Exploring Spatial Relationship Between Electrical Conductivity and Spectral Salinity Indices in the Mekong Delta

Abstract

The negative impact of salinization concurrent with drought is a severe problem that creates challenges for agriculture in deltas and coastal lowlands. This study aims to investigate the spatial relationship among the field measured electrical conductivity (EC) and spectral salinity indices derived from Remote sensed data in the Mekong Delta using Geographically Weighted Regression (GWR). A wide range of Landsat 8 Operational Land Images (OLI) products, including single bands, band ratios, vegetation indices (NDVI and EVI), intensity indices (INT), and brightness indices (BI) were employed for computing salinity indices. The Kriging and Co-kriging interpolation techniques were used to estimate the spatial pattern of the field measured EC. Additionally, the Ordinary Least Square (OLS) regression were employed to characterize the relationship between single bands and EC measurement before applying the GWR for exploring the spatial correlation among the indices. There was a gradually increased of EC value from inland to coastal area. A significant relationship between EC measurement and spectral salinity indices and the highest correlation coefficient with p value less than 0.05 was found in EVI (r2 = 0.736). This study demonstrated that the GWR is germane to analyse the spatial correlation among the mentioned variables in the study area. Moreover, it also revealed that spectral salinity indices could be an alternative option for EC measurement in monitoring salt water intrusion at coastal areas.