Hoque et al. (2026) Extreme drought dynamics in the northern part of Bangladesh: Long-term spatiotemporal characteristics and forecasted changes using SPI and RAI-based dual index-driven Mann-Kendall and ANN approach

Identification

• Journal: H2Open Journal
• Year: 2026
• Date: 2026-09-09
• Authors: Md Abrarul Hoque, Md. Abul Kalam Azad, Rubinoor Islam, Md. Jahir Uddin, Md. Zakir Hossain, Abdullah All Noman, Meree Sarkar, Nupur Akter, Sathi Akter
• DOI: 10.1016/j.htopen.2026.100080

Research Groups

• Department of Civil Engineering, European University of Bangladesh
• Department of Civil Engineering, Khulna University of Engineering & Technology

Short Summary

This study aims to understand extreme drought dynamics in northern Bangladesh by integrating drought indices (SPI and RAI), machine learning-based forecasting, and statistical trend analysis. The results indicate that ANN models perform better for RAI, with a correlation coefficient R = 0.89.

Objective

• To develop a comprehensive framework for understanding extreme drought dynamics in northern Bangladesh.
• To characterize historical and future drought conditions using SPI and RAI across multiple timescales.
• To compare the performance of SPI and RAI in capturing drought frequency, intensity, and duration.
• To identify present and future drought hotspots in northern Bangladesh.

Study Configuration

• Spatial Scale: Northern region of Bangladesh, with a focus on 8 selected stations.
• Temporal Scale: Long-term daily rainfall data (1994–2023) for the selected stations.

Methodology and Data

• Models used: Standardized Precipitation Index (SPI), Rainfall Anomaly Index (RAI), Artificial Neural Network (ANN)-based forecasting models.
• Data sources: Bangladesh Meteorological Department (BMD), daily rainfall data from 1994 to 2023.

Main Results

• The study found that ANN models perform better for RAI, with a correlation coefficient R = 0.89.
• SPI-based analysis showed that 46.07% of droughts were severe or extreme.
• Forecasted results indicate a 5.12% increase in extreme drought and 22.58% increase in post-monsoon drought onset.

Contributions

• This study contributes to a more comprehensive understanding of extreme drought dynamics in monsoon-dominated regions.
• The findings highlight the complementary roles of SPI and RAI in capturing drought behavior.
• The study demonstrates the potential of data-driven forecasting for supporting proactive drought risk management.

Funding

• This research was funded by [project name], reference code [reference code].

Citation

@article{Hoque2026Extreme,
  author = {Hoque, Md Abrarul and Azad, Md. Abul Kalam and Islam, Rubinoor and Uddin, Md. Jahir and Hossain, Md. Zakir and Noman, Abdullah All and Sarkar, Meree and Akter, Nupur and Akter, Sathi},
  title = {Extreme drought dynamics in the northern part of Bangladesh: Long-term spatiotemporal characteristics and forecasted changes using SPI and RAI-based dual index-driven Mann-Kendall and ANN approach},
  journal = {H2Open Journal},
  year = {2026},
  doi = {10.1016/j.htopen.2026.100080},
  url = {https://doi.org/10.1016/j.htopen.2026.100080}
}