Accurate water-table depth estimation using seismic refraction in areas of rapidly varying subsurface conditions
A common approach for estimating water-table depth simply and reliably is using seismic refraction. Typical layered solutions from seismic refraction are successful in areas where the water table does not fluctuate rapidly. However, there are many areas around the world where water tables rise and fall rapidly in response to intense rainfall events, especially in tropical and semi-arid regions. These areas are frequently heavily populated, and the shallow unconfined aquifers, used for drinking water and irrigation, are under increasing threat of overexploitation and pollution. To help mitigate these concerns, a reliable method to accurately determine water-table depth in these environments is required. We present a unique analysis method based on a calculated reference refraction velocity that gives not only accurate predictions of water-table depth for cases of rapidly fluctuating water tables but also accurate predictions of largest historical recorded depth to water. Data used came from a series of 60 refraction surveys carried out at 15 locations with monitoring bores in a tropical area. In a time-lapse approach, four surveys were performed at each of the bore locations over a nine-week period during both wet and dry conditions. Using a typical forward/reverse profile approach, we first calculate forward and reverse velocities. These layer velocities are then averaged using a scheme based on moisture conditions to give a reference velocity. Finally, the predicted depth for the water table is determined by identifying the depth at which this reference velocity occurs on a separately calculated refraction velocity tomogram. Results for predicting current water-table depth for the 60 surveys gave water-table depth predictions between 0.38 m above and 0.13 m below the measured water table using a 99% confidence interval. Conventional two-layer solution predictions resulted in only 28% of predictions lying within 1 m of the measured water table. Using a modification to this approach, we are also able to accurately determine the maximum historical recorded depth to water table using information on soil lithology/texture, records for the greatest recorded depth to water table, and refraction velocity tomograms at known locations. This method can then be applied to locations with similar but known lithology but without monitoring bores.