Near real-time soil erosion mapping through mobile gamma-ray spectroscopy

Soil erosion has been associated with various negative environmental impacts foremost of which is the potential pressure it could impose on global food security. The poor conditions of our agricultural soil can be attributed to years of unsustainable farming practices occurring throughout history that has placed significant pressure on the environment. Moreover, climate change scenarios indicate further intensification which is likely making prediction and assessment of erosion processes critical for long term agricultural sustainability. This study demonstrates the potential of mobile gamma-ray spectrometry with large volume NaI(Tl) detectors to identify, at high spatial resolution, changes in ¹³⁷Cs soil concentration within the ploughed layer of soil and enabling the soil erosion processes to be quantified. This technique represents a significant advantage over conventional spatially-isolated point measurements such as soil sampling as it offers real time mapping at the field scale. However, spectral signal derived from measurements in the field are highly dependent on the calibration procedure used and are particularly sensitive to source-detector changes such as the presence of a vehicle, ground curvature and soil moisture content. Conventional calibration procedures tend to not consider these potential sources of uncertainty potentially leaving the system vulnerable to systematic uncertainties, especially when 137Cs concentrations are low. This study used Monte Carlo simulations to investigate such changes utilising additional information including a high-resolution digital terrain model. The method was demonstrated on a ploughed site in Scotland, revealing a mixture of tillage and water erosion patterns supported by soil core data. Findings showed that the sites topography had relatively little effect (<10%) on calculated erosion rates, but moisture content could be the determining factor, albeit very difficult to measure reliably throughout a survey.