BUILDING 3D HYDROGEOLOGIC FRAMEWORK MODELS IN AREAS OF DATA SPARSENESS

Computer-based hydrogeologic framework models (HFMs) are constructed from geospatially registered surface and subsurface geologic data. Limitations in HFMs exist because of the difficulties in the representation of complex geometry and spatial variability of heterogeneous hydrogeologic materials and geologic structures.

The information presented in an HFM include interpretations of the lithology in well logs and cross sections, geophysical interpretations, and isopach data for units. In some parts of the world, there are areas that are “data rich”, while in others (such as the developing world with restricted access to water) lithologic logs are sparse.

The Death Valley regional groundwater flow system (DVRFS) represents a data-rich HFM constructed in an area of about 80,000 km². Existing and new data were collected, analyzed, interpreted, and published as information products such as datasets and reports. The HFM contained 194 rows, 160 columns, and 27 hydrogeologic units (HGUs). Horizontal grid spacing was 1,500 m.

HFMs for areas in the developing world can be hampered by sparse and/or poor quality data. As an example, for an HFM constructed for northwest Kenya, only 67 of 3000 well logs from relatively shallow water-supply wells contained lithologic data and were properly located in an area of almost 140,000 km². The HFM contained 599 rows, 533 columns, and 5 HGUs. Horizontal grid spacing was 1,000 m. The overall volume of the Kenya HFM was greater than the DVRFS HFM even though it contained considerably less data for construction.

Accommodating these data limitations requires professional experience and judgement and an acknowledgement of the inherent uncertainty. Cross sections from various sources can be used (or created), and (if available) data from proprietary petroleum databases (lithologic as well as geophysical) could be utilized. However, much of the interpretation of the hydrogeology relies on tacit knowledge of the local geology and reliance on local geoscientists. Examples of HFMs constructed from data-rich and data-sparse areas are presented, as well as a “path forward” for data-sparse” areas.