Identifying dissolved phosphorus source areas and predicting transport from an urban watershed using distributed hydrologic modeling

[1] A reduction in surface water quality in urban watersheds due to nonpoint source phosphorus (P) loading has prompted municipalities to consider management practices to reduce P loss from landscapes. However, locating P source areas can be time consuming and expensive. Use of distributed models allows delineation of P source areas and focused management strategies. Using the spatially distributed soil moisture distribution and routing model, we adapt and validate a dissolved P (DP) loading model for application to an urban watershed, in Ithaca, New York, to identify P source areas. The model calculates DP loss separately for base flow, impervious surfaces, plant-soil complex, and fertilized areas. The load at the outlet is the sum of P loss from the four components distributed throughout the watershed. Both stream and distributed DP loss were well predicted as indicated by comparison with measured data. The model predicted the largest contribution from plant-soil complexes (36%). Impervious surfaces contributed 10% of the total load but as much as 17% in the winter. More important, the impervious surfaces increased DP losses from the adjacent areas due to runoff from the impervious surfaces saturating the soil, thus increasing runoff losses. Fertilizer contributed substantially following application but decreased rapidly thereafter, a result of conversion from soluble to insoluble P. However, fertilization increased soil P levels, and thus DP losses were higher as a whole (19%). Results demonstrate that correctly predicting the coincidence of P and runoff source areas can be a powerful tool to identify and mitigate contamination of surface waters.