You are here:
Watershed Nitrogen Modeling: Benefits of Diverse Approaches Using a Case Study from New York State
Citation:
GOLDEN, H. E., C. D. KNIGHTES, AND E. W. BOYER. Watershed Nitrogen Modeling: Benefits of Diverse Approaches Using a Case Study from New York State. Presented at National Water Quality Monitoring Conference, Atlantic City, NJ, May 18 - 22, 2008.
Impact/Purpose:
Presentation at the National Water Quality Monitoring Conference, May 2008.
Description:
Watershed-scale models have evolved as an important tool for estimating the sources, transformation, and transport of contaminants to surface water systems. A wide variety of modeling approaches exist for estimating inputs, fate, and transport of constituents but most are broadly mathematically categorized as either statistical or process-based. Selection of either approach is often determined by the level of computational detail required for the intended output, spatiotemporal scales of the project or research, and the skills and desired outcomes of the end-user audience. Our work aims to evaluate the efficacy of statistical versus process-based modeling for assessing exposures in surface water systems. Here we focus on a specific driver of watershed nutrient models, presenting a statistical model for estimating contemporary atmospheric inputs of nitrogen to multiple-scale watersheds across New York State and comparing this approach to process-based atmospheric deposition models (e.g., CMAQ). Movement toward hybrid approaches for fate and transport modeling (e.g., SPARROW) in some modeling communities suggests that these two ostensibly disparate approaches can be complementary and that neither modeling approach needs to operate exclusively of the other. For example, statistical models can potentially provide a distribution of important input parameters for some process-based models. Further, we suggest that implementation of monitoring networks - both atmospheric and surface water - at spatial and temporal scales appropriate for model calibration is critical to the success of either separate or integrated statistical and process-based modeling approaches.