Evaluating tryptophan-like fluorescence to quantify E. coli concentrations in a combined sewer overflow impaired...

This field study evaluates the effectiveness of an optical indicator parameter called Tryptophan-like fluorescence (TLF) combined with other water quality parameters to predict E. coli concentrations. Commercially available multi-parameter sondes measuring TLF were deployed upstream and downstream, of five active combined sewer overflow regulators located within a 1.6-km section of the Lower Rouge River in Michigan, United States. E. coli grab samples were collected during three storms of varying size and across seasons: 3/8/22, 28/4/23 and 14/8/23 (D/M/YY) and analyzed by Colilert-18®. These data were paired with TLF and other parameters measured by the sonde to build algorithms that predict E. coli concentrations. Two sets of algorithms were created. The first algorithm was based on a single storm monitored on 14/8/23, and the algorithm was validated against data collected during the 3/8/22 and 28/4/23 storms. An additional algorithm was created for explorational purposes using the results from all three storms to test the hypothesis that data from additional storms would improve the accuracy of E. coli predictions. The manufacturer retained proprietary ownership of the algorithms, so they are not disclosed in this publication. The lab measured and predicted E. coli concentrations were compared on 1:1 plots and R² values were generated for each location and each storm. The R² values generated from the 14/8/23 algorithm range from 0.5705 to 0.0005. The explorational algorithm based on all three storms resulted in R² values of 0.4953 to 0.0045.