Repository logo

Measuring and modeling geosmin removal from Horsetooth Reservoir water by powdered activated carbon for selected contact times




Koester, Kirk, author
Omur-Ozbek, Pinar, advisor
Carlson, Ken, committee member
Goemans, Chris, committee member

Journal Title

Journal ISSN

Volume Title


Presence of geosmin, an odorous algal metabolite produced by cyanobacteria and actinomycetes, has been an issue in drinking waters in Northern Colorado. Geosmin does not pose a health threat; however, it imparts an earthy taste and odor to the finished drinking water even at very low concentrations (4 -10 ng/L), resulting in consumer complaints and dissatisfaction. Geosmin cannot be removed by conventional water treatment processes, so further treatment is required to achieve concentrations below detection limits. This study investigated the geosmin removal from the raw water obtained from the Horsetooth Reservoir, in Fort Collins, CO by powdered activated carbon (PAC). Raw water samples were spiked with stock geosmin solution to obtain concentrations from 10 to 50 ng/L and stock PAC solution to obtain concentrations from 5 to 20 mg/L. Thirteen different geosmin/PAC concentrations were tested for 90 minutes contact time (and up to 6 hours for selected combinations) and the geosmin removal was determined by headspace solid phase microextraction and gas chromatography/mass spectrometry. Results indicated that 50 to 70% removal was achieved for lower doses of geosmin/PAC combinations, and 80 to 97% removal was achieved for higher PAC dose combinations. Most (65%) of the geosmin removal was achieved within the first thirty minutes. For 54% of the samples, geosmin concentrations in the treated water were lowered below 4 ng/L after 90 minutes, which is a low enough concentration to prevent consumer complaints. Furthermore, 38% of the samples lowered geosmin concentrations below 4 ng/L after 45 minutes. Additionally three PAC/geosmin combinations that contained 8 mg/L of total organic carbon were examined and a decrease in geosmin removal by 8-12% after 90 minutes was observed. Further analysis with Stat-Ease® Design Expert® (v.8) was used to model geosmin removal to predict required PAC dosages for geosmin levels not tested in this study. The model developed by the Design Expert® considered the initial geosmin concentration, PAC dosing and contact time, and a simple equation was obtained to predict the remaining geosmin concentrations in the treated water. Results from the model fit the data obtained from laboratory measurements and reliably predicted geosmin removal for concentrations not tested.


Rights Access


drinking water treatment
taste and odor
powdered activated carbon


Associated Publications