Bayesian nitrate source apportionment to individual groundwater wells in the Central Valley by use of elemental and isotopic tracers. Issue 7 (31st July 2016)

• Record Type: Journal Article
• Authors: Ransom, Katherine M.; Grote, Mark N.; Deinhart, Amanda; Eppich, Gary; Kendall, Carol; Sanborn, Matthew E.; Souders, A. Kate; Wimpenny, Joshua; Yin, Qing‐zhu; Young, Megan; Harter, Thomas
• Is Part Of: Water resources research. Volume 52:Issue 7(2016:Jul.)
• Page Start: 5577
• Page End: 5597
• Publication Date: 2016-07-31
• Subjects: groundwater -- nitrate -- isotopes -- tracers -- Bayesian -- iodine; Hydrology -- Periodicals; 333.91
• DOI: 10.1002/2015WR018523
• Languages: English
• ISSNs: 0043-1397

Abstract

Abstract: Groundwater quality is a concern in alluvial aquifers that underlie agricultural areas, such as in the San Joaquin Valley of California. Shallow domestic wells (less than 150 m deep) in agricultural areas are often contaminated by nitrate. Agricultural and rural nitrate sources include dairy manure, synthetic fertilizers, and septic waste. Knowledge of the relative proportion that each of these sources contributes to nitrate concentration in individual wells can aid future regulatory and land management decisions. We show that nitrogen and oxygen isotopes of nitrate, boron isotopes, and iodine concentrations are a useful, novel combination of groundwater tracers to differentiate between manure, fertilizers, septic waste, and natural sources of nitrate. Furthermore, in this work, we develop a new Bayesian mixing model in which these isotopic and elemental tracers were used to estimate the probability distribution of the fractional contributions of manure, fertilizers, septic waste, and natural sources to the nitrate concentration found in an individual well. The approach was applied to 56 nitrate‐impacted private domestic wells located in the San Joaquin Valley. Model analysis found that some domestic wells were clearly dominated by the manure source and suggests evidence for majority contributions from either the septic or fertilizer source for other wells. But, predictions of fractional contributions for septic and fertilizer sources were often of similar magnitude, perhaps because modeled uncertainty about the fraction of each was large. For validation of the Bayesian model, fractional estimates were compared to surrounding land use and estimated source contributions were broadly consistent with nearby land use types. Key Points: Of the four nitrate sources, our model shows the manure source is the most distinctive Septic and fertilizer waste were more difficult to distinguish Model predictions broadly correlate to historical land use