Improving Modified Perpendicular Drought Index in order to Estimate Drought Severity in Arid Central Regions of Iran

Drought is a severe dilemma which has influence on economy, agronomy, environment and society. In the recent decades, many indices based on satellite images have been presented by researchers. The objective of this study was to improve Modified Perpendicular Drought Index (MPDI) by improving its main input parameter, which is the vegetation fraction. Calculation of vegetation fraction was considered by using an ALOS imagery.Sheitoor basin located in Yazd province (part of the central arid area of Iran) was selected as the study area. The canopy cover percentage of 52 sample plots (50m by 50m) was measured in the field in June 2008. Seventeen vegetation indices and different types of neural networks algorithms have been used. Regression models were used to assess the relationship between the field data and the calculated reflectance data. Results revealed that the vegetation indices which are based on soil line parameters can estimate the fraction of vegetation more accurately compared to other indices (R2>0.63, RMSE<4%). On the other hand, neural network results showed that these algorithms could estimate the vegetation fraction even more accurate than vegetation indices(R2> 0.74,RMSE < 2%), since it benefits from having more input parameters. Finally, Drought severity of the area was mapped using the estimated soil line parameters and the vegetation fraction which was estimated using the neural network algorithm.