New Mexico Geological Society Annual Spring Meeting — Abstracts

This summary describes a portion of an ongoing stochastic groundwater modelling study of the Columbus Basin, New Mexico. Transmissivity estimates and the spatial correlation structure of transmissivity (T) are required inputs for the stochastic model.

There are two T values and 38 specific-capacity (C) values available for the portion of the Columbus Basin under study. Because many factors as sociated with the C-data values are unknown (e.g., the length of the pumping period), two linear regressions were performed in order to determine whether, there was a meaningful correlation between the variables T and C. One regression was for T vs. C and one was for log(T) vs. log(C). The two T values in the Columbus Basin had corresponding C values, and 14 additional pairs of data points were used from the adjoining Mimbres Graben and Deming Basin regions to the north and northwest of the Columbus Basin, respectively. The correlation coefficients for the standard and log-transformed regressions were 0.78 and 0.71, respectively. The uncertainty of a T value calculated from a C value using either of the regressions above can be quantified, using the variance of the regression.

Once the 38 C values in the Columbus Basin were converted to T's, a spatial correlation structure for the T values was sought assuming an isotropic statistical field. The variogram estimate for the non-transformed T values exhibited no correlation structure. In contrast. the variogram estimate for the log-transformed T data exhibited a clearly discernible correlation structure, and this structure can be used as input to the stochastic groundwater model.