Water quality assessment in the Cherry Creek watershed: Patterns of nutrient runoff in an agricultural watershed

Abstract

Access to safe, high quality water for consumption, agriculture, industry, and recreation is critically important. Continuous agricultural and mining activities have impaired the waters of the Grand Lake watershed in the central Great Plains region of the United States. The Grand Lake watershed encompasses portions of southeast Kansas, southwest Missouri, northwest Arkansas, and northeast Oklahoma, and drains into Grand Lake in northeast Oklahoma. The Cherry Creek watershed drains approximately 882.2 km² (218,000 ac) of land in southeast Kansas and is a contributor of water to the Grand Lake watershed. This paper presents a water quality assessment in the Cherry Creek watershed, with an end toward mitigation of nonpoint source pollutants that are a major contributor to sediment and nutrient contaminants in Grand Lake. A hydrological model was developed using the Hydrological Simulation Program Fortran code and was updated, calibrated, and verified with measured data reported by US Geological Survey (USGS) and the Kansas Department of Health and Environment (KDHE). The model was extended to simulate water quality within the study area. Nitrate (NO₃⁻-N), total ammonia (TAM), total phosphorus (TP), and orthophosphate (PO₄³⁻-P) concentrations measured at the USGS stations were used to calibrate the model, and concentrations reported by KDHE at downstream locations were used to verify the model. Results indicate good performance of the hydrological model as tests of fitness were within levels established in previous studies (root-mean-squared-error to standard deviation ratio [RSR] < 0.75; −30% < percentage bias [PBIAS] < 30%; R² > 0.6; NS > 0.5). Nutrient measurements below the minimum quantifiable limit (MQL) hampered precise simulation of nutrient changes, though simulated values were acceptable in terms of ranges of contaminant concentration values and seasonal trends. The calibrated model was used to estimate the probability that nutrient levels would exceed established water quality criteria for rivers and streams. Concentration values of NO₃⁻-N and TAM are shown to be low for an agricultural watershed: 75% of the NO₃⁻-N and TAM concentrations are lower than 0.41 mg L⁻¹. The probability of NO₃⁻-N and TAM concentrations being toxic for aquatic communities is lower than 9%. Although model-estimated PO₄³⁻-P concentrations were low in numerical value (ranging from 0 to 2.60 mg L⁻¹), they could still promote eutrophication according to the accepted 0.05 mg L⁻¹ criteria for maximum PO₄³⁻-P concentrations in streams. Concentrations of PO₄³⁻-P in Cherry Creek have a 30.4% probability of exceeding that threshold. Extensive use of animal manure (primarily poultry) or manure from cattle grazing on pasture may account for the elevated levels of PO₄³⁻-P observed in the watershed. Although nutrient runoff from agricultural watersheds is anticipated to be a major contributor to elevated stream nutrient loads, these results indicate minimal contamination from Cherry Creek.