Regional soil salinity spatiotemporal dynamics and improved temporal stability analysis in arid agricultural areas

Monitoring and evaluating spatiotemporal dynamics of soil salinity over large areas for an extended period is important for keeping crop yield in salt-affected areas, but difficult due to its high variability. In this study, measurements of soil salinity with 68 sampling sites at different depth from top soil to 1.8 m were carried out in 2017–2018 to understand soil salinity variability and temporal stability at an agricultural area (> 80 km2). The spatial variability and mean of soil salinity was estimated by the geostatistical analysis and temporal stability analysis, respectively. Then, improved temporal stability analysis was proposed by dividing samples into 7 groups, and mean soil salinity in each group was estimated by temporal stability analysis. Lastly, monitoring network was recommended to evaluate long-term soil salinity. Strong spatial dependency of soil salinity was found with most degree of spatial dependence smaller than 25%. The temporal stability analysis was difficult to choose the representative sites due to large range of mean relative difference and standard deviation of relative difference of soil salinity. The predictions of improved temporal stability analysis were significantly improved with mean relative error of soil salinity means ranging from − 2.72 to 1.61%, and determination coefficient more than 0.90. Spatial distribution of soil salinity determined by 32 long-term soil salinity monitoring locations was consistent with that of all 68 sampling locations. The improved temporal stability analysis combined with geostatistical analysis can obtain spatial pattern and spatial mean of regional soil salinity, and improve monitoring efficiency greatly.