Key Descriptors of Nutrient Concentrations in Flood-Impacted Surface Waters

During major flood events, surface waters connect to landscapes that normally are not flooded and may contain pollutants. The connectivity of flood waters with different water quality pollutant sources introduces environmental and public health concerns. Several studies have characterized the magnitude of flood water quality contaminant concentrations; however, few have attempted to explain the response as a function of watershed characteristics and weather conditions. To better understand flood water quality dynamics after a major storm event, we modeled nutrient concentrations in surface waters following Hurricane Florence (2018). Our objectives were to (1) quantify the water quality response during flooded conditions as well as baseline conditions, (2) explain the observed flood water quality trends as a function of landscape characteristics, and (3) evaluate the predictor variables that best explain trends in flood water quality. To accomplish these objectives, water samples were collected across eastern NC, an area dominated by concentrated animal feeding operations and row crops, at four separate time points in the year following Hurricane Florence, which caused widespread flooding in the region. Samples were analyzed for total Kjeldahl nitrogen, ammonium, nitrate and nitrite, total phosphorus, and orthophosphate. A multivariate Bayesian statistical model was then trained to predict flood water quality responses as a function of precipitation, watershed characteristics, and euclidean distances to pollutant point sources. The results highlight that variables related to relative locations of concentrated animal feeding operations (CAFOs) were among the top three most important variables in explaining each nutrient response. For nitrate/nitrite, total phosphorus, and orthophosphate, distance to nearest CAFO was a top predictor regardless of sampling phase, thus suggesting that CAFOs may have chronic impacts on nutrient concentrations in surface waters outside of flood conditions. Other important variables included precipitation and wetland to cropland area ratio. These results indicate that improved waste management may reduce specific nutrient concentrations and reduce water quality threats associated with flooding.