Journals Proceedings

Abstract

This paper focuses on the utilization of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data to predict soil salinity. After the necessary geometric and radiometric corrections were implemented, it is aimed to model the changes of electrical conductivity of soil samples, with related spectral profiles extracted from the satellite data. An electrical conductivity meter was employed for measuring the salinity levels of extracts acquired from the saturated soil samples. Statistical methods (stepwise regression – backward elimination) were used to reveal the significant bands for soil salinity and to diagnose the soil salinity related variances through the spectrum. The correlation between the selected soil samples having highest & lowest soil salinity levels was found to be consistent with each other. There seems to be a negative relation with the increased electrical conductivity content and spectral responses in shortwave infrared region. Furthermore, multivariate regression analysis showed that second and sixth band of the ASTER imagery make a significant contribution to the equation (p<.001) that models salinity and spectral profiles