Assessment and Prediction of Soil Salinity Using Remote Sensing (Case Study: Bueinzahra

Soil salinity is one of the great problems that affects the growth of crops. The rate of salinization is more than 2 million hectares a year in the worldwide. Remotely sensed data and GIS are the useful tools for mapping saline soils.The present study was performed in Bueinzahra area in Ghazvin province (21506 ha). 42 surface soil samples were collected and the surface electric conductivity (EC) of soil was calculated. TM data, that cover the area, were acquired in 2010 and geo referenced.With regards to factor analysis, two components selected. The first component including NDMI, BI, SI1, SI2, SI3 with the variance 47% and the second component including main bands with the variance 40%.The results of OLS analysis showed the first model is better, and had a stronger relationship to ECe (R2.adj of 0.6) and the model is EC1 = -12.078 + 60 NDMI - 1.99 BI+ 0.62 SI1 + 1.8 SI2 - 0.24 SI3. Eight soil samples for investigation of map accuracy were applied (20% samples).Moran‘s I of residuals was 0.02. The accuracy assessment of estimations using a validation set of 8 samples showed ME and RMSE of 0.08 and 2.53 dS/m respectively.

Keywords:

images , Electrical conductivity , Soil salinity prediction