Forest yield prediction under different climate change scenarios using data intelligent models in Pakistan

Area of Science:

• Forestry science
• Climate change impact assessment
• Ecological modeling

Background:

• Forest ecosystems are vital for Pakistan's economy and environment.
• Climate change poses significant risks to forest productivity and health.
• Accurate yield prediction is crucial for sustainable forest management.

Purpose of the Study:

• To develop and evaluate data-driven models for predicting forest yield under climate change scenarios.
• To compare the performance of Random Forest (RF) and Kernel Ridge Regression (KRR) models for forest yield prediction.
• To provide a reliable tool for forest management and planning in Pakistan.

Main Methods:

• Developed and evaluated RF and KRR models using forest yield data (Blue pine, Silver fir) and climate data (temperature, humidity, rainfall, wind speed).
• Assessed prediction accuracy using metrics such as RMSE, MAE, correlation coefficient, RRMSE, LM, WI, and NSE.
• Utilized Gallies forest division, Abbottabad, Pakistan, as the study area.

Main Results:

• The Random Forest (RF) model demonstrated higher accuracy in forecasting forest yield compared to the Kernel Ridge Regression (KRR) model.
• RF model's superior performance indicates its suitability for predicting forest growth and yield.
• Both models were evaluated using multiple statistical metrics to ensure robust comparison.

Conclusions:

• The RF model is recommended for application in other regions of Pakistan for predicting forest growth and yield.
• Implementing the RF model can significantly aid in the management and future planning of forest productivity.
• Data-driven modeling provides a valuable approach for understanding and mitigating climate change impacts on forests.