SOFTWARE TOOLS FOR CONSTRAINING GROUND WATER MODELS USING GEOLOGY TO IMPROVE RESOURCE MANAGEMENT

Groundwater management can be improved by using geologic processes and characteristics to constrain ground-water models. Application of such constraints can be facilitated through appropriately designed software. Here we discuss software with five relevant capabilities. First is the ability to define a geologic framework independently of numerical grid considerations. In its most general sense, this involves the advances in GIS and visualization capabilities of recent decades. Ground-water models can be designed to coordinate with these capabilities. For example, the Hydrogeologic-Unit Flow (HUF) Package of MODFLOW allows hydrogeologic units to be defined independently of model layers. An example shows the effects of this capability on time of travel calculations. Second, the use of a full hydraulic-conductivity tensor can be important to representing the effects of fractures and bedding planes. The layer variable-direction anisotropy (LVDA) capability of the MODFLOW HUF Package supports a full tensor within each model layer; SUTRA supports a fully three-dimensional hydraulic conductivity tensor. Third, the hydromechanics of geologic materials can result in subsidence, which can be simulated using the Subsidence (SUB) Package in MODFLOW. Fourth, geologic-based estimates of system properties can be incorporated into model calibration using the prior information capabilities of, for example, UCODE_2005. In addition, information on relative property values (such as that a silt deposit would be expected to have a smaller hydraulic conductivity than a sand deposit) can be included to evaluate alternative conceptual models using the multi-model analysis (MMA) software. Fifth, constructed models can be used to identify additional geologic data that would be most useful to predictions of interest using sensitivity analysis measures such as the parameter-predictions (PPr) statistic. Additional efforts not discussed here are transition probability methods. Taken together, the methods discussed and mentioned provide the beginning of a ground-water modeling paradigm in which greater constraints produce more trustworthy models for use in public policy and management decisions.