Processes like evapotranspiration and infiltration are closely linked to the wetness of the soil, and soil moisture is therefore a key variable for water balance studies. Catchment scale hydrological modeling is used for weather and climate prediction and for estimating fluxes and variables of the hydrological system important for managing the water resources. Soil moisture is highly variable in time and space, and the variability changes with scale. Soil moisture measurements at a scale comparable to the discretization of catchment scale models are therefore of great importance for validation and calibration. Yet, soil moisture measurements are traditionally provided on either point or kilometer scale from electromagnetic based sensors and satellite retrievals, respectively.
Above the ground surface, the cosmic-ray neutron intensity (eV range) is inversely correlated to all hydrogen present in the first few hectometers in the atmosphere above the land surface and the upper few decimeters of the subsurface. Neutron intensity detection is suitable for soil moisture estimation since it often forms the major dynamic pool of hydrogen. Furthermore, the scale of measurement compares well with the discretization of many catchment scale models. Within the neutron detector footprint hydrogen is also present in atmospheric water vapor, snow, canopy interception, biomass, litter and soil organic matter. The uncertainty of the cosmic-ray neutron soil moisture estimates is larger at field locations featured by several pools of hydrogen as more variables affect the neutron intensity. The method may therefore be
improved if the signals of the other featuring components are identified and adjusted for. Identifying the different signals may furthermore be used to extent the application of neutron intensity detection (e.g. canopy interception and biomass quantification).
The neutron transport at the lower part of the atmosphere is examined using measurements and modeling. Both thermal and epithermal neutron intensity at field sites of varying environmental settings (farmland, forest plantation, heathland and brackish water) is included in the examination. Initially, a method enabling comparability of measured and modeled neutron intensity is developed. Both the scale and energy ranges are different, and the exact neutron energy response of the neutron detectors is unexplained. The method involves correction models for the estimation of pure thermal and pure epithermal neutron intensities from measurements using bare and moderated neutron detectors. The method also includes conversion factors to obtain modeled neutron intensities at a scale comparable to measurements. The correction models are found to be non-universal and must therefore be determined for the specific neutron detector systems. Site-specific models provided neutron intensities in satisfactory agreement with measurements. The sitespecific
models are following used to test the sensitivity of soil moisture, forest canopy conceptualization, complexity of soil chemistry, litter, soil bulk density, canopy interception and biomass.
The sensitivity of soil moisture on neutron intensity is dependent on neutron energy and land cover. The sensitivity of soil moisture is greatest for epithermal neutron intensity, however, it decreases with the presence of litter and biomass. The opposite is valid for thermal neutrons. Here, the sensitivity of soil
moisture becomes increasingly greater with the presence of litter and biomass. The site-specific models are furthermore used to determine site-specific calibration functions for soil moisture estimation at the farmland, forest and heathland field sites. The soil moisture estimates of the site-specific calibration
functions are in relation to the estimates provided by the standard N₀-calibration function not improved at the three field sites. The signal of biomass on neutron intensities was found to be significant for both the
shape of the neutron intensity height profiles and the ground level thermal-to-epithermal neutron ratio. The ratio increases considerably with increasing amounts of biomass and satisfactory agreement of measurements and modeling is obtained at the field sites of varying amounts of vegetation. A slight
increase in ground level thermal neutron intensity was provided with canopy interception. The change is for the surveyed field location within the measurement uncertainty when hourly time scales are regarded. Yet, canopy interception quantification may be possible at field locations of higher neutron intensity and with canopy interception of longer residence time.