Measured and Modeled Soil Moisture Compared with Cosmic‐Ray Neutron Probe Estimates in a Mixed Forest

Abstract

Soil moisture is a key variable in most environmental processes, and the cosmic-ray neutron probe (CRNP) fills a niche for intermediate-scale soil moisture measurements. In this study, the CRNP estimated soil moisture was compared with a soil moisture measurement network including 108 time domain transmissometry (TDT) probes. We also used a Hydrus-1D numerical model of measured soil moisture at targeted locations by inversely fitting soil hydraulic parameters used to simulate soil moisture in the near surface (0.03 and 0.05 m) during the growing seasons of 2011 and 2012. Both simulated and TDT-measured soil moisture were used in constructing the depth-weighted mean areal soil moisture for comparison with the CRNP estimates. The results showed that near-surface soil moisture estimated by the numerical simulation improved the correlation between the sensor network and the CRNP estimation, especially during rainfall events. The CRNP estimates of soil moisture exhibited a dry bias under relatively wet conditions at the beginning of the snow-free period because of the almost binary spatial distribution of soil moisture. Using a combination of soil moisture measurements and near-surface simulations, the CRNP output was recalibrated to capture the wetter conditions, resulting in a RMSE (0.011 m3/m3) of less than half the original calibration RMSE (0.025 m3/m3). The calibration was validated using CRNP data from the 2013 growing season against independent soil moisture values.