MODELING THE IMPACT OF ANTHROPOGENIC AND CLIMATIC DYNAMICS ON FOREST AREAS USING ROBUST STATISTICAL METHODS

Abstract

Background: Forests play a central role in climate regulation, biodiversity conservation, and ecosystem stability, yet they are increasingly threatened by anthropogenic and climate-related pressures. In developing countries such as Pakistan, low forest cover combined with rapid population growth, rising emissions, and land-use change has intensified forest degradation. Understanding how socio-economic and climatic variables jointly influence forest area is essential for evidence-based environmental planning and climate-adaptive forest management.

Objective: To quantitatively assess the effects of key anthropogenic and climatic variables on forest area in Pakistan and to identify the most robust predictors of forest change.

Methods: A quantitative, retrospective analytical design was adopted using 60 annual observations derived from a national-level dataset. Variables included forest area, carbon dioxide emissions, total greenhouse gas emissions, urban population, cereal yield, renewable energy consumption, rainfall, and agricultural land. Statistical analysis involved Pearson correlation, multiple linear regression, model reduction, and ANOVA-based model comparison. Diagnostic evaluation identified heteroscedasticity, non-linearity, and influential observations, prompting the application of robust regression with log-transformation of the dependent variable to ensure stable and reliable estimates.

Results: Correlation analysis revealed strong interrelationships among emissions, agricultural productivity, and urbanization. In the reduced regression model, carbon dioxide emissions, total greenhouse gas emissions, and urban population emerged as statistically significant predictors, jointly explaining 95.01% of the variation in forest area (adjusted R²). Robust log-linear regression confirmed that both carbon dioxide and total greenhouse gas emissions were negatively associated with forest area, while urban population showed a statistically significant positive association. Diagnostic plots demonstrated improved variance stability and reduced influence of outliers following robust estimation.

Conclusion: The findings provide strong empirical evidence that emission-related factors are the principal drivers of forest decline in Pakistan, while urbanization exhibits a context-dependent relationship shaped by governance and land-use practices. The study underscores the urgency of emission mitigation and integrated urban environmental planning to support sustainable forest conservation.