PS 34-223 - Using the Agricultural Policy/Environmental eXtender (APEX) model to evaluate riparian nutrient retention in the Pacific Northwest

Abstract: Mitigating eutrophication requires understanding local nutrient dynamics. Numerical models, such as the Agricultural Policy/Environmental eXtender (APEX) model, can help researchers understand nutrient transport across a landscape and the potential effects of mitigation strategies, including riparian restoration. While APEX has been used extensively in the Midwest, different ecosystems in the Pacific Northwest require new calibration. Bellingham Bay, part of the Salish Sea in northwestern Washington, frequently experiences eutrophication, with the agricultural lowlands in the Nooksack River basin contributing approximately half of the culpable nutrients. Previous work attempted to use APEX to model hydrology and nutrient cycling in the predominantly agricultural watershed of Kamm Creek, a small tributary of the Nooksack River. This effort (the “baseline”) used non-local weather data and constant atmospheric CO₂ concentrations, and ultimately showed mixed results for discharge predictions and dramatically underpredicted nitrogen fluxes. As a first step toward assessing the effects of riparian buffers on nutrient retention, we sought to refine this previous work in Kamm Creek. We hypothesized that using local weather data and historic atmospheric CO₂ concentrations would improve hydrologic and nitrogen flux predictions by APEX. We updated these inputs and compared the predictions to the measured fluxes and baseline predictions.

Our updated data inputs moderately improved the model’s predictive ability for both discharge and nitrogen flux. The baseline results for hydrologic discharge tracked the pattern of the measured data well (R² = 0.87), but substantially underpredicted (percent bias (PBIAS) -45%). Using updated data improved the fit overall: model predictions were slightly more variable (R² = 0.72), and still low, but improved (PBIAS -39%). The Nash-Sutcliffe Efficiency for discharge increased from -1.82 to -0.16 (ideal = 1) with the updated data. The baseline results did not capture the magnitudes or seasonality of the measured nitrogen fluxes, underpredicting by an order of magnitude (R² = 0.31, PBIAS -89%). The updated inputs better captured the seasonality of nitrogen fluxes, despite lower R² (0.28). While the updated results still underpredicted nitrogen fluxes (PBIAS -69.5%), they were ~3-fold greater than the baseline predictions. We will expand on this work by incorporating groundwater contributions, which should increase model predictions for both hydrologic discharge and nitrogen fluxes. While using updated inputs led to more accurate predictions of discharge and nitrogen fluxes, as expected, these results suggest that APEX needs more refinement before we can apply it as an effective riparian restoration planning tool in the Pacific Northwest.