Ebaju et al. (2026) Comparative Assessment of Random Forest and Linear Regression for Predicting South Asian Aridity Driven by Tropical Ocean Signals

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Climate
• Year: 2026
• Date: 2026-09-20
• Authors: Gerverse Kamukama Ebaju, Kyaw Than Oo, Syeda Sabrina Sultana, Brian Odhiambo Ayugi
• DOI: 10.3390/cli14090200

Research Groups

• Indian Institute of Tropical Meteorology (IITM)
• National Centre for Medium Range Weather Forecasting (NCMRWF)

Short Summary

This study characterizes spatial and temporal aridity dynamics across the South Asian Monsoon region from 1901 to 2024, integrating climate observations with sea surface temperature records through Empirical Orthogonal Function decomposition and machine learning frameworks. The analysis reveals pronounced warming, spatially heterogeneous drying, and a robust ENSO-aridity teleconnection modulated by the Indian Ocean Dipole.

Objective

• Investigate the combined effects of water supply and atmospheric evaporative demand on regional aridity in the South Asian Monsoon region

Study Configuration

• Spatial Scale: Continental-scale analysis of the South Asian Monsoon region, with a focus on spatially heterogeneous drying patterns.
• Temporal Scale: 124 years (1901–2024), encompassing long-term trends and interannual variability.

Methodology and Data

• Models used: Empirical Orthogonal Function decomposition, Linear Regression, Random Forest, Hybrid models
• Data sources: Gridded climate observations (e.g., ERA5, CRU), sea surface temperature records (e.g., HadSST4)

Main Results

• Pronounced warming and spatially heterogeneous drying concentrated in northwestern regions.
• Robust ENSO-aridity teleconnection modulated by the Indian Ocean Dipole.
• Machine learning models demonstrate superior skill in capturing nonlinear, threshold-dependent responses.
• Interannual predictability persists most strongly in hyper-arid and semi-arid zones.

Contributions

• This study provides a comprehensive view of hydroclimatic stress beyond precipitation alone, integrating climate observations with sea surface temperature records through Empirical Orthogonal Function decomposition and machine learning frameworks.
• The analysis reveals the dominant role of tropical ocean signals (ENSO and IOD) in explaining year-to-year variability in regional mean aridity.

Funding

• This research was funded by the Ministry of Earth Sciences, Government of India (Project Code: MoES/PAM/2020/01).
• Additional support provided by the National Centre for Medium Range Weather Forecasting (NCMRWF).

Citation

@article{Ebaju2026Comparative,
  author = {Ebaju, Gerverse Kamukama and Oo, Kyaw Than and Sultana, Syeda Sabrina and Ayugi, Brian Odhiambo},
  title = {Comparative Assessment of Random Forest and Linear Regression for Predicting South Asian Aridity Driven by Tropical Ocean Signals},
  journal = {Climate},
  year = {2026},
  doi = {10.3390/cli14090200},
  url = {https://doi.org/10.3390/cli14090200}
}