Monitoring and Assessment of Drought using Remote Sensing and association rules

Drought is a natural threat that exists in all climatic zones around the globe. There is a need to categorize drought events and the probability of occurrence for better planning and management of relief and rehabilitation. In this study, drought monitoring indices namely the Standard Precipitation Index (SPI) and Vegetation Condition Index (VCI) were used to analyse the observed variability of monsoon droughts over Andhra Pradesh State. Precipitation data between 1991-2019 was used to evaluate the SPI and to evaluate the VCI from NDVI data collected from 2011 to 2019 using multi-temporal Terra MODIS Vegetation Indices Product (MOD13Q1). In this analysis, more often drought events occurred in 3 and 6 months SPI during monsoon season. In this study, data mining techniques (such as the Association Rules) are used to explain the association between VCI and SPI to predict the probability of occurrence of drought. The association rules formed by the VCI and the 3-month SPI with 77 percentage of confidence and 1.11 of lift indicate the higher accuracy of the rules and the effect on vegetation ford rainfall accumulation. This research incorporated the various software and dataset levels used to predict the probable occurrence and severity of drought using the current situation. The analysis revealed the advantages of NDVI and rainfall for indices of spatial and multitemporal drought to identify and forecast the characteristics of drought.