Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 Applications of the NRI Data to Inventory, Monitor, and Appraise Offsite Erosion Damage Lee A. Christensen There are two general impacts of soil erosion from agricultural land. Onsite effects, those occurring at the field or farm level, are primarily reflected in soil productivity changes associated with erosion. Offsite impacts occur primarily when soil and chemicals are carried from fields and farms in runoff, causing water pollution and deposition problems downstream and groundwater infiltration. This paper addresses the applicability and use of the U.S. Department of Agriculture (USDA) National Resources Inventory (NRI) to investigate such offsite damages. Historically, public attention and funds have con- centrated on reducing the adverse onsite impacts of soil erosion. Increased concerns about water quality degrada- tion and the associated cleanup costs, combined with passage of clean water legislation during the 1970s, focused attention on offsite impacts. Negative offsite impacts have broadened public concern about soil erosion to include more than soil productivity issues. Soil leaving a field due to water erosion represents costs not only to the farmer and consumer in terms of lost profits and high food prices, but also to those downstream, for associated cleanup costs. Solving offsite problems has been complicated by the lack of information and capabilities to explain the diffuse and complex nature of the physical relationships. Recent studies in this area have vastly improved the ability to explain and model the physical processes (Bailey and Swank, 1983). This work is being linked with economic data to develop more complete assessments of the offsite damages and the alternatives for their control and reduction. The NRI conducted by the Soil Conservation Service (SCS) in 1977 and 1982 is a data source that might help 237
238 assess offsite damages associated with soil erosion (USDA, 1984). This paper considers such a use of the NRI data. It begins with a brief discussion of terms and the nature of offsite pollution, followed by an examination of some applications of the NRI data. Possible uses of the data in conjunction with available water quality models to address water quality questions are explored. The final section focuses on potential uses of the NRI and presents some suggestions for future inventories. THE NATURE OF OFFSITE DAMAGES Water bodies receive pollution loads from point sources, such as municipalities and industrial plants, and nonpoint sources, including agriculture. Nonpoint source pollution originates from ill-defined and diffused sources r such as urban areas, cultivated fields, forests, and pastures. Most agricultural nonpoint source pollution is caused by sediment and sediment-transported chemicals (Bailey and Waddell, 1979). It does not include runoff from urban areas, mining and construction activities, highways, logging activities, or streambank erosion. The offsite effects of agricultural nonpoint source pollution are diverse and complex. Each type of pollutant has unique characteristics, both with respect to the mode of transport through the water course, and the fate of the pollutant as it moves from field to stream to lake to river or reservoir. Heavy sediment loads can fill reser- voirs and cause channel siltation, which raises the costs of water treatment and channel dredging. Excess levels of nitrogen or phosphorus in streams, lakes, or estuaries can cause eutrophication. Sediment and chemicals can have adverse effects on fish and wildlife, greatly reducing the economic and recreational value of streams and lakes. It is important to understand the distinction between problems with surface water quality and those related to groundwater or contamination. Agriculturally related groundwater questions arise primarily with regard to the leaching of nitrates and to soluble persistent pesticides. Solutions to surface water quality problems are not necessarily answers to groundwater problems.
239 APPLICABILITY OF NRI DATA TO WATER QUALITY ISSUES Design Constraints The NRI is a tremendously rich source of information, but with serious limitations for addressing water quality questions. It is designed to systematically develop information on the condition of the nation's agricultural land base every 5 years. Its area of primary application is the onsite, or farm, level. The NRI data base can be used to address some of the water quality issues influenced by activities based on agricultural land, primarily sediment loadings. However, it contains no direct data for the analysis of water quality problems attributable to nonagricultural sources. Levels of Detail and Aggregation Analysis of water quality problems due to agricultural sediment is facilitated if data can be aggregated along hydrologic boundaries, such as a watershed or river basin, rather than Major Land Resource Area (MLRA) boundaries. However, since the NRI is considered reliable at the BRA level, information for small watersheds that are fractions of MLRAs must be used with care. For the larger basins that consist of one or more MLRAs, however, the accuracy of the expanded NRI data should be adequate. This assump- tion needs some further testing and examination. Components Affecting Water Quality Assessment of the water quality impacts of agricultural activities requires land-based information and practices that can be linked with hydrologic and toxicological information. The NRI provides data that can be used to estimate sediment movement, and stream loading by inference, but it provides no time-sequenced hydrologi data or direct information on either the amount of fertilizers or pesticides applied to the fields or that transported by sediment or carried in solution. An assessment of the data and factors most applicable for the assessment of offsite impacts is shown in Table 1. Information from these fields can be used to estimate gross sediment movement, but not deposition. Estimating movement beyond the edge of a field, although possible, ,c
240 TABLE 1 Impacts Data File Fields in 1982 NRI Related to Offsite Field Name 6 10 12 20-22 24-27 28-30 32-37 39 Hydrologic unit T factor Degree of erosion Land use and cover Cropping history Conservation practice USLE factors Average annual tons of soil 40-41 Average annual tons of soil loss due to sheet and rill and wind erosion SOURCE: USDA (1984). requires the development and use of sediment delivery ratios, a complex and difficult task. By comparing 1977 and 1982 NRI data for a particular area, it may be feasible to determine some trends in sediment loading rates, particularly in areas where there have been significant changes in land use or cropping or tillage practices. However, although the NRI can be used to estimate partial sediment loads, it provides no information on particle size distribution, which is very important for assessing such offsite impacts as fish reproduction. Soils-5 data contain particule size information which can be combined with the NRI. SOME WATER QUALITY APPLICATIONS OF THE DATA The NRI data base can be used in conjunction with pollution loading models and more complex water quality models. This section examines some generalized loading model considerations, some specific applications of NRI data with other data bases, and some possible uses of the NRI data with existing water quality models.
241 Gross Load Estimation Generalized Procedure The quantity of a given pollutant passing through a system at any time is the interaction of the process of supply and transport. Several generalized procedures can use NRI data to estimate pollutant loads, which can be useful indicators of the water quality impacts of erosion. One approach is to multiply the Universal Soil Loss Equation (USLE) estimates of soil movement at the water shed or subwatershed level by a sediment delivery ratio and a potency factor, which yields an estimate of the pounds/acre/year of a pollutant (such as sediment, nitrogen, or phosphorus) moving into streams. The potency factor measures the amount of pollutant associated with each unit of sediment (Dean, 1983). The challenge is devising the proper sediment delivery ratio and potency factors for this procedure. There are several ways to predict or simulate agricultural pollutant loads. These range from simple sediment loading functions to physical processes requiring simulation of chemical reaction, transformation, and dynamic transport. A pollutant load is defined as a mass of pollutant moving to a receiving water body in a given period of time. If the pollutant is assumed to be linearly correlated with the amount of sediment moving from the watersheds, the pollutant load (or loading function) can be estimated as the product of the amount of sediment delivered to a receiving water body and a potency factor P. which is a factor relating the load of pollutant associated with each unit loading of sediment. P is very complex and difficult to estimate. As an empirical approximation, the potency factor can be envisioned as the product of the average concentration of a pollutant in the surface layer of the soil, the enrichment ratio of the pollutant of interest, and the ratio of the mean particle density of surface soil to the mean particle density of the eroded sediment (Dean, 1983). The enrichment ratio represents the effect of several processes that cause the ratio of the mass of pollutant to sediment to be higher at stream edge rather than at the source, back at the watershed. Estimates of the sediment delivery ratios and potency factors for various pollutants have been developed for specific studies, but there are few generalized sets
242 available to apply in large area studies (Gianessi et al., 1981a). Loading rates are related indirectly to the various tillage practices through the interactions of the management components in the USLE calculations. If some figures on changes in use of conservation tillage are available for a given watershed, changes of likely pollutant loadings can be estimated. The gross loading information from the 1982 NRI provides some information for estimates. However, the data needed for state-of-the-art models to assess offsite effects of agricultural activities are much greater and more complex than those provided by the NRI. Examples of Loading Models Haith and Tubbs (1981) developed three loading models to estimate nutrient and pesticide losses from cropland. The models range from simple loading functions to detailed computer simulation models for the soil environment. They share several common attributes. All have a daily time step and are based on the SCS Curve Number Runoff Equations and the USLE. None requires calibration, and each model was tested. Simple planning models or loading functions can provide straightforward means of estimating nonpoint source pollution. Simple nonpoint source models (loading functions) have several deficiencies. The USLE was not designed to evaluate nonpoint source pollution. It can be used to calculate average annual soil loss, but not loss from single storm events. Although loading functions have been used extensively for this purpose, their ability to provide reasonable estimates of agricultural nonpoint source pollution in large watersheds has not been established. The problem often overlooked in the use of simple loading models is that different categories of potential pollutants are transported in different fashions. For example, dissolved chemicals move with runoff water while most phosphorus and some nitrogen and hydrophobic chemicals are associated with sediment. Loading functions for sediment-associated chemicals should be based on soil loss estimates; dissolved chemicals require runoff-based loading functions.
243 Specific Applications Resources for the Future Resources for the Future (RFF) used the 1977 NRI data base to analyze the relative importance of nonpoint source pollution control options at the national level (Gianessi et al., 1981a,b). This national network model linked point and nonpoint sources of pollution to evaluate agricultural sediment control policies in conjunction with point source controls. The model linked pollution- y~erac~ng aches In each county to a detailed network of rivers, lakes, and bays. It provides general estimates of the impact of sediment and sediment-bound pollutants on water quality in specified bodies of water. However, it does not evaluate the transport and impact of soluble pollutants. RFF is updating its national model to incorporate the 1982 NRI data and refining it to include a sediment transport component. This will help in revising estimates for gross sediment and associated pollutants reaching streams. Economic Research Service The Economic Research Service of the USDA is using NRI data to estimate offsite benefits associated with soil conservation (Ribaudo, 1984). ma_ -- ' ~ fine NHl Information is being used in conjunction with other data sets to relate the levels of pollution associated with erosion parameters to specific impaired water uses. Water quality data from the National Stream Quality Accounting Network (NASQUAN) was used to estimate the ambient water quality levels, total suspended solids, total phosphorus, and total nitrate for each of the 99 watershed units defined by the Water Resource Council as aggregated subareas (ASA). These levels were then compared with standards reflecting impacts on water use. Pollutant loads in the various watercourses from all sources were estimated using the National Water Discharge Inventory developed by RFF. Sediment discharges from cropland, pastureland, rangeland, and forestland were based on erosion estimates provided by the 1977 NRI. Estimates of streambank, gully, construction site, and other erosions come from other sources. A sediment delivery ratio was calculated for each ASA by RFF and
244 used to estimate the amount of eroded soil reaching waterways. The amounts of total suspended solids, total phosphorus, and total nitrogen in the discharge were estimated using coefficients based on the characteristics of the major soil groups contained in each ASA. The final step was to compare the pollutant loadings from agriculture with the uses made of the streams in the affected areas and to identify regions where agricultural erosion has significant impact on offsite water uses. Thirty-eight ASAs were identified as having a water quality problem due to agriculture, but only 15 were intensive use regions. The estimates are being updated by incorporating the 1982 NRI and other information into the RFF model. Linkages with Water Quality Models Several models have been developed to evaluate the impacts of alternative management strategies on water quality and the influence of specific management practices on the levels of particular pollutants. Some use NRI data, but since most models have gone beyond the gross loading stage, they use the NRI data as one input among several, not as the primary data set. A number of these models may be able to take the erosion estimates as input, but others need only the USLE coefficients in the NRI. Selected examples of models that can use some of the NRI data are described in this section. A pesticide root zone model (PRZM) being developed by the U.S. Environmental Protection Agency (EPA) at the Athens Environmental Research Laboratory simulates the vertical movement of pesticides in the unsaturated soil within and below the plant root zones (Carsel et al., 1984). The model consists of hydrology and chemical . . . . . . transport components that simulate runoff, erosion, plant uptake, leaching, decay, or surface washoff and vo~at~zation of a pesticide. The hydrology component for calculating runoff and erosion is based on the SCS Curve Number technique and the USLE. PRZM can be used to estimate frequency distributions of the mass of pesticide leaching from the plant root zone to investigate the risks of pesticide use, par- ticularly pertaining to groundwater pollution. The model uses the Modified Universal Soil Loss Equation (Williams and Berndt, 1977). This modification replaces the R (rainfall erosivity) term with an energy term and allows
245 estimation of the volume of event runoff and peak storm runoff. The model requires all the other USLE factors. A comprehensive basin-scale simulation model developed to predict water quality arising from both point source and agricultural nonpoint source pollution is the Hydro- logical Simulation Program in For tran (HSPF) (Donigian et al., 1983, 1984). The goal of this model is to go beyond the prediction of the quantity and quality of runoff from agricultural lands and to predict instream water quality effects of the best management practices. However, runoff models by themselves are not sufficient to do this, since instream transport and transformations are usually not represented. Using a model of this type requires far more data than the NRI provides. Yet, applying it allows simulation of the movement of pollutants and assessment of the likely impacts of changes in management practices on water quality through time. Models like this are indicative of the state of the art in water quality modeling. Linkage with the Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS) model developed by USDA is another possible application of the NRI data (Knisel, 1980; Knisel et al., 1983). A major use of CREAMS is evaluation of alternative management practices for control or minimization of runoff of sediment and chemicals. It has three components (hydrology, erosion/ sedimentation, and chemistry) and describes the movement of runoff, sediment, and plant nutrients and pesticides from field-sized areas. It is a continuous simulation model that operates efficiently to allow consideration of long-term records (20 years). The model can be used to evaluate the impact of management practices on the yield of sediment and chemical pollutants from field-sized areas at specific sites. It is also being expanded to address questions at the watershed level. The model's erosion/sedimentation component could use some of the NRI data. It considers the basic processes of soil detachment, transport, and deposition. Detachment is described by a modification of the USLE for a single storm event. The calculation of a rate of interrill detachment and the rate of detachment by rill erosion uses the USLE factors.
246 Uses of the MRI Data for Offsite Analysis: A Summary As noted, some inherent constraints in the NRI data base influence its usefulness and applicability for inventorying, monitoring, and appraising offsite erosion damages and adverse impacts on water uses. It provides information on land use and gross erosion estimates, which are useful for loading estimates and models, but it does not have complete data for water quality analysis. Nevertheless, there are several opportunities for its use. The NRI is a source of coefficients for estimating sediment load from agricultural nonpoint sources, which can be combined with other sources to estimate combined point and nonpoint sediment stream-pollution loads. The applications by RFF (Gianessi et al., 1981a,b) and the water quality assessments by the Economic Research Service (Ribaudo, 1984) are examples. As a screening tool, the NRI data can be used as part of a system to identify areas where sediment and associated pollutant loads in streams really impair stream usage. Comparisons between subbasin characteristics are useful for isolating exceptional situations. The USLE coefficients in the NRI can be used directly in regional water quality modeling efforts. For example, PRIM needs USLE coefficients to operate, and CREAMS uses USLE factors as part of the input data. Data from the 1977 and 1982 NRIs has limited use for trend assessment. For particular areas, changes in land use and conservation and tillage practices can be used to estimate changes in gross pollution loads from agricul- ture, provided sediment delivery ratios and pollution loading coefficients are available. Using the same system, projected changes such as significant shifts to conservation tillage can be analyzed for impact on pollution loads, provided the distinctions are maintained. For example, studies have found a reduction in sediment- transported nitrogen and phosphorus with fluted courter conservation tillage, and control of both sediment- transported and solution nitrogen and phosphorus with in-row chisel tillage (Langdale and Leonard, 1983). The NRI can also be part of the data base needed for water quality assessments in specific regions, as done by Ribaudo (1984). It can be coupled with the vast amounts of information already developed in areas of intensive study, such as the Chesapeake Bay (EPA, 1982; Northern Virginia Planning District Commission, 1983). Loading rates for the entire Bay have been estimated. The NRI -
247 can be used to estimate loadings from specific basins or subbasins to aid in planning and analysis. CONSIDERATIONS FOR FUTURE NRIs Before advocating changes to remedy the NRI's inability to address water quality questions, it is important to remember that the main objective of inventory is to assess the nation's soil and water resources. As such, it was not designed or specifically charged with a responsibility to assess offsite damages. Yet, those charged with designing the next NRI might consider the extent to which it should address offsite questions. Although few data bases are intended to be universal in scope and applicability, with some minor modifications to the NRI, a more complete set of data for water quality analysis could be assembled. Some suggestions for consideration are offered in this section. A critical missing link in the NRI as far as offside damage assessment goes is its lack of linkages with hydrologic, toxicological, or meteorological data bases. To get good estimates of sediment load, nitrogen runoff concentration, or phosphorus runoff concentration, the timing of tillage practices and fertilizer applications needs to be tied to meteorological data, particularly rainfall. Time-series data on meterological and hydrologic data are needed for water quality simulation models such as the nonpoint source model (Donigian and Crawford, 1976b, 1977), the agricultural runoff model (Donigian and Crawford, 1976a), HSPF, and CREAMS. Water quality problems are time-based, but the NRI provides no time-variance loading information. It is impossible to predict water quality accurately from an average annual estimate provided by the USLE. It has validity only for cases where the retention time is 1 year or greater. Linkages are needed on a storm-by-storm basis to rainfall, runoff, soil loss, pollutant concen- tration, utilization, infiltration, percolation, and movement of soluble pollutants to groundwater. The feasibility of collecting such information in the future needs to be evaluated. Better links between movement on the field and deposition in the field and streams are also needed. The USLE generates sediment load information, but the sediment delivery ratios and potency factors needed to
248 better assess the amount actually moving to streams are generally inadequate. Information on management practices and application rates of pesticides and fertilizer would help assess runoff problems as well as the enrichment of sediment. Very little is known about the composition of runoff, either in terms of quantity or quality. Data on fertilizer and pesticide management practices and the properties of the chemicals are particularly important in assessments of whether potential pollutants are moving overland to a stream in solution or bound to soil particles, or whether they are moving down through the soil profile in solution. Research has indicated greater efficiencies in the use of nitrogen in corn and soybeans with no-till than with conventional till (Hoyt et al., 1983). In the Southeast, changes in tillage practices have resulted in greater concentrations of phosphorus and nitrogen, but in reductions of the transported mass (Langdale and Leonard, 1983). The NRI does not now address the question of ephemeral gully erosion, in which there is considerable interest (see Foster, this volume). Whether it should or could do so needs to be assessed, first to explain the total erosion process better, and second to assess impacts on water quality that fall between rill and gully erosion. Given the rapid adoption of conservation tillage and no-till, future NRIs need to consider more explicit measurement of these practices (as well as related practices such as fertilizer, insecticide, and herbicide use) and to ensure that they are properly accounted for in the gross erosion calculations. Explicit linkages between the type of tillage and the impact on a specific pollutant and its pathway to stream or groundwater are needed. There are several excluded sources of erosion needing consideration in efforts to a~ the imp of "rna;^n ~ ...= ~ ~ · ~ on water quality. Ephemeral gully erosion, which occurs between rill and gully erosion, is one such source. There is considerable interest in better explaining ephemeral gully erosion as part of the total erosion process; its inclusion in the NRI needs to be assessed. Other erosion not measured in the NRI includes streambank erosion, erosion trom federal lands, and erosion from construction sites. Lastly, assessment of the offsite water quality impacts of soil erosion needs to involve major federal and state agencies with capabilities and responsibilities
249 in the area. Coordination with departments, such as the EPA, is needed if water quality becomes a major emphasis of NRI efforts. SUMMARY There are limits on the usefulness of NRI data for assessing the offsite effects of soil erosion. Selected data can be used to estimate pollutant loads, primarily sediment, and thus help identify and inventory potential sources of offsite damages. Data from the 1977 and 1982 NRIs can provide points for assessing trends in changes in the resource use, thus assisting in monitoring factors that influence offsite erosion damages. Data and coefficients from the NRI can be linked with other data bases and water quality models to appraise offsite impacts. It must be remembered that the NRI was not designed with water quality as its primary focus. Thus, it must be viewed as an important source of information, but useful for addressing water quality questions primarily in conjunction with other models and data. The next NRI could be modified to be more directly applicable to water quality and offsite damage. Issues to consider include improved linkages with hydrologic data bases, linkages with time-based pollutant loads, field-to-stream linkages, linkages to management and tillage practices, the role of ephemeral gully erosion and streambank erosion, and possible coordination with other water quality agencies. . . . . . . . REFERENCES Bailey, G. W., and R. R. Swank, Jr. 1983. Modeling agricultural nonpoint source pollution: A research perspective. Pp. 29-41 in Proc. of the National Conference on Agricultural Management and Water Quality. Ames: Iowa State University Press. Bailey, G. W., and T. E. Waddell. 1979. Best management practices for agriculture and silviculture: An integrated overview. Pp. 33-36 in Best Management Practices for Agriculture and Silviculture, Proc. of the 1978 Cornell Agricultural Waste Management Conference. Ann Arbor, Mich.: Ann Arbor Science Publishers.
250 Carsel, R. F., C. N. Smith, L. A. Mulkey, D. Dean, and P. Jowise. 1984. Users Manual for the Pesticide Root Zone Model (PRZM)--Release 1. (Draft.) Athens, Gal: U.S. Environmental Protection Agency. Dean, J. D. 1983. Potency factors and loading functions for predicting agricultural nonpoint source pollution. Pp. 155-177 in Proc. of the National Conference on Agricultural Management and Water Quality. Ames: Iowa State University Press. Donigian, A. S., and N. H. Crawford. 1976a. Modeling Pesticides and Nutrients on Agricultural Lands. EPA 600/2-76-043. Athens, Gal: U.S. Environmental Protection Agency. Donigian' A. S., and N. H. Crawford. 1976b. Modeling Nonpoint Pollution from the Land Surface. EPA 600/3-76-083. Athens, Gal: U.S. Environmental Protection Agency. Donigian, A. S., and N. H. Crawford. 1977. Simulation of Nutrient Loadings in Surface Runoff with the NPS Model. EPA 600/3-77-065. Athens, Gal: U.S. Environmental Protection Agency. Donigian, A. S., Jr., J. C. Imhoff, and B. Bicknell. 1983. Predicting water quality resulting from agricultural nonpoint source pollution via simulation--HSPF. Pp. 200-249 in Proc. of the National Conference on Agricultural Management and Water Quality. Ames: Iowa State University Press. Donigian, A. S., Jr., J. C. Imhoff, B. R. Bicknell, and J. L. Kittle, Jr. 1984. Application Guide for Hydrological Simulation Program in For tran (HSPF). EPA 600/3-84-065. Athens, Gal: U.S. Environmental Protection Agency. EPA (Environmental Protection Agency). 1982. Chesapeake Bay Program Technical Studies: A Synthesis. Washington, D.C.: U.S. Environmental Protection Agency. Gianessi, L. P., H. M. Peskin, and G. K. Young. 1981a. Analysis of national water pollution control policies: 2. Agricultural sediment control. Water Resources Res. 17:803-821. Gianessi, L. P., H. M. Peskin, and G. K. Young. 1981b. Analysis of national water pollution control policies: 1. A national network model. Water Resources Res. 17:796-801. Haith, D. A., and L. J. Tubbs. 1981. Operational Methods for Analysis of Agricultural Nonpoint Source Pollution. Ithaca: Cornell Agricultural Experiment Station, State University of New York at Cornell.
251 Hoyt, G. D., R. L. Todd, R. A. Leonard, L. Asmussen, and E. D. Threadgill. 1983. Effect of tillage on nutrient cycling in Southeastern Coastal Plain agroecosystems. In Nutrient Cycling in Agricultural Ecosystems. Special Publication No 23. Athens: University of Georgia College of Agriculture Experiment Stations. Knisel, W. G., ed. 1980. CREAMS: A Field Scale Model for Chemicals, Runoff, and Erosion from Agricultural Management Systems. Conservation Research Report No. 26. Washington, D.C.: U.S. Department of Agriculture. Knisel, W. G., G. R. Foster, and R. A. Leonard. 1983. CREAMS: A system for evaluating best management practices. In Nutrient Cycling in Agricultural Ecosystems. Special Publication No. 23. Athens: University of Georgia College of Agriculture Experiment Stations. Langdale, G. W., and R. A. Leonard. 1983. Nutrient and sediment losses associated with conventional and reduced-tillage agricultural practices. In Nutrient Cycling in Agricultural Ecosystems. Special Publication No. 23. Athens: University of Georgia College of Agriculture Experiment Stations. Northern Virginia Planning District Commission. 1983. Chesapeake Bay Basin Model--A Final Report. (Draft.) Annandale: Northern Virginia Planning District Commlsslon. Ribaudo, M. 1984. Water resource problems. Economic Research Service (ERS) draft materials. Washington, D.C. November 29. Photocopy. USDA (U.S. Department of Agriculture). 1984. National Resources Inventory--A Guide for Users of 1982 NRI Data Files. (Draft.) Washington, D.C.: Soil Conservation Service. Williams, J. R., and H. D. Berndt. 1977. Sediment yield reduction based on watershed hydrology. Trans. ASAE 20:1100-1104. Discussion Ronald B. Outen The National Resources Inventory (NRI), as Christensen's paper points out, is just a piece of the puzzle. It does provide some useful information, but alone it is not yet sufficient to make the critical connection between land use practices and water quality
252 When water quality problems associated with nonpoint source pollution were first raised years ago, the discussion almost always turned to agriculture, and almost immediately to soil erosion. Yet as Christensen notes, sediment per se is not necessarily the biggest problem in all or perhaps even most of the watersheds, strictly in terms of water quality. Very often the issues most discussed are nutrients from fertilizer and animal waste runoff and, also, pesticides. Most people make a brief reference to groundwater and then move on to talk about surface water runoff, which is understandable because over the last few years the Clean Water Act has dealt almost exclusively with surface water. There may be a bit of a contradiction in national policy goals in terms of protecting groundwater and surface water. On a given piece of land with an excessive amount of nitrogen, phosphorus, other fertilizer materials, or pesticides, measures to prevent runoff might exacerbate the groundwater problem. More coordination is needed between these programs. Ultimately, groundwater, like nonpoint source pollution of surface water, must be dealt with on an areawide basis in terms of aquifer recharge areas, at least for those pollutants that tend to be dispersed across the landscape. Moreover, an integrated hydrologic regime that includes both groundwater and surface water must be considered, as well as an integrated land/water network. Furthermore, management practices must be broadly defined. There soon will be a federal law calling on states to develop implementation programs to apply best management practices in large areas of the country for purposes of water quality. This new nonpoint source management program, which will be added to the Clean Water Act, has received strong support in both houses of Congress.
