A simple statistical method for correcting and standardizing heat flows and subsurface temperatures derived from log and test data

Abstract

Data from subsurface temperature measurements provide widely used and vital input data for maturity modeling_ Because maturity calculations are very sensitive to thermal history, and because reconstruction of the thermal history begins with the modern temperature profile, accurate knowledge of true formation temperatures is vital. Temperature data used in maturity modeling come from a variety of sources, including BHTs derived from single logging runs, BHTs obtained from multiple logging runs at the same depth and corrected using the Horner plot method, RFTs, DSTs, and production tests (PTs)_ However, temperatures obtained using most of these techniques require some correction before they represent true formation temperatures. Unfortunately, the need for these corrections is not generally recognised, leading most modelers to consistently underestimate modern subsurface temperatures_ Such errors can lead to major errors in subsequent calculations of hydrocarbon generation and cracking, and can thus have profound effects on exploration decisions_ In an effort to evaluate the accuracy of data from single logging runs, Horner plots, RFTs, and DSTs or PTs, an extensive temperature data base was developed for the Malay Basin. Basal heat flows calculated for many wells using each type of temperature data were compared. It was found that all other temperature data considerably underestimated subsurface temperatures compared to DSTIPT data, which were assumed to represent true formation temperatures. Results were analyzed using two different statistical approaches, which gave quite consistent conclusions, and average correction factors for each type of temperature data were developed. To be equivalent to heat flows calculated from DSTI PT temperatures, heat flows calculated using single BHT data already subjected to a standard 10% correction had to be corrected upward by an additional 16%, those calculated from Horner plot extrapolations by an additional 14%, and those obtained from uncorrected RFT data by 9%. Measured subsurface temperatures were corrected using a more complex set of equations that take surface temperature (T.) into account. The corrected subsurface temperature To is given by one of the following formulas, where Tb is the uncorrected temperature from a single logging run, Th is the extrapolated temperature obtained from a Horner plot, and Tr is the uncorrected RFT temperature. To = (1.1 eTb - T.)e1.16 + T. To = (Th - T.)e1.14 + T. To = (Tr - T.)e1.09 + T. Although these correction factors were developed for the Malay Basin, evidence presented by other workers suggests that corrections are needed in other basins as well, and that the magnitude of the corrections suggested here is reasonable for other areas. Future work should test these hypotheses and extend this calibration to ·other types of basins in other parts of the world. The correction factors established in this study only represent statistical averages, and cannot be expected to work well in all cases. Whenever DST or PT data are available, they should be weighted considerably more heavily than any other type of data, even those corrected by the best methods available. In the absence ofDST or PT data, however, these correction methods will greatly increase our confidence in subsurface temperatures from RFTs or from wireline logs.