Characterizing the spatial variability of soil salinity in lake urmia basin by applying geo-statistical methods

Land degradation by salinity is one of the main environmental hazards threatening soil sustainability especially in arid and semi-arid regions of the world characterized by low precipitation and high evaporation. Geo-statistical approaches and remote sensing (RS) techniques have provided fast, accurate and economic prediction and mapping of soil salinity within the last two decades. Obtaining multi-temporal data via satellite images in different spatial domains with various scales is one of the key developments of monitoring spatial variability of soil salinity. In addition, geo-statistical methods have the capability of producing prediction surfaces from limited sample data. This study aims to map spatial distribution of soil salinity in the selected pilot area which is located in the western part of Urmia Lake Basin, Iran, by applying geo-statistical methods. A kriging based map and three different co-kriging based maps were produced using electrical conductivity (EC) measurements as primary variable and three different soil salinity index values as secondary variable. Three soil salinity indices were created by using Sentinel-2A image that were acquired in the same date of field measurements to generate 3 various soil salinity prediction maps. Salinity maps obtained from geo-statistical methods were compared and validated to understand the performance of these approaches for soil salinity prediction. The results of this study demonstrated that co-kriging can provide promising estimation of spatial variability of soil salinity especially when there is relevant and abundant set of secondary data derived from satellite images.