Analysis of the scale mismatch in measuring the electrical conductivity and soil moisture using a multi-sensor approach.

Soil moisture patterns are influenced by both vertical heterogeneity and horizontal spatial variability. Capacitance sensors for estimating point-scale soil moisture remains predominant in most studies. The measurement of bulk electrical permittivity, temperature, and electrical conductivity in a soil volume of a few hundred cubic cm still tends to be considered as ground truth with high temporal resolution and accuracy. Recently, non-invasive proximal soil moisture sensors such as the cosmic-ray neutron probes (CRNPs) have gained importance in estimating field-scale integrated soil moisture over a footprint of hundreds of meters in diameter and several decimeters of soil depth. However, ground-based and proximal sensors gain information on soil moisture dynamics only in the upmost layers of the soil profile. By contrast, sensitivity to changes in soil water storage can be obtained by using sporadic geophysical surveys which are able to measure soil electrical conductivity over the three-dimensional spatial domain. The methods mentioned above are sensitive to soil moisture but suffer from scale mismatch and therefore are differently affected by the spatio-temporal variability of soil properties, terrain features, and vegetation patterns. This study combined point-scale capacitance sensors, cosmic-ray neutron probes (CRNPs), and geophysical methods for estimating soil moisture spatial patterns in two experimental sites in the Upper Alento River Catchment (southern Italy). A CRNP was installed in each of the experimental sites and capacitance sensors (GS3 of METER Company) were installed on both sites around the CRNP to monitor soil moisture over 20 positions and at soil depths of 15 cm and 30 cm. Time-lapse electrical resistivity tomography (ERT) and electromagnetic induction (EMI) surveys were sporadically performed over the two areas from August 2020 to August 2021 by using portable devices to capture the soil moisture seasonality. The spatially-explicit geophysical conductivity maps were evaluated with co-located information on electrical conductivity monitored by the point-scale capacitance sensors. The comparison revealed significant offsets in the measured electrical conductivity values from ERT, EMI, and point-scale capacitance sensors. We explored the causes inducing the observed scale mismatch. Taking advantage of the ERT scale flexibility, high-resolution surveys were rescaled by using process-oriented electrical conductivity models to gain further insight into the lateral and vertical sensitivity of the ERT and EMI methods. The numerical simulations evidenced the sensitivity of electrical conductivity (and soil moisture) patterns to the soil heterogeneity. They proved that the multi-method approach is necessary to upscale and downscale the information from the point sensors to CRNPs and vice versa.