یکپارچه سازی مدل سازی اقتصادی، زیست محیطی و GIS به هدف قرار دادن هزینه موثر اراضی بازنشستگی در حوزه های چندگانه

An integrated framework of economic, environmental and GIS modeling is developed to study cost-effective retirement of cropland within and across multiple watersheds to achieve environmental goals. This framework is applied to 12 contiguous agricultural watersheds in the Illinois Conservation Reserve Enhancement Program region of the United States. A key goal of this program is to reduce sediment loadings in the Illinois River by 20% by retiring land from crop production. The characteristics of land parcels to be targeted for retirement within each watershed and the criteria for cost-effective allocation of abatement responsibility across watersheds are analyzed. Our analysis suggests that program costs are minimized when the abatement standard is set for the region rather than uniformly for each watershed. For both policy scenarios, the land parcels targeted for retirement should be those that are highly sloping and adjacent to a water body.

Increasing emphasis on protecting water quality from nonpoint pollution sources has led to a shift in the policy objectives of conservation programs, such as, the Conservation Reserve Program (CRP), from on-site erosion control towards controlling off-site sediment loadings. More recently, with the development of the Conservation Reserve Enhancement Program (CREP) the focus of these programs has shifted towards achieving explicitly defined goals for water quality in locally identified environmentally sensitive river basins. The goals of the CREP in Illinois include a 20% reduction in sediment loadings, a 10% reduction in nitrogen and phosphorus loadings and an increase in aquatic and wildlife populations in the Illinois River (USDA, 1998). The program seeks to achieve these goals by retiring about 94 thousand hectares of cropland in the Illinois River Basin for at least 15 years and possibly for 35 years or permanently through offering rental payments to landowners based on soil productivity. Eligible land for enrollment is defined as cropland within the basin's 100-year floodplains, highly erodible cropland adjacent to riparian areas, and cropland where landowners are willing to create wetlands. A secondary program restriction requires that 85% of the land enrolled be within the basin's environmentally sensitive riparian areas defined above as 100-year floodplains. The amount of eligible land is spread out over 100 sub-watersheds1 and is more than six times larger than the program's acreage enrollment goal of 94 thousand hectares. This raises two program implementation issues on which the program description itself is silent. First, should land enrollment be targeted to achieve the environmental goals of the program at the sub-watershed level or in the aggregate for the river basin as a whole? Second, what types of land parcels should be selected for enrollment in each sub-watershed to achieve the environmental goals cost-effectively? The literature in environmental economics suggests that when polluters are heterogeneous, an aggregate abatement goal can be achieved cost-effectively if each polluter undertakes pollution abatement such that the cost of the last unit of abatement is equalized throughout the river basin (Baumol and Oats, 1971). This concept implies that more abatement will occur in sub-watersheds with low abatement costs, and less abatement will occur in high abatement cost sub-watersheds; implying that a non-uniform performance standard will be more cost effective than a uniform standard. However, a social planner may prefer a uniform performance standard where the same abatement goal must be achieved in every sub-watershed, for equity reasons (Schleich et al., 1996), or to reduce transaction costs. The first objective of this paper is to estimate and compare differences in costs, hectares retired and land retirement patterns between the two performance standards on sediment loadings. The second objective is to develop criteria for the land parcels that should be selected for enrollment to meet the sediment abatement goals of the program cost-effectively. Land parcels even within a watershed are heterogeneous in their characteristics and differ in their opportunity costs of retirement (that is the forgone profits from crop production) and in the extent to which their retirement would reduce off-site sediment loadings in the water body. Babcock et al., 1996 and Babcock et al., 1997 suggest that land with the highest abatement benefits to opportunity cost ratio should be selected to achieve abatement goals efficiently. Determining the contribution of each land parcel to off-site sediment loading requires estimation of its sediment-trapping coefficient, and those of the downslope parcels. Each parcel's coefficient depends on that parcel's site-specific characteristics (slope, soil characteristics, and distance from the water body) and land use decision (crops, trees, pasture or grass) as well as highly complex interdependencies between land use and sediment trapping efficiencies of upslope and downslope land parcels. Thus, the sediment deposition coefficient of a land parcel needs to be determined endogenously with the land use decisions upslope and downslope parcels. To address these issues we develop an integrated economic, environmental and GIS modeling framework that includes an endogenously determined sediment transport process and incorporates the micro-economic decision by a farmer to retire land or continue crop production and the environmental impacts of such decisions. This framework extends the model development in Khanna et al. (2003) and applies it to compare cost, acreage, location and parcel differences between a non-uniform and uniform standard across twelve sub-watersheds in the Illinois River Basin. Our research builds upon previous studies of environmental programs on land use and farming practices in several ways. A few studies have compared the efficiency and distributional implications of setting environmental standards at a field/farm/watershed level or analyzed crop management strategies at a watershed level to reduce pollution loadings. However, these studies assumed that the relationship between on-site pollution generation and off-site pollution loadings is dependent on exogenously given site-specific factors (Kramer et al., 1983, Ribaudo, 1986, Ribaudo, 1989, Prato and Wu, 1996 and Carpentier et al., 1998). Braden et al. (1989) assumed a more complex pollution transport process in which the portion of pollution trapped depends not only on the site-specific characteristics but also on the management practices of downslope parcels. However, they did not incorporate the effect of the volume of runoff flowing from upslope parcels on the trapping capacity of downslope land parcels and therefore ignored the impact of land use changes upslope on sediment transport coefficients. Lintner and Weersink (1999) was one of the few studies incorporating economic, environmental and spatial analyses to examine policies to reduce nitrogen and phosphorus runoff. They incorporated the interdependence between sediment deposition coefficients and fertilizer-use decisions of all parcels in a flow path, but they assumed that all parcels were identical. This assumption allowed the sediment deposition ratio of a parcel to be dependent only on its own characteristics and reduced the true complexity of the problem. The approach we present in this paper offers a more realistic simulation of the sediment generation and transport process by simultaneously determining heterogeneous land use decisions and off-site sediment abatement for all parcels in a flow path. In Section 2, we present the conceptual framework underlying the empirical model described in Section 3. In 4 and 5, we describe the data used for 12 sub-watersheds in the study area targeted under Illinois' CREP and the results obtained under both non-uniform and uniform performance standards. In Section 6, we discuss the policy implications and conclusions from this research.

In the U.S., policies to reduce non-point source pollution have generally taken the form of incentives rather than taxation and other ‘polluter pays’ programs. The USDA's Conservation Reserve Enhancement Program (CREP), for example, seeks to achieve state-specified numeric environmental goals within an environmentally sensitive region by providing financial incentives to farmers to retire cropland. Development of appropriate criteria for selecting the land to be enrolled in the program to achieve the program's goals is a research problem of theoretical and practical significance. In this paper, we develop an integrated framework that combines economic, environmental and GIS modeling to examine cost-effective targeting of land for retirement to achieve environmental goals in a region with multiple watersheds. This framework is used to determine the cost-effective non-uniform sediment performance standards for 12 contiguous agricultural watersheds in the Illinois CREP region and to identify land parcels that should be enrolled across watersheds to achieve a sediment abatement goal of 20%. The framework is also used to compare the costs of abatement and pattern of land retirement under a uniform standard and the same sediment abatement goal of 20% for each of the 12 watersheds. We also examine the criteria that should be used to select land parcels for retirement in each watershed under both types of performance standards. Consistent with the theoretical framework, a non-uniform standard results in a lower cost and smaller amount of cropland retired. The cropland selected for retirement tends to be adjacent to streams, sloping, erodible, less productive, and less profitable than the cropland not selected. The runoff channels of the selected parcels also tend to have greater upland sediment inflow than the runoff channels where no parcels were selected. Under a uniform standard policy, selected land parcels differ from those under a non-uniform standard policy in that the number of parcels retired increases in watersheds with higher marginal costs of sediment abatement and it decreases in watersheds with lower marginal costs. Characteristics of land parcels selected for retirement mirror those under the non-uniform standard. Several important policy implications can be stated given our theoretical framework and empirical analysis. First, the results confirm the superiority of a non-uniform standard in minimizing costs and land removed from production for any given sediment abatement goal. With only slight modifications, however, costs and land retirement patterns for a uniform standard can be estimated and compared to the non-uniform standard, thus giving policy makers the opportunity to assess the costs of policies such as uniformly applied water quality standards on streams. Second, our use of a generic 274-m buffer also produced some interesting conclusions regarding eligibility requirements of CREP in Illinois. The least-cost solution shows that highly sloping, less productive cropland adjacent to streams with significant sediment inflow from upland parcels should be targeted for retirement. The Illinois CREP program for the most part targets floodplain cropland, which is flatter, more productive and costly. Our results instead suggest that a buffer program which targets either sloping land or any cropland within a narrow buffer would likely decrease costs of sediment abatement and acreage retired compared to a floodplain buffer program. This study should be viewed as the first step in the development and use of spatial economic and environmental models for improving the design and implementation of public conservation programs. Our integrated framework could be extended in several ways. First, the binary choice of cropping or retirement could be increased to include continuous cropping choices with conservation practices. Second, our modeling framework could be extended to examine pollutant trading. The least-cost non-uniform allocation of land retirement could be achieved by a program that allows trading of pollution permits across watersheds. Our results show the cost-savings that could be achieved by allowing trading instead of requiring each watershed to meet the same environmental standard. Third, the hydrologic component could be expanded to include tile drainage as well as stream bank and in-stream erosion. There is considerable debate among researchers regarding the effectiveness of buffers in watersheds that have been extensively tiled and watersheds where sediment from overland sediment flow is considerably less compared to stream bank and in-stream erosion. Fourth, further research could be conducted to study the implications of characterizing or simplifying the endogenous pollution transport process in model estimation. The investigation will shed light on the errors involved in using exogenously fixed pollution delivery coefficients and developing simpler models that could be used for policy analysis on a larger scale. Fifth, the institutional barriers to targeting conservation programs and the transactions costs of targeting could be incorporated to improve policy analysis. Finally, more research is needed to examine how the criteria for land enrollment would change if the other objectives of CREP are included in our analysis.