GAMMA RADIOMETRIC SURVEYS OF SOIL DERIVED FROM INTERBEDDED GEOLOGIC MATERIALS ON THE FIELD SCALE

Gamma radiometric data consist of the passive gamma radiation that is naturally emitted from rocks and the soil. A sensor, with either a sodium iodide or bismuth germinate oxide crystal inside, is used to detect concentrations of four elements (five isotopes): ¹³⁷Cs, ²³²Th, ⁴⁰K, ²³⁸U and ²³⁵U. Gamma rays have variable energies which are characteristic of the radioactive isotope. Gamma rays are known to be related to mineralogy and geochemistry, such that it has been used to distinguish differences in (i) soil parent materials, (ii) weathering and pedogenesis, (iii) geomorphic activity, and (iv) erosion and deposition. However, gamma ray emissions are random due to the excitation of the elections, which suggests that at any given time the pattern of emissions coming from the land surface could be different. To investigate this, three surface surveys were completed by attaching a gamma instrument to a vehicle, and was driven at approximately 5 mph, with transects 3 m apart across a field (32 ha) in Lehighton, PA. Here the beds are vertical and geology varies horizontally across a field in less than a kilometer in this area of Pennsylvania. Randomly stratified soil samples (50 samples) to a depth of 30 cm were collected. Soil analysis consist of: pH, particle size distribution, organic carbon (loss on ignition), plant available P, K, Ca, Mg, Na and S (Mehlich III), and elemental composition (portable X-ray fluorescence). The data from these three surveys were interpolated using inverse distance weighting then subtracted from each other to assess the differences. The calculated differences between the surveys will be evaluated for significance using one way ANOVA. We also investigated if gamma radiometrics could capture soil differences on a field scale. Using a semivariogram/covariance cloud to acquire the threshold distance for the hot spot analysis to find the significant differences across the field, the soils samples will be used for training and validation of this analysis.