ENVIRONMENTALASSESSMENTOF AGRICULTUREAT A REGIONAL SCALE ... et al 2002.pdf · environmentalassessmentof agricultureat a regional scale in the pampas of argentina e. f. viglizzo∗,

ENVIRONMENTAL ASSESSMENT OF AGRICULTURE AT AREGIONAL SCALE IN THE PAMPAS OF ARGENTINA

E. F. VIGLIZZO∗, A. J. PORDOMINGO, M. G. CASTRO and F. A. LERTORAINTA/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 Pam-pas there are neither sufficient knowledge on environmental issues, nor clear perception of environ-mental alterations across space and time. The general objective of this work was to provide decisionmakers with a scientifically sound set of indicators aiming at the assessment of current status andfuture 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 applic-able 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 useefficiency, (4) nitrogen (N) balance, (5) phosphorus (P) balance, (6) nitrogen contamination risk,(7) phosphorus contamination risk, (8) pesticide contamination, (9) soil erosion risk, (10) habitatintervention, (11) changes in soil carbon stock, and (12) balance of greenhouse gases. Indicators weregeographically referenced using a geographic information system (GIS). The strength of this study isnot in the absolute value of environmental indicators, but rather in the conceptualization of indicatorand the identification of changing patterns, gradients and trends in space and time. According to ourresults, we can not definitely say that agriculture in the Pampas, as a whole, tends to be sustainableor not. While some indicators tend to improve, others keep stable, and the rest worsen. The relativeimportance among indicators must also be considered. The indicators that showed a negative netchange are key to the identification of critical problems that will require special attention in the closefuture.

Keywords: agro-environmental assessment, Argentine Pampas, regional scale, sustainabilityindicators

1. Introduction

Environmental decision- and policy-makers require approximate quantification ofenvironmental properties such as vulnerability, potential risk, conservation statusor ability of ecosystems to recover after a perturbation (Villa and McLeod, 2002).Knowledge of such properties is essential, but its precise characterization requiresinvestment in data collection, research and modeling that is not always possiblein developing countries. As a result, measures and indicators are often calculatedwith poor scientific justification, arousing a controversy about their utility for sounddecision making. Thus, a scientific-based guidance on environmental assessmentsuitable 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 oftime. This is a demand that needs to be matched in the Pampas of Argentina.

Like any other economic activity involving nature, agriculture affects and isaffected by the environment. Mutual effects cut across different levels and scales(Allen and Starr, 1982). In the Pampas region of Argentina, there are neither suf-ficient knowledge on environmental issues, nor clear perception of environmentalalterations across space and time (Viglizzo et al., 1997). We have learnt that landuse and technology were both major drivers of environmental alteration (Viglizzoet al., 2001). However, some critical questions are still unanswered: How haveland use change and technology incorporation impacted on the rural environment?What critical environmental trends are detectable? What impacts can we projectto the short- and mid-term? Certainly, the right answers are not simple, but thisinformation is necessary to address the problem.

Although indicators to measure social and economic changes are abundant inArgentina, proper indicators for assessing environmental changes are scarce. How-ever, it is increasingly recognized that they are essential to: (a) assess changes inthe rural environment at different geographic scales (farm, ecosystem, region andcountry), (b) guide preventive strategies and corrective tactics, and (c) deliver prac-tical recommendations for users that operate at those different scales (producers,scientists, technicians, consultants, and policy makers that operate in national andinternational organizations).

Governments need good information to design policies, and science and tech-nology organizations are appropriate to provide it. At present, the conservation ofenvironmental goods and services (e.g., water cleaning, air purification, soil erosionprevention, natural decontamination, mineral cycling, pollination, recreation, etc.)is not still considered a cost in public accounting. However, an appropriate valu-ation of affected goods and services that aims at internalizing environmental costswill be unavoidable in the near future. Pressures to do so increase in developedcountries, and will grow in developing countries as well.

The general objective of this work was to provide decision makers with a sci-entifically sound set of indicators that aim at assessing changes in the rural en-vironment of the Pampas in Argentina. Specific objectives were: (a) to propose astandard monitoring system to describe and quantify changes (progress, stabilityor regress) in the rural environment; (b) to inform policy makers on changes inthe rural environment, in particular in areas of higher risk, and (c) to facilitate thedevelopment 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 that50 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. Bothecoregions were mostly native rangelands until the end of the 19th century andthe beginning of the 20th, and lands were then allocated to crop (cereal cropsand oil seeds) and cattle production under dryland conditions. Provided that landcultivation was accomplished with unsuited tillage systems and machinery, theyboth were affected by heavy erosion episodes (dust bowls) especially on the fragilelands 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 andrainfall patterns, the region can be divided into five homogeneous agroecologicalareas (Figure 1) as follows: (1) Rolling pampas, (2) Central pampas (which canbe subdivided in Subhumid on the East and Semiarid on the West), (3) Southernpampas, (4) Flooding pampas, and (5) Mesopotamian pampas.

Rainfall and soil organic matter, granular structure, and nutrient contents de-cline from East to West. According to FAO (1989) criteria, deep and well-drainedsoils predominate on the Rolling pampas, which provides conditions for continu-ous farming. In the Central pampas, despite increased wind erosion towards theWest, most of the lands are suited for cultivation. The Flooding and Mesopotamianpampas are mostly devoted to cattle farming on native and introduced perennialpastures. Limitations for crop production in these areas are associated with shallowsoil depth, frequent flooding, soil salinity, poor drainage, and water erosion.

The mixed grain crop-cattle production systems have extended nowadays overmost of the Pampas. Grain crops rotate among them and cattle pastures are in-tegrated in mixed production programs under different land use schemes, all ofwhich depend on environmental limitations. Cattle production activities, on theother hand, vary from cow-calf to cattle finishing. Rainfall regimes vary in timeand space, causing occasional droughts and flood episodes that transitorily affectboth crop and cattle production (Viglizzo et al., 1997).

3. Materials and Methods

Different sources of information have been utilised in this study: (1) two nationalgeneral censuses (years 1960 and 1988) that comprised the totality of farms scatter-ed in 147 political districts, (2) one national survey for 1996, that comprised asample of farms in different areas, (3) a variety of production and yield statisticsregularly published by the Secretary of Agriculture of Argentina, and ((4) energy,nitrogen (N), and phosphorus (P) concentration of inputs and outputs as determinedby 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 andnatural grasslands with respect to the total area devoted to farming activities. Theanalysis was based only on predominant (crop and beef production) activities suchas wheat (Triticum aestivum L.), maize (Zea mays L.), soybean (Glycine max L.


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 termdata, beef production was estimated from equations for each ecologically homo-geneous area (Viglizzo, 1982) that relate stocking rate (available data) to meatproduction per hectare.


3.1. SUSTAINABILITY 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 notprovide specific information, or that could become hard to interpret or use. Indicat-ors here referred to significant ecological structures and functions, avoiding thosethat had low environmental significance.

We faced the lack of sufficient data, variable quality of data and heterogeneityof data sources. Emphasis were put on the homogenization of data basis for allstudy areas. Thus, twelve basic indicators were defined: (1) land use, (2) fossilenergy (FE) use, (3) fossil energy use efficiency, (4) nitrogen (N) balance, (5) phos-phorus (P) balance, (6) nitrogen contamination risk, (7) phosphorus contaminationrisk, (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 informationsystem (GIS). Dot-density maps were built for most indicators. The procedureallowed a graphic comparison of attributes on different areas. The detection ofspatial patterns and density gradients for each study environmental parameter wasthe main outcome of this procedure.

3.2. LAND USE (INDICATOR NO. 1)

Land use is an indicator of high relevance in agro-ecology. Because they affectthe functionality of agro-ecosystems, both land use and technology application aredeterminants of sustainability in the rural environment (Viglizzo et al., 2001). Landuse refers to the purposes (Van Latesteijn, 1993; Rabbinge et al., 1994) by whichlands are allocated in agriculture. Changes in land use generate spatial patterns andtemporal trends of environmental alteration (International Geosphere-BiosphereProgram/Human Dimensions of Global Environmental Change Program, 1995).This indicator was basic for calculating the rest of the indicators that were calcu-lated here. In our case, different land use patterns in time and space were estimatedin terms of the proportional allocation (%) of the land to: (a) native rangeland,(b) introduced perennial pastures, and (c) annual crops.

3.3. FOSSIL ENERGY USE (INDICATOR NO. 2) AND FOSSIL ENERGY USE

EFFICIENCY (INDICATOR NO. 3)

The use of fossil energy (FE) highly correlates with the intensification of agri-culture. Being usually linked to contamination episodes and to greenhouse gasesemission, the increasing use of FE is frequently associated with environmentaldegradation (Agriculture and Agri-Food Canada, 2000).

This indicator was calculated by considering the annual energy costs per hectareof predominant inputs (fertilizers, seeds, concentrates, pesticides) and practices


(tillage, planting, weeding, harvesting, etc.) expressed in joules per hecatare. Thefossil energy cost of inputs and practices were obtained from different sources ofestimations (Reed et al., 1986; Stout, 1991; Conforti and Giampietro, 1997; Pimen-tel, 1999). Although it was not possible to check in detail the original proceduresto estimate such figures, we assume they provide an acceptable estimation of allinvolved fossil energy costs.

The fossil energy use efficiency was calculated by considering the amount ofmegajoules (Mj) of FE used to get one Mj of product. Calculations were made onannual basis taking into account the proportional participation of each analyzedactivity. Under this scheme, the larger the amount of FE used to produce one unitof energy, the less efficient the production process was. When the input/output ratiodecreases over time, the FE use efficiency increases and the relative environmentalimpact decreases in equivalent proportion.

3.4. BALANCE OF NITROGEN (INDICATOR NO. 4) AND PHOSPHORUS

(INDICATOR NO. 5)

Among other nutrients, the adequate supply of nitrogen (N) and phosphorus (P) isessential to the plant growth and development. The consecutive cultivation of landthrough many years, alters the original nutrient endowment of arable soils. Theaccumulation of effects generates imbalances that worsen over time. If extractionexceeds supply over the years, the accumulation of negative balances may finallyprovoke a severe nutrient depletion, reduce crop yields and lower economic returns.Conversely, if supply exceeds extraction, the accumulation of positive balancescan overload the soil with nutrients, causing potential for both soil and water con-tamination. Ideally, a well-balanced input/output ratio of nutrients is essential forachieving and maintaining soil sustainability.

A simplistic procedure was followed here. The annual average of availableN and P in soil was estimated as the difference between inputs and outputs perhectare and per year, in each study area. A simplistic assumption was adopted:(a) if extraction exceeds supply, soil fertility declines, and (b) if supply exceedsextraction, the excess is cause of potential contamination. The export of N an P byagricultural products was the only way of loss considered in this work. On the otherhand, the considered ways of nitrogen gain were (a) precipitation, (b) fertilizers,(c) biologic fixation by legumes, and (d) purchased feed, later excreted as cattleurine and manure. The predominant phosphorus gains are fertilizers and purchasedfeed.

3.5. CONTAMINATION RISK BY NITROGEN (INDICATOR NO. 6) AND

PHOSPHORUS (INDICATOR NO. 7)

The assessment of contamination risks by N and P is key for assessing the sustain-ability of intensive agriculture. For example, nitrate leaching into ground water canbe risky for human and animal health. On the other hand, the runoff of water with


nitrates and P can increase the eutrophication risk on ponds and lakes. The rapidincrease of algae and aquatic plants depletes the oxygen in water and alters the biotacomposition in surface water. Given that the balance of P was always considerednegative in the Argentine Pampas, it was empirically assumed that contaminationrisk was low. But considering that phosphorus fertilization has rapidly increased inthe 1990’s, the risk of contamination has probably increased, particularly in certainareas.

The N contamination risk was calculated considering only the residual N whenthe N balance was positive. Dividing the amount of residual N by the amount ofwater available for nitrogen dilution (water excess), the N concentration in watercan be estimated. The excess of water was calculated on annual basis from a waterbalance estimation, which takes into account the water gain by rainfall (mm yr−1)less the real evapotranspiration in the same period. Besides, the contamination riskcalculation proceeded only in areas where the excess of water exceeded the waterholding capacity of soils. We have utilized default values for water holding capacityof average soils cited by McDonald, 2000. They were: (a) 100 mm for a sandy ora 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 – realevapotranspiration) was less than the water holding capacity of the soil, saturationdid not happen, leaching was absent and the water contamination risk was low. Asimilar procedure was used to calculate the P contamination risk.

3.6. PESTICIDES CONTAMINATION RISK (INDICATOR NO. 8)

The pesticides contamination risk was described here by a relative index. This in-dicator seems to be useful for doing temporal and geographic comparisons. Strictlyspeaking, absolute values are not meaningful. Calculation included the most com-mon insecticides, herbicides and fungicides that predominated in different decadesfor the main agricultural activities. The actual toxicity values (LD-50) used wereprovided by manufacturers, and obtained from a well-known current pesticidesguide (CASAFE, 1997).The pesticide contamination risk arises concern on: (a) wa-ter and soil degradation by pesticides residues, (b) air quality degradation by thevolatile fraction of pesticides, and (c) the negative impact on biodiversity.

This indicator was based on the assessment of the relative toxicity of predomin-ant pesticides packages used in different decades for various farming activities indifferent areas of the Argentine Pampas. The land use allocation was the base layerof information on which pesticide packages, and their corresponding toxicity val-ues, were superimposed. Commercially recommended application rates of activeproducts were used for each area. The relative index for pesticide contaminationwas the result of summing-up each pesticide contribution per hectare, after mul-tiplying the proportion (%) of land allocated to each analyzed crop, by the toxicityof each product. Using a GIS to superimpose various information layers, maps


of relative toxicity were produced for different ecological areas in the Pampas, indifferent decades.

3.7. SOIL EROSION RISK (INDICATOR NO. 9)

Soil quality was defined here as the soil capacity to sustain agricultural activitiesover time without affecting its productivity and the quality of the environment.Erosion has negative effects at the field level, and also can impact larger areasoutside the field. Normally, erosion is expressed in terms of lower chemical fertility,yield falls, reduced soil infiltration and water retention, increased soil compactionand runoff, shift in soil pH, and ditch formation. Outside the field boundaries,the principal consequences are the sediment deposition on adjacent fields, sanddunes formation, sedimentation and embankment of drainage ditches and naturalsurface waters (ponds, lakes, creaks and rivers), water quality degradation, andwater contamination.

Unsuitable land use schemes and tillage practices can cause severe erosion infragile lands. In general, the high quality soils tolerate some degree of erosionwithout affecting their productivity. In those cases, a natural soil forming processseems to compensate erosion losses, at least in the short-term. However, temporaltrends in different areas need to be assessed and monitored to prevent undesirableconsequences.

In this work, the erosion risk was assessed by means of a relative index. Themethod was based on the relative quantification of main responsible factors thatgenerate water or wind erosion. Four factors were multiplied: (a) land use (cropand pasture allocation in lands), (b) soil fragility (valued through the organic car-bon content), (c) tillage practices (e.g., conventional, minimum tillage, or no-till),and (d) slope. Although digital elevation models could be very useful for gettingreliable slope data, this factor was considered irrelevant for our analysis becausethe Argentine Pampas is very large flat plain with no steep slopes. Accepted criteria(Agriculture and Agri-Food Canada, 2000) classify water erosion risk into fivecategories: 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 (morethan 33 tons). The method used here allows only for estimation of a relative erosionrisk. 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 ofsuch tool did not allow us to get quantitative comparisons.

3.8. HUMAN PRESSURE ON HABITATS (INDICATOR NO. 10)

Over the centuries, agriculture has transformed natural habitats into human-de-signed habitats. Despite the fact that agriculture has historically benefited frombiodiversity, habitat intervention by agriculture greatly reduces natural biodiversity.Biodiversity has provided agriculture (a) germplasm and gene variability, (b) pol-linators, (c) beneficial insects and control agents, and (d) waste disposal and recyc-


ling pools. On the other hand, diversified agriculture could also benefit biodiversity,through integrated farming schemes that favor the recovery of environments thatwere 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 ag-gression as well. The aggressive agronomic practices can reinforce the negativeimpact of intensive land use on habitat and biodiversity.

The estimation of a habitat quality index (HQI) that we have used here, re-sembles the above described method for calculating the risk erosion indicator. Un-der the assumption that human action affects habitat and biodiversity, this methodwas used to generate a relative HQI to assess the degree of human intervention onthe habitat through (a) land use, (b) tillage practice, and (c) pesticide contamina-tion. 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 erosionrisk. The same coefficients obtained from the estimation of pesticide contaminationrisks were used here as well. The combined HQI factor was the result of the simplemultiplication of those three intervention factors. Thus, the higher the proportion ofannual crops, the aggressiveness of tilling practices, and the toxicity of pesticides,the greater the detrimental effect of humans on habitats.

3.9. CHANGES IN SOIL CARBON STOCKS (INDICATOR NO. 11)

Carbon (C) is the main component of organic matter, and therefore a factor thatstrongly determines soil quality. Organic matter decays are associated with soilfertility and soil structure losses, and also with a higher soil erosion risk. Dependingon organic matter gain or loss, soils can respectively act as a sink or a source ofatmospheric 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 originalorganic 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 in-corporation of C is possible. However, it is unlikely we can achieve a full recoveryup to levels of pristine stages.

The procedures followed in this work were based on the IPCC (1996) reviewedguideline methodology. An initial C stock that varied according to the study areawas estimated. Calculation of this stock was achieved assuming that the soil Ccontent for 1960 was 65% of the initial content of native soils. As the soil C contentin 1960 varied according to the study ecological area, so did the estimation of theoriginal value.

Past information on land use was the basic factor utilized for calculating changesin 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-


idation rates. While minimum- and no-till practices favored soil C sequestration,the conventional tillage was cause of C emission. Furthermore, the IPCC methodprovided default coefficients values for organic matter enrichment of soils via cropresidue accumulation.

3.10. GREENHOUSE GASES (GHG) BALANCE (INDICATOR NO. 12)

Atmospheric gases, including carbon dioxide (CO2), methane (CH4), nitrous oxide(N2O), ozone (O3) and water vapor, are normal components of the atmosphere thathave, at the same time, a greenhouse effect. Such gasses are transparent to shortwave radiation that reaches the earth, but they are opaque to earth emission oflong wave radiation. The consequence is that emissions bounce, and the trappedradiation increases the global temperature of the Earth. As a natural phenomenon,the global warming caused by the greenhouse gases is known as natural g reen-house effect. The greenhouse gases concentration has remained rather constantover 10 000 yr. But anthropogenic emissions increased dramatically during the last200 yr, raising the global temperature. Fossil fuels consumption for industry andcommerce, transportation, household comfort, deforestation and soil cultivation areconsidered major causes of increased GHG emissions, triggering a human-drivenglobal warming that had no antecedents in the Earth history.

Carbon dioxide and CH4 account for 90% of the human-driven greenhouseeffect. Atmospheric concentration of these gases has respectively increased 30, 15and 145% for CO2, N2O and CH4 (Desjardins and Riznek, 2000). Grain crops andcattle are both significant sources of greenhouse gases that need to be estimated forthe Pampas plain.

The GHG balance for rural areas was estimated following the standard guidelinesof IPCC (1996). All gases were converted into CO2 equivalent (ton ha−1 yr−1). Thecalculations included emission and sequestration due to land use change, graincropping and cattle production activities.

The use of fossil fuels was another major source of CO2. The method includedfuels used in rural activities and fuels used for manufacturing fertilizers, herbicidesand machinery. Because of the heterogeneity of cases and procedures, emissionsdue to rural transportation were not included in our calculations.

Ruminants were a significant source of GHG. Ruminants emit methane fromenteric fermentation and fecal losses. Methane has a greenhouse power that is 21times greater than CO2. This figure was used to convert CH4 into CO2 equivalents.Nitrogen excreted in feces and distributed with fertilizers was another significantsource of nitrous oxide (N2O) emission. N2O has a greenhouse power 310 timesgreater than CO2. Losses of N2O 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 valuessuggested by the IPCC (1996) were used for estimating gains and losses of carbon.


The methodology proposed by IPCC (1996) estimated emission or sequestration ofCO2 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 offorests and prairies into arable land (CO2-CTBP), (4) abandonment of intervenedlands (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 cropemissions (CO2-EA). Our study took into account the first 2. Because rice is not apredominant crop in the Pampas, it was not included in the study. Emissions fromruminants were calculated from beef cattle only.

The emission of N2O was the most difficult to estimate because of the com-plexity of determinations. Emission sources are: (1) Feces and urine (CO2-EDHO)from domestic animals, (2) volatilization, runoff and infiltration (CO2-EIVLI) fromsynthetic 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 gradi-ents, and the interpretation of temporal trends, more than on the absolute value ofthe estimated indicators. We aimed at determining if environmental conditions areimproving, worsening or keeping stable over time. We consider that comparisonsamong areas and decades are not invalidated by some methodological constraintsmentioned above.

4.1. LAND 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 landuse and land use change patterns. In comparison to the rest of the agrecologicalareas, grain crops had their faster expansion in the Rolling pampas, replacingthe pristine natural lands. A cropland profile supported by a high environmentalpotential for crop production, has characterized the Rolling pampas during the lastcentury.


The Flooding and the Mesopotamian pampas, on the other side, have the lowercapacity for crop production. Due to soil quality and topography limitations, theproportion of land devoted to cereal crops and oil seeds has remained low andconstant across time. A net predomination of introduced and native pastures definea net orientation to cattle production in both areas.

In terms of agricultural productivity, the Central and the Southern pampas areboth in the middle between the natural conditions of the areas mentioned above.They are characterized by predominant mixed, grain crop-cattle production sys-tems. Soil quality and climatic limitations have prevented a larger conversion ofgrazing lands to croplands. During the 1990’s, however, technologies that aimed atincreasing productivity and reducing environmental impacts (e.g., no-till planting,fertilization to minimize nutrient depletion) have favored the expansion of graincrops.

Maps in Figures 2 and 3 show, in a time-space dimension, land-use changesduring the period 1960–2000. Figure 2 illustrates the expansion of wheat, maize,sunflower, soybeans, and winter and summer small-grain pastures. Color intensityshows a consistent increase of croplands in the Rolling pampas and in the mixedsystems of Central and Southern pampas. Croplands expansion, however, did nothappen by forcing livestock into marginal lands, as it was frequently assumed.Furthermore, legume-based perennial pastures have also increased, primarily onthe Central, Southern and Flooding pampas (Figure 3), as an indication of a sim-ultaneous intensification of crop and cattle production in these areas. The plantingof improved annual and perennial pasture species, the increased use of concentratefeeds and the management of a higher animal stock density in smaller areas, havecharacterized the predominant livestock production system in the Pampas.

Land use patterns and land use changes were key factors for explaining theenvironmental behavior of the Pampas. All indicators, from fossil energy consump-tion to contamination risk, from erosion risk to GHG emission, were particularlysensitive to land use. Technology was the second important factor. Results clearlysuggest that the 1990’s decade was the starting point in land transformation, mainlyin the Rolling pampas, but also in the Central and the Southern pampas.

4.2. AN INCREASING ENERGY STATUS

Some limitations in the calculation of fossil energy (FE) consumption and FE useefficiency need to be stated. Firstly, the FE cost of inputs and activities was notalways well documented in literature, and accuracy of data can be argued in somecases. Secondly, some inputs or activities were underestimated due to their smallparticipation in the general calculation, and also due to insufficient data. Thirdly,data reported in the literature were not always updated in response to technologychange. And fourthly, the imprecise description of calculations reported in theliterature raises the risk of double accounting in some cases.


Fig

ure

2.C

hang

esin

the

area

allo

cate

dto

annu

alcr

ops

(whe

at,m

aize

,so

ybea

n,su

nflow

er,

win

ter

and

sum

mer

annu

alpa

stur

es)

inth

eA

rgen

tine

pam

pas

duri

ngth

epe

riod

1880

–200

0.


Fig

ure

3.C

hang

esin

the

alan

dal

loca

ted

tocu

ltiv

ated

pere

nnia

lpas

ture

sdu

ring

the

peri

od18

80–2

000

inth

eA

rgen

tine

pam

pas.


TABLE I

Fossil energy (FE) efficiency use in the Argentine pampas during the period1960–2000. 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 0.17 0.11 0.16

Central subhumid 0.35 0.18 0.17

Central semiarid 0.45 0.44 0.41

Southern 0.18 0.14 0.15

Flooding 0.13 0.12 0.15

Mesopotamian 0.18 0.17 0.23

Figure 4 shows changes in energy productivity (Gj ha−1 yr−1) and FE consump-tion (Gj ha−1 yr−1) in the five study areas between 1960 and 1996. The resultsindicate that the energy status and the energy fluxes have increased all over theregion, and primarily in a belt defined by the Rolling, the Central sub-humid andthe Southern pampas. The increased proportion of grain crops mostly explains thisincreased energy budget.

Table I provides figures on FE use efficiency over the three analyzed histor-ical periods. Although they have higher FE consumption, the areas where graincrops have largely expanded show energy yields that greatly exceed those wherecattle production predominates. Consequently, croplands show a higher FE ef-ficiency use. Furthermore, efficiency tended to increase in parallel with the in-creasing energy yield of annual grain crops. The larger adoption of minimum andno-tillage practices that demand less FE use, was probably highly associated withthat increased efficiency.

4.3. NUTRIENT BALANCE AND NUTRIENT CONTAMINATION: BETWEEN THE

RISK AND THE CHALLENGE

The complexity of the N and P cycles in the Pampas was clearly demonstrated byour own research (Bernardos et al., 2001) supported by the EPIC model (Williamset al., 1989). It was quite evident that the elementary input-output approach that wehave used here does not properly represent real cycles simply because the complexdynamics between nutrients pools and fluxes were not properly captured. Ideally,rates of leaching, run-off, sediment and volatilization should have been consideredin our estimations. However, time and cost limitations to gather such information in


Fig

ure

4.C

hang

esin

ener

gy(M

jha

−1yr

−1)

harv

este

dth

roug

hpr

edom

inan

tag

ricu

ltur

alpr

oduc

ts(w

heat

,co

rn,

soyb

ean,

sunfl

ower

and

beef

)w

ithi

nth

epe

riod

1969

–200

0in

the

Arg

enti

nepa

mpa

s.


TABLE II

Estimation of nitrogen (N) and phosphorus (P) balances in ecologically homogeneous areas of theArgentine pampas during the period 1960–2000

Pampas N balance (kg ha−1 yr−1) P balance (kg ha−1 yr−1)

1960 1988 1996 1960 1988 1996

40 70 100 40 70 100

Regional average 0.23 1.76 9.0 16.3 26.8 –2.03 – 5.07 – 5.9 –4.2 –2.3

Rolling 2.40 –6.03 –0.5 14.6 29.7 –3.66 –12.71 –12 –9.2 –6.2

Central subhumid –1.04 –5.29 11.8 18.3 24.8 –2.11 – 5.66 – 7.1 –5.6 –4.1

Central semiarid 2.43 3.20 12.5 16.6 20.8 –0.91 – 1.56 – 5.1 –4.4 –3.9

Southern –2.65 2.83 6.8 12.8 18.7 –1.86 – 3.84 – 4.6 –3.1 –1.6

Flooding 1.01 8.35 11.4 5.7 19.9 –1.25 – 1.48 – 1.8 –0.3 –1.2

Mesopotamian 1.16 6.83 6.5 10.7 14.8 –1.03 – 1.22 – 3.2 –1.7 –0.2

In 1996, the estimations include 3 fertilization hypothesis: 40, 70 and 100% of the area wasfertilized following commercial recommended dose.

each analyzed geo-political district, were not overcome. In this context, the simpleinput/output approach was adopted as the best option available, assuming thatnegative balances indicated nutrient loss, and the positive ones indicated potentialcontamination. Beyond such limitations, the method seems to be useful for makinggeographic and temporal comparisons.

Likewise, the lack of accurate information on evapotranspiration and water soilretention capacity at the district level was an additional limitation. The method hadlow resolution for detecting contamination risk in areas of high animal concen-tration (beef and dairy feedlots, poultry and pig farms), or in irrigated areas forintensive fruit and vegetable production. Neither was the method sensitive for de-tecting contamination risk under unexpected meteorological events, such as heavyrains that result in flooding or excessive leaching.

Table II shows a comparison of changes in N and P balances. A significantincrease of N and P extraction from 1960 to 1996, consistent with cropland ex-pansion, was detected especially in the Rolling pampas. But such extraction alsooccurred in less productive areas like the Central Sub-humid pampas. Because ofthe lack of precise data on fertilizers use during the 1990 decade, three hypothet-ical fertilization schemes based on commercial recommendations were considered:100, 70 and 40% of the area devoted to grain crops, respectively. This allowed asimplistic interpretation about varying degrees of fertilizer adoption in differentareas.

The risk of contamination was closely related to positive balances of nutrientsand water. In the case of P, the whole region would have underwent a net negativeP balance, even in the 1990’s (Table III). In the 1990’s, positive balances were only


TABLE III

Estimation of the relative soil erosion risk in the Argentine pam-pas and its ecologically homogeneous areas during the period1960–2000

Pampas Coefficient of relative erosion risk

1960 1988 1996

Regional average 0.11 0.09 0.09

Rolling 0.08 0.08 0.08

Central subhumid 0.16 0.11 0.12

Central semiarid 0.21 0.20 0.26

Southern 0.13 0.11 0.10

Flooding 0.06 0.03 0.05

Mesopotamian 0.11 0.09 0.10

detected where extraction is very low (grazing lands), and when P fertilizationbecomes a massive practice (in 100% of lands).

The systematic fertilization of crops and prairies has consolidated in the Pampasduring the 1990’s. Considering the three hypothetical fertilization levels previouslydiscussed, if fertilization were reduced to 70% or to 40% of the total area, posit-ive nitrogen balances tended to disappear, and negative phosphorus balances wereinevitable. Despite the three levels of fertilization discussed, we assumed that the40% level has probably represented the more realistic scenario for the 1990’s.

4.4. PESTICIDE CONTAMINATION RISK: UPS, DOWNS AND UNCERTAINTIES

The calculation method applied was based on several assumptions. Firstly, al-though the utilized pesticides varied largely, pesticide packages were uniform in allareas for each particular crop. We have to recognize, however, that alternative pack-ages for similar activities have been utilized, especially during the 1990’s whenthe offer of commercial pesticides has multiplied. Secondly, our model assumedthat 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 thatthe larger the combined potential toxicity of applied pesticides, the larger the im-pact on biodiversity. Field measurements were ideally required, but no direct fieldmeasurements of pesticides impact on biodiversity was possible for this study.

Figure 5 shows the estimates of relative indices of contamination risk due topesticides application between 1960 and 2000. The expansion of croplands and theincreasing use of pesticides over the decades would explain the significant increaseof pesticide contamination risk, especially during the 1990’s. The effect seems to


be particularly dramatic in the most productive areas, for example in the Rollingpampas, but also in areas where cropland expanded. However, field studies arerequired to confirm this estimation. We suppose that this warning trend may reversein the near future because a family of new, soft pesticides are being introduced atpresent in the commercial market. Some common and highly toxic pesticides usedduring part of the 90’s were banned nowadays.

4.5. SOIL EROSION AND HABITAT THREATS: CONTRASTING TRENDS

The relative coefficients used to evaluate the soil erosion risk were calculated tak-ing into account the predominant soil quality, land use and tillage practices in eachperiod.

Table III depicts the estimations of relative erosion risk for the different Pam-pas areas in the period 1960–2000. The higher risk tended to concentrate on theWestern lands of the Central pampas, where the conversion of grazing lands intocroplands 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 areasduring the 1990 decade. Despite the expansion of the cropping area, this stabilizedperformance was probably the result of the increased adoption of soil conservationtillage, which would have compensated the more intensive use of land.

Similarly, the relative coefficients used to assess the degree of habitat interven-tion by humans were estimated taking into account three driving factors: land use,tillage operations and pesticide contamination risk. A combined factor arose frommultiplying the three factors. Although no long-term data is available on the effectsof agriculture on wildlife in the Pampas, we assumed that a relative interventionindex would provide an indirect estimation of agricultureþs impact on biodiversityacross space and time. Nevertheless, this indicator says nothing about the actualimpact of man on biodiversity. Ideally, this method should include the assessmentof habitat fragmentation on key species.

Figure 6 shows the relative index of habitat intervention across ecological areasand over time. Estimations demonstrated that the greatest intervention would havetaken place over the most productive lands of the Rolling pampas. Aggressionwould have been particularly strong during the 1990’s. The lowest degree of inter-vention was detected both in the Flooding and the Mesopotamian pampas. Theseareas still conserve large proportions of land devoted to cattle production on nativeand improved pastures. Tillage operations and pesticides application are minimum.This agrees with results of a previous work of Viglizzo et al. (2001).

4.6. CARBON STOCKS AND GHG EMISSIONS: IMPROVING TRENDS

The method applied was rather simplistic if we accept that the soil C dynamicsis very complex. Because our assumptions were strong, an unavoidable degreeof uncertainty about the accuracy of estimations had to be accepted. Land use,


Fig

ure

5.E

stim

atio

nof

the

rela

tive

risk

ofpe

stic

ides

cont

amin

atio

nin

diff

eren

tare

asof

the

Arg

enti

nepa

mpa

sw

ithi

nth

epe

riod

1960

–200

0.


Fig

ure

6.E

stim

atio

nof

the

rela

tive

risk

onha

bita

tsan

dbi

odiv

ersi

tydu

eto

hum

anin

terv

enti

onin

diff

eren

tare

asof

the

Arg

enti

nepa

mpa

sw

ithi

nth

epe

riod

1960

–200

0.


climate, soil and tillage interactions, added to agronomic practices, were majorsources of variation in soil C stock responses. Furthermore, the lack of data toestimate the original soil C content in some areas has introduced an additional con-straint. Likewise, soil sampling errors and the quality of laboratory analysis in thepast are other sources of uncertainty that could not be overcome. Therefore, resultscomprise average estimates that were considered the best information available.

According to the results, the whole Pampas plain has experienced a significantloss of soil C between 1960 and 2000 (Figure 7). Over the 40 yr study period, thelargest 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 increaseof the cultivated area in time, the C loss rate has declined in time. Rates of C losswere significantly lower in the 1990’s than in the 1960’s. The quick expansion ofminimum-and no-tillage may be the main factor explaining this behavior.

On the other hand, the procedures suggested by the IPCC (1996) to estimateGHG emissions are relatively new. Then, some uncertainty about results has arisenwhen models and suggested default values were applied. Ideally, experimentaland field data would be necessary to adjust procedures and coefficients to eachenvironment under analysis.

Our results suggest that the whole Pampas has behaved as a net emitter of GHGbetween 1960 and 2000 (Figure 8). The larger net emissions were found on theRollling pampas and the Central pampas during two (1960 and 1988) of the threeanalyzed periods. A generalized decline in emissions for the whole region wasestimated for 1996. However, differences among the five study areas have persisted.

We have interpreted that minimum and no-till practices were the main respons-ible for these positive changes. The reduced fossil fuel consumption due to min-imum or no-tillage operations has reduced CO2 emissions and favors C retentionin soil organic matter over time.

5. Conclusions

Probably, this work represents a pioneer effort for getting, at a regional scale, anintegrated view of the environmental sustainability of agriculture in the ArgentinePampas. It can also be viewed as a first contribution to get a permanent system ofenvironmental monitoring for rural areas in the country.

Certainly, both strengths and weaknesses should be taken into account. Weak-nesses were in turn pointed out when each indicator was considered in particular.One obvious conclusion is that our calculation methods and techniques need tobe perfected. An extra effort is needed to capture not only geographical patternsand gradients, and time tendencies, but also to get a realistic quantification of suchchanges. The strengths of this study are not in the absolute value of environmentalindicators, but in the identification of patterns, gradients and trends in space andtime. This allowed us to determine if critical environmental parameters are keeping


Fig

ure

7.E

stim

atio

nof

carb

onlo

sses

(tC

ha−1

yr−1

)in

diff

eren

tare

asof

the

Arg

enti

nepa

mpa

sdu

ring

the

peri

od19

69–2

000.


Fig

ure

8.E

stim

atio

nof

gree

nhou

sega

ses

emis

sion

(tC

ha−1

yr−1

)in

diff

eren

tare

asof

the

Arg

enti

nepa

mpa

sdu

ring

the

peri

od19

60–2

000.


stable, improving or worsening across time. We must also consider the relativeimportance of factors. Since they depend on the use of an index, they should havedifferent weights.

Beyond the results, two questions have arisen: is the Pampas agriculture sus-tainable? Are their production trajectories sustainable in time? Indicators differedin their behavior and trends. Nowadays we can not say that agriculture, in the wholePampas, is sustainable or not sustainable. While some indicators tend to improve,others keep stable, and the rest worsen. Furthermore, the projections of trends areno homogeneous in all areas. The transition towards a more intensive model ofagricultural production, both in terms of land use and technology application, hasclearly characterized the 1990’s. The production and environmental trajectoriesindicate that many farming systems in the Pampas are resembling some intensivemodels that are common in the Northern Hemisphere. In response, we need a newproductive and environmental view to replace the traditional one.

The trajectory of indicators differed in different ecological areas. Positivechanges 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) aconsistent decline in greenhouse gases emissions. No-till and conservation tillagehave apparently driven those beneficial changes.

Some negative changes that deserve especial attention were observed on theother hand. Some issues that should be taken into account are the following: (a) theregional N balances tended to become positive in many cropping areas, deliveringresidual N that can potentially contaminate soils and water. The contamination risktended to increase as the N fertilization increased. However, given that fertilizationlevels were neither high nor homogeneous in all areas, we believe that the risk ofcontamination is still low in the Pampas as a whole, (b) provided that P balanceswere negative in most of the studied areas, the risk of P contamination still appearsto be rather far in the future, and (c) a more careful assessment of the increasingrisk of pesticides contamination during the 1990’s needs to be considered.

The indicators that showed a negative trend are key to identify critical prob-lems that will require special attention in the close future. Putting aside the geo-graphic heterogeneity, an environment degradation belt is emerging. Starting fromthe Rolling pampas, the belt is extending over many districts in the subhumid Cent-ral pampas. Thus, a potentially trouble area that would require close attention wasdetected. An environmental policy, and a well-defined set of research priorities, arenecessary to face future challenges on this environmentally critical belt.

Finally, the long-term monitoring system for the rural environment through thesupport of reliable indicators, appears to be a key tool to guide environmentalpolicies, to drive research priorities, and to orient communication strategies. Aperiodic updating of information is crucial to detect critical changes in space andtime, and to fit environmental strategies and tactics that match vital needs of theregion.


Acknowledgements

We thank the Secretary of Science and Technology of Argentina for the financialsupport to this research. We also acknowledge the contribution of various publicinstitutions for providing key data for this study, and many colleagues for theirvaluable 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-EnvironmentalIndicador Project.

Allen, T. F. H. and Starr, T. B.: 1982, Hierarchy Perspectives of Ecological Complexity, Universityof 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 farmingtransformation in the Argentine pampas’, Agricult. Syst. 69, 215–234.

CASAFE: 1997, ‘Guía de Productos Fitosanitarios para la República Argentina’, Cámara de SanidadAgropecuaria y Fertilizantes, Buenos Aires.

Cole, C. V., Stewart, J. W. B., Ojima, D. S., Parton, W. J. and Schimel, D. S.: 1989, ‘ModellingLand 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 compar-ison’, Agricult. Ecosyst. Environ. 65, 231–243.

Covas, G.: 1989, ‘Evolución del manejo de suelos en la región pampeana semiárida’, Actas de lasPrimeras 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 UnitedNations (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 SciencePublishers B.V., Amsterdam.

IGBP/HDP: 1995, ‘Land-Use and Land-Cover Change: Science/Research Plan’, Report No. 35 ofThe International Geosphere-Biosphere Programme (IGBP), and Report No. 7 of The HumanDimensions 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 ofthe Agri-Environmental Indicador Project, Agriculture and Agri-Food Canada, Ottawa, Ontario,pp. 161–170.


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 Searchfor 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 Areasin 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 inCalifornia’, J. Agron. Crop Sci. 157, 99–104.

Sala, O. E., Stuart Chapin, F., Armesto, J. J., Berlow, E. et al.: 2000, ‘Global biodiversity scenariosfor the year 2100’, Science 287, 1770–1774.

Satorre, E.: 2000, ‘Production Systems in the Argentine Pampas and their Ecological Impact’, inO. T. Solbrig, F. di Castri and R. Paarlberg (eds), The Impact of Global Change and Informationon 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 RegiónPampeana 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 thepampas of Argentina’, Agricult. Ecosyst. Environ. 81, 65–81.

Viglizzo, E. F., Roberto, Z. E., Lértora, F., López Gay, E. and Bernardos, J.: 1997, ‘Climate andland-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 planningand 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.

Documents

ENVIRONMENTALASSESSMENTOF AGRICULTUREAT A REGIONAL SCALE ... et al 2002.pdf · environmentalassessmentof agricultureat a regional scale in the pampas of argentina e. f. viglizzo∗,