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.
Improving Federal Resource Assessment Efforts Major surveys on natural resource use patterns and conservation needs have been conducted by the USDA nearly every five years for several decades. The NRI series has been expanded to include new data and refined to include more precise definitions and sampling proce- dures. In addition to improving the statistical validity of the sampling, a major recent methodological advance was use of the Universal Soil Loss Equation (USLE) and Wind Erosion Equation (WEE) to collect information about erosion conditions and conservation practices on U.S. nonfederal lands in a consistent, reliable manner. The 1982 NRI represents a significant improvement over earlier resource inventories. Field procedures; sampling methods; and den- sity, quality control, data accuracy, presentation, and documentation were of a higher quality in 1982 than in 1977. The SCS has consistently upgraded its resource inventories over the past 25 years. To some extent, successive refinements in areas including inventory design, procedures, and definitions have made direct comparisons among inventories, and subsequently assessment of resource trends over time, difficult. In general, however, the committee believes that the SCS has struck a favorable balance between the often contradictory goals of providing consistent inventory data over time and improving the procedures and reliability of successive inventories. There will always be a need for resource inventories. The U.S. Con- gress has already expressed interest in a program of periodic resource inventory and appraisal in the Soil and Water Resources Conservation 19
20 SOIL CONSERVATION Act of 1977, which calls for another inventory in 1987. In this chapter the committee identifies ways to improve future inventories. The next generation of NRIs will be influenced by several develop- ments. As scientific understanding and public concern about resources and environmental quality evolve, it is likely that the types of resource data collected in successive inventories will change. For example, the committee believes that information directly relevant to water pollu- tion problems derived from nonpoint sources will become an impor- tant goal of future resource assessments. The scale of these future inventories- the number of sample points and the number of observations made at each point- will be condi- tioned by cost considerations. In the past few years, USDA officials have expressed a desire to lower the cost of the NRIs by reducing their scale and by conducting them every 10 instead of every 5 years, supple- mented perhaps with smaller special surveys. Many experts believe that future advances in sensory technology can replace at least some of the costly on-the-ground data collection characteristic of past invento- ries. In addition, supplemental funds might be sought from states, particularly if a state is interested and willing to support sampling at smaller units, at the county level, for example. Government agencies should begin to take these developments into account now if resource inventories are to be even more reliable, useful, affordable, and timely in the future. As the principal agency involved in the planning and execution of NRIs, the SCS will have to upgrade the status, financial support, and procedures of its inventory and monitor- ing functions. A more formal degree of interagency coordination and planning will be needed at the earliest stage of inventory development to ensure that the next generation of resource inventories serves even broader purposes than the 1982 NRI. Federal agencies involved in resource management will need to experiment with such new survey techniques as remote sensing for the NRI and to refine established procedures such as use of the USLE. Completed inventories will have to be compiled and distributed in forms that reflect the most up-to-date concepts of land classification and at the same time can be used interac- tively with other data bases. The committee stresses that all of these steps will require a sustained and systematic effort over many years. A comprehensive review of federal resource and environmental inventories is beyond the scope of this report. However, the committee has identified the need to improve the overall planning and design of future NRIs in four areas: (1) compi- lation and dissemination of data, (2) use of remote sensing and related technologies, (3) coverage of federally owned land, and (4) coverage of parameters related to water quality.
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 21 Compilation and Dissemination of NRI Data The committee sees a need to improve several aspects of NRI data compilation and dissemination: the inventory and monitoring function of the SCS, statistical documentation, presentation of published data, and distribution of NRI computer tapes. Upgrading Inventory and Monitoring Functions of the SCS Given the massive scale of every phase of the 1982 NRI, the SCS has performed in a timely and exemplary manner in completing the survey, checking its quality, and releasing inventory results. The committee believes, however, that certain delays in processing NRI data might have been avoided if the inventory and monitoring functions within the agencyat its national headquarters, in particular had been given higher levels of support, permitting larger staff or greater latitude in seeking assistance outside the agency. Given the scale of the NRIs and the need to provide inventory results as soon as possible to thef~eld staffof the SCS and other users, S(:S should periodically increase the size andfinancial support level of its inventory and monitoring staff within the national headquarters. The committee believes that it would be especially useful to increase the number of technical staff with training in statistics, computer sci- ence, computer-aided mapping, and remote sensing to permit more timely completion of quality control checks, inventory documentation, and analyses. The committee recognizes that budgetary constraints and the desire to increase field staff have led to a reduction in staff at the SCS national headquarters in recent years. The committee believes, however, that inventory and monitoring will become central functions of the agency in the future. Therefore, an upgrading of the inventory and monitoring functions should begin immediately to ensure that the many statistical and analytical tasks remaining for the 1982 NRI are completed as soon as possible and to begin preparing for subsequent inventories. Statistical Documentation The experiences of researchers working with NRI data have shown that better basic documentation on the 1982 NRI is needed (Brown, 1983; U.S. Congress, House, 1981~. Documentation is needed for both the 1982 and 1977 NRIs. It is important that SCS take the lead in foster- ing enhanced sophistication in the use of NRI survey data. A diversity of resource issues can be addressed with NRI data. The
22 SOIL CONSERVATION full benefits to society of additional applications of NRI data will come about in conjunction with the increasing range, complexity, and geo- graphic specificity of NRI-based analyses. Such applications, however, will stretch and could exceed the statistical reliability of the underly- ing NRI data unless steps are taken to inform NRI users about the appropriate statistical confidence limits that should be ascribed to vari- ous aggregations of NRI sample points. The committee believes that the agency's desire to provide the high- est quality NRI data might have led to excessive hesitancy in applying data to further analyses that might involve questions of statistical relia- bility. Rather than forego these types of analyses, the SCS should take steps to objectively quantify the degree of statistical reliability associ- ated with them. Often analysts can gain insights by assessing the rela- tive values and distribution of values of certain NRI parameters in a given geographic area or by comparing regions, even if the absolute value of the parameter for some primary sampling units may be vari- able and of limited statistical reliability at a particular level of aggrega- tion. Documented procedures for testing the statistical reliability of most NRI data points and applications are readily available, primarily through calculation of confidence intervals. However, the committee strongly urges SCS to place a high priority on the development of appropriate statistical procedures and other tools needed to carry out more sophisticated tests of significance and reliability. Procedures addressing statistical reliability tests for the 1982 and 1977 NRIs should be developed and disseminated to users as a first order of priority in conjunction with new applications of the survey data. Relia- bility issues that arise from joint applications of the 1982 and 1977 NRIs in time series studies also deserve special technical attention. To address documentation needs in past NRIs and to provide an optimal basis for structuring and reporting results of future NRIs, the committee recommends the following: Documentation of NRIcomputer tapeformats and contents should be accom- panied by a detailed, technical statistical guide. Such a statistical guide should accomplish the following: · Describe the sampling method and the way expansion factors are calculated and used in extrapolating from the 3 percent NRI sample of data points to nationwide coverage of nonfederal lands; · Include instructions for the calculation of confidence intervals for any specific data element at various levels of aggregation; · Offer step-by-step recommendations showing how analysts can
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 23 incorporate statistical reliability checks into specific applications of the raw data. Because NRI data will be applied to a diversity of research studies requiring varying degrees of accuracy, instructions are needed to build specific criteria for confidence intervals into modeling exercises. SCS should develop, refine, apply, and explain such criteria in the course of compiling the special publication of 1977 and 1982 NRI results described above. SCS should designate an individual or office within the agency that can offer assistance to outside users in resolving technical questions related to statistical reliability of NRI applications. Assistance in responding to such inquiries could also be provided by the Iowa State University Statistical Labora- tory at Ames and SCS National Technical Centers. SCS should publish a special report that describes applications of the NRI surveys and earlier inventories to time series analyses of conservation needs and accomplishments and land resource use patterns. Such a report would address the implications for time series studies of the changing defini- tions, statistical design, and data elements in these surveys. Special attention should be directed toward ways to use the 1982 NRI for retro- spective studies of conservation needs and accomplishments. Presentation of Published Data For many NRI users, tabular materials published in reports or other forms by SCS are the only NRI data consulted, used, or needed. Tradi- tionally, published data pertaining to soil erosion have been organized using the Land Capability Class System (LCCS). Using this system, Roman numerals I through VIII indicate progressively greater limita- tions and narrower choices for practical use of the land. For example, class I soils have few limitations that restrict their use; class VIII soils have limitations that preclude their use for commercial plants and restrict use to recreation, wildlife habitat, water supply, or esthetic purposes. The additional letters w, s, c, and e are designated subclasses that indicate, respectively, whether the problem is caused by wetness; shallowness, drought, or stoniness; climate; or erodibility. In focusing on erosion and its potential effects, tables that report erosion rates by land capability class are of more limited value than classifications now available from the NRI based on characteristics spe- cifically related to potential erodibility. Reporting in tables should reflect the appropriate focus and scale. Thus, the LCCS, reflecting a number of influences, may be the appropriate reporting formfor
24 SOIL CONSERVATION example, where data refer to the rate of adoption and effect of conserva- tion practices, especially practices routinely recommended specifically for soils within certain land capability class and subclass units. Me SCS should publish tables of 1982 NRI data that report erosion rates, quantities of eroded soil, effectiveness of erosion control practices, and other measures of erosion control needs and accomplishments according to the ranges of inherent erodibility of land or susceptibility to erosion-induced productivity losses. Conservation needs and the effectiveness of conservation practices can be analyzed and understood most reliably through use of the USLE and its individual factors. Based on runoff plot experiments, the USLE is used to predict the longtime average soil losses in runoff from specific field areas in specified cropping and management systems. (The WEE similarly was developed as a method for estimating the potential for wind erosion in the field.) Reports containing 1982 NRI results with tables based primarily on the LCCS should be supplemented with additional tabular material. Supplemental 1982 NRI reports and subsequent surveys and analyses should include tabulations reporting erosion rates and gross erosion- estimated by the USLE, where appropriate, and an improved WEE by ranges of erosion and by ranges of the inherent potential for erosion measured by the RKLS-product values from the USLE (see Chapter 3 for discussion of RKLS). Providing 1982 NRI data in tables using the LCCS and RKLS-based classification systems will facilitate an orderly transition to the use of appropriate alternative land classification schemes, a task the committee sees as essential and now technically feasible. The SCS should encourage the transition within the conservation and land use research community toward improved presentation of NRI survey results by publishing a special statistical report that includes data from both the 1977 and 1982 NRIs. The SCS report should include discussion of several issues that will be raised in subsequent chapters of this report, including statistical reliability, the effects of definitional and other changes on time series use of NRI data, and the relationship between alternative land classifi- cation systems. The report should also include the new tables described above, in which acreage is partitioned according to ranges of inherent erodibility. The report should discuss and contrast various land classifi- cation schemes, with emphasis on the most appropriate and inappro- priate applications of each. The analysis and experience gained in producing this special report might provide further insights for struc- turing and reporting results of future inventories.
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 25 In developing the content of statistical reports on future NRIs, the agency should seek opportunities to extend NRI data to users involved in water quality research and planning and other aspects of land use analysis outside the basic mission of the SCS. Special aggregations of NRI data by watersheds or water resource regions would be of interest to water quality researchers. To improve such aggregations, additional water resource area geographic codes could be added to future NRIs. In addition to these broad areas, the specific problems of salinity in land and water have been called to the attention of the committee and warrant consideration in future resource inventories. Other agencies might agree to. assist in supporting and executing future NRIs if the USDA can demonstrate a willingness and capability to tailor publications, data, raw data, tapes, and perhaps surveys to the study of some of the analytical problems confronting other agencies. Distribution of NR7 Computer Tapes Efforts by the SCS to disseminate data tapes of the 1982 NRI have been exemplary. The potential value of these inventories to society depends on the widespread use of their basic data in analyzing conser- vation problems and evaluating alternative solutions. Wide dissemina- tion of the NRI data tapes to land grant universities, experiment stations, Agricultural Research Service (ARS) research sites, and pri- vate institutions helps to ensure that the full potential of the NRI is realized. The committee believes that the positive experiences to date from outside use of the NRI confirm that further steps are now warranted to make NRI data tapes available to a wider range of users. Several rela- tively simple steps should be taken to broadly extend the use and dissemination of NRI survey results and coordinate the work of inde- pendent analysts using the NRI series. In NRI data tapes, the SCS should includefor each sample point the soil series identification code needed to link the NRI data file to Soils-5, the cooperative, computerized soil survey file. Soils-5 is an important but limited source of additional data on the physical and chemical properties of individual soil series. Linkage involving soil properties, soil mapping units (where available), and NRI data points can be helpful in carrying out many types of research tasks, including quantification of the relationships among erosion, land use, and crop productivity. If demand warrants, the USDA should consider making available for
26 SOIL CONSERVATION researchers a combined computer tape that includes the following for each point in the 1977 and 1982 NRIs: · Identification codes for the primary sampling unit, including the soil series code; · Pertinent data from the Soils-5 file for each point; · Individual data elements on land use, erosion, and conservation practices in 1982 and 1977 for those points covered by the 1977 NRI; · Individual factors of the WEE. The SCS should have the capability to provide, upon request, these data tapes for any user in else United States and its common geographic divisions. This capability would aid in making cost-effective responses to requests for specific data in defined geographic regions. By developing this capability and responding to requests, the USDA can guarantee that NRI data are widely and properly used by analysts. In particular, SCS could recommend against the release of data of ques- tionable statistical reliability detailed land use data by county, for example and ensure that data users understand the definitions and statistical reliability of specific data. The USDA should implement a means to provide researchers with geo- graphic coordinates for the location of NR] sample points to facilitate linkages with remote sensing and computer-based geographic information systems. Progress in mapping, education, and analysis has been made by linking NRI data in selected states to geographic information systems. lohannsen (1986) has described the potential uses for the NRI in state and local decision making in Missouri, for example. Concerns regard- ing the need to preserve the confidentiality of landowners have been successfully resolved in situations involving the NRI and other sur- veys. These experiences should be drawn upon in developing strate- gies for sharing geographic coordinate information with other researchers who pursue sound research objectives and agree to abide by procedures designed to protect confidentiality. SCS should continue to develop and implement strategies to improve the accessibility and application of the NRl senes. This task could be assigned to an existing group or to a new working group composed of federal agencies involved with resource management and environmental qual- ity. Representatives of state governments, the private sector, and aca- demia should be involved. Use of Sensing Technologies in Future NRIs Improvements in most existing natural resource data bases, includ- ing the NRI, generally would entail additional costs. Current federal
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 2 7 A VALUES (USING MINIMUM LS) , ~ a. ~ ~ _ A VALUES (USING MAXIMUM LS) 1 mile _ T VA LU ES A (MAXIMUM LS)T VALUES The four unages were derived from a combination of aerial photography and digitized soil survey maps of the Little Washita River watershed near Okla- homa City, Oklahoma, to evaluate nonpoint pollution. The A values represent the average annual soil loss per acre predicted by the Universal Soil Loss Equation (A = RKLSCP). In these images the values for L (slope length) and S (slope steepness) vary. T is the soil loss tolerance value, and A - T shows the soil loss beyond tolerance. (The white line on the T-value map indicates the river channel; little erosion is occurring in the floodplain.) White indicates a value of zero; values increase from light to dark shades of grey, the darkest indicating most severe erosion. Resolution is 30 meters. Credit: R. E. Pelletier, National Aeronautics and Space Administration. expenditures for data collection are low. New sensory technologies, including remote sensing, in time might offset budget constraints for some resource inventory applications by replacing costly on-the- ground surveys. The committee is aware that promotion of remote sensing has exceeded the capacity of the new technology to define those features on the ground at the level of specificity defined by particular users. Never-
28 SOIL CONSERVATION (A) EROSION CLASSIFICATION (RKLS) n . NONE ROSIVE MODERATELY EROSIVE HIGHLY EROSIVE ~ - _ . ;~ or 1 mile (B) EROSION CLASSIFICATION (RKLSCP) C] BELOW T AT ALL CP LEVELSRKLS ~ 7 BELOW T FOR LOW CP LEVELSRKLS ~ 7; A G T ~ ABOVE T FOR HIGH CP LEVELS7 ~ RKLS ~ 50; A ~ T · ABOVE T. AT ALL CP LEVELSRKLS > 50; ~ A ~ T Images derived by using remotely sensed data for the cover and management factor (C) with additional digitized maps show alternative erosion classifica- tions based on (A), the physical (RKLS) components of the equation only, or (B), the conventional equation (RKLSCP) classification. Credit: R. E. Pelletier, National Aeronautics and Space Administration.
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 29 theless, improvements in technologies for both sensing and data han- dling continue to hold promise and demand coordination in development, evaluation, and use. The committee's comments are directed toward this promise and to the necessity of taking a national view of resource inventories. No single federal agency will be able to take full advantage of the new developments anticipated from subsequent generations of civilian remote sensing technology. Ideally, however, coordination among the many resource inventory activities of the federal government covering the country as a whole could be enhanced to take advantage of remote sensing, computer-assisted mapping, and related survey methods without a major overlap in effort. Integration of data collection among the federal agencies is difficult. Different information needs and survey procedures and a reluctance to share data collection functions are common barriers to coordination. Yet, because sensing technologies have a variety of applications and hold promise for upgrading current resource assessment activities at a reasonable cost, the committee believes that federal coordination is essential. Because of the potential for public benefit, it might be appro- priate for defense agencies to reevaluate policies that in the past have precluded civilian applications of sensing technologies and other capa- bilities developed for military uses. The NRI, because of its scope and cost, is an ideal way to pursue methods of replacing or supplementing on-the-ground procedures with sensing technologies. The SCS has made some use of remote sensing and has investigated ways of integrating it into the NRIs. How- ever, the agency could benefit by a more concerted and fully funded effort. Innovative steps have already been taken to combine remote sensing, various forms of aerial photography, and NRI data. These combined techniques are likely to be of particular value at the state and local levels and will enhance progress in mapping, conservation program adminis- tration, and other analytic activities. Sensing technologies can also expand the coverage and usefulness of NRI data to include water qual- ity and groundwater use assessments. The committee believes that a fundamental transition in the basic method of collecting much of the NRI-type data may occur within decades. Future inventories will be based on remotely sensed data and coded to computer-based cartographic systems that will be equipped with highly refined analytic capabilities. The resource data system of the future, like contemporary NRIs, will be dependent on periodic field surveys to provide ground verifications. Field surveys will be pivotal in
30 SOIL CONSERVATION verifying and refining the analytic tools used in translating remotely sensed images and high-altitude aerial photographs into geographic information systems and models. Ad hoc committees and special groups have been convened within and among agencies at times to review topics specific to special areas, digital cartography or agency-specific remote sensing research needs, for example. A formal group that provides a continuing assessment of technologies and applications would benefit the design and manace- ment of future resource data systems. The committee recommends thatafederal interagencyworkinggroup, coordi- nated by the Office of Science and Technology Policy or another appropriate agency, be formed to formally evaluate potential applications of existing and future sensing technologies to natural resource inventories. The working group should proactively determine types of new sensing technologies that could be developed to support USDA in carrying out resource management responsibilities. The working group should possess the scientific capability to assess and exploit technological advances occur- ring in remote sensing, computer-assisted mapping, and other survey and analytic methods. The committee recommends that the appropri- ate agencies of the USDA, the Department of the Interior, the Environ- mental Protection Agency, and the National Oceanic and Atmospheric Administration be represented in this working group. In addition, appropriate agencies of the Department of Defense should be repre- sented on an ad hoc basis to provide information about sensing technol- ogies that could be useful for future resource assessments. ~ _ ~ Expansion of NRI Coverage to Federal Lands The U.S. government owns large tracts of land, particularly in the western United States. A majority of the acreage of some states is federally owned. Traditionally, SCS inventories have not included federally owned lands. However, surveys are needed of resource conditions and trends on federal lands, substantial portions of which are in grazing and for- estry uses (Sampson, 1986; Renard, 1986~. The absence of data on erosion, range conditions, and conservation needs for large portions of certain states, particularly in the West, compromises the NRI as a source of information for Public and Private management of land resources. A committee representing the Forest Service, the Bureau of Land Manage- ment within the Department of the Interior, the U. S. Geological Survey, the Department of Defense, the Department of Energy, and other relevant federal
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 31 agencies should be formed to evaluate the quality of current information about resource conditions and trends on federally owned land. The primary purpose of this evaluation would be to determine if a supplemental inventory effort is warranted. The evaluation would identify the resources and parameters that should be considered in the design of a new inventory and the methods and institutional responsi- bilities that will be required. The committee strongly recommends inclusion of information on rangeland conditions and soil erosion esti- mates in any inventory of federally owned lands. This proposed com- mittee should be coordinated with the interagency working group that assesses applications of sensing technology to NRI survey needs. The supplemental inventory should be designed and conducted to be consistent with the 1982 NRI. Parameters collected on federal lands will include some of those collected for all sample points on private lands. Some data, however, will not be appropriate for the majority of federal lands, for example, cropping history and conservation practices. Other data not collected from private lands might be needed to characterize the uses and conservation needs of federal lands, such as whether lands managed for multiple uses are subject to state or federal conserva- tion programs or initiatives. The appropriate level of detail and density of geographic sampling are yet to be determined. The detail achieved in the 1982 NRI, upon analysis, might not be necessary. Efficient natural resource data collection, compilation, and dissemi- nation at the federal level will require the cooperation and collaboration of many agencies that are now responsible for specific surveys and data bases. To the user seeking comprehensive data, shortcomings related to coverage in USDA's NRI series are analogous to those in other natu- ral resource data bases. Inclusion of Data Related to Water Quality The effects of agricultural production practices and other human activities on water quality are viewed as major environmental problems in many parts of the country. The committee believes that agriculture's potential contribution to improved groundwater and surface water quality will be an important consideration in future research, farm policies, and regulatory and educational programs. The increasing emphasis on water quality reflects two reinforcing trends. First, substantial progress has been made in identifying major point sources of water pollution and initiating control. Agricultural opera- tions in many regions are now recognized as the most important non- point sources of water pollutants. Second, the potential hazards of
32 SOIL CONSERVATION agricultural pollutants to health and ecological systems have changed dramatically in the last 20 years with increasing rates of application of fertilizers and changes in the toxic properties and use patterns of crop protection chemicals. Runoff from cultivated cropland into surface waters often transports pesticides and fertilizers, in addition to sedi- ments. Water percolating from surface soil into aquifers can carry pollu- tants into underground water supplies. Currently, detailed monitoring data are limited for agricultural pollu- tants in surface water and groundwater. Data are particularly sparse for pesticides. Whether extremely low levels of contamination are of toxi- cological significance, steps to minimize the flow of agricultural pollu- tants into water are viewed as imperative and prudent. Devising long-term solutions for reducing the volume and hazards of these water pollutants will require better information on the pollutant loadings and hydrology of agricultural watersheds. Such information is not collected as part of the NRI, although a number of U. S. Geological Survey water quality network stations gather some information of this type. Insights can be gained from research on the effects of agricultural management practices on water quality. The Iowa Geological Survey (1984), for example, has combined sophisticated hydrologic monitor- ing studies with assessments of on-farm cropping, conservation, and agronomic management practices in several parts of the state. These studies have documented many complex interactions between conser- vation and other management practices. For example, some conserva- tion practices that help control soil erosion may increase the potential hazard for surface water contamination by fertilizers and pesticides. Contamination of groundwater by fertilizers or herbicides has been documented in areas of more permeable soils and rock formations (Iowa Geological Survey, 1984~. Runoff of water carrying sediment, nutrients, herbicides, and other materials from agricultural land has a major influence on the quality of surface water in many regions of the country (Baker, 1984; Gianessi and Peskin, 1981~. In the future, conser- vation systems on the farm may be tailored to the dual goals of control- ling erosion and mitigating the potential impact of dissolved constituents and sediments on the quality of surface water and ground- water. To achieve these objectives, continuing study is needed to develop or to improve, test, and evaluate models of water quality. This process itself helps to identify the types of field data that may be useful for monitoring those activities that influence water quality. In turn, data
IMPROVING FEDERAL RESOURCE ASSESSMENT EFFORTS 33 from the NRI can contribute to model development and evaluation and, subsequently, monitoring. After giving consideration to the costs of data collection, other infor- mation sources, and the competing uses of natural resource inventory funds, the following kinds of data should be evaluated for potential inclusion in future inventories: · Additional data needed to estimate sediment loads to streams from estimates of erosion; the parameters needed to calculate the ratio of soil leaving a farm field and entering waterways for sheet, rill, and ephem- era-1 gully erosion; and the distribution of sediment offsite, particularly related to lakes and reservoirs; · Types, sources, and magnitudes of nonpoint pollution from such nonagricultural sources as uncovered landfills, construction sites, and surface mining operations; · Deposition of wind-borne soil particles into surface water; · Farming activities that may be related to specific groundwater pol- lutants. There are practical and financial limits to the amount of data that might profitably be included in future NRIs. Yet the NRI already includes observations relevant to water quality, and in many cases no practical and accurate survey technique exists for broad-scale invento- ries such as the NRI. In view of the importance of water quality prob- lems nationwide, the committee believes that it is useful to evaluate further ways in which the NRI can be effectively related to the federal network of water quality data collection, and vice versa. Future NRIs can play an important role in a broader, better coordinated federal data collection effort to measure and analyze water quality trends. Other agencies may agree to assist in supporting future NRIs if the USDA can demonstrate a willingness and capability to include new data. It would be logical to incorporate into future NRIs new survey meth- ods for assessing nonpoint water pollution problems and their causes. The committee believes that the collection of data to assist in reliably assessing the impact of agricultural practices on water quality should be the next extension of the NRI data series.
