STATISTICAL ANALYSIS OF FACTORS CONTROLLING NONPOINT SOURCE POLLUTANT LOADS IN A SMALL URBAN WATERSHED, SAN MARCOS, TX, USA

We monitored, quantified, and modeled runoff and associated Nonpoint Source (NPS) pollutants during storm events from the Sessom Creek Watershed into the Upper San Marcos River, San Marcos, TX. Sessom Creek is the smallest and steepest of four sub-basins that comprise The Upper San Marcos River watershed, but is highly urbanized. Consequently, discharge and NPS pollutant transport during storm events can be extremely high. Impairment of the San Marcos River is a concern due to the presence of several threatened and endangered species and intense recreational use of the river.

Twelve storm events of varying magnitude were sampled between March and September of 2018. An ISCO automatic sampler was used to collect 24 water samples per storm event. NPS pollutants (Total/volatile/non-volatile suspended sediments, and total and dissolved forms of nitrogen and phosphorous) were analyzed in all samples using standard methods. Results show that loads of these variables are strongly and positively correlated with each other (Pearson coefficient > 0.8), except for Total Nitrogen (TN), which is less strongly correlated with other measured pollutants (Pearson coefficient between 0.5 and 0.78). Since Total Suspended Sediments (TSS) was the variable with the strongest correlation, it was used to predict dissolved forms of N (ammonium and nitrate), total phosphorous, and soluble reactive phosphorus (SRP). Multiple Linear Regression models were developed to predict TSS and TN using antecedent conditions, rainfall intensity and magnitude, and runoff. These models suggest that TSS loads are dependent on antecedent environmental conditions (prior evapotranspiration and rain), total rain during each storm event, and maximum runoff volume. Although TN loads seem to be dependent on the same conditions as TSS, rain intensity was an additional factor significant in predicting TN loads.