Abstract
Cyanobacteria cause a multitude of water-quality concerns, including the potential to produce toxins and taste-and-odor compounds. Toxins and taste-and-odor compounds may cause substantial economic and public health concerns and are of particular interest in lakes, reservoirs, and rivers that are used for drinking-water supply, recreation, or aquaculture. The Kansas River is a primary source of drinking water for about 800,000 people in northeastern Kansas. Water released from Milford Lake to the Kansas River during a toxic cyanobacterial bloom in late August 2011 prompted concerns about cyanobacteria and associated toxins and taste-and-odor compounds in downstream drinking-water supplies. During September and October 2011 water-quality samples were collected to characterize the transport of cyanobacteria and associated compounds from upstream reservoirs to the Kansas River. This study is one of the first to quantitatively document the transport of cyanobacteria and associated compounds during reservoir releases and improves understanding of the fate and transport of cyanotoxins and taste-and-odor compounds downstream from reservoirs.
Milford Lake was the only reservoir in the study area with an ongoing cyanobacterial bloom during reservoir releases. Concentrations of cyanobacteria and associated toxins and taste-and-odor compounds in Milford Lake (upstream from the dam) were not necessarily indicative of outflow conditions (below the dam). Total microcystin concentrations, one of the most commonly occurring cyanobacterial toxins, in Milford Lake were 650 to 7,500 times higher than the Kansas Department of Health and Environment guidance level for a public health warning (20 micrograms per liter) for most of September 2011. By comparison, total microcystin concentrations in the Milford Lake outflow generally were less than 10 percent of the concentrations in surface accumulations, and never exceeded 20 micrograms per liter.
The Republican River, downstream from Milford Lake, was the only Kansas River tributary with detectable microcystin concentrations throughout the study period, and concentrations exceeded 1 microgram per liter for most of September 2011. Microcystin was detected periodically in other tributaries, but concentrations were low (less than 0.3 micrograms per liter). In contrast, the taste-and-odor compounds geosmin and 2-methylisoborneol (MIB) were detected in all tributaries located immediately downstream from reservoirs and total concentrations generally exceeded the human detection threshold (5 to 10 nanograms per liter) from September through mid-October. Microcystin, geosmin, and MIB were not detected in the Smoky Hill River upstream from the confluence with the Republican River that forms the Kansas River.
Within a week after initial reservoir releases, microcystin, geosmin, and MIB were detected throughout a 173-mile reach of the Kansas River; these compounds remained detectable throughout the reach until mid-October. Losses to groundwater when streamflows in the Kansas River were increasing indicate the potential for reservoir releases to affect groundwater quality as well as surface-water quality. Total microcystin concentrations in the Kansas River generally were highest within about 24 miles of the confluence of the Smoky Hill and Republican Rivers, and decreased downstream; concentrations exceeded 1 microgram per liter in the Kansas River upstream from Topeka during the first 2 weeks of September. Patterns in microcystin occurrence and concentration at Kansas River tributary and main-stem sites indicate that Milford Lake was the source of microcystin in the Kansas River; however, the source of taste-and-odor compounds was not as evident, possibly because multiple tributaries contributed taste-and-odor compounds to the Kansas River.
Microcystin and taste-and-odor compounds co-occurred in 56 percent of samples collected, indicating co-occurrence was common. Despite frequent co-occurrence, the spatial and temporal patterns in microcystin, geosmin, and MIB were unique and did not necessarily match patterns in cyanobacterial abundance. Use of a single compound or cyanobacterial abundance alone cannot necessarily be used as an indicator of the presence or concentration of these compounds.
Measured concentrations of cyanobacteria and associated compounds were substantially higher than expected concentrations based on simple dilution models at some sites and substantially lower at others, though spatial and temporal patterns were unique for individual compounds. Data were not collected in such a way to determine whether differences between measured and expected concentrations were statistically significant. Results, however, indicate that simple dilution models were not sufficient to describe the downstream transport of cyanobacteria and associated compounds in the Kansas River.