Estimating Organic Carbon of the surface soil via integration of Factor Analysis and Multiple Linear Regressions in Semi-Steppic Rangelands, Lazour, Firoozkooh

   Organic carbon is one of the most important soil quality indices, affecting almost all physical, chemical and biological properties of the soil. The purpose of this study was to investigate soil spectral and morphometric factors to estimate the organic carbon of topsoil, using factor analysis and multiple regression methods in semi-steppe rangelands of Lazour. Soil samples were taken with a stratified random method. For this purpose, 157 training sites were selected in homogeneous units. Of these, 127 sites were used to calibrate the model and 30 sites were used to validate the model. In each of the training site in a random manner, a soil sample including nine observations was taken from a depth of 0 to 20 cm of soil surface. Soil Organic Carbon (SOC) was measured using Walkley-Black titration method. The results showed that the variables of Albedo, Clay Index (CI), NDVI, Relative Relief and Tasseled-Cap's Brightness and Greenness indices had a significant correlation with the SOC (p<0.05). Also, the result of factor analysis by Principal Component Analysis (PCA) method with eigen-values greater than one indicated that the total cumulative variance, explained by the six variables, was equal to 81.1%.This variance was explained by two components. Using multiple regression model, an appropriate regression equation was calculated to predict SOC (R2=0.789). The Root Mean Square Error and the Mean Absolute Relative Error of the proposed model were calculated as 0.24 and 0.10, respectively. Due to the direct relationship between the SOC and the factors such as soil fertility and sustainability against erosion, a spatial distribution model of SOC could be an important sub-model to design other complex models such as the terrestrial ecosystems biomass and soil erosion models.