Soil moisture time series using gamma-ray spectrometry detection representing a scale of tens-of-meters

Recently, stationary gamma-ray spectrometry (GRS) detection was introduced in the field of hydrology for continuous soil moisture estimation at a unique spatial scale of tens-of-meters. This method offers promise for applications such as early warning systems for emerging droughts and floods, ecosystem health evaluation, and optimizing agricultural practices. Initial research has primarily focused on K-40 concentrations measured at agricultural fields, with limited validation against independent soil moisture measurements. The objective of this study is to perform a comprehensive evaluation of the GRS soil moisture method. We estimate soil moisture with three conversion functions using the apparent K-40, Th-232, and Cs-137 concentrations from three field sites. The performance is assessed using three independent soil moisture products of different spatial scale and temporal resolution. We obtain GRS soil moisture estimates with high accuracy (average rmse-value = 0.036 [-]). Across all sites, the most accurate soil moisture estimates, rmse-value = 0.026 [-], were found for the K-40 concentrations using the modified Baldoncini equation, here named the Becker equation. This equation introduces an extra fitting parameter, that allows to adjust the slope of the conversion function. A high accuracy was also found for Th-232 and Cs-137 estimated soil moisture time series using the Becker equation, with rmse-values of 0.033 [-] and 0.040 [-], respectively. An important finding was the strong correlation between cosmic-ray neutron intensities and GRS nuclide concentrations across all sites. This is promising as the two methods rely on completely different physics, and thereby co-validate each other.