Effect of Lake Nitrogen Fixation on Watershed Export under Loading Reduction Scenarios

Coastal eutrophication is an important problem, and nitrogen input from the watershed is a primary cause. Consequently, management efforts often focus on lowering watershed N export by reducing the N loading (e.g.,from point sources), and models are used to evaluate alternative reduction scenarios at the watershed scale. Existing models generally assume lake export decreases proportionally with loading. However, reducing N loading lowers the N:P ratio, and N fixation by cyanobacteria may increase, which may change the effective N export fraction (export/external input). This paper explicitly considers how lake N fixation may change when N loading is reduced and estimates the reduction in lake and watershed N export. A simple model, Fixation and Export of Nitrogen from Lakes (FENL) is developed; implemented into a watershed model, Spatially Referenced Regression on Watershed Attributes (SPARROW) of the Chesapeake Bay Watershed; and applied to predict steady-state N export under various management scenarios. The results suggest that lake and watershed N export will not be reduced proportionally with N loading. Under a 60% N-only reduction scenario, the model predicts that the majority of the lakes in the watershed are pushed into the N-fixation regime, which increases lake N fixation rate by a factor of 5.6 and may result in an equal increase in the concentration and blooms of N-fixing cyanobacteria. The model-predicted response is heterogeneous and suggests there are more-vulnerable and less-vulnerable regions, which may be exploited for more cost-effective management. The added N by fixation counteracts the N loading reduction, but the effect on the watershed export is small for Chesapeake Bay. In contrasts to a watershed load reduction of N alone, a balanced N+P reduction (say, of 60%) will not change the N-fixing state (fixing N or not fixing N) of lakes and result in the same percentage reduction in watershed N export (i.e.,also 60%) as the watershed load reduction.