Total soluble P in surface runoff water delivered to the stream edge is determined by estimating runoff volume (hectare centimeters) and P concentration in the runoff (milligrams per liter).
Runoff Phosphorus Concentration
Initially, the U.S. Environmental Protection Agency suggested 1 mg/L dissolved P as a guideline for agricultural runoff to protect surface water quality (U.S. EPA, 1986). However, processes controlling runoff P concentration and ultimate transport to surface water are influenced by many factors within the field and between the field edge and edge of the stream or water body. It is more appropriate to evaluate contributions of dissolved P in runoff relative to factors specific to each site. For example, in one watershed or field, 0.05 mg P/L could be considered too high whereas in another field 2 mg P/L in runoff may not significantly degrade surface water quality.
One easily measurable parameter that is directly related to dissolved P in runoff is soil test P. Three soil tests are commonly used in the United States and were originally designed to provide indices of plant available P for purposes of making P recommendations for crop production. The Mehlich (Mehlich, 1984) and Bray (Bray and Kurtz, 1945) soil tests are typically used for acid soils in which Al and Fe phosphate minerals control soil solution P concentration and P availability, whereas the Olsen soil test (Olsen et al., 1954) is used in neutral and calcareous soils, in which Ca phosphate minerals predominately control soil solution P concentration. The advantage of the Mehlich soil test over the Bray soil test is its ability to extract other plant nutrients in addition to P.
The Mehlich-3 P soil test has been established as a useful index of the potential environmental risk of increasing P in soils (Sharpley et al., 1996; Khiari et al., 2000; Sims et al., 2002). The soil solution P concentration is related to percentage of phosphorus saturation of the soil, which in turn influences Mehlich-3 P level. Numerous studies have established relationships between soil test P and dissolved P concentration in runoff (Sharpley et al., 1994; Pote et al., 1999; Weld et al., 2001; Torbert et al., 2002). In these studies, the Mehlich-3 P level that results in 1 mg P/L in surface runoff varied from 150 to 625 mg P/kg soil. Because P is adsorbed more strongly in clay soils, higher Mehlich-3 P levels are associated with high clay soils. Therefore, higher soil test P is required for a given runoff P concentration in clay soils compared with sandy soils.
Similar results were obtained from studies on the relationship between Mehlich-3 P and soluble P in runoff in North Carolina soils (Cox, 1994; Cox and Hendricks, 2000). For a given concentration of Mehlich-3 P, there was more dissolved P in runoff in sandy soil than clay soil (Figure 6). A concentration of 1 mg P/L in runoff from a sandy soil was associated with a Mehlich-3 P soil test of 253 mg/kg, whereas it was estimated by extrapolation to be about 500 mg/kg Mehlich-3 P for a clay soil (Figure 6). This difference is related to P being held less tightly in the sand compared with the clay soil because of differences in P adsorption capacity.
McDowell and Sharpley (2001) showed that in a Pennsylvania watershed (clay and clay loam soils) dissolved P in runoff increased with Mehlich-3 P (Figure 7). These results also showed that a Mehlich-3 P level of 500 mg/kg resulted in a dissolved P concentration of approximately 1 mg/L. Based on many studies, the effects of Mehlich-3 P on runoff P concentrations for the major soil groups were established for North Carolina (Figure 8). Mehlich-3 P levels producing 1 mg/L soluble P for the organic, sand, loam, and clay soils were 50, 100, 200, and 500 mg/kg, respectively (Novais, 1977; Reddy et al., 1980; Cox and Hendricks, 2000).
The quantity of runoff water associated with an individual storm event depends on characteristics of the rainfall (quantity, intensity, and duration), surface soil conditions that influence infiltration (residue cover and soil physical properties, including water content), subsoil properties that influence hydrologic conductivity (soil structure, texture, water content), and water table depth (USDA-NRCS, 1989). During rainfall, water enters the soil through large, surface-connected macropores under a positive hydraulic head. Water then diffuses vertically and horizontally into a network of micropores by capillary action or soil moisture tension (SMT). Water flow into the subsoil volume is predominately governed by SMT. After soil macropores are filled, water moves through micropore areas toward the highest SMT. The water infiltration or transport rate is governed by the number, size, and continuity of the pore network. The presence of old root channels, earthworm borrows, and natural subsoil structural macropores can substantially increase water infiltration and transport of dissolved P through the profile; however, their presence and influence are difficult to quantify.
Presence of a shallow water table can reduce water transport in the subsoil and increase runoff. In regions where shallow water tables are prevalent, artificial drainage is commonly installed to improve infiltration and soil productivity. Models designed to estimate runoff volume should account for both drained and undrained soil conditions (Skaggs et al., 1982; Evans et al., 1995). In addition, whatever runoff model is incorporated into a practical P loss assessment tool, it must be based on annualized runoff instead of individual runoff events, and should be capable of routine use by field technicians.
These criteria are met for well-drained, upland soils, with the empirical curve number approach used to estimate runoff volume (USDA-NRCS, 1989). The curve number method is also used in several process-based simulation models to predict daily runoff (Knisel, 1993; Sharpley and Williams, 1990; Arnold et al., 1998). The curve number method relates runoff potential to land use and soil characteristics. The runoff depth is determined from the total rainfall adjusted for estimated infiltration determined from surface and subsoil physical characteristics. The method assumes that an accumulated rainfall depth of >0.2 x maximum soil water retention must occur before generating runoff. The maximum soil water retention and curve number are determined for major soil hydrologic groups based on soil physical properties, vegetation, land use, and antecedent moisture conditions or soil water content at time of runoff producing rainfall (Table 3). Curve numbers are established to reflect variations in soil properties and antecedent moisture that influence infiltration and runoff. For a specific rainfall amount, runoff would increase with increasing curve number (Figure 9). In practice, relationships similar to Figure 9 are developed for each county using long-term rainfall records, various cropping systems, and hydrologic soil conditions to estimate runoff volume.
For artificially drained, shallow-water-table soils, the curve number approach does not accurately estimate runoff volume. Most shallow-water-table soils similar to the coastal plain regions of the United States are not suitable for agricultural production unless artificial drainage is installed. Calculation of runoff for drained soils requires information on the drain depth, the estimated distance between drains, and the soil transmissivity (square centimeters per hour), which represents the ability of the soil profile to transmit water laterally when the water table is 30 cm below the soil surface. These values have been established for all soils in areas where artificial drainage is commonly used. The model widely accepted for this purpose is DRAINMOD, which has been thoroughly field-calibrated and is widely used to estimate runoff volume and water transport through the profile in high-water-table soils (Skaggs, 1999; Skaggs and Chescheir, 1999).
Soluble Phosphorus Retaining Practices
Unlike sediment P, there are few conservation practices that can reduce soluble or runoff P before reaching the stream edge. However, any soil management practice that increases infiltration and decreases runoff (conservation tillage or water control structures) can reduce the transfer of soluble P to surface water. Riparian buffers, for example, do not reduce transfer of soluble P, but they are very effective in reducing transfer of sediment P to surface water (Nash and Murdoch, 2000). Studies in Oklahoma show that permanent residue cover (pasture) resulted in much less dissolved P in runoff compared with row-cropped fields, predominately due to reduced contact between runoff water and surface soil (Sharpley et al., 1991; Smith et al., 1991) (Figure 10). Sharpley et al. (2002) summarized data over a wide geographical area in the United States and showed that increasing surface residue cover decreased dissolved P in runoff (Figure 11).