A multivariate analysis of covariance to determine the effects of near-stream best management practices on nitrogen and phosphorus concentrations on a dairy farm in the New York Conservation Effects Assessment Project watershed

Abstract

Near-stream best management practices (BMPs), such as exclusionary fencing or cattle crossings, are often recommended to improve water quality, but quantification of their impacts is limited. Surface and subsurface processes in these near-stream areas impact the contribution of nitrogen and phosphorus to watercourses and the effectiveness of near-stream BMPs. To test the impact of the near-stream BMPs, groundwater samples from 30 piezometers and streamwater samples along two adjacent creeks (one control and one with BMP treatment) were collected over a three-year period before and after installation of exclusionary fencing with a cattle crossing in the Catskill Mountains of New York State. Samples were analyzed for nitrate nitrogen (NO₃⁻-N), soluble reactive phosphorus (SRP), dissolved oxygen, and dissolved organic carbon. Analysis results and other ancillary variables (rainfall and groundwater table depth) were incorporated into a multivariate statistical model to evaluate the impact of the BMPs on the concentrations of NO₃⁻-N and SRP in the treatment creek. Results of the analysis indicate that the installation of the near-stream BMPs resulted in a 27% yearly (34% during the growing season) reduction in treatment creek SRP concentrations, while there was little impact on NO₃⁻-N concentrations. Incorporating the SRP concentrations measured in a nearby control creek and controlling for the effects of groundwater SRP levels and groundwater hydrology (water table height) had a significant effect on the overall analytical model performance. These results indicate that protecting near-stream areas from potentially pollution-causing practices can be an important means of controlling phosphorus levels in water bodies.