Soil variables as auxiliary information in spatial prediction of shallow water table levels for estimating recovered water volume

Abstract

Spatial data became increasingly utilized in many scientific fields due to the accessibility of monitoring data from different sources. In the case of hydrological mapping, measurements of external environmental conditions, such as soil, climate, vegetation, are often available in addition to the measurements of water characteristics. An integrated modelling approach capable to incorporate multiple input data sets that may have heterogeneous geometries and other error characteristics can be achieved using geostatistical techniques. In this study, different physical hydric properties of soils extensively sampled and topography were used as auxiliary information for making optimal, point-level inferences of water table depths in forest areas. We used data from 48 wells in the Bauru Aquifer System in the Santa Barbara Ecological Station (EEcSB), in the municipality of Aguas de Santa Barbara in Sao Paulo State, Brazil. Using the resistance of soil to penetration and topography as auxiliary variables helped reduce prediction errors. With the generated maps, it was possible to estimate the volumes of water recovered from the water table in two periods during the monitoring period. These values showed that 30% of the recovered volume would be sufficient for a three-month supply of water for a population of 30,000 inhabitants. Therefore, this raises the possibility of using areas such as the EEcSB as strategic supplies in artificial recharging management.