The success rate for Devonian shale wells in southwestern West Virginia is greater than 90%, yet the geologic factors determining hydrocarbon distribution are poorly understood. Initial potentials (IP) vary, but are generally low. Siting a new well may have little more justification than proximity to successful wells with acceptable to high open flows. Such a procedure can be made objectively through the use of geostatistics that measure the degree of randomness in the distribution of dry penetrations and the spatial variation in IP's.

Semivariograms were calculated to show (1) average difference in success between wells relative to the distance separating them, and (2) average difference in open flow between wells relative to distance. The semivariogram for gas IP showed a large degree of noise, but some spatial autocorrelation. Similarly, dry penetrations were found to be clustered. Success probability and gas IP contoured from kriged estimates exhibit clustering of dry penetrations and nonrandom patterns in IP's, in particular, linear highs and lows paralleling structural features. Highest flows correspond to flanks of known anticlines. One linear trend in IP follows a lineament that might represent a cross-strike structural discontinuity. Significant correlation between IP at each well site with that predicted f om surrounding wells demonstrates the efficacy of the geostatistical approach.

End_of_Article - Last_Page 266------------