Towards identifying data needs for a regional hydrogeologic contamination study using multiple realization simulations

This paper describes a stochastic, distributed parameter simulation approach which is being used to identify/prioritize data collection activities for a 250 km2 region containing numerous potential contamination sites. The region is located in the southeast part of the Albuquerque Basin in central New Mexico, USA. The Basin is part of the Rio Grande Trough, a large graben with large vertical displacements between the central basin and the adjacent highlands. Numerous potential human receptors are located around the periphery of the region, and it is the desired to have a groundwater monitoring well network.which can help provide early detection of contamination plumes as well as provide relevant data on the regional hydrogeologic framework. A 2D numerical model of the regional basin-fill aquifer is developed with explicit recognition of uncertainties in flow parameter spatial distributions. We account for uncertainty in the parameter field through Monte Carlo simulation. To reduce the computational burden of multiple realization simulation, we employ a linearized stochastic model which permits cosimulation of transmissivity and head fields, conditioning on both transmissivity and head. The locations of selected contamination sites within the region then are overlaid on the simulation results, and flow paths and groundwater travel times from the contamination sites are assessed using sensitivity analysis to identify preferred locations for collection of additional data and the types of data which should be collected (e.g., head, transmissivity, or geologic cores)