Development of ADI, the Aggregate Drought Index, Based on Principle Component Analysis for Monitoring Agricultural Drought in Golestan Province, Iran

Univariate indices can be useful for certain purposes but are unable to comprehensively monitor drought characteristics as a complex climate phenomenon. In this study, to monitor agricultural drought conditions in Golestan province, the multivariate index ADIhas been developed based on Principle Component Analysis (PCA) technique using the results of drought indices as inputs. For this purpose the daily metrological variables from 10 synoptic and evaporative stations were obtained for 33 hydrological years to compute the values of 5 drought indices i.e. Standardized Precipitation Index (SPI), Standardized Evapotranspiration–Precipitation Index (SPEI), Palmer Drought Severity Index (PDSI), Z-Index, and Soil Moisture Index (SMI) in monthly timescale. Then by applying PCA technique and considering all established components, the ADI drought index was developed based on linear combinations of derivation components for each of the studied stations. Finally the ADI results were compared to common indices such as SPEI and SPI. The results showed a strong correlation coefficient between ADI and SPEI indices in wet regions and months of November to March based on Kendal –Tau (greater than 0.8). The degree of correlation were reduced during warm months of April to June. Comparative study on the results of ADI and SPEI showed that the ADI index as a multivariate index could reflect the effects of input drought indices such as SMI and Z-Index and present a more comprehensive drought monitoring