Water-quality variability and constituent transport and processes in streams of Johnson County, Kansas, using continuous monitoring and regression models, 2003-11

Abstract

The population of Johnson County, Kansas increased by about 24 percent between 2000 and 2012, making it one of the most rapidly developing areas of Kansas. The U.S. Geological Survey, in cooperation with the Johnson County Stormwater Management Program, began a comprehensive study of Johnson County streams in 2002 to evaluate and monitor changes in stream quality. The purpose of this report is to describe water-quality variability and constituent transport for streams representing the five largest watersheds in Johnson County, Kansas during 2003 through 2011. The watersheds ranged in urban development from 98.3 percent urban (Indian Creek) to 16.7 percent urban (Kill Creek). Water-quality conditions are quantified among the watersheds of similar size (50.1 square miles to 65.7 square miles) using continuous, in-stream measurements, and using regression models developed from continuous and discrete data. These data are used to quantify variability in concentrations and loads during changing streamflow and seasonal conditions, describe differences among sites, and assess water quality relative to water-quality standards and stream management goals. Water quality varied relative to streamflow conditions, urbanization in the upstream watershed, and contributions from wastewater treatment facilities and storm runoff. Generally, as percent impervious surface (a measure of urbanization) increased, streamflow yield increased. Water temperature of Indian Creek, the most urban site which is also downstream from wastewater facility discharges, was higher than the other sites about 50 percent of the time, particularly during winter months. Dissolved oxygen concentrations were less than the Kansas Department of Health and Environment minimum criterion of 5 milligrams per liter about 15 percent of the time at the Indian Creek site. Dissolved oxygen concentrations were less than the criterion about 10 percent of the time at the rural Blue River and Kill Creek sites, and less than 5 percent of the time at the other sites. Low dissolved oxygen at all sites generally coincided with lowest streamflow and warmer water temperatures. Hourly dissolved oxygen concentrations less than 5 milligrams per liter were measured at all sites every year, indicating that even under normal climate conditions in non-urban watersheds such as Kill Creek, dissolved oxygen concentrations may not meet State aquatic-life criterion. Specific conductance was nearly always highest in Indian and Mill Creeks, which were the most urban streams with the largest upstream discharges from wastewater treatment facilities. The largest chloride concentrations and variability were recorded at urban sites and during winter. Each winter during the study period, chloride concentrations in the most urban site, Indian Creek, exceeded the U.S. Environmental Protection Agency-recommended criterion of 230 milligrams per liter for at least 10 consecutive days. The U.S. Environmental Protection Agency-recommended ecoregion criterion for turbidity was exceeded 30 (Indian Creek) to 50 (Blue River) percent of the time. The highest average annual streamflow-weighted suspendedsediment concentration during the study period was in Mill Creek, which has undergone rapid development that likely contributed to higher sediment concentrations. One of the largest suspended-sediment load events in Indian Creek was recorded in early May 2007 when about 25 percent of the total annual sediment load was transported during a period of about 2.25 days. A simultaneous load event was recorded in Kill Creek, when about 75 percent of the total annual sediment load was transported. Sediment yields generally increased as percent impervious surface increased. Computed hourly total nitrogen and total phosphorus concentrations and yields and streamflow-weighted concentrations generally were largest in Indian and Mill Creeks. Annual percent contribution of total nitrogen in the Blue River from wastewater treatment facility discharges ranged from 19 percent in 2010 to 60 percent in 2006. Annual percent contribution of total nitrogen in Indian Creek from wastewater treatment facility discharges ranged from 35 percent in 2010 to 93 percent in 2006. The largest percent nutrient contributions from wastewater discharges coincided with the smallest annual precipitation and streamflow volume, resulting in less contribution originating from runoff. Fecal indicator bacteria Escherichia coli density at the urban Indian Creek site was usually the largest of the five monitoring sites, with an annual median density that consistently exceeded the State primary contact criterion value but was less than the secondary contact criterion. Less than 1 percent of the total annual bacteria load in the Blue River and Indian Creek originated from wastewater discharges, except during 2006 when about 6 percent of the Indian Creek load originated from wastewater. Continuous water-quality monitoring provides a foundation for comprehensive evaluation and understanding of variability and loading characteristics in streams in Johnson County. Because several directly measured parameters are strongly correlated with particular constituents of interest, regression models provide a valuable tool for evaluating variability and loading on the basis of computed continuous data. Continuous data are particularly useful for characterizing nonpoint-source contributions from stormwater runoff. Transmission of continuous data in real-time makes it possible to rapidly detect and respond to potential environmental concerns. As monitoring technologies continue to improve, so does the ability to monitor additional constituents of interest, with smaller measurement error, and at lower operational cost. Continuous water-quality data including model information and computed concentrations and loads during the study period are available at http://nrtwq.usgs.gov/ks/ .