Optimizing gamma-ray spectrometers for UAV-borne surveys with geophysical applications.

Heavy duty unmanned aerial vehicles (UAVs) have made it possible to fly with large gamma-ray spectrometers that weigh several kilograms. Moreover, they can be purchased at an affordable price. These large UAV-borne gamma-ray detection systems are used to map the naturally occurring radionuclides K, U, Th. Such platforms have the advantage that they can be deployed over terrain that is difficult to access, while still maintaining a high spatial resolution. In contrast to UAV-borne radioactive pollution studies, the naturally occurring radionuclides have a much lower activity and therefore require longer integration time, slower flying speed or a larger detector, in order to effectively determine the spatial radionuclide distribution. Therefore, the question arises: what is the minimum practical detector size required to successfully map K, U and Th concentrations from UAV platforms. In this study an agricultural field has been mapped with three different scintillator-based gamma-ray spectrometers: a 2000 ml, 1000 ml, and 350 ml detector. They were mounted together on the same UAV. At a flying height of 20 m and a speed of 5.6 m s the field was mapped. The various aerial measurements were compared to each other and to the ground-based measurements. The field had a low spatial variation in the K concentration (relative standard deviation (RSD) = 9%) and a larger variation for U and Th concentrations (RSD = 24% and 31% respectively). Radionuclide concentrations have been extracted from the survey data by Full Spectrum Analysis (FSA). Uncertainties and variances of the radionuclides have been determined by using two methods. Firstly, they are calculated directly from the FSA output and secondly they are extracted from a variogram. The latter incorporates spatial variation and was shown to provide a lower uncertainty. When using small detectors, the former approach could lead to the conclusion that the uncertainty is larger than the variance, while the variogram approach does capture the spatial variation. All three detectors were able to characterize the spatial distribution of the Th concentration. It is shown that the Th concentration is a good predictor of the sand and clay fraction of the topsoil in the field. By comparing the UAV-borne measurements to the ground-based measurements it is found that UAV-borne measurements at 20 m height are less sensitive to extreme values than ground-based measurements and they have the tendency to shift to the mean concentration of the area. The results of this study can be used to optimize the detector volume, survey height, and survey speed to maintain an acceptable accuracy for gamma-ray studies with small UAV-borne detectors.