Calibration of Cosmic Ray Neutron Sensors in different land-use/land covers: A modeling approach

The scale gap between in situ and satellite soil moisture dataset has persisted. Cosmic ray neutron sensors (CRNS) bridges this gap by providing soil moisture data at hectometer scale. CRNS calculates the number of thermalized neutrons created by collision between cosmic ray and hydrogen atoms present in a water molecule. However, extensive calibration is required to filter out the slow neutron count from all other hydrogen pools such as aboveground biomass and atmospheric water content. The objective of this study is to calibrate the CRNS by calculating the area average soil moisture in different land use/land covers using a sub-surface model. Three cosmos sensors were installed in the Brazos river basin representing different land use i) Traditional agriculture (Riesel) ii) Native Prairie (Stiles farm) and iii) Managed Prairie (Texas A&M experimental farm). All sites have a comprehensive network of distributed in-situ soil moisture sensors, eddy covariance towers, rain gauges and phenocams that were used to set up boundary conditions. Using profile soil moisture sensors at three different locations in each land cover, inverse-modeling was performed using the Shuffled Complex Evolution Algorithm to determine soil hydraulic parameters in Hydrus_1D. The hydraulic parameters from the 3 locations were interpolated to populate Hydrus 2D model domain which was used to simulate the soil moisture distribution in the fields. The HYDRUS_2D model output was evaluated against distributed gravimetric soil moisture, and biomass water gathered from multiple field campaigns. The COSMOS sensor is calibrated against the area average of modeled soil moisture distribution in the field obtained using HYDRUS 2D. Validation is done i) using point scale measurements from soil moisture sensors present in the field and ii) calculating the water budget using different sensors mentioned above. The CRNS dataset seems to agree with the soil moisture data obtained from point scale sensors and water budget method. The method reduces the labor in the regions where conducting field campaigns is difficult and provides a continuous calibration dataset for CRNS in the Lower Brazos river basin.