A GIS-BASED WORKFLOW FOR BUILDING A 3-D AQUIFER MODEL FROM GEOLOGIC DATA

Developing physically realistic geometric configurations for aquifer model development is a challenging and time-consuming task. Here, we combine GIS, structure contour, and injection well data to develop a model mesh for simulating regional-scale oilfield wastewater disposal operations. The study area comprises ~120,000 km² of north-central Oklahoma and southeastern Kansas. The model extends to ~9 km below ground surface. Surfaces are prepared for the top and bottom of the late-Cambrian to early-Ordovician Arbuckle Group using data conversion, interpolation, and raster tools within the ESRI ArcGIS software package. Additional model layers, including Arbuckle overburden and Precambrian basement, are derived from USGS digital elevation models and extended from the Arbuckle bottom surface, respectively. Injection well data for the years 2011-2018 are formatted with geoprocessing tools and custom shell scripts, and then incorporated into the PetraSim numerical simulation code as well table input files. The model geometry is then discretized within PetraSim using Voronoi tessellation, which results in a mesh with bipartite resolution and time-dependent mass generation terms for grid cells containing injection wells. This visual mesh generation workflow permits high-resolution discretization within the study area to facilitate enhanced radial flow accuracy, as well as a lower resolution buffer area that mitigates nonphysical boundary condition feedbacks and optimizes computational load. The methodology presented herein can be used to develop high accuracy 3-D subsurface models for a wide variety of energy, environmental, and engineering applications.