A STATISTICAL APPROACH TO ASSESSING HANDHELD GAMMA RAY SPECTROMETERS ON OUTCROP DATA

Handheld gamma-ray (GR) units are often used to collect information to determine a variety of geologic parameters. However, factors such as collector type, environmental conditions, calibration, stabilization methods, processing techniques, etc. may affect data quality and therefore could skew interpretations. Two different models of handheld GR spectrometers were used to collect three spectral assays (WBU, PH, and TTU) over the same outcrop section. A statistical approach was used to examine the variation in unit response variables (Total GR, Computed GR, K, Th, U, and elemental ratios) and determine if there was a significant difference between the assays, which could affect the interpretation of the outcrop. There was a significant difference between GR spectrometers using means from the previously mentioned variables (p<0.05, ANOVA). In addition, there are statistically significant differences in a majority of variables from the TTU survey, with essentially no difference between the WBU and PH surveys. The strongest variation was noted in the TTU Thorium counts. Correlation coefficients of Total GR for all surveys range from 0.22 to 0.95, however the correlation between WBU and PH surveys result in mostly strong, positive relationships. Experience from previous and recent surveys show environmental conditions, such as temperature appear to be a major issue with these units. Variation between units may also be caused by calibration or stabilization related issue. Invalid responses such as those recorded by the TTU instrument may result in incorrect lithological interpretation. For example the “error-reduced” Th values would inflate the Th/U ratio, indicating a more U rich environment and would similarly increase the Th/K ratio resulting in possible incorrect clay type interpretations. When used for correlation to regional or local markers the bad Th values would cause incorrect associations. Evaluation of differences in unit responses can provide an insight for adjustments that may be needed during interpretation and application of outcrop GR data.