Development of Multiple Regression Models to Predict Nitrate Concentrations in Nebraska Surface Waters

Abstract. Nitrate in Nebraska streams contribute to excess nitrogen loads in the Mississippi River and Gulf of Mexico. Eighty-four sites across Nebraska were sampled an average of 202 times from 2002 to 2014. The objectives of this study were to (1) evaluate the spatial distribution of stream water nitrate concentrations ([NO₃^-]) in watersheds across Nebraska and (2) develop a multiple-regression equation to explain [NO₃^-] as a function of variables for which large datasets are easily accessible. Of the 84 sampling locations, 47 satisfied our criteria to be included in the development of the multiple-regression models. Analysis was conducted on all 47 watersheds and on subsets of “small” (<1,000 km², n = 22) and “large” (>1,000 km², n = 25) watersheds. Of the 14 variables evaluated individually, six to eleven were significant predictors, which were then used to develop multiple-regression models for the three data sets. The best model for each dataset was selected by maximizing R² and the adjusted R². While land use was a significant predictor in each data set, precipitation and soil properties were significant predictors for small and large watersheds, respectively. The adjusted R² was 0.44, 0.77 and 0.54 for the small, large, and combined watersheds, respectively. The simple and multiple-regression models developed in this study can be applied throughout Nebraska to obtain good estimates of [NO₃^-] within un-sampled watersheds, or for large watersheds where sampling is sparse. Furthermore, the methodology developed and applied in this study can be easily replicated in other regions.