You are here:
Using Model Comparisons to Understand Sources of Nitrogen Delivered to US Coastal Areas
Citation:
McCrackin, M. L., J. A. Harrison, AND J. E. COMPTON. Using Model Comparisons to Understand Sources of Nitrogen Delivered to US Coastal Areas. Presented at AGU Fall, San Francisco, CA, December 13 - 17, 2011.
Impact/Purpose:
model comparisons provide strong evidence that anthropogenic activities have a profound effect on N delivered to coastal areas of the US, especially along the Atlantic coast and Gulf of Mexico
Description:
Nitrogen loading to water bodies can result in eutrophication-related hypoxia and degraded water quality. The relative contributions of different anthropogenic and natural sources of in-stream N cannot be directly measured at whole-watershed scales; hence, N source attribution estimates at scales beyond a small catchment must rely on models. Although such estimates have been accomplished using individual N loading models, there has not yet been a comparison of source attribution by multiple regional- and continental-scale models. We compared results from two models applied at large spatial scales: Nutrient Export from WatershedS (NEWS) and SPAtially Referenced Regressions On Watersheds (SPARROW). Despite widely divergent approaches to source attribution, NEWS and SPARROW identified the same dominant sources of N for 65% of the modeled drainage area of the continental US. Human activities accounted for over two-thirds of N delivered to the coastal zone. Regionally, the single largest sources of N predicted by both models reflect land-use patterns across the country. Sewage was an important source in densely populated regions along the east and west coasts of the US. Fertilizer and livestock manure were dominant in the Mississippi River Basin, where the bulk of agricultural areas are located. Run-off from undeveloped areas was the largest source of N delivered to coastal areas in the northwestern US. Our analysis shows that comparisons of source apportionment between models can increase confidence in modeled output by revealing areas of agreement and disagreement. We found predictions for agriculture and atmospheric deposition to be comparable between models; however, attribution to sewage was greater by SPARROW than by NEWS, while the reverse was true for natural N sources. Such differences in predictions resulted from differences in model structure and sources of input data. Nonetheless, model comparisons provide strong evidence that anthropogenic activities have a profound effect on N delivered to coastal areas of the US, especially along the Atlantic coast and Gulf of Mexico.