Modeling Nitrate in Texas Groundwater at Regional and Aquifer Scales

Nitrate is considered the most prevalent contaminant in groundwater. Owing to its high solubility and mobility, it is prone to leaching through soils with infiltrating water. High nitrate concentration in groundwater is a health concern, particularly for infants. Predicting existing nitrate conditions can help focus regional and national water-management goals and monitoring strategies on the most susceptible areas. We present here a study to predict nitrate concentration in Texas groundwater and to characterize the uncertainty. This study is conducted at the High Plains Aquifer, Seymour Aquifer, and the Trinity Aquifer. Nitrate transport is described using a finite difference groundwater flow model (MODFLOW combined with MT3D). Flow paths and travel times of nitrate to reach the groundwater are analyzed using MODPATH. Despite advances in computational abilities and inverse modeling techniques, aquifer characterization remains an extremely difficult problem due to underlying complexities such as spatial heterogeneity, temporal variability, and coupled chemical, physical, and biological processes. We present here a study to incorporate the above mentioned complexities in aquifer characterization using the wavelet transform technique. Wavelet transformation enabled us to identify dominant spatial scales across temporal scales and to estimate the model parameters at each spatial scale. We also used the Ensemble Kalman Filter (EnKF) approach to continuously update these model parameters at each spatial scale while simultaneously providing an estimate of the uncertainty through data fusion. The results of our study demonstrate the capability of wavelet analysis in identifying dominant spatial scales while analyzing contamination of groundwater by nitrates. We also show that the performance of EnKF in terms of model prediction is a function of spatial scale.