Nitrogen Balance in and Export from an Agricultural Watershed

Surface water nitrate (NO~-) pollution from agricultural production is well established, although few studies have linked field N budgets, NO~loss in tile drained watersheds, and surface water NO~loads. This study was conducted to determine field sources, transport, and river export of NO~from an agricultural watershed. The Embarras River watershed at Camargo (48 173 ha) in east-central Illinois was investigated. The watershed is a tile-drained area of fertile Mollisols (typical soil is Drummer silty clay loam, a fine-silty, mixed mesic Typic Haplaquoll) with primary cropping of maize (Zea mays L.) and soybean (Glycine max L.). Agricultural field N sources and sinks, tile drainage NO~concentrations and fluxes, and river NO~export were estimated for the entire watershed. Large pools of inorganic N were present following each harvest of maize and soybean (average of 3670 Mg N yr-1 over a 6-yr period). The source of most of the inorganic N was divided between N fertilizer and soil mineralized N. High concentrations of NO~ were found in four monitored drainage tiles (5-49 mg N L-l), and tile concentrations of NO~were synchronous with Embarras River NO~concentrations. High flow events contributed most of the yearly NO~loss (24.7 kg N ha-1 yr-1) from tile drained fields in the 1995 water year (1 Oct. 1994 through 30 Sept. 1995) where high rainfall events occurred in a low overall precipitation year (in one tile 21% of the annual load was exported in 1 d). During the 1996 water year, NO~ export in tiles was much higher (44.2 kg N ha-~ yr-j) due to greater precipitation, and individual days were less important. On average, about 49% (average of 1688 Mg N yr -1 over a 6-yr period) of the field inorganic N pool was estimated to be leached through drain tiles and seepage and was exported by the Embarras River, although depending on weather and field N balances this ranged from 25 to 85% of the field N balance over the 6-yr period. It seems likely that agricultural disturbance (high mineralization inputs of N) and N fertilization combined with tile drainage contributed significantly to NO~ export in the Embarras River. N ITRATE contamination of surface and groundwaters is of environmental concern throughout agricultural areas of the USA. High inputs of N fertilizer are required to support intensive row-crop agriculture, particularly for corn in the Midwest where fertilizer application rates are typically 100 to 200 kg N ha-I yr-1. It is difficult to maintain the fine balance of available N required to satisfy crop needs and at the same time minimize leaching losses, even though fertilization combined with soil mineralization can provide large amounts of inorganic N (Keeney and DeLuca, 1993). Under optimal growing season conditions and fertilizer N application rates, the crop grain yield contains typically only about 50% of the added fertilizer N (Oberle and Keeney, 1990). Throughout many areas of the Midwest and in particular in much of Illinois, agricultural fields are drained with subterranean tiles (perforated pipe) to allow farmM.B. David, L.E. Gentry, and K.M. Smith, University of Illinois, Department of Natural Resources and Environmental Sciences, W503 Turner Hall, 1102 South Goodwin Av., Urbana, IL 61801; D.A. Kovacic, University of Illinois, Department of Landscape Architecture, 101 Temple Buell Hall, 611 East Lorado Taft Drive, Champaign, IL 61820. Received 21 Oct. 1996. *Corresponding author (m-david@ uiuc.edu). Published in J. Environ. Qual. 26:1038-1048 (1997). ing to be practical and economically viable (Fausey et al., 1995). In Illinois alone, about 4 000 000 ha are tile drained, representing 35% of all Illinois cropland (USDA, 1987). As early as the late 1960s, studies tile drainage waters showed high concentrations of N, mostly as NO~(Willrich, 1969). Kohl et al. (1971) 15N proportions in various N pools to estimate that 55 to 60% of the NO;found in surface waters of a typical Illinois corn/soybean watershed was from applied fertilizer not used by the crop. Many studies since these early reports have clearly shown that tile drainage waters contain high concentrations of N as NO~-, often much greater than the USEPA drinking water standard of 10 mg N L-1 (e.g., Lowrance et al., 1984; Logan et al., 1994; Fausey et al., 1995). Few studies have examined N transport from undisturbed, grassland ecosystems in the Midwest to estimate precultivation concentrations and loss of N to surface waters. Recently, however, Dodds et al. (1996) found that the annual N transport averaged 0.16 kg N ha-~ yr-1 from several tallgrass prairie watersheds in Kansas, with NO;concentrations ranging from 0.01 to 0.39 mg N L-1. This demonstrated the tight cycling of N in the grasslands that once dominated much of the central USA, despite their high soil organic C and N contents (Dodds et al., 1996), and clearly contrasts the large increase in NO;concentrations and flux due to intensive agriculture. Recent studies have estimated both farm N budgets (Barry et al., 1993) and large watershed N sources (Keeney and DeLuca, 1993) to evaluate NO;inputs groundand surface waters, respectively. Barry et al. (1993) used a simplified N budget where they looked at many inputs and outputs of N, and assumed total soil-N remained constant from one rotation to the next. They found this method gave useful estimates of potential NO3leaching, despite the major assumptions made. One of their major uncertainties was atmospheric deposition of NH3 from on-farm sources (Barry et al., 1993). Also, high levels of denitrification (up to 62 kg N ha -t yr-1 for continuous corn) were estimated by examining differences between N present at harvest and the amount of N leached. Keeney and DeLuca (1993) took a different approach to estimating N sources. They attempted to use flow and NO~measurements in the Des Moines River to identify the effect of agricultural practices on NO;concentrations. Concentrations and agricultural practices in the 1980s were compared to the 1940s in a 3 194 000ha basin. They concluded that intensive agricultural activities in 1945 and 1980 to 1990 were the major source of river NO~-, and not entirely N fertilizer (Keeney and DeLuca, 1993). The primary source of N from intensive agricultural activities was thought to be mineralization from soil disturbance coupled with tile drainage. Since most of the land was already under cultivation and tiled by 1945, the major disturbance had been made leading Abbreviations: USGS, U.S. Geological Survey; ANI, added N interaction.