Kinetics and Modeling of Dissolved Phosphorus Export from a Tile-Drained Agricultural Watershed

Agricultural runoff can be a source of P, a limiting factor for freshwater eutrophication. To develop a simple method to estimate P export from the cropland, we studied 1.2-p.m filtered dissolved phosphorus (DP) output from four tiles draining areas ranging from 8 to 25 ha, and from a river draining a 48 173 ha watershed in eastcentral Illinois during 1993 to 1996. The land was under maize (Zea mays L.)-soybean (Glycine max L.) rotation. The tiles were estimated to contribute more than 86% of the river flow and 65 to 69% of the river DP export during 1995 to 1996. The DP load from tiles followed consecutive pseudo first-order kinetics in terms of tile flow (DP load depended on the amount of DP remaining in the soil matrix). The kinetic curves indicated a soluble-inorganic-P pool that was quickly depleted and replenished. In contrast, for DP export from the river at the watershed scale we observed pseudo zero-order kinetics based on river flow (DP export was independent of how much DP remained in the watershed). The contribution from numerous tiles and surface runoff to the river may have stabilized DP export at the watershed scale and therefore could explain the different kinetic orders. For the study watershed, a one parameter equation could estimate watershedwide DP export: k’ × (surface water discharge from the watershed) × (watershed area), with k’ being 3.94 × 10-6 mg P L-1 ha-1. Our approach should be tested in watersheds with different geographic and agricultural characteristics. p ’HOSPHORUS is known to be a limiting factor in fresh water eutrophication. Because point sources have generally been controlled, increased attention is being directed at the impact of nonpoint P sources from agricultural runoff on the quality of receiving water bodies (Sharpley et al., 1994; Tiessen, 1995). Various models have been constructed to estimate P export from agricultural runoff (Dillon et al., 1991; Frink, 1991; Sharpley and Smith, 1992). However, many of these prediction models require numerous parameters and variables that are not readily available. For example, some regression models require geological, hydrological, and meteorological parameters. Examples of these parameters include humidity, slope, peat area percentage, minor till plain area, carbonate till area, exposed bedrock, flow in spring, and area of small open waters (Dillon et al., 1991); the determination of these parameters is tedious. As a result, Sharpley et al. (1995) pointed out that the practical usefulness of these complicated models is often limited. First-order kinetics have been used to describe the export of pollutants such as chemical oxygen demand Y. Xue, M.B. David, and L.E. Gentry, Dep. of Natural Resources and Environmental Sciences, Univ. of Illinois, Urbana, IL 61801; and D.A. Kovacic, Dep. of Landscape Architecture, Univ. of Illinois, Champaign, IL 61820. Received 19 Sept. 1997. *Corresponding author ([email protected]). Published in J. Environ. Qual. 27:917-922 (1998). and total P from residential, commercial, and highway land (Wanielista et al., 1997): -dX/dt : ~ x [1] where X is the amount of a pollutant remaining in a domain, t is time, and k is a constant independent of X or t. By analogy, P export from agricultural land might follow the same kinetics. When the output rate of X from a domain is small as compared to the amount of X in the domain, X could in fact remain constant with respect to time. Then the k X term in Eq. [1] could be combined as k’. Some studies have found weak or no relationships between soil P content and P concentration in runoff (Daniel et al., 1993), and between P loss from leaching and fertilization (Vighi et al., 1991). Therefore, we considered zero-order kinetics for P export as a possibility: -dX/dt : k’ [2] The two kinetic models are illustrated in Fig. 1, where the variable t in Eq. [1] and [2] has been substituted by flow simply because there will be no P export without flow. For the zero-order kinetics in Fig. 1, a straight line results because the export rate of X is the constant k’ (Eq. [2]), whereas the first-order kinetic curve in Fig. reaches a plateau because the amount of X remaining in the land is proportional to the export rate and when X is depleted the output rate diminishes (Eq. [1]). The objectives of this investigation were therefore to (i) elucidate the pattern of P export as related to flow from a typical Midwest tile-drained agricultural watershed; and (ii) estimate P output from the cropland using simple and easily measured parameters. We focused on 1.2-~m filtered DP, because DP is immediately available for algal growth (Steenbergen et al., 1993). MATERIALS AND METHODS Camargo Watershed The upper segment of the Embarras River drains a 48 173 ha watershed with a U.S. Geological Survey (USGS) gauging station located at Camargo, IL, (39°47’30"N, 88°11’10"W). The watershed was under a maize and soybean rotation in eastcentral Illinois. Besides farmland, 4.5% of the watershed is urban, 0.6% woodland, 0.5% grassland, 0.6% water bodies, 0.3% home sites, and 2.9% roads. There was no large source of sewage effluent into the river. The Mollisols in the area formed in 100 to 150 cm of loess over medium to fine-textured till. Drummer (fine-silty, mixed mesic Typic Haplaquolls) silty clay loams and closely related soils (Flanagan-Catlin) are dominant in the nearly flat watershed where tile drainage is essential for agricultural production. Subterranean drainage tiles Abbreviations: DP, dissolved phosphorus passed through 1.2 ~m filter; NTS, non-tile seepage; PR, precipitation; SR, surface runoff; TD, tile drainage.