Hydrostratigraphic Drilling Record Assessment (HyDRA): Assessing the Consistency and Quantitative Utility of Water Well Drillers’ Logs

Thursday, 18 December 2014
Geoffrey Bohling, Univ Kansas, Lawrence, KS, United States, Carolyn F Helm, University of Kansas, Lawrence, KS, United States and James J Butler Jr, University of Kansas, Kansas Geological Survey, Lawrence, KS, United States
The Hydrostratigraphic Drilling Record Assessment (HyDRA) project is a three-year study to develop improved methods for building groundwater flow models from drillers’ logs. Lithologic logs recorded by water well drillers represent a voluminous source of information regarding hydrostratigraphy. However, developing quantitative models from drillers’ logs is challenging due to the idiosyncratic nature of each driller’s approach to describing sediments and lithologies as well as variability in the amount of care invested in the description process. This presentation uses three approaches to assess the consistency and utility of drillers’ logs from 250 wells in the vicinity of a continuously monitored “index” well in the High Plains Aquifer in Thomas County, Kansas. The first assessment procedure will examine logs from wells in the vicinity of the index well to determine whether they show evidence of lateral confinement of a region immediately surrounding the index well, as seems to be indicated by the index well hydrograph. The second will apply a cross-validation procedure to determine the degree of consistency among logs at different wells and identify logs that are most out of keeping with logs at nearby wells. The logs are cast in quantitative terms by first representing the sediment descriptions using 72 standardized lithology terms, further categorizing the standardized lithologies into five hydraulic property categories, and then computing the proportions of the hydraulic property categories over regular ten-foot-intervals in each well. The cross-validation procedure involves using a cross-entropy measure to compare the actual category proportions in each well to those interpolated from neighboring wells. Finally, results of a groundwater flow model using property fields developed from the drillers’ logs will be briefly discussed. Comparisons between observed and simulated water levels at the index well and other continuously and annually monitored wells in the area serve as a basis for assessing the utility of the drillers’ logs in development of groundwater flow models.