Elsevier

Ecological Economics

Volume 55, Issue 4, 1 December 2005, Pages 467-484
Ecological Economics

Analysis
Using GIS-based ecological–economic modeling to evaluate policies affecting agricultural watersheds

https://doi.org/10.1016/j.ecolecon.2004.12.006Get rights and content

Abstract

This paper has three purposes. The first is to conceptualize agricultural watersheds as complex adaptive human ecosystems that co-produce agricultural goods and ecosystem services. The second is to demonstrate a generalizable framework for the spatial modeling of ecosystem service production in watersheds based on this conceptualization. The third is to examine the policy implications of the analysis conducted using this spatial decision support system (SDSS).

Analyses using the SDSS show that restrictions on soil loss to the “tolerance level” (T) cause average farm income to decline by only 4%, a reduction that is nearly eliminated if the Conservation Reserve Program (CRP) is available to farmers as an income-generating alternative. The spatially variable response of farmers to soil loss restrictions and the CRP creates a complex pattern of winners and losers and a markedly different land use pattern and crop mix than occurs without these programs. The land use pattern associated with T restrictions and the CRP yields about 64% lower erosion rates and 43% lower sediment yields than the pattern without T restrictions or CRP. These and other results from the SDSS analysis point out that ecosystem service-based subsidies, such as CRP, improve the joint production of farm income, soil conservation and water quality in agricultural watersheds. These subsidies could perhaps receive greater funding by shifting agricultural subsidies from income supports tied to yield and price as well as other crop-based programs. In this way, public expenditures on agriculture would produce a valuable public benefit in the form of load reductions in a TMDL context, and an augmentation of ecosystem services now in decline in many agricultural watersheds. Further methodological developments now underway using evolutionary algorithms can find near-optimal solutions for farms over time and for landscape patterns over whole watersheds.

Introduction

Ecosystem services such as nutrient cycling, regulation of atmospheric gases, soil formation and binding, sediment trapping, energy fixation, and expansion of wildlife habitat are increasingly recognized as essential to society and of great economic value (Costanza et al., 1997, Daily, 1997, United Nations Development Program, United Nations Environmental Program, World Bank, World Resources Institute, 2000). With the considerable successes that have been achieved in controlling industrial pollution, especially point-source water pollution, the scale of management for pollution control and improving ecosystem services is increasingly at the landscape and watershed scale. Improvement in biodiversity and control of polluted runoff, for example, are issues that must be addressed by managing landscapes. While fossil fuel combustion is the primary source of greenhouse gases emitted, managing landscapes to foster carbon sequestration is also a significant part of the potential response to increased greenhouse gases and consequent global warming (Caspersen et al., 2000, Schulze et al., 2000).

Agricultural landscapes, which constitute about 50% of the land in the contiguous U.S. and similar proportions in other inhabited regions of the world (Vitousek et al., 1997), are distinct from other rural landscapes by their focus upon production of food and fiber commodities. While agricultural landscapes harbor natural capital and, thus, produce ecosystem services, forest, prairie, wetland, riparian and other ecosystems are often capable of producing far greater flows of ecosystem services per hectare than the agriculture fields that replaced them (Costanza et al., 1997). Tilman et al. (2002) outline the steep challenges that must be met if global food production is to fulfill the demands of a growing and increasingly affluent 21st Century population sustainably. Increasing demand for meat and use of fertilizers and water for irrigation pose substantial threats to ecosystem services flows from agricultural lands. Conversely, the potential to restore the production of ecosystem services lies greatly in private agricultural lands, especially grazing lands and croplands that are marginal due to wetness, dryness, steepness, or erodibility. For example, the vast majority of U.S. sites suitable for wetland restoration are now farmland (McCorvie and Lant, 1993) and working farmland has a considerable capacity to sequester atmospheric carbon (Lal et al., 1998).

In Costanza's (2001) expanded model of the ecological–economic system, natural capital, in association with human, social, and manufactured capital, is responsible for maintaining the productive base upon which future economic productivity, as well as individual and community well-being, are absolutely dependent. If this positive model of the manner in which society, nature, and the economy interact is adopted, with the normative goal of sustainability, the conclusion is that society should make greater investments in natural capital to ensure greater delivery of ecosystem services in the present and the future. This greater investment might come at the expense of agricultural commodities produced in the present; however, it is possible such investments would result in greater future delivery of agricultural commodities. Unfortunately, as Zimmerman (1951), Ciriacy-Wantrup (1952), Firey (1960), Hardin (1968), Randall (1983), Lee (1992), Gottfried et al. (1996) and other social scientists have articulated over the past half-century, there is only a narrow range of social circumstances under which resource managers such as farmers are willing to make substantial personal investments in the present to achieve even more substantial public benefits in the future. One critical implication is that in private sector markets, ecosystem services and the natural capital generating them will be under-produced relative to agricultural commodities. This is a consequence of their non-excludable or public good nature and the resulting lack of markets for either the services or the natural capital itself (Randall, 1983). Empirical studies of modern agricultural systems bear out the conclusions of these social scientists. For example, the flow of ecosystem services from agricultural landscapes in Sweden is declining (Bjorklund et al., 1999). Negative environmental externalities (i.e. damage to ecosystem services) in UK agriculture are large—over $300/ha/year (Pretty et al., 2000).

In the U.S., agricultural conservation policy influences considerably the land use choices farmers make and therefore the ecosystem services produced on farms (Lant et al., 2001). For example, the rate at which wetlands have been drained for agricultural production dropped 87% from 237,000 acres/year in the decade 1974–1983 to 30,900 acres/year in the decade 1982–1992 (Wiebe et al., 1996). Farm Bills since 1985 have utilized (1) cross compliance in the form of Conservation Compliance, Sodbuster and Swampbuster and (2) economic incentives in the form of the Conservation Reserve Program (CRP), Wetland Reserve Program (WRP), Grassland Reserve Program (GRP), Wildlife Habitat Improvement Program (WHIP), and Environmental Quality Incentives Program (EQIP) as policy tools to influence land use behavior with considerable effect in reducing soil erosion and wetland drainage (Esseks and Kraft, 1991, Esseks and Kraft, 1993). The CRP, a focus of this study, pays farmers an annual rental fee, often between $50 and 100 per acre per year, for 10 or more years to replace cropping with one of a number of conservation practices. Eligible lands include riparian filter strips, wetlands and highly erodible lands. The CRP was initiated in the 1985 Farm Bill and has been reinstated by the 1990, 1996 and 2002 Farm Bills. Enrollment has been in the range of 30–36 million acres since the late 1980s, an area about the size of Illinois, with the emphasis shifting slightly from cost-effective erosion control to support of a more varied selection of ecosystem services including water quality and wildlife habitat reflected in an Environmental Benefit Index (Lant et al., 2001).

Beyond institutional factors, the second difficulty in increasing ecosystem service flows from agricultural landscapes is that spatial units of land management (farms) are not generally at the spatial scale at which ecosystem services are generated, even though it is on farms that actions to maintain natural capital and improve ecosystem services must be taken if cumulative impacts on ecosystem services are to be avoided. Water quality, for example, is produced at a watershed scale, biodiversity at a landscape scale, and carbon sequestration at a small scale, but with global additive effects. Freyfogle (1998) terms this problem the “tragedy of fragmentation,” as the converse of Hardin's (1968) “tragedy of the commons”; Gottfried et al. (1996) identifies “economies of configuration” where ecosystem services can be fostered by maintaining landscape patterns such as riparian corridors. Using complexity theory (see for example, Capra, 1996), we can conceptualize ecosystem services as “emergent properties” of landscapes at spatial scales generally greater than farms. Each of these concepts means that the geographical pattern of land uses, as well as the total land allocated to crops, pasture, forest, urban and other uses, is responsible for producing many ecosystem services. While the value of a ton of sequestered carbon is independent of where that ton is sequestered due to global mixing of atmospheric carbon, the discipline of landscape ecology has demonstrated the importance of landscape pattern for maintaining biodiversity, among other services. Watersheds have increasingly come to the fore as spatial units for which natural capital and ecosystem services are analyzed (Montgomery et al., 1995, Napier, 1997) and managed (Duram and Brown, 1999, Lant, 1999). The U.S. EPA lists over 2000 watershed-based groups and some 20,000 water bodies and stream segments are in violation of water quality requirements. The Clean Water Act requires about 40,000 Total Maximum Daily Load (TMDL) plans to be written for these water bodies—mostly for sediment, N and P, common agricultural sources of polluted runoff. However, management practices that produce clean water may also facilitate, for example, biodiversity, soil formation, and carbon sequestration; ecosystem services are often complementary.

When interaction effects attributable to landscape patterns exist, thereby resulting in nonlinear relationships between land use and ecosystem services produced, prediction of future states becomes much more difficult, and “planning,” which presumes predictive capacity, evolves into “adaptive management” (after Holling, 1978). Due to these difficulties, water resources and land-use planning in multiple-owner, largely private watersheds has been fragmented and subject to a variety of forces originating both within and outside the watershed (see for example Rogers, 1993). The results of this style of watershed management have been particularly detrimental to the aquatic ecosystems of the U.S. in the 20th Century, despite large expenditures to build, operate and administer pollution control facilities. Broad cross-sectional data indicate a general decline in riverine ecosystem health attributable to excessive water withdrawals, channel modifications, erosion and sedimentation, deterioration of substrate quality, chemical contaminants, over fishing, increases in impervious surface, and the introduction of exotic species (Adler et al., 1993, Doppelt et al., 1993). Thus, we can conclude that the value of natural capital capable of generating ecosystem services at a watershed scale is declining and that this decline is likely due to incentives in the decision environment of land and water resource managers.

This paper has three purposes. The first is to conceptualize agricultural watersheds as complex adaptive human ecosystems that co-produce agricultural goods and ecosystem services. The second is to demonstrate a generalizable framework for the spatial modeling of ecosystem services and the production of agricultural goods in watersheds based on this conceptualization. The third is to examine the policy implications of the analysis conducted using this spatial decision support system (SDSS).

Section snippets

Watersheds as complex-adaptive human ecosystems

Checkland (1981) provides a thorough guide to the practice of “soft-systems” analysis that focuses on human decision making. We apply the method of systems thinking by conceptualizing problems in a rigorous way with flows, stocks, feedbacks, inputs and outputs. The goal of such an analysis, which may or may not include the construction of simulation models, is the pragmatic one of “improving” rather than “optimizing” the real-world operation of the system by exposing bottlenecks, thresholds,

A spatial decision support system for managing watershed dynamics

Spatial decision supports systems (SDSS) were created to support the analysis of complex spatial problems where it is not possible to completely define a problem or fully articulate the objectives of the solution in mathematical terms. In recent years, several SDSS developments have been reported in the GIS literature that integrate GIS and modeling software to provide support to decision-makers for natural resources management (Leavesley et al., 1996, Watson and Wadsworth, 1996). Here we

Results

Table 1 provides the overall results for these scenarios. Under the base scenario (no soil loss regulation and no CRP), the value of crop sales is highest; 72% of agricultural area is allocated to conservation tillage corn and soybeans and 25% is allocated to grazing. Soil loss averages 1.87 T, with 87% of farms losing soil at over T. Sediment fluxes are correspondingly large (Table 1, column 1). When the CRP enrollment option is introduced, gross margin increases by about 1% as 466 ha are

Discussion and conclusions: policy implications

If we place the analyses conducted using the SDSS (Table 1; Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8) in the context of agricultural watersheds as complex adaptive systems (Fig. 1) we arrive at a number of implications for farm policy, watershed planning in a TMDL context, and the benefits of integrating these two policy arenas.

Discussion and conclusions: methodological implications

The research presented here presents a hard systems model of watersheds as complex adaptive systems in the form of an SDSS, but there is a long way to go in model development as well as in better understanding those components of human decision-making and behavior that are beyond models. In fact, models such as those presented here are tools to be used in that larger social response context. GIS-based modeling frameworks such as the Distributed Intelligent Geographical Modeling Environment

Acknowledgements

The authors wish to thank the U.S. Department of Agriculture for supporting the project “Understanding the social context of ecological restoration in multiple ownership watersheds” through the Water and Watersheds program. We would also like to thank the Illinois Council for Food and Agricultural Research (C-FAR) for their support of the research project “Decision support for water quality planning in multiple-ownership watersheds.”

References (76)

  • W.A. Brock et al.

    Resilience and sustainability: the economic analysis of nonlinear dynamic systems

  • F. Capra

    The web of life: a new scientific understanding of living systems

    (1996)
  • J.P. Caspersen et al.

    Contributions of land-use history to carbon accumulation in U.S. forests

    Science

    (2000)
  • P. Checkland

    Systems thinking, systems practice

    (1981)
  • R.J. Chorley et al.

    Physical geography: a systems approach

    (1971)
  • S.V. Ciriacy-Wantrup

    Resource conservation economics and policies

    (1952)
  • R. Costanza

    Visions, values, valuation, and the need for an ecological economics

    Bioscience

    (2001)
  • R. Costanza et al.

    The value of the world's ecosystem services and natural capital

    Bioscience

    (1997)
  • M. Doppelt et al.

    Entering the watershed: a new approach to save America's river ecosystems

    (1993)
  • L.A. Duram et al.

    An assessment of public participation in federally funded watershed planning initiatives

    Society and Natural Resources

    (1999)
  • V. Esat et al.

    Water resources system optimization using genetic algorithms

  • J.D. Esseks et al.

    Land user attitudes toward implementation of conservation compliance farm plans

    Journal of Soil and Water Conservation

    (1991)
  • J.D. Esseks et al.

    Midwestern farmers' perceptions of monitoring for conservation compliance

    Journal of Soil and Water Conservation

    (1993)
  • Farm Services Agency, U.S. Department of Agriculture, 2004. Conservation Reserve Program Reports. Accessed 31 August...
  • J. Ferber

    Multi-agent systems: an introduction to distributed artificial intelligence

    (1999)
  • W. Firey

    Man, mind, and land: a theory of resource use

    (1960)
  • E.T. Freyfogle

    Bounded people, boundless lands: envisioning a new land ethic

    (1998)
  • D.E. Goldberg

    Genetic algorithms in search, optimization, and machine learning

    (1989)
  • G. Hardin

    The tragedy of the commons

    Science

    (1968)
  • A.P. Hellman et al.

    Optimal design of storm water inlets for highway drainage

  • A.C. Hilton et al.

    Constraint handling for genetic algorithms in optimal remediation design

    Journal of Water Resources Planning and Management, ASCE

    (2000)
  • J.H. Holland

    Adaptation in natural and artificial systems

    (1975)
  • Illinois Department of Agriculture

    State erosion and sediment control guidelines: as adopted 1980

    (1980)
  • S.E. Kraft et al.

    Soil degradation and land-use changes: agro-ecological data acquired through representative farm and linear programming analysis

    Journal of Soil and Water Conservation

    (1984)
  • R. Lal et al.

    The potential of U.S. cropland to sequester carbon and mitigate the greenhouse effect

    (1998)
  • Cited by (0)

    View full text