ABSTRACT
The microbial quality of irrigation water is typically assessed by measuring the concentrations of E. coli in irrigation water reservoirs that are variable in space and time. E. coli concentrations are affected by water quality parameters that co-vary with E. coli concentrations and may be easily measured with currently available sensors. The objective of this work was to identify the most influential environmental covariates affecting E. coli concentrations during a three-month biweekly monitoring period within two irrigation ponds in Maryland during the summer of 2017. E. coli levels as well as sensor-based water quality parameters including turbidity, pH, dissolved oxygen, dissolved fluorescent organic matter, conductivity, and chlorophyll were measured at 23 and 34 locations in ponds 1 and 2, respectively. Regression tree analyses were used to determine the most influential water quality parameters for the prediction of E. coli levels. Correlations between E. coli and water quality covariates were not strong and were inconsistently significant. Shoreline sample locations had higher E. coli concentrations than interior pond samples and significant differences were observed when comparing these two groups. Regression trees provided fairly accurate predictions of E. coli levels based on water quality parameters with R2 values ranging from 0.70 to 0.93. Factors identified via the regression trees varied by sampling date but common leading covariates included cyanobacteria, organic matter, and turbidity. Results indicated environmental covariates, sensed either remotely or in situ, could be useful to delineate areas with different E. coli survival conditions across irrigation ponds and potentially other water bodies such as lakes, rivers, or bays.
