Saline soil mapping using ETM+ data in the Qazvin plain

Salinity is the major factors of soil degradation in semi arid and arid regions. The mainaim of this study was to evaluate the capability of Landsat ETM+ data for soil Salinitymapping in the selected part of the Qazvin plain, an area of arid environment. In thisstudy spectral classes carried out on remotely sensed data and with the help of fieldobservation and soil analysis were regrouped to soil salinity classes to prepare soilsalinity map.. Soil sampling was implemented using stratified random samplingmethod, depending on landscape complexity and homogeneity as well as on therepresentativeness of Landsat ETM+ data. Also in each soil map unit at least one profilewas studied for subsoil salinity variations. Field samples taken by using augur andprofiles were analyzed in laboratory for Na+, Ca2+, Mg2+ cations, as well as soil texture,ECe and pH. We have analyzed the effectiveness of additional data such as digitalelevation model to improve the accuracy of classification. Also NDVI, SRVI, PVI,SAVI, SI, BI and NDSI indices, PCA and Tasseled cap were analyzed. Soil salinity mapof each selected bands produced and with ground truth map crossed. The resultsindicated that combination of DEM with ETM+ bands has highest accuracy. This studyaddressed that thermal band of ETM+ can increase the classification accuracy whichillustrated its effective role to classify the soil salinity. Tasseled cap and other indiceshad almost high accuracy among studied image processing techniques. The SI and BIindices had the highest correlation with EC and could distinguish the saline and nonsaline soils while the optimum index factor had overall low accuracy.