Assessment of Probabilistic Multi-Index Drought Using a Dynamic Naive Bayesian Classifier

The proper consideration of all plausible feature spaces of the hydrological cycle and inherent uncertainty in preceding developed drought indices is inevitable for comprehensive drought assessment. Therefore, this study employed the Dynamic Naive Bayesian Classifier (DNBC) for multi-index probabilistic drought assessment by integrating various drought indices (i.e., Standardized Precipitation Index (SPI), Streamflow Drought Index (SDI), and Normalized Vegetation Supply Water Index (NVSWI)) as indicators of different feature spaces (i.e., meteorological, hydrological, and agricultural) contributing to drought occurrence. The overall results showed that the proposed model was able to account for various physical forms of drought in probabilistic drought assessment, to accurately detect a drought event better than (or occasionally equal to) any single drought index, to provide useful information for assessing potential drought risk, and to precisely capture drought persistence in terms of drought state transition probability in drought monitoring. This easily produced an alternative method for comprehensive drought assessment with combined use of different drought indices.This research was supported by a grant (17AWMP-B079625-04) from Water Management Research Program funded by Ministry of Land, Infrastructure and Transport of Korean government