You are here:
USING MUSSEL ISTOPE RATIOS TO ASSESS ANTHROPOGENIC NITROGEN INPUTS TO FRESHWATER ECOSYSTEMS
Citation:
McKinney, R A., J L. Lake, M Charpentier, AND S A. Ryba. USING MUSSEL ISTOPE RATIOS TO ASSESS ANTHROPOGENIC NITROGEN INPUTS TO FRESHWATER ECOSYSTEMS. ENVIRONMENTAL MONITORING AND ASSESSMENT 74:167-192, (2002).
Description:
Stable nitrogen isotope ratios ( 15N) of freshwater mussels from a series of lakes and ponds were related to watershed land use characteristics to assess their utility in determining the source of nitrogen inputs to inland water bodies. Nitrogen isotope ratios measured in freshwater mussels from 19 lakes and ponds in Rhode Island, U.S.A., ranged from 4.9?12.6 0/00 and were found to significantly correlate with the fraction of residential development in 100 and 200 m buffer zones around the ponds. Mussel 15N values in 12 of the 19 ponds also showed significant correlation with average dissolved nitrate concentrations, which ranges from 23?327 g L-1. These observations, in light of previous studies which link elevated 15N values of nitrogen derived from septic wastewater with those seen in biota, suggest that mussel isotope ratios may reflect nitrogen source in freshwater ecosystems. We followed an iterative approach using multiple regression analysis to assess the relationship between mussel 15N and the land use categories fraction residential development, fraction feedlot agriculture, fraction row-crop agriculture, and fraction natural vegetation in 100 and 200 m buffer zones and pond watersheds. From this we developed a simple regression model to predict mussel 15N from the fraction of residential development in the 200 m buffer zone around the pond. Subsequent testing with data from 16 additional sites in the same ecoregion led us to refine the model by incorporating the fraction of natural vegetation. The overall average absolute difference between measured and predicted 15N values using the two-parameter model was 1.6 0/00. Potential sources of error in the model include differences in the scale and categorization of land-use data used to generate and test the model, differences in physical characteristics, such as retention time and range of residential development, and exclusion of sources of enriched nitrogen such as runoff from feedlot operations or increased nitrogen loading from inefficient or failed septic systems.