Potential P transport to surface and groundwater increases when waste P application rates exceed crop requirement (Sharpley et al., 1996; Sims et al., 2000). The primary factors responsible for potential P delivery from an agricultural field to surface or groundwater include 1) P adsorbed to eroding sediments, 2) soluble P in runoff water, 3) soluble P in leaching water, and 4) P losses related to the type of waste P applied. Practical P loss assessment methods or tools must include all four components to accurately assess the risk of P loss from agricultural fields (Lemunyon and Gilbert, 1993; Sharpley et al., 1996). Interestingly, the first P loss assessment tool provided by USDA-NRCS as a framework from which states could build individual P loss assessment tools did not include P leaching potential or waste source effects on P loss (USDA-NRCS, 1994). Under the revised USDA-NRCS standard, P application rates can be based on the following: 1) an agronomic interpretation of soil test P levels, 2) an environmental interpretation of soil test P, or 3) a P index (site-specific assessment of potential P delivery). In the case of using an agronomic soil test interpretation, the critical soil test P level represents that level above which crop yield is not increased with additional P applied, and, therefore, no fertilizer or manure P would be recommended. Using an environmental interpretation requires establishing the relationship between soil test P level and runoff P concentration. Soil test P levels that produce, for example, >1 mg/L dissolved P in runoff (or some other level), would be established for various soil types. Once this level is attained, no fertilizer or manure P would be recommended. The P index approach should assess the primary P loss pathways as influenced by soil properties, land forms, and management parameters and provides an estimate of the potential risk of P to water bodies (USDA-NRCS, 1994). Based on the P loss assessment, nutrient management plans would be adjusted to reduce P rates and the potential for P loss to surface and groundwater. Adjustments would include animal waste rates ranging from those entirely based on P requirements of the crop to cessation of waste applications. The most reliable method must account for all sources and mechanisms of P transport off a field to the edge of the water body.
While several states decided to use an agronomic or environmental P soil test interpretation to assess risk of P loss, most states recognized that this approach does not adequately reflect P transport, P source, and P management factors primarily contributing to P loss. Although more difficult to develop, a robust P index tool that accurately accounts for P transport pathways and inherent variability in environment, crops and cropping systems, soil types and management, and many other factors should provide accurate assessments of potential P delivery to surface and groundwater.
Several process-based models have been developed to quantify P transport (Sharpley et al., 2002). For example, AGNPS (Agricultural Nonpoint Pollution Source) assesses nutrient loading on watershed scales of >20,000 ha using individual runoff events (Young et al., 1995) or annual runoff (Croshley and Theurer, 1998). The Soil and Water Assessment Tool (SWAT) was developed to predict the impact of land and water management on sediment and chemical transport in large watersheds (Arnold et al., 1998). The SWAT uses daily time steps capable of 1- to 100-yr simulations. The Erosion-Productivity Impact Calculator (EPIC) has also been used to quantify sediment and nutrient transport (Sharpley and Williams, 1990). Although these and other process-based models have been adapted to assess P transport and loss on a field basis, all require detailed data on soil properties, hydrology, crop management, and many other parameters. However, the greatest limitation to adapting these models for use in a P index tool is the skill required by the user. The USDA-NRCS mandated that a P loss assessment tool was intended to be used by public or private county-level technical service providers on a field basis. Thus, states electing to develop a P index approach for P loss assessment must utilize routine methods and models that are well established to facilitate their use by practitioners.