Spatial Variability Analysis of Soil Salinity and Alkalinity in an Endorreic Volcanic Watershed

Abstract

Three areas with different soils were recognized from an automated electromagnetic survey, at detailed scale, within a desiccated alkaline-salt-lake, using an EM38 instrument connected to a GPS and a datalogger, mounted in vertical mode on a mobile platform at 25 cm aboveground. Secondary magnetic field, surface temperature and position of measurements were recorded. GPS accuracy was evaluated before and during the survey and positioning was optimized using Wide Area Augmentation System, as well as performing the survey during the period with the lower expected Degradation Of ositioning. The raw data corresponding to the apparent soil electrical conductivity, measured from the secondary magnetic field induced by the EM instrument, were grouped statistically into three clusters using a partitioning-around-medoids algorithm. Subsequent soil sampling within each cluster was performed for characterization purposes, and main soil properties were analyzed. Each cluster representative was found to be different from others in terms of soil depth, texture, pH and Electrical Conductivity. The semivariogram adjusted to the signal data, a Gaussian-type with low nugget effect, suggest that the lag between measurements was adequate for capturing the spatial structure of the apparent electrical conductivity variable. Also can be derived that the salinity values change smoothly along the transect of the alkaline-salt-lake according to 2nd order stationarity. The instrumental arrangement provided a combined salinity+texture gradients mapping at very detailed scale.