STATISTICAL ESTIMATION OF THE GROUDWATER TABLE BENETAH THE ST. LOUIS METROPOLITAN AREA USING GIS METHODS

The elevation of the permanent groundwater table and its position beneath the existing ground surface are important factors in engineering and environmental decision making for waste disposal sites and natural hazards, such as shaking induced liquefaction. Water table contouring is also used to estimate the preferred paths of groundwater flow as well as velocity.

The study area encompasses 29 7.5-minute quadrangles in the St. Louis Metropolitan area of Missouri and Illinois, with a land area of 4,432 km2. Groundwater elevation data were collected and analyzed to prepare a contour map illustrating the estimated elevation of the permanent groundwater table. The input data consisted of the following components: 1) 1046 well logs obtained from the Missouri and Illinois geological surveys, and; 2) 469 points along rivers interpolated from 10m grid digital elevation models (DEM) of each quadrangle, stitched together. The predicted elevation of the groundwater table are proportional to those of the undulating ground surface, and influenced by standing bodies of water, such as lakes and perennial river channels.

Using ArcGIS software, groundwater levels were interpolated by using geostatistical methods, such as ordinary kriging and cokriging, as well as power regression. Ordinary cokriging accounts for the influence of undulating ground elevation data measured at specific data points (wells), and extracted from 100m grid DEM as a second variable.

The power regression model is a simple mapping method that compares the similarity between the shapes of groundwater table and overlying ground surface. Ordinary cokriging provides the lower mean error between measured and estimated values than ordinary kriging, based on cross validation. The authors found that ordinary cokriging provides a more statistically accurate estimation of the local perturbations of a regional groundwater table than ordinary kriging.