Assessing Recharge and Hydrostratigraphic Model Uncertainty in the Climax Mine Area of the Nevada Test Site

Hydrologic analyses are commonly based on a single conceptual model. Yet hydrologic environments are open and complex, rendering them prone to multiple interpretations and conceptualizations. Considering conceptual model uncertainty is a critical process for the assessment of hydrologic uncertainty. This study assesses recharge and geologic model uncertainty for the Climax Mine area of the Nevada Test Site, Nevada, where recharge processes and site geology are described by five recharge models and five hydrostratigraphic framework models (HFMs). Combining the recharge models and HFMs, a total of 25 conceptual models are formulated and incorporated into the Death Valley Regional Flow System modeling framework developed by the U.S. Geological Survey. The conceptual model uncertainty is quantified using a model averaging method. The method estimates model relative plausibility (measured by model probabilities) by incorporating expert judgment through expert elicitation and field observations of head and flow through inverse modeling. Parametric uncertainty of each model is assessed using Monte Carlo simulation, and each model’s mean predictions and associated predictive uncertainty for hydraulic head and flow are averaged using the model probabilities as weights. The posterior (averaged) variance, incorporating both parametric and conceptual model uncertainty, is compared with those of individual models. It is shown that model averaging provides larger uncertainty than individual models, indicating that more uncertainty is incorporated and model predictions are more scientifically defensible.