversity.
The estimation of a habitat quality index (HQI) that we have used here, resembles the above described method for calculating the risk erosion indicator. Under the assumption that human action affects habitat and biodiversity, this method
was used to generate a relative HQI to assess the degree of human intervention on
the habitat through (a) land use, (b) tillage practice, and (c) pesticide contamination. Land use was the proportion (%) of land cultivated annually with grain crops.
The tillage practice factor was the same that was used for estimating soil erosion
risk. The same coefficients obtained from the estimation of pesticide contamination
risks were used here as well. The combined HQI factor was the result of the simple
multiplication of those three intervention factors. Thus, the higher the proportion of
annual crops, the aggressiveness of tilling practices, and the toxicity of pesticides,
the greater the detrimental effect of humans on habitats.
3.9. C HANGES IN SOIL CARBON STOCKS (I NDICATOR NO . 11)
Carbon (C) is the main component of organic matter, and therefore a factor that
strongly determines soil quality. Organic matter decays are associated with soil
fertility and soil structure losses, and also with a higher soil erosion risk. Depending
on organic matter gain or loss, soils can respectively act as a sink or a source of
atmospheric C. Thus, C stocks are dynamic and highly sensitive to human action.
After about 20 yr under cropping, soils can lose up to 35% of their original
organic matter endowment. Most soils seem to reach equilibrium at that point,
and further changes depend on land use, tillage and other agronomic practices.
Well-designed agronomic strategies can convert losses into gains, and a sizable incorporation of C is possible. However, it is unlikely we can achieve a full recovery
up to levels of pristine stages.
The procedures followed in this work were based on the IPCC (1996) reviewed
guideline methodology. An initial C stock that varied according to the study area
was estimated. Calculation of this stock was achieved assuming that the soil C
content for 1960 was 65% of the initial content of native soils. As the soil C content
in 1960 varied according to the study ecological area, so did the estimation of the
original value.
Past information on land use was the basic factor utilized for calculating changes
in the C stock. In the long term, factors associated with crops and pasture vary:
while perennial pastures improved the C stocks, annual crops depleted such stock.
Tillage practices were associated with coefficients that represented different C ox-
178
E. F. VIGLIZZO ET AL.
idation rates. While minimum- and no-till practices favored soil C sequestration,
the conventional tillage was cause of C emission. Furthermore, the IPCC method
provided default coefficients values for organic matter enrichment of soils via crop
residue accumulation.
3.10. G REENHOUSE GASES (GHG) BALANCE (I NDICATOR NO . 12)
Atmospheric gases, including carbon dioxide (CO2 ), methane (CH4 ), nitrous oxide
(N2 O), ozone (O3 ) and water vapor, are normal components of the atmosphere that
have, at the same time, a greenhouse effect. Such gasses are transparent to short
wave radiation that reaches the earth, but they are opaque to earth emission of
long wave radiation. The consequence is that emissions bounce, and the trapped
radiation increases the global temperature of the Earth. As a natural phenomenon,
the global warming caused by the greenhouse gases is known as natural g reenhouse effect. The greenhouse gases concentration has remained rather constant
over 10 000 yr. But anthropogenic emissions increased dramatically during the last
200 yr, raising the global temperature. Fossil fuels consumption for industry and
commerce, transportation, household comfort, deforestation and soil cultivation are
considered major causes of increased GHG emissions, triggering a human-driven
global warming that had no antecedents in the Earth history.
Carbon dioxide and CH4 account for 90% of the human-driven greenhouse
effect. Atmospheric concentration of these gases has respectively increased 30, 15
and 145% for CO2 , N2 O and CH4 (Desjardins and Riznek, 2000). Grain crops and
cattle are both significant sources of greenhouse gases that need to be estimated for
the Pampas plain.
The GHG balance for rural areas was estimated following the standard guidelines
of IPCC (1996). All gases were converted into CO2 equivalent (ton ha−1 yr−1 ). The
calculations included emission and sequestration due to land use change, grain
cropping and cattle production activities.
The use of fossil fuels was another major source of CO2 . The method included
fuels used in rural activities and fuels used for manufacturing fertilizers, herbicides
and machinery. Because of the heterogeneity of cases and procedures, emissions
due to rural transportation were not included in our calculations.
Ruminants were a significant source of GHG. Ruminants emit methane from
enteric fermentation and fecal losses. Methane has a greenhouse power that is 21
times greater than CO2 . This figure was used to convert CH4 into CO2 equivalents.
Nitrogen excreted in feces and distributed with fertilizers was another significant
source of nitrous oxide (N2 O) emission. N2 O has a greenhouse power 310 times
greater than CO2 . Losses of N2 O occur via volatilization, leaching and runoff.
Arable soils were also a direct source of greenhouse gases through fertilizers,
biological N fixation and crop residues.
When data from direct field measurements were unavailable, default values
suggested by the IPCC (1996) were used for estimating gains and losses of carbon.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
179
The methodology proposed by IPCC (1996) estimated emission or sequestration of
CO2 through the following components: (1) CO2 stock exchange in soils over time
(CO2 -SC), (2) CO2 stock exchange in timber biomass (CO2 -BL), (3) conversion of
forests and prairies into arable land (CO2 -CTBP), (4) abandonment of intervened
lands (CO2 -aband), and (5) emission of CO2 from fossil fuels burning (CO2 -CF)
in different agricultural activities. Items 2, 3 and 4 were not relevant in our study,
and were not included in our estimations.
The procedure estimates CH4 emission from 3 sources: (1) enteric fermentation
(CO2 -FE) from domestic animals, (2) fecal emissions (CO2 -EF), and (3) rice crop
emissions (CO2 -EA). Our study took into account the first 2. Because rice is not a
predominant crop in the Pampas, it was not included in the study. Emissions from
ruminants were calculated from beef cattle only.
The emission of N2 O was the most difficult to estimate because of the complexity of determinations. Emission sources are: (1) Feces and urine (CO2 -EDHO)
from domestic animals, (2) volatilization, runoff and infiltration (CO2 -EIVLI) from
synthetic fertilizers and animal excrements (urine and feces), and (3) arable soils
(CO2 -EDSA), through chemical fertilizers, biological N fixation and crop residues.
Therefore, the final equation for estimating the CO2 balance was:
CO2 balance = (CO2 -SC + (CO2 -BL + CO2 -CTBP + CO2-aband) + CO2 -CF) +
((CO2 -FE + CO2 -EF) × 21) + ((CO2 -EDHO + CO2-EIVLI + CO2 -EDSA) × 310)
4. Results and Discussion
The strength of this work lies in the identification of geographic patterns and gradients, and the interpretation of temporal trends, more than on the absolute value of
the estimated indicators. We aimed at determining if environmental conditions are
improving, worsening or keeping stable over time. We consider that comparisons
among areas and decades are not invalidated by some methodological constraints
mentioned above.
4.1. L AND USE AND LAND USE CHANGE : TOWARDS A MORE INTENSICE
MODEL
A previous analysis based on national censuses from 1881 to 1988 (Viglizzo et al.,
2001), has demonstrated that the five homogenous study areas differ both in land
use and land use change patterns. In comparison to the rest of the agrecological
areas, grain crops had their faster expansion in the Rolling pampas, replacing
the pristine natural lands. A cropland profile supported by a high environmental
potential for crop production, has characterized the Rolling pampas during the last
century.
180
E. F. VIGLIZZO ET AL.
The Flooding and the Mesopotamian pampas, on the other side, have the lower
capacity for crop production. Due to soil quality and topography limitations, the
proportion of land devoted to cereal crops and oil seeds has remained low and
constant across time. A net predomination of introduced and native pastures define
a net orientation to cattle production in both areas.
In terms of agricultural productivity, the Central and the Southern pampas are
both in the middle between the natural conditions of the areas mentioned above.
They are characterized by predominant mixed, grain crop-cattle production systems. Soil quality and climatic limitations have prevented a larger conversion of
grazing lands to croplands. During the 1990’s, however, technologies that aimed at
increasing productivity and reducing environmental impacts (e.g., no-till planting,
fertilization to minimize nutrient depletion) have favored the expansion of grain
crops.
Maps in Figures 2 and 3 show, in a time-space dimension, land-use changes
during the period 1960–2000. Figure 2 illustrates the expansion of wheat, maize,
sunflower, soybeans, and winter and summer small-grain pastures. Color intensity
shows a consistent increase of croplands in the Rolling pampas and in the mixed
systems of Central and Southern pampas. Croplands expansion, however, did not
happen by forcing livestock into marginal lands, as it was frequently assumed.
Furthermore, legume-based perennial pastures have also increased, primarily on
the Central, Southern and Flooding pampas (Figure 3), as an indication of a simultaneous intensification of crop and cattle production in these areas. The planting
of improved annual and perennial pasture species, the increased use of concentrate
feeds and the management of a higher animal stock density in smaller areas, have
characterized the predominant livestock production system in the Pampas.
Land use patterns and land use changes were key factors for explaining the
environmental behavior of the Pampas. All indicators, from fossil energy consumption to contamination risk, from erosion risk to GHG emission, were particularly
sensitive to land use. Technology was the second important factor. Results clearly
suggest that the 1990’s decade was the starting point in land transformation, mainly
in the Rolling pampas, but also in the Central and the Southern pampas.
4.2. A N INCREASING ENERGY STATUS
Some limitations in the calculation of fossil energy (FE) consumption and FE use
efficiency need to be stated. Firstly, the FE cost of inputs and activities was not
always well documented in literature, and accuracy of data can be argued in some
cases. Secondly, some inputs or activities were underestimated due to their small
participation in the general calculation, and also due to insufficient data. Thirdly,
data reported in the literature were not always updated in response to technology
change. And fourthly, the imprecise description of calculations reported in the
literature raises the risk of double accounting in some cases.
Figure 2. Changes in the area allocated to annual crops (wheat, maize, soybean, sunflower, winter and summer annual pastures) in the Argentine pampas
during the period 1880–2000.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
181
Figure 3. Changes in the aland allocated to cultivated perennial pastures during the period 1880–2000 in the Argentine pampas.
182
E. F. VIGLIZZO ET AL.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
183
TABLE I
Fossil energy (FE) efficiency use in the Argentine pampas during the period
1960–2000. Comparison among ecologically homogeneous areas
Pampas
Efficiency of FE use
(Gj of FE consumed per Gj of energy produced)
1960
1988
1996
Regional average
0.22
0.16
0.17
Rolling
Central subhumid
Central semiarid
Southern
Flooding
Mesopotamian
0.17
0.35
0.45
0.18
0.13
0.18
0.11
0.18
0.44
0.14
0.12
0.17
0.16
0.17
0.41
0.15
0.15
0.23
Figure 4 shows changes in energy productivity (Gj ha−1 yr−1 ) and FE consumption (Gj ha−1 yr−1 ) in the five study areas between 1960 and 1996. The results
indicate that the energy status and the energy fluxes have increased all over the
region, and primarily in a belt defined by the Rolling, the Central sub-humid and
the Southern pampas. The increased proportion of grain crops mostly explains this
increased energy budget.
Table I provides figures on FE use efficiency over the three analyzed historical periods. Although they have higher FE consumption, the areas where grain
crops have largely expanded show energy yields that greatly exceed those where
cattle production predominates. Consequently, croplands show a higher FE efficiency use. Furthermore, efficiency tended to increase in parallel with the increasing energy yield of annual grain crops. The larger adoption of minimum and
no-tillage practices that demand less FE use, was probably highly associated with
that increased efficiency.
4.3. N UTRIENT BALANCE AND NUTRIENT CONTAMINATION : BETWEEN THE
RISK AND THE CHALLENGE
The complexity of the N and P cycles in the Pampas was clearly demonstrated by
our own research (Bernardos et al., 2001) supported by the EPIC model (Williams
et al., 1989). It was quite evident that the elementary input-output approach that we
have used here does not properly represent real cycles simply because the complex
dynamics between nutrients pools and fluxes were not properly captured. Ideally,
rates of leaching, run-off, sediment and volatilization should have been considered
in our estimations. However, time and cost limitations to gather such information in
Figure 4. Changes in energy (Mj ha−1 yr−1 ) harvested through predominant agricultural products (wheat, corn, soybean, sunflower and beef) within the
period 1969–2000 in the Argentine pampas.
184
E. F. VIGLIZZO ET AL.
185
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
TABLE II
Estimation of nitrogen (N) and phosphorus (P) balances in ecologically homogeneous areas of the
Argentine pampas during the period 1960–2000
Pampas
N balance (kg ha−1 yr−1 )
P balance (kg ha−1 yr−1 )
1960
1960
1988
1996
40
Regional average
0.23
1.76
70
1988
100
9.0 16.3 26.8
Rolling
2.40 –6.03 –0.5 14.6 29.7
Central subhumid –1.04 –5.29 11.8 18.3 24.8
Central semiarid
2.43 3.20 12.5 16.6 20.8
Southern
–2.65 2.83 6.8 12.8 18.7
Flooding
1.01 8.35 11.4 5.7 19.9
Mesopotamian
1.16 6.83 6.5 10.7 14.8
1996
40
70
100
–2.03 – 5.07 – 5.9 –4.2 –2.3
–3.66
–2.11
–0.91
–1.86
–1.25
–1.03
–12.71
– 5.66
– 1.56
– 3.84
– 1.48
– 1.22
–12
– 7.1
– 5.1
– 4.6
– 1.8
– 3.2
–9.2
–5.6
–4.4
–3.1
–0.3
–1.7
–6.2
–4.1
–3.9
–1.6
–1.2
–0.2
In 1996, the estimations include 3 fertilization hypothesis: 40, 70 and 100% of the area was
fertilized following commercial recommended dose.
each analyzed geo-political district, were not overcome. In this context, the simple
input/output approach was adopted as the best option available, assuming that
negative balances indicated nutrient loss, and the positive ones indicated potential
contamination. Beyond such limitations, the method seems to be useful for making
geographic and temporal comparisons.
Likewise, the lack of accurate information on evapotranspiration and water soil
retention capacity at the district level was an additional limitation. The method had
low resolution for detecting contamination risk in areas of high animal concentration (beef and dairy feedlots, poultry and pig farms), or in irrigated areas for
intensive fruit and vegetable production. Neither was the method sensitive for detecting contamination risk under unexpected meteorological events, such as heavy
rains that result in flooding or excessive leaching.
Table II shows a comparison of changes in N and P balances. A significant
increase of N and P extraction from 1960 to 1996, consistent with cropland expansion, was detected especially in the Rolling pampas. But such extraction also
occurred in less productive areas like the Central Sub-humid pampas. Because of
the lack of precise data on fertilizers use during the 1990 decade, three hypothetical fertilization schemes based on commercial recommendations were considered:
100, 70 and 40% of the area devoted to grain crops, respectively. This allowed a
simplistic interpretation about varying degrees of fertilizer adoption in different
areas.
The risk of contamination was closely related to positive balances of nutrients
and water. In the case of P, the whole region would have underwent a net negative
P balance, even in the 1990’s (Table III). In the 1990’s, positive balances were only
186
E. F. VIGLIZZO ET AL.
TABLE III
Estimation of the relative soil erosion risk in the Argentine pampas and its ecologically homogeneous areas during the period
1960–2000
Pampas
Regional average
Rolling
Central subhumid
Central semiarid
Southern
Flooding
Mesopotamian
Coefficient of relative erosion risk
1960
1988
1996
0.11
0.08
0.16
0.21
0.13
0.06
0.11
0.09
0.08
0.11
0.20
0.11
0.03
0.09
0.09
0.08
0.12
0.26
0.10
0.05
0.10
detected where extraction is very low (grazing lands), and when P fertilization
becomes a massive practice (in 100% of lands).
The systematic fertilization of crops and prairies has consolidated in the Pampas
during the 1990’s. Considering the three hypothetical fertilization levels previously
discussed, if fertilization were reduced to 70% or to 40% of the total area, positive nitrogen balances tended to disappear, and negative phosphorus balances were
inevitable. Despite the three levels of fertilization discussed, we assumed that the
40% level has probably represented the more realistic scenario for the 1990’s.
4.4. P ESTICIDE CONTAMINATION RISK : UPS , DOWNS AND UNCERTAINTIES
The calculation method applied was based on several assumptions. Firstly, although the utilized pesticides varied largely, pesticide packages were uniform in all
areas for each particular crop. We have to recognize, however, that alternative packages for similar activities have been utilized, especially during the 1990’s when
the offer of commercial pesticides has multiplied. Secondly, our model assumed
that all crops have systematically received the commercially recommended doses.
And thirdly, for each study period, the method has calculated a ‘potential toxicity’
that was the toxicity summation of all applied pesticides. It was also assumed that
the larger the combined potential toxicity of applied pesticides, the larger the impact on biodiversity. Field measurements were ideally required, but no direct field
measurements of pesticides impact on biodiversity was possible for this study.
Figure 5 shows the estimates of relative indices of contamination risk due to
pesticides application between 1960 and 2000. The expansion of croplands and the
increasing use of pesticides over the decades would explain the significant increase
of pesticide contamination risk, especially during the 1990’s. The effect seems to
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
187
be particularly dramatic in the most productive areas, for example in the Rolling
pampas, but also in areas where cropland expanded. However, field studies are
required to confirm this estimation. We suppose that this warning trend may reverse
in the near future because a family of new, soft pesticides are being introduced at
present in the commercial market. Some common and highly toxic pesticides used
during part of the 90’s were banned nowadays.
4.5. S OIL EROSION AND HABITAT THREATS : CONTRASTING TRENDS
The relative coefficients used to evaluate the soil erosion risk were calculated taking into account the predominant soil quality, land use and tillage practices in each
period.
Table III depicts the estimations of relative erosion risk for the different Pampas areas in the period 1960–2000. The higher risk tended to concentrate on the
Western lands of the Central pampas, where the conversion of grazing lands into
croplands has exposed their fragile soils to increasing wind and water erosion.
With the exception of the semiarid Central pampas where the risk increased,
the estimates show that the erosion risk kept rather stable in the rest of the areas
during the 1990 decade. Despite the expansion of the cropping area, this stabilized
performance was probably the result of the increased adoption of soil conservation
tillage, which would have compensated the more intensive use of land.
Similarly, the relative coefficients used to assess the degree of habitat intervention by humans were estimated taking into account three driving factors: land use,
tillage operations and pesticide contamination risk. A combined factor arose from
multiplying the three factors. Although no long-term data is available on the effects
of agriculture on wildlife in the Pampas, we assumed that a relative intervention
index would provide an indirect estimation of agricultureþs impact on biodiversity
across space and time. Nevertheless, this indicator says nothing about the actual
impact of man on biodiversity. Ideally, this method should include the assessment
of habitat fragmentation on key species.
Figure 6 shows the relative index of habitat intervention across ecological areas
and over time. Estimations demonstrated that the greatest intervention would have
taken place over the most productive lands of the Rolling pampas. Aggression
would have been particularly strong during the 1990’s. The lowest degree of intervention was detected both in the Flooding and the Mesopotamian pampas. These
areas still conserve large proportions of land devoted to cattle production on native
and improved pastures. Tillage operations and pesticides application are minimum.
This agrees with results of a previous work of Viglizzo et al. (2001).
4.6. C ARBON STOCKS AND GHG EMISSIONS : IMPROVING TRENDS
The method applied was rather simplistic if we accept that the soil C dynamics
is very complex. Because our assumptions were strong, an unavoidable degree
of uncertainty about the accuracy of estimations had to be accepted. Land use,
Figure 5. Estimation of the relative risk of pesticides contamination in different areas of the Argentine pampas within the period 1960–2000.
188
E. F. VIGLIZZO ET AL.
Figure 6. Estimation of the relative risk on habitats and biodiversity due to human intervention in different areas of the Argentine pampas within the period
1960–2000.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
189
190
E. F. VIGLIZZO ET AL.
climate, soil and tillage interactions, added to agronomic practices, were major
sources of variation in soil C stock responses. Furthermore, the lack of data to
estimate the original soil C content in some areas has introduced an additional constraint. Likewise, soil sampling errors and the quality of laboratory analysis in the
past are other sources of uncertainty that could not be overcome. Therefore, results
comprise average estimates that were considered the best information available.
According to the results, the whole Pampas plain has experienced a significant
loss of soil C between 1960 and 2000 (Figure 7). Over the 40 yr study period, the
largest loss of C has taken place in areas where grain cropping is the main activity,
primarily the Rolling pampas. It should be noted, however, that despite the increase
of the cultivated area in time, the C loss rate has declined in time. Rates of C loss
were significantly lower in the 1990’s than in the 1960’s. The quick expansion of
minimum-and no-tillage may be the main factor explaining this behavior.
On the other hand, the procedures suggested by the IPCC (1996) to estimate
GHG emissions are relatively new. Then, some uncertainty about results has arisen
when models and suggested default values were applied. Ideally, experimental
and field data would be necessary to adjust procedures and coefficients to each
environment under analysis.
Our results suggest that the whole Pampas has behaved as a net emitter of GHG
between 1960 and 2000 (Figure 8). The larger net emissions were found on the
Rollling pampas and the Central pampas during two (1960 and 1988) of the three
analyzed periods. A generalized decline in emissions for the whole region was
estimated for 1996. However, differences among the five study areas have persisted.
We have interpreted that minimum and no-till practices were the main responsible for these positive changes. The reduced fossil fuel consumption due to minimum or no-tillage operations has reduced CO2 emissions and favors C retention
in soil organic matter over time.
5. Conclusions
Probably, this work represents a pioneer effort for getting, at a regional scale, an
integrated view of the environmental sustainability of agriculture in the Argentine
Pampas. It can also be viewed as a first contribution to get a permanent system of
environmental monitoring for rural areas in the country.
Certainly, both strengths and weaknesses should be taken into account. Weaknesses were in turn pointed out when each indicator was considered in particular.
One obvious conclusion is that our calculation methods and techniques need to
be perfected. An extra effort is needed to capture not only geographical patterns
and gradients, and time tendencies, but also to get a realistic quantification of such
changes. The strengths of this study are not in the absolute value of environmental
indicators, but in the identification of patterns, gradients and trends in space and
time. This allowed us to determine if critical environmental parameters are keeping
Figure 7. Estimation of carbon losses (t C ha−1 yr−1 ) in different areas of the Argentine pampas during the period 1969–2000.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
191
Figure 8. Estimation of greenhouse gases emission (t C ha−1 yr−1 ) in different areas of the Argentine pampas during the period 1960–2000.
192
E. F. VIGLIZZO ET AL.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
193
stable, improving or worsening across time. We must also consider the relative
importance of factors. Since they depend on the use of an index, they should have
different weights.
Beyond the results, two questions have arisen: is the Pampas agriculture sustainable? Are their production trajectories sustainable in time? Indicators differed
in their behavior and trends. Nowadays we can not say that agriculture, in the whole
Pampas, is sustainable or not sustainable. While some indicators tend to improve,
others keep stable, and the rest worsen. Furthermore, the projections of trends are
no homogeneous in all areas. The transition towards a more intensive model of
agricultural production, both in terms of land use and technology application, has
clearly characterized the 1990’s. The production and environmental trajectories
indicate that many farming systems in the Pampas are resembling some intensive
models that are common in the Northern Hemisphere. In response, we need a new
productive and environmental view to replace the traditional one.
The trajectory of indicators differed in different ecological areas. Positive
changes were detected in most areas: (a) a higher efficiency in fossil energy use,
(b) a lower risk of soil erosion, (c) a decreasing risk of soil carbon loss, and (d) a
consistent decline in greenhouse gases emissions. No-till and conservation tillage
have apparently driven those beneficial changes.
Some negative changes that deserve especial attention were observed on the
other hand. Some issues that should be taken into account are the following: (a) the
regional N balances tended to become positive in many cropping areas, delivering
residual N that can potentially contaminate soils and water. The contamination risk
tended to increase as the N fertilization increased. However, given that fertilization
levels were neither high nor homogeneous in all areas, we believe that the risk of
contamination is still low in the Pampas as a whole, (b) provided that P balances
were negative in most of the studied areas, the risk of P contamination still appears
to be rather far in the future, and (c) a more careful assessment of the increasing
risk of pesticides contamination during the 1990’s needs to be considered.
The indicators that showed a negative trend are key to identify critical problems that will require special attention in the close future. Putting aside the geographic heterogeneity, an environment degradation belt is emerging. Starting from
the Rolling pampas, the belt is extending over many districts in the subhumid Central pampas. Thus, a potentially trouble area that would require close attention was
detected. An environmental policy, and a well-defined set of research priorities, are
necessary to face future challenges on this environmentally critical belt.
Finally, the long-term monitoring system for the rural environment through the
support of reliable indicators, appears to be a key tool to guide environmental
policies, to drive research priorities, and to orient communication strategies. A
periodic updating of information is crucial to detect critical changes in space and
time, and to fit environmental strategies and tactics that match vital needs of the
region.
194
E. F. VIGLIZZO ET AL.
Acknowledgements
We thank the Secretary of Science and Technology of Argentina for the financial
support to this research. We also acknowledge the contribution of various public
institutions for providing key data for this study, and many colleagues for their
valuable comments and recommendations.
References
Agriculture and Agri-Food Canada: 2000, ‘Environmental Sustainability of Canadian Agriculture’,
in T. J. McRae, C. A. S. Smith and L. J. Gregorich (eds), Report of the Agri-Environmental
Indicador Project.
Allen, T. F. H. and Starr, T. B.: 1982, Hierarchy Perspectives of Ecological Complexity, University
of Chicago Press, Chicago.
Bernardos, J. N., Viglizzo, E. F., Jouvet, V., Lértora, F. A., Pordomingo, A. J. and Cid, F. D.:
2001, ‘The use of EPIC model to study the agroecological change during 93 years of farming
transformation in the Argentine pampas’, Agricult. Syst. 69, 215–234.
CASAFE: 1997, ‘Guía de Productos Fitosanitarios para la República Argentina’, Cámara de Sanidad
Agropecuaria y Fertilizantes, Buenos Aires.
Cole, C. V., Stewart, J. W. B., Ojima, D. S., Parton, W. J. and Schimel, D. S.: 1989, ‘Modelling
Land Use Effect of Soil Organic Matter Dynamics in the North American Great Plains’, in M.
Clarholm and L. Bergstrom (eds), Ecology of Arable Lands, Kluwer Academic Publishers, NY.
Conforti, P. and Giampietro, M.: 1997, ‘Fossil energy use in agriculture: An international comparison’, Agricult. Ecosyst. Environ. 65, 231–243.
Covas, G.: 1989, ‘Evolución del manejo de suelos en la región pampeana semiárida’, Actas de las
Primeras Jornadas de Suelos en Zonas Aridas y Semiáridas, Santa Rosa (LP), Argentina, pp. 1–
11.
Desjardins, R. L. and Riznek, R.: 2000, ‘Agricultural Greenhouse Gas Budget’, in T. J. McRae,
C. A. S. Smith, and L. J. Gregorich (eds), Environmental Sustainability of Canadian Agriculture:
Report of the Agri-Environmental Indicador Project, Agriculture and Agri-Food Canada, Ottawa,
Ontario, pp. 133–140.
FAO: 1989, Guidelines for Land Use Planning, Food and Agriculture Organization of the United
Nations (FAO), Rome.
Hall, A. J., Rebella. C. M., Ghersa, C. M. and Culot, J. Ph.: 1992, ‘Field-crop Systems of the Pampas’,
in C. J. Pearson (ed.), Field Crop Ecosystems, Serie: Ecosystems of the World, Elsevier Science
Publishers B.V., Amsterdam.
IGBP/HDP: 1995, ‘Land-Use and Land-Cover Change: Science/Research Plan’, Report No. 35 of
The International Geosphere-Biosphere Programme (IGBP), and Report No. 7 of The Human
Dimensions of Global Environmental Change Programme (HDP), Stockholm.
IPCC: 1996, IPCC Guidelines for National Greenhouse Gas Inventories: Reference Manual,
Intergovernmental Panel on Climate Change, IPCC Technical Support Unit, Bracknell, U.K.
Lal, R.: 1994, ‘Sustainable Land Use Systems and Soil Resilience’, in D. J. Greenland and I. Szabolcs
(eds), Soil Resilience and Sustainable Land Use, C.A.B. International, Wallingford, U.K.
McDonald, K. B.: 2000, ‘Risk of Water Contamination by Nitrogen’. in T. J. McRae, C. A. S.
Smith and L. J. Gregorich (eds), Environmental Sustainability of Canadian Agriculture: Report of
the Agri-Environmental Indicador Project, Agriculture and Agri-Food Canada, Ottawa, Ontario,
pp. 161–170.
ENVIRONMENTAL ASSESSMENT OF AGRICULTURE IN ARGENTINA
195
Pimentel, D.: 1999, ‘Environmental and Economic Benefits of Sustainable Agriculture’, in J. Kohn,
J. Gowdy, F. Hinterberger and M. A. Northampton (eds), Sustainability in Question: The Search
for a Conceptual Framework, New York.
Rabbinge, R., Van Diepen, C. A., Dijsselbloem, J., De Konig, G. H. J., Van Latesteijn, H. C., Woltjer,
E. and Van Zijl, J.: 1994, ‘Ground for Choices: A Scenario Study on Perspectives for Rural Areas
in the European Community’, in L. O. Fresco, L. Stroosnijder, J. Bouma and H. van Keulen (eds),
The Future of Land: Mobilising and Integrating Knowledge for Land Use Options, John Wiley
& Sons Ltd., New York, pp. 95–121.
Reed, W., Shu, G. and Hills, F. J.: 1986, ‘Energy input and output analysis of four field crops in
California’, J. Agron. Crop Sci. 157, 99–104.
Sala, O. E., Stuart Chapin, F., Armesto, J. J., Berlow, E. et al.: 2000, ‘Global biodiversity scenarios
for the year 2100’, Science 287, 1770–1774.
Satorre, E.: 2000, ‘Production Systems in the Argentine Pampas and their Ecological Impact’, in
O. T. Solbrig, F. di Castri and R. Paarlberg (eds), The Impact of Global Change and Information
on the Rural Environment, Harvard University Press, Cambridge, MA.
Stout, B. A.: 1991, Handbook of Energy for World Agriculture, Elsevier, New York.
Van Latesteijn, H. C.: 1993, ‘A Methodological Framework to Explore Long Term Options for Land
Use’, in F. W. T. Penning de Vries et al. (eds), Systems Approaches for Agricultural Development,
Kluwer Academic Publishers, pp. 445–455.
Viglizzo, E. F.: 1982, ‘Los Potenciales de Producción de Carne en la Región Pampeana Semiárida’,
in E. F. Viglizzo (ed.), Primeras Jornadas Técnicas sobre Producción Animal en la Región
Pampeana Semiárida, Universidad Nacional de La Pampa, Santa Rosa, La Pampa, Argentina.
Viglizzo, E. F., Lértora, F. A., Pordomingo, A. J., Bernardos, J., Roberto, Z. E. and Del Valle, H.:
2001, ‘Ecological lessons and applications from one century of low external-input farming in the
pampas of Argentina’, Agricult. Ecosyst. Environ. 81, 65–81.
Viglizzo, E. F., Roberto, Z. E., Lértora, F., López Gay, E. and Bernardos, J.: 1997, ‘Climate and
land-use change in field-crop ecosystems of Argentina’, Agricult. Ecosyst. Environ. 66, 61–70.
Villa, F. and McLeod, H.: 2002, ‘Environmental vulnerability indicators for environmental planning
and decision-making: Guidelines and applications’, Environ. Manage. 29, 335–348.
Williams, J. R., Jones, C. A., Kiniry, J. R. and Spanel, D. A.: 1989, ‘The EPIC crop growth model’,
Transact. ASAE 32, 497–511.