Test applications of a predictive model for delineation of fluvial sand bodies using geologically based calculations

A model for minimizing data needed to define a discontinuous sand body, specifically a fluvial, meandering sand body, was applied to hypothetical exploration programs. The model used geologically based statistical methods to guide the exploration program. These methods included establishing environment of deposition from stratigraphic data, which was then used in conjunction with predictive formulas to estimate sand body width. With an estimate of sand body width, probability of encountering the sand body again was established, which was combined with errors of sand body thickness estimates produced by kriging, to establish data point locations. ocedures of the model were followed until the sand bodies were defined. The model was applied to an existing data set at the Rocky Mountain Arsenal in Colorado. The sand body at the Rocky Mountain Arsenal was defined to within the standard deviation of the sand thicknesses less than that of the whole data set. However, the exploration already conducted at the Arsenal had not defined the sand body in this way, thus, a planned comparison for accuracy could not be made. This resulted in a supplemental site being used for comparison purposes. the Arsenal site did support the stratigraphic determination portion of the model, the supplemental site was chosen as a modern floodplain whose width could be easily established, and used as a hypothetical sand body. The site chosen was a portion of the Brazos River floodplain in Texas. Hypothetical exploration programs were then conducted to define the sand body, using the modelʹs method of exploration and a typical grid method of exploration. The results of each method were then compared. The comparison showed that the modelʹs method of exploration had defined the hypothetical sand body with fewer data points, but with similar accuracy to that of the grid method of exploration. Thus, the modelʹs method for exploration can be used to select data point locations in defining a sand body deposited by a fluvial meandering system by following the sand body into and through a site and with fewer data point locations.