A novel Bi-weight Mid Correlation Coefficient Divergence (BMCCD) approach for multi-model ensemble-based drought...

Area of Science:

• Environmental Science
• Climate Science
• Data Science

Background:

• Droughts pose significant risks to ecosystems, water resources, and agriculture.
• Global climate models (GCMs) are crucial for forecasting but exhibit inter-model variability.
• Multi-model ensemble (MME) approaches offer more robust climate predictions than single models.

Purpose of the Study:

• Introduce a novel weighting scheme, Bi-weight Mid Correlation Coefficient Divergence (BMCCD), for enhanced MME reliability.
• Compare BMCCD's performance against Simple Model Averaging (SMA) and weighted ensemble (WE) methods.
• Develop the Standardized Bi-weight Divergence Index (SBDI) using BMCCD-aggregated data for drought analysis.

Main Methods:

• Developed and applied the Bi-weight Mid Correlation Coefficient Divergence (BMCCD) weighting scheme.
• Compared BMCCD with Simple Model Averaging (SMA) and weighted ensemble (WE) using referenced data.
• Utilized BMCCD-aggregated data to create the Standardized Bi-weight Divergence Index (SBDI).
• Projected drought characteristics from 2015 to 2100 under three Shared Socio-economic Pathways (SSP) using linear regression.

Main Results:

• BMCCD achieved the highest average correlation (0.749) and the lowest mean error (1.332) compared to SMA and WE.
• The Standardized Bi-weight Divergence Index (SBDI) was developed for drought assessment.
• Analysis across seven time scales and three SSPs revealed low probabilities for extreme drought (ED) and extreme wet (EW) events.

Conclusions:

• The BMCCD weighting scheme significantly enhances the reliability and efficiency of MME for drought prediction.
• While rare, extreme drought and wet events require policy attention for risk mitigation strategies.
• The SBDI provides a valuable tool for long-term drought characteristic evaluation under various future climate scenarios.