Regional estimates of the amount of U.S. agricultural land located in watersheds with poor water quality

Abstract

An estimated 71 percent of U.S. cropland (nearly 300 million acres) is located in watersheds where the concentration of at least one of four common surface water contaminants exceeded generally accepted criteria in 1989. This figure serves as an approximate measure of the efficiencies to be gained by restricting agricultural pollution controls to watersheds with water quality. Introduction It is widely believed that achievement of national water quality goals will require increased control of nonpoint-source pollution from agricultural land (Adler and others, 1993). One of the more contentious issues related to the design of nonpoint-source pollution controls is whether, in the interest of fairness, controls should be imposed more or less uniformly on all agricultural land of a given type, or whether, in the interest of economic efficiency, controls should be imposed only on agricultural land in areas of water quality (Knopman and Smith, 1993). Debate on this issue would be improved by better information on the efficiencies to be gained by limiting controls to areas of water quality. This report presents estimates of the amount of agricultural land in the conterminous United States that is in watersheds where contaminant concentrations in streams exceed generally accepted water-quality criteria (Smith and others, 1993). Estimates are developed separately for four types of agricultural land, seven regions of the Nation, four water contaminants, and for the early and late 1980's. It should be noted that the extent to which agricultural activity contributes to water quality in a given watershed is not at issue in this analysis. Rather, the focus here is on the savings in pollution control costs to be achieved by limiting agricultural controls to areas with water quality regardless of the causes. Study Methods The water-quality and land-use data used in this analysis are described and illustrated in detail in Smith and others (1993). Watersheds are defined as the 2111 U. S. Geological Survey hydrologic cataloging units in the conterminous United States (Seaber and others, 1987). Water quality conditions for a sample of hydrologic units were determined on the basis of concentrations of dissolved nitrate, total phosphorus, fecal coliform bacteria, and suspended sediment measured at a stream sampling station located near the outflow of each hydrologic unit in the sample. Groundwater quality in the watersheds is not considered in this study. Water-quality conditions were determined separately for the early and latter 1980's. Conditions for the latter 1980's were based on mean concentrations of contaminants in 1989; conditions for the early 1980's were based on mean concentrations for 1980 in the case of dissolved nitrate, fecal coliform bacteria, and suspended sediment, and for 1982 in the case of total phosphorus. Recent quality assurance data indicate that measurements of total phosphorus for 1980 and 1981 are less reliable than for 1982 (Alexander and others, 1993). Water quality conditions in the sampled watersheds were classified as either poor or good on the basis of whether mean contaminant concentrations exceeded generally accepted water quality criteria (see discussion of water quality indicators and criteria in Smith and others, 1993). Contaminant criteria were as follows: dissolved nitrate, 3.0 mg/1; total phosphorus, 0.1 mg/1; fecal coliform bacteria, 200 colonies per 100 ml; suspended sediment, 500 mg/1. Classifications were made separately for the four contaminants as well as jointly for the suite of contaminants. In the latter case, watersheds were classified as having water quality if the mean concentration of any contaminant exceeded its criterion. Definition of Regions For purposes of statistical analysis, the conterminous United States was divided into five regions on the basis of aggregations of the regional drainage basins described in Smith and others (1993) [see Figure 44, p 129). The correspondence between the regions used in the two analyses is as follows: This Analysis North Atlantic South Atlantic/Gulf Upper Midwest