Forest fire expansion under global warming conditions: multivariate estimation, function properties and predictions for 29 countries

This study investigates the average relative burned area, as a function of different conditions, in 29 countries. Detailed international statistics of forest fires, published by FAO and European Commission, are used as empirical data. A multivariate fire area function with empirically very convincing statistical properties is defined, tested, and estimated. A set of hypotheses was created based on three fundamental factors. The hypotheses could not be rejected on statistical grounds, and the estimated parameters obtained the expected signs with very low P-values. The residual analysis supports the selected functional form. Future fire areas are predicted for 29 countries, conditional on three alternative levels of global warming conditions. The estimated fire area function can explain the forest fire areas in different countries via three fundamental factors that are 1) The average area of forest fires divided by the total forest area is increasing the average temperature. Hence, global warming is expected to make future forest fire problems even more severe, 2) The average forest fire area divided by the total forest area is an increasing function of the total forest area, and finally 3) The average area of forest fires divided by the total forest area is a decreasing function of the population's size.