Analysis of the future potential impact of environmental and climate changes on wildfire spread in Ghana's ecological zones using a Random Forest (RF) machine learning approach

Climate change is increasing the hazards and catastrophes in many parts of the world, without sparing the natural zones, which are likewise growing more vulnerable in the face of human pressure. This study looked at how climate change affects wildfire spread via prediction in the Guinean savannah (GSZ) and the forest-savannah mosaic (FSZ) in Ghana. Random forest regression was employed to attain this goal. As a result, the CCLM4-8-17 model outputs were employed and analyzed under two Representative Concentration Pathway (RCP4.5 and RCP8.5), since they better forecast the climate of the biological zones studied with less error (R2, RMSE, Bias, Spearman Correlation). As a result, there is a decrease in rainfall over the simulated period (2022–2100) in both zones (FSZ, GSZ), considering both scenarios (RCP4.5; RCP8.5), and an increase in temperature in FSZ (+0.08 °C for RCP4.5 and 0.1 °C for RCP8.5) and GZS (+0.04 °C for RCP4 and +0.09 °C for RCP8.5). Under RCP4 and RCP8 scenarios, the average number of active fires (HS) on their part shows that the average near future (2022–2060) will have more areas favourable to fire activities (active points) in both ecological zones: RCP4_GSZ_HS (316 n), RCP4_FSZ_HS (252 n), RCP8_GSZ_HS (333 n), RCP8_FSZ_HS (285 n), followed by the average far future (2061–2100) The average burnt area (BA) under RCP4.5 and RCP8.4 that average near future (2022–2060) will also contain a more burnt area in both ecological zones: RCP4_GSZ_BA (25290.96 ha), RCP4_FSZ_BA (20758.51 ha), RCP8_GSZ_BA (25766.66 ha), RCP8_FSZ_BA (20699.91 ha) followed by average for far future (2061–2100): RCP4_GSZ_BA (14849.20 ha), RCP4_FSZ_BA (11782.11 ha), RCP8_GSZ_BA (19,299.07 ha), RCP8_FSZ_BA (1072.01 ha). Thus, under the RCP4.5 and RCP8.5 scenarios, tmax_RP4, SM, and ETA contribute more to fire activity and spread in the GSZ. However, in the FSZ, RH, tmax_RP4, and SM contribute more to fire activity and spread in the RCP4.5 and RCP8.5 scenarios, with varying amounts. Because of the above, decision-makers and academics must take tangible steps to develop novel techniques to minimise the future impacts of fires.