Water Resources Research Act Program

Details for Project ID 2020AL353B


Institute: Alabama
Year Established: 2020 Start Date: 2020-03-01 End Date: 2021-02-28
Total Federal Funds: $24,976 Total Non-Federal Funds: Not available

Principal Investigators: Brendan Higgins

Abstract: During the late spring, summer, and fall, drinking water in Alabama often takes on an earthy, musty odor caused by high concentrations of MIB and geosmin in source waters. These are the same compounds that contribute to “muddy†flavor in fish and lead to frequent phone calls to water utilities by upset rate payers. Even for consumers who do not drink tap water, the smell of these compounds pervades other uses of tap water including showering and cooking. Many researchers and utilities have sought to predict when “taste and odor†episodes will occur in order to more rapidly deploy mitigation measures such as the addition of activated carbon to treatment works. Unfortunately, existing predictive tools have not been particularly effective. Dzialowski et al. developed empirical regression models for several reservoirs in the state of Kansas.3 In some cases, the models could predict taste and odor episodes but only in certain reservoirs. Even when the models had acceptable predictive power, they were unique to the reservoir – application of the model to a neighboring reservoir resulted in no predictive capability. The reason for the apparent randomness of these empirical models is that they predict taste and odor events through factors correlated with algal blooms even though the vast majority of algae do not make MIB or geosmin.2 Even within a particular genus of cyanobacteria, some strains produce odor compounds and others do not.4 To overcome this challenge, a better understanding of the underlying mechanism for taste and odor is needed: in this case, investigation of the genes that actually synthesize MIB and geosmin. The goal of this project is to develop a better prediction tool for taste and odor episodes that relies on quantification of the MIB and geosmin synthase genes in water samples. By overlaying this genetic dataset on top of traditional water quality parameters (e.g. temperature, turbidity, chlorophyll a), we hypothesize that we can develop a model with better predictive capability within and across reservoirs. This hypothesis is supported by results from other researchers who have shown that cell count data on known MIB and geosmin producers was the single most important predictive factor in successful empirical models.5,6 We will utilize Classification and Regression Trees (CART) models to predict conditions and genetic states that contribute to critical thresholds for taste and odor compounds. The project goal aligns with area 3C and 4A of the “major water