Compressional-wa ve velocity variation in the upper Maca ´ e formation: A well-log regression analysis study

In this contribution, a regression analysis methodology allows prediction of P-wave velocities in locations where sonic logs are for some reason unavailable. The proposed methodology uses fundamental geophysical well logs (i.e., nuclear and resistivity logs) for least-squares estimation of regression coefficients. Fractional effective porosity and clay volume as well as bulk density form the set of assumed independent variables in the corresponding regression models. Considering P-wave velocities varying according to linear, parabolic and multivariate linear models, the calibration process confirms previous results showing porosity as the most important dependence parameter in the regressions. When calibration is extended to a well far away from the original location, it reveals that combinations of certain obtained regressions can predict P-wave velocities in the formation under study with reasonable accuracy.