Preliminary assessment of a water-quality monitoring program for total maximum daily loads in Johnson County, Kansas, January 2015 through June 2016

Abstract

Municipalities in Johnson County in northeastern Kansas are required to implement stormwater management programs to reduce pollutant discharges, protect water quality, and comply with applicable water-quality regulations in accordance with National Pollutant Discharge Elimination System permits for stormwater discharge. To this end, municipalities collect grab samples at streams entering and leaving their jurisdiction to determine levels of excessive nutrients, sediment, and fecal bacteria to characterize pollutants and understand the factors affecting them. In 2014, the U.S. Geological Survey and the Johnson County Stormwater Management Program, with input from the Kansas Department of Health and Environment, initiated a 5-year monitoring program to satisfy minimum sampling requirements for each municipality as described by new stormwater permits issued to Johnson County municipalities. The purpose of this report is to provide a preliminary assessment of the monitoring program. The monitoring program is described, a preliminary assessment of the monitoring program design is provided using water-quality data collected during the first 2 years of the program, and the ability of the current monitoring network and sampling plan to provide data sufficient to quantify improvements in water quality resulting from implemented and planned best management practices is evaluated. The information in this initial report may be used to evaluate changes in data collection methods while data collection is still ongoing that may lead to improved data utility. Discrete water-quality samples were collected at 27 sites and analyzed for nutrients, Escherichia coli ( E. coli ) bacteria, total suspended solids, and suspended-sediment concentration. In addition, continuous water-quality data (water temperature, pH, dissolved oxygen, specific conductance, turbidity, and nitrate plus nitrite) were collected at one site to characterize variability and provide a basis for comparison to discrete data. Base flow samples indicated that point sources are likely affecting nutrient concentrations and E. coli bacteria densities at several sites. Concentrations of all analytes in storm runoff samples were characterized by substantial variability among sites and samples. About one-half of the sites, representing different watersheds, had storm runoff samples with nitrogen concentrations greater than 10 milligrams per liter. About one-third of the sites, representing different watersheds, had storm runoff samples with total phosphorus concentrations greater than 3 milligrams per liter. Six sites had samples with E. coli densities greater than 100,000 colonies per 100 milliliters of water. Total suspended solids concentrations of about 12,000 milligrams per liter or greater occurred in samples from three sites. Data collected for this monitoring program may be useful for some general assessment purposes but may also be limited in potential to fully inform stormwater management activities. Valuable attributes of the monitoring program design included incorporating many sites across the county for comparisons among watersheds and municipalities, using fixed-stage samplers to collect multiple samples during single events, collection of base flow samples in addition to storm samples to isolate possible point sources from stormwater sources, and use of continuous monitors to characterize variability. Limiting attributes of the monitoring program design included location of monitoring sites along municipal boundaries to satisfy permit requirements rather than using watershed-based criteria such as locations of tributaries, potential pollutant sources, and implemented management practices. Additional limiting attributes include having a large number of widespread sampling locations, which presented logistical challenges for predicting localized rainfall and collecting and analyzing samples during short timeframes associated with storms, and collecting storm samples at fixed-stage elevations only during the rising limb of storms, which does not characterize conditions over the storm hydrograph. The small number of samples collected per site resulted in a sample size too small to be representative of site conditions, including seasonal and hydrologic variability, and insufficient for meaningful statistical analysis or site-specific modeling. Several measures could be taken to improve data utility and include redesigning the monitoring network according to watershed characteristics, incorporating a nested design in which data are collected at different scales (watershed, subwatershed, and best management practices), increasing sampling frequency, and combining different methods to allow for flexibility to focus on areas and conditions of particular interest. A monitoring design that would facilitate most of these improvements would be to focus efforts on a limited number of watersheds for several years, then cycle to the next set of watersheds for several years, eventually returning to previously monitored watersheds to document changes. Redesign of the water-quality monitoring program requires considerable effort and commitment from municipalities of Johnson County. However, the long-term benefit likely is a monitoring program that results in improved stream conditions and more effective management practices and efficient expenditure of resources.