Earticle

초록

영어

Milicia excelsa (iroko) is an economically important indigenous timber species in tropical Africa whose persistence may be affected by climate change. This study assessed current and future climatic suitability of M. excelsa across West and Central Africa using Random Forest species distribution modelling and explainable machine learning. A total of 307 spatially thinned occurrence records and 307 pseudo-absences were used. Following multicollinearity screening, four predictors: precipitation of the driest month (BIO14), precipitation seasonality (BIO15), precipitation of the coldest quarter (BIO19), and elevation, were retained. The final model achieved a test ROC-AUC of 0.8917, with an optimal classification threshold of 0.5824. Current suitable habitat covered 476,444.99 km2, declining to 431,335.14 km2 in 2050 and 434,938.59 km2 in 2070 under the SSP245 scenario. Despite these reductions, 80.85% and 80.64% of current suitable habitat were retained in 2050 and 2070, respectively. Persistent climatic refugia comprised 17,248 cells (368,088.30 km2) that remained suitable across all three periods. SHAP analysis identified BIO14 as the most influential predictor and revealed a nonlinear relationship between dry-season precipitation and predicted suitability. Environmental novelty increased from 0.79% under current conditions to 2.28% in 2050 and 2.86% in 2070, while approximately 99% of suitable cells remained environmentally supported. The findings indicate that climate change is likely to redistribute rather than uniformly eliminate climatically suitable habitat for M. excelsa, providing evidence to support climate-informed conservation and sustainable forest management.

목차

Abstract
Introduction
Materials and Methods
Data
Model training
Validation and model interpretation
Results
Occurrence records and environmental predictor selection
Model performance
Current and future habitat suitability
Habitat loss, gain, and retention
Persistent climatic refugia
Model interpretation using SHAP analysis
Environmental novelty and projection robustness
Discussion
Model performance and reliability
Climatic and environmental controls on the distribution of Milicia excelsa
Climate-induced habitat redistribution
Implications for conservation, study limitations, and future research
Conclusion
Author Contributions
Data Availability
Funding
Conflict of Interest
References

키워드

저자

• Peter Edemewe Ugege [ Informatics and ICT Resource Centre, Forestry Research Institute of Nigeria, Ibadan 200272, Nigeria ] Corresponding Author
• John Oseyomon Aigbokhan [ Environmental Modeling and Biometric, Forestry Research Institute of Nigeria, Ibadan 200272, Nigeria ]
• Olaremi Rebecca Olatidoye [ Forest Conservation Department, Forestry Research Institute of Nigeria, Ibadan 200272, Nigeria ]
• David Ogbe Agbenu [ Information and Documentation Department, Forestry Research Institute of Nigeria, Ibadan 200272, Nigeria ]
• Mercy Akoun [ Information and Documentation Department, Forestry Research Institute of Nigeria, Ibadan 200272, Nigeria ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  Journal of Forest and Environmental Science [산림과학연구]

• 간기

  계간

• pISSN

  2288-9744

• eISSN

  2288-9752

• 수록기간

  1981~2026

• 등재여부

  KCI 등재

• 십진분류

  KDC 526 DDC 634