Reduction of Errors in Hydrological Drought Monitoring – A Novel Statistical Framework for Spatio-Temporal Assessment of Drought

Continuous and accurate drought monitoring has an important role in early warning drought mitigation policies. This study aims to provide an accurate standardized drought monitoring indicator by enhancing the representative characteristics of precipitation data using advanced statistical methods. We proposed a two-phase statistical procedure index – the Regional Multi-Component Gaussian Hydrological Drought Assessment (RMcGHDA) – for accurate drought monitoring under a multi-auxiliary variable-based sampling estimator and K-Component Gaussian Mixture Distribution (CGMD) model. The first phase of our proposed method increases the regional representativeness of the data under Spatio-temporal settings and the second phase describes the use of the Twelve-Component Gaussian Mixture Distribution (CGMD) model in the standardization stage of SDIs. We applied the proposed framework to 52 meteorological stations in Pakistan and compared the RMcGHDA performance with existing methods using Pearson correlation (r) and spatial patterns of various drought categories. We found significant differences between RMcGHDA and existing methods (i.e., Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI)) for drought assessment. By the rationale of the data improvement under-sampling estimator and the use of multi-component Gaussian function, these differences indicate that RMcGHDA provides a practical and accurate way for drought assessment.