Phosphorus Budgets

        Because nutrient concentrations have such a large influence on pond trophic state, much attention has been devoted to means of assessing the sources and fates of nutrients. These include a) point source inputs from specific, identifiable inflows, b) nutrients contained in direct precipitation, and c) nutrients originating as runoff from non-point sources such as agriculture, septic fields or fertilized lawns within the watershed. The latter category is typically the most important, and unfortunately also the most difficult to quantify or control.
        A variety of nutrient budget models have been formulated to estimate the relative contributions of various nutrient inputs. These differ in part according to their complexity and data requirements. Very simple models, such as the one described here, are easily compiled and understandable, but are not sensitive to year-to-year or shorter term variation in weather, whereas the use of more complex models often requires daily rainfall and much more detailed land use information.
        Phosphorus budgets for the 13 ponds were prepared following Reckhow and Chapra (1983):

[P] = L /[Vs + qs]

where [P] = the predicted mean phosphorus concentration in the pond, L = the estimated annual phosphorus loading to the pond, vs = net P settling velocity and qs = the estimated annual water loading to the pond.
        A nutrient budget for one of the 13 ponds (GF) is shown in Table 2. Total loading for each land use is the product of its area multiplied by a loading coefficient that estimates P export per unit area for that land use. Export coefficients were derived from available literature (e.g., Reckhow et al. 1980). Nutrients also reach the pond through direct precipitation. These inputs are summed to obtain predictions of total annual nutrient influx from the watershed (W). Water loading (qs) is calculated from precipitation data for the region. Finally, predicted concentrations of P are compared with actual values to determine the likely effects of other watershed or pond features.
        As seen in Table 2, the watershed of pond GF is dominated by cropland (37%) and forest (35%), with smaller amounts of residential housing and pasture. Cropland typically yields more phosphorus per unit area than does forest, and is estimated to provide more than half (11.11 kg/yr / 25.14 kg/yr, or 52%) of total phosphorus loading to GF. Watershed management efforts to reduce phosphorus loading to the pond might thus reasonably focus on agricultural practices.

Table 2. Nutrient Budget for a pond in East Bradford Township. Export (loading) coefficients estimated from previous studies were multiplied by areas of each land use (determined from aerial photographs taken in 2000). Direct precipitation inputs were based on pond surface area.

Table 2

        Predicted values of total phosphorus concentration in the water column deviated widely from actual mean values in the 13 study ponds. BO was notable in having much less phosphorus than predicted, while several ponds (especially CH, WA) had higher concentrations than predicted. The lack of fit suggests that other environmental variables may influence actual phosphorus concentrations. A more complete description of the nutrient budget model used and the degree of concordance between observed and predicted phosphorus concentrations is presented in Anderson (2003).
Figure 29

Fig. 29. Fit of TP predicted by Reckhow and Chapra model to actual TP based on three visits to each of 13 ponds. Line represents a 1:1 fit.