A Bayesian Approach to Infer Nitrogen Loading Rates from Crop and Landuse Types Surrounding Private Wells in the Central Valley, California

Abstract. Nitrate contamination of alluvial aquifers in agricultural areas is a typical and major problem around the world. Nitrogen applied to crops, in the form of synthetic fertilizers or manure, in excess of plant uptake, largely leaches to groundwater in the form of nitrate, which is stable and highly mobile in oxygen-rich groundwaters. Increased awareness of the impact that excess nitrogen has had on groundwater and major health concerns about nitrate are prompting new regulations for farmers, e.g., in Europe and California, USA. This study is focused in the Central Valley, California, USA, an intensively farmed region with high agricultural crop diversity. Though nitrogen loading rates for several crop and landuse types in the Central Valley have been estimated or measured in a handful of studies, nitrogen loading rates for specific crop or landuse types and their impact to groundwater quality remain largely unknown. Knowledge of crop or other landuse specific groundwater nitrate impact may aid future regulatory decisions. Nitrogen loading rates for specific crop or landuse types are expected to vary depending on individual landuse practices; and interactions with hydrogeologic parameters that may promote or inhibit nitrate leaching. In this study, we developed a novel Bayesian regression model that allowed us to estimate crop or other landuse-specific groundwater nitrogen loading rate probability distributions from surveys of private wells, each of which is likely impacted by more than one landuse. We used recent nitrate measurements from 2149 wells in the Central Valley. We estimated nitrogen loading rate distributions for 15 crop and landuse groups. These were shown to compare favorably with prior mass-balance estimates of loading rates based on agronomic estimates of nitrogen loading.