On the Information Content of Cosmic‐Ray Neutron Data in the Inverse Estimation of Soil Hydraulic Properties

CORE IDEAS: Cosmic‐ray neutron data are used to inversely estimate soil hydraulic properties. The forward neutron operator COSMIC is coupled with the vadose zone model HYDRUS‐1D. Bayesian analyses confirm the information content of cosmic‐ray neutron data. Observations of soil moisture content from remote sensing platforms can be used in conjunction with hydrological models to inversely estimate soil hydraulic properties (SHPs). In recent years, cosmic‐ray neutron sensing (CRNS) has proven to be a reliable method for the estimation of area‐average soil moisture at field scales. However, its use in the inverse estimation of the effective SHPs is largely unexplored. Thus, the main objective of this study was to assess the information content of aboveground fast‐neutron counts to estimate SHPs using both a synthetic modeling study and actual experimental data from the Rollesbroich catchment in Germany. For this, the forward neutron operator COSMIC was externally coupled with the hydrological model HYDRUS‐1D. The coupled model was combined with the Affine Invariant Ensemble Sampler to calculate the posterior distributions of effective soil hydraulic parameters as well as the model‐predictive uncertainty for different synthetic and experimental scenarios. Measured water contents at different depths were used to assess estimated SHPs. The analysis of both synthetic and actual CRNS data from homogenous and heterogeneous soil profiles, respectively, led to confident estimations of the shape parameters α and n, while higher uncertainty was observed for the saturated hydraulic conductivity. Furthermore, results demonstrated that neutron data are less influenced by local sources of uncertainty compared with near‐surface point measurements. The simultaneous use of CRNS and water content data further reduced the overall uncertainty, opening up new perspectives for the combination of CRNS with other remote sensing techniques for the inverse estimation of the effective SHPs.