GIS-based Soil Salinity Mapping in Forest Soils of Telangana using Interpolation Techniques

Abstract

Soil salinity poses a significant challenge to soil fertility, crop productivity, and sustainable agricultural development in India. This study focuses on evaluating soil salinity in forest soils of Telangana using Geostatistical Models. The study area spans five different districts in Telangana, India: Nizamabad, Kamareddy, Medak, Siddipet, and Sangareddy, encompassing various forest types, including dry deciduous scrub, dry teak forest, southern dry mixed deciduous forest, and non-forest regions. A total of 174 soil samples were collected from both non-degraded (Moderate Dense Forest) and degraded (Scrub, Open, and Non-Forest) forest regions at two different depths (0-30 cm and 30-60 cm). These samples were then analyzed to determine their soil electrical conductivity (EC) levels. In the study area, measured EC values ranged from 1 to 20 dS m-1 at depths of 0-30 cm and 0.5 to 14.5 dS m-1 at depths of 30-60 cm. Inverse Distance Weighting (IDW), Ordinary Kriging (OK), and EBK interpolation techniques were used for spatial distribution and mapping of soil salinity. Semivariograms and cross-validation results indicated that Ordinary Kriging, Empirical Bayesian Kriging, and Inverse Distance Weighting were statistically significant in generating spatial distribution maps of soil salinity at both soil depths. However, the Ordinary Kriging and Empirical Bayesian Kriging models were the best-fit models, with higher R² and lower RMSE values compared to Inverse Distance Weighting, for generating spatial distribution maps of soil salinity in the forest soils of Telangana.