Modeling of Two Common Taste and Odor Compounds, 2-Methylisoborneol and Geosmin, for Settling Reservoirs in a Drinking Water Treatment Plant

Abstract

Harmful algae blooms are a worldwide concern that has recently regained attention due to the concern of climate change impacts. While algal bloom concerns have historically been centered around toxin producing algae, focus has shifted to include blooms of any nuisance species including those that release taste and odors. 2-methylisoborneol and geosmin are the two most studied taste and odor compounds due to their low detection threshold (5-10 ng/L). As a water utility, it is important to remove taste and odors, not because of harmful effects to human health, but rather due to public perception of water. Water that has a taste or odor is often perceived as poor water quality, making it important to properly manage taste and odor events. Most utilities study taste and odors prior to treatment in source water reservoirs. This study models MIB and geosmin production within a settling reservoir post primary treatment (coagulation and flocculation) using quantile regression models. Quantile regression models the relationship between predictor variables and specific percentiles of target variables allowing for a comprehensive view of how predictor variables affect different parts of distributions. Seasonal, biological, and water quality parameter variables were evaluated within the model. These models were able to successfully explain MIB and geosmin concentration variations at their highest concentrations, when consumers complain (95th quantile). The MIB model was also able to explain variations within the 75th and 90th quantiles. These models will be used to more efficiently monitor taste and odor production by allowing the utility to recognize seasonal patterns and predict concentrations where treatment changes may be required.