Geology topics

Matthew D. Mahoney

Publications and source records attributed to Matthew D. Mahoney.

2 recordsLinked to original sources

USGS2021

Sediment concentrations and loads upstream from and through John Redmond Reservoir, east-central Kansas, 2010–19

Streambank erosion and reservoir sedimentation are primary concerns of resource managers in Kansas and throughout many regions of the United States and negatively affect flood control, water supply, and recreation. The Cottonwood and upper Neosho Rivers drain into John Redmond Reservoir, and since reservoir completion in 1964, there has been substantial conservation-pool sedimentation and storage loss in John Redmond Reservoir, causing storage capacity losses more rapidly than most other Federal reservoirs in Kansas. The U.S. Geological Survey (USGS), in cooperation with the Kansas Water Office, has monitored water quality (temperature, specific conductance, and turbidity) on the Cottonwood River (upstream from the reservoir) and Neosho River (upstream and downstream from the reservoir) since 2007 with additional sites added in 2009. The purpose of this report is to quantify suspended-sediment concentrations, loads, and yields entering and exiting John Redmond Reservoir during January 1, 2010, through December 31, 2019. Three water-quality monitoring sites were upstream from the reservoir (Cottonwood River near Plymouth, Kansas [USGS site 07182250; hereinafter referred to as “Cottonwood”]; Neosho River at Burlingame Road near Emporia, Kans. [USGS site 07179750; hereinafter referred to as “Burlingame”]; and Neosho River at Neosho Rapids, Kans. [USGS site 07182390; hereinafter referred to as “Neosho Rapids”]), and one water-quality monitoring site was downstream from the reservoir (Neosho River at Burlington, Kans. [USGS site 07182510; hereinafter referred to as “Burlington”]). The Neosho Rapids streamgage is downstream from the confluence of the Cottonwood and upper Neosho Rivers and has a contributing drainage area accounting for 91 percent of the total contributing drainage area to John Redmond Reservoir. Continuously measured streamflow, water quality, and discrete water-quality data were used to develop updated regression models to compute suspended-sediment concentrations, loads, and yields upstream and downstream from John Redmond Reservoir in east-central Kansas. Several turbidity sensors were deployed during the analysis period, and there are no established relations between the sensors; therefore, individual models for each sensor were developed. Model statistics for the turbidity and suspended-sediment concentration linear regression models were better (based on the coefficient of determination, root mean square error, and model standard percentage error) than the streamflow and suspended-sediment concentration linear regression models, indicating better model performance. Computed concentrations, loads, and yields do not account for the ungaged 9 percent of the drainage basin downstream from the Neosho Rapids streamgage. Mean daily suspended-sediment loads upstream from the reservoir were largest at Neosho Rapids (2,250 tons), second largest at Cottonwood (2,180 tons), and smallest at Burlingame (624 tons). Streamflow at Burlington was predominately regulated by reservoir releases, and mean daily suspended-sediment loads were smaller (286 tons) than at upstream sites. Among the upstream sites, Cottonwood had the largest mean daily suspended-sediment concentration (179 milligrams per liter [mg/L]), followed by Neosho Rapids (162 mg/L), and Burlingame (108 mg/L). Burlington had the smallest mean daily suspended-sediment concentration of all sites (46 mg/L). Annual reservoir trapping efficiency ranged from 82 to 94 percent, and the largest sediment mass trapped was during 2019 (2,230,000 tons). Reservoir storage decreased an estimated 7,750 acre-feet during 2010 and 2014–19. Using the mean trapping efficiency to estimate suspended-sediment loads during years with missing data (2011–13), the total estimated reservoir storage lost to sedimentation for the analysis period (2010–19) was 8,690 acre-feet, about 17 percent of the remaining storage space reported in 2007. The mean annual sedimentation rate during the analysis period (747 acre-feet per year) was about 85 percent larger than the design sedimentation rate (404 acre-feet per year) originally projected during construction. Different reservoir outflow management strategies, including operating near normal capacity as opposed to higher flood pool levels, could reduce the total reservoir storage lost by 3 percent (about 261 acre-feet), which is equal to 14 percent of the total sediment removed during the dredging operation in 2016. During the study period, about 56 percent of the total suspended-sediment load was transported during streamflows greater than the National Weather Service flood action stage at the upstream sites (0.1–5 percent of the record; Cottonwood mean: 48 percent; Burlingame mean: 40 percent; Neosho Rapids mean: 78 percent). Disproportionately large sediment loads were delivered during short periods of time, and localized efforts of stream erosion protection (streambank stabilization, riparian buffers) were likely to be overwhelmed. Precipitation frequency and intensity are projected to continue to increase in this region; therefore, future sediment reduction strategies that account for extreme episodic events may be beneficial. Changes to reservoir outflow management could also minimize sediment accumulation while still preserving flood control. Continued investigation of sediment reduction measures is necessary for future mitigation with the understanding that sedimentation rate is largely driven by high flows. Results from this study can be used to calibrate sediment models, explore sediment reduction strategies, highlight the importance of continued water-quality monitoring to determine effectiveness and changes in sediment transport, and assess the ability of John Redmond Reservoir to support designated uses into the future.

Kansas↗

USGS2018

Water-quality conditions with an emphasis on cyanobacteria and associated toxins and taste-and-odor compounds in the Kansas River, Kansas, July 2012 through September 2016

Cyanobacteria cause a multitude of water-quality concerns, including the potential to produce toxins and taste-and-odor compounds that 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. Extensive cyanobacterial blooms typically do not develop in the Kansas River; however, reservoirs in the lower Kansas River Basin occasionally develop blooms that may affect downstream water quality. During July 2012 through September 2016, continuous and (or) discrete water-quality data were collected at several sites (Wamego, De Soto, and three main reservoir-fed tributaries) on the Kansas River to characterize the sources, frequency and magnitude of occurrence, and causes of cyanobacteria, cyanobacterial toxins, and taste-and-odor compounds and to develop a real-time notification system of changing water-quality conditions that may affect drinking-water treatment. Algal biomass, as estimated by chlorophyll, was consistently higher at the downstream De Soto site than the upstream Wamego site. Higher algal biomass at the De Soto site likely was caused by algal growth during downstream transport without major losses due to grazing by aquatic organisms or other processes. Algal biomass at the Wamego and De Soto sites was negatively correlated with streamflow and total and bioavailable nutrient concentrations. The negative association between algal biomass and nutrients in the Kansas River likely reflects the relatively strong positive association between nutrient concentrations and streamflows. Cyanobacteria were relatively common in the Kansas River but rarely dominated the algal community. Like overall algal biomass, cyanobacterial abundances typically were higher at the De Soto site than the Wamego site. Cyanobacterial abundances generally peaked in late summer or early fall (July through October), with smaller peaks occasionally observed in spring (April through May). Cyanobacteria in the Kansas River rarely exceeded 20,000 cells per milliliter, the abundance at which cyanobacteria may become a concern for drinking-water treatment. Relations between cyanobacterial abundance and streamflow, turbidity, and nutrients in the Kansas River were similar to those between chlorophyll and total phytoplankton abundance, indicating the same processes that influence overall algal biomass and dynamics also are influencing cyanobacteria. The cyanotoxin microcystin was detected in about 27 percent of the samples collected from Kansas River tributary and main-stem sites. Cylindrospermopsin was detected in one sample from the De Soto site. Microcystin occurrence and concentration were similar between the Wamego and De Soto sites. Concentrations exceeded the U.S. Environmental Protection Agency health advisory guidance values for finished drinking water of 0.3 (for bottle-fed infants and pre-school children) and 1.6 micrograms per liter (μg/L; for school-age children and adults) in 6 percent or less of samples collected. These guidance values are for finished drinking water and are not directly applicable to observed environmental concentrations but do provide a benchmark for comparison. Microcystin was detected most often and had the highest concentrations during summer. Though seasonal patterns in microcystin occurrence were generally consistent, seasonal maxima varied by an order of magnitude across years. The taste-and-odor compounds geosmin and 2-methylisoborneol (MIB) were detected in about 78 and 43 percent of samples, respectively, collected across all sites (main stem and tributaries). Geosmin and MIB occurrence and concentration varied considerably between the Wamego and De Soto sites. Geosmin was detected in about 67 percent of Wamego samples and 81 percent of De Soto samples. The human detection threshold of 5 nanograms per liter (ng/L) was exceeded for geosmin in about 11 and 17 percent of the samples collected at the Wamego and De Soto sites, respectively. Geosmin was detected during all months of the year at both sites, and there were no clear seasonal patterns. MIB was detected less frequently in the Kansas River than geosmin and was observed in about 42 percent of Wamego samples and 33 percent of De Soto samples. Concentrations exceeded 5 ng/L in about 7 and 5 percent of samples from the Wamego and De Soto sites, respectively. As observed for geosmin, there were no clear seasonal patterns in MIB occurrence or concentration. There seems to be a connection between microcystin detections in the Kansas River and occurrence of microcystin in upstream reservoirs (and tributary streams). Microcystin concentrations greater than 0.3 μg/L may be likely during the summer when streamflow is less than 3,000 cubic feet per second (ft 3 /s) and contributions from Milford Lake exceed about 30 percent of total flow in the Kansas River. Observed microcystin concentrations typically were higher at the De Soto site than the Wamego or tributary sites during 2012 through 2016, indicating cyanobacteria may continue to grow and produce microcystin once introduced to the Kansas River. The spatial and temporal patterns in geosmin and MIB occurrence and concentration were more complex than microcystin. There were no clear connections between geosmin and MIB occurrence in the Kansas River and potential upstream reservoir (or tributary stream) sources. Likewise, there was not a clear relation between algal biomass, cyanobacteria, or actinomycetes bacteria and taste-and-odor events in the Kansas River. Geosmin and MIB were not strongly correlated with any measured environmental variable at either Kansas River site. Continuous water-quality data may be used independently or in combination with regression models to provide information on changing water-quality conditions that may affect drinking-water treatment processes or recreational activities on the Kansas River. For example, logistic regression model outputs and continuous water-quality data may both be indicative of the potential for microcystin events. Logistic regression models that are estimating a high probability of microcystin occurrence at concentrations above 0.1 μg/L can be used as one indicator. Streamflows less than 3,000 ft 3 /s during upstream reservoir releases during periods with low turbidity and increased chlorophyll fluorescence, specific conductance, and pH values may also be indicative of microcystin events. Advanced or near-real-time notification may inform proactive, rather than reactive, management strategies when water-quality conditions are changing rapidly or are likely to cause cyanobacteria-related events.

Kansas↗