Development of predictive geosmin models in northern Colorado lakes, reservoirs, and rivers

Abstract

Geosmin is a taste and odor (T&O) compound that is naturally occurring, produced by bacteria, and released into drinking water source waters. Geosmin in many parts of the country is a seasonal issue, so drinking water providers often look for temporary solutions to the T&O caused by geosmin. Being able to predict when geosmin will be an issue is vital if drinking water providers are going to succeed in using temporary mitigation methods. Therefore research is being performed to develop predictive models. This study is a broad sampling of Northern Colorado water bodies investigating the role of watershed and elevation, as well as biotic and abiotic water quality parameters. Water quality and zooplankton samples were collected from 20 different lakes and reservoirs as well as 20 sites on 4 rivers in Northern Colorado. Statistical models were developed using Multiple Linear Regression and Principal Component Analysis. Models show significant correlations between geosmin and zooplankton, particularly the species Nauplii and Daphnia in the lakes and reservoirs data. Modeling of the river data revealed geosmin relationships with elevation and dissolved oxygen, but did not show a significant correlation with stream flow. As expected from previous studies month of the year was also shown to be a significant factor.