Abstract: Survey Design and Data Integration When Using Surface Geochemistry

James M. Fausnaugh

Surface geochemistry can be a valuable extension to any exploration project when applied properly. Staging the exploration program so that samples are acquired judiciously is the most often prescribed survey practice. This minimizes cost while maximizing sample placement. However, due to the noisy nature of geochemical data, staging sample collection can present some problems. How can separate data sets, both reconnaissance and detailed, be integrated into one cohesive sample population?

Many of the geochemical methods currently in use can yield spurious anomalies. False anomalies can be detected if (1) the sample spacing is sufficiently close to determine the discontinuous nature of the anomaly, (2) a sufficient number of samples are taken to obtain an adequate population to differentiate an anomaly from background, (3) the method used meets or exceeds an expected signal to background ratio.

There are several statistical options available to test the significance of geochemical anomalies. Significance testing is related to frequency distribution and probability. Because geochemical data is usually log-normally distributed, data transformations are required so that inferences regarding a survey can be easily interpreted. Normalization procedures allow the integration of multiple data sets. Parametric statistical methods, such as z-transforms and linear regression, are often used so that the magnitude of the data is retained. Non-parametric methods, such as data ranking, can also be utilized and performs satisfactorily when delineating general trends of anomalies.

AAPG Search and Discovery Article #90950©1996 AAPG GCAGS 46th Annual Meeting, San Antonio, Texas