Modeling stream discharge and nitrate loading in the Iowa-Cedar River Basin under climate and land use change

A Soil and Water Assessment Tool (SWAT) model was developed for the Iowa-Cedar River Basin (ICRB), a 32,660 km2 watershed dominated by agricultural land cover (~70%) to simulate hydrology and water quality for the prediction of stream discharge, nitrate loads, and nitrate concentration under climate and land use change scenarios. Iowa exports as much as 20% of the nitrogen entering the Gulf of Mexico at the mouth of the Mississippi, contributing to Gulf hypoxia as well as local threats to water quality in the ICRB. The model utilized a combined autocalibration and sensitivity procedure incorporating Sequential Uncertainty Fitting (SUFI) and generalized additive models. This procedure resulted in Nash-Sutcliffe Efficiency (NSE) goodness-of-fit statistics that met literature guidelines for monthly mean stream discharge (NSE≥0.60) and daily nitrate load (NSE≥0.50). Artificial neural networks coupled with SWAT stream discharges aided in the simulation of daily mean nitrate concentrations that met the literature guideline (NSE≥0.50).;The North American Regional Climate Change Assessment Program (NARCCAP) provided an ensemble of 11 climate change scenarios. NARCCAP is a multi-institutional effort to simulate climate change at the mesoscale by downscaling global circulation models (GCM) with regional climate models (RCM). The resulting GCM-RCM produced synthetic precipitation and temperature time-series that drove the SWAT simulations and scenarios. The land use scenarios were a collaboration with the U.S. Army Corps of Engineers, using a rule-based GIS method to generate scenarios that (1) maximized agricultural productivity, (2) improved water quality and reduced flooding, and (3) enhanced local biodiversity. The SWAT simulations and ensemble climate change scenarios resulted in a warmer and wetter climate with greater and more extreme discharge in all seasons except summer where the models indicate a somewhat higher probability of extreme low flows (p-value<0.05). The land use scenarios for SWAT showed that nitrate load and discharge positively and linearly scale with percent of agricultural land area (p-value<0.05).