RESUMO
Several statistical techniques have been adopted in the analysis and interpretation of radiometric survey data set with little attention to the geostatistical approach. In this study, the geostatistical technique was used to analyse an airborne uranium data set in the North-western part of Ogun state to deduce the autocorrelation level of uranium with the corresponding geological composition through the geostatistical model fitting. Twelve uranium lithological (UL) units were identified and analysed using variogram and kriging. Three variogram models namely; spherical, exponential and gaussian, were utilised. The variogram results showed that the three models fit in the interpretation of uranium data in the order of Spherical < Gaussian < Exponential models for all the UL units except at UL6 which had range (R) values of 4227.3 m, 5318.2 m and 5636.4 m for Gaussian, Spherical and Exponential models, respectively. The Nugget-Sill ratios (NSR) were in the range of 3.57-18.97% for the three models across the 12 UL units. These results indicate strong spatial autocorrelation/dependency of uranium concentration on geological composition. The ordinary kriging estimator revealed that the weathered basement formation has the highest uranium concentration compared to basement and sedimentary formations across the study area. The coefficient of variation results also revealed homogeneity in the uranium source across the UL units. The geostatistics technique is a useful mathematical tool in the analysis and interpretation of uranium concentration in the North-western part of Ogun State.