Assessing trends and drivers of burned areas in forest areas in the Kurdistan Region

Wildfires pose an escalating threat to the oak-dominated forests of the Kurdistan Region of Iraq, where long-term fire trends and predictors have remained poorly quantified. This study assessed interannual variability and long-term trends in total and forest-specific burned area from 2001 to 2024, examined spatial differences, and identified primary climatic predictors of fire extent using MODIS MCD64A1 Version 6.1 burned-area data masked to a 2024 NDVI-based forest mask (~ 10,660 km²). Across the entire Kurdistan Region, burned area averaged 687 km² year⁻¹ (SD = 640 km²), totalled 16,486 km² over the 24-year period, and exhibited a statistically significant upward trend of 31 km² year⁻¹ (Theil-Sen slope; Mann-Kendall p = 0.026). Forest burned area averaged 356 km² year⁻¹ (equivalent to an annual burn rate of 3.3% of forest cover), displayed a significant increasing trend of 15 km² year⁻¹ (Mann-Kendall p = 0.021), and reached a cumulative 8541 km², with Duhok and Sulaymaniyah together accounting for 77% of cumulative forest burned area and showing the strongest upward trends. Maximum temperature and drought severity were the dominant climatic predictors: each 1 °C rise in monthly maximum temperature increased expected burned area by 14.3% (incidence-rate ratio [IRR] = 1.143, p < 0.001). Drought severity (negated PDSI, positive values = greater drought severity) demonstrated a dual effect on fire dynamics. In the count model predicting fire magnitude, drought was significantly and negatively associated with burned area (IRR = 0.787, p < 0.001), reducing expected fire size by 21.3% per unit increase in drought severity; the zero-inflation component showed a similar but non-significant trend (p = 0.417). Model results showed pronounced non-linear escalation of predicted fire activity above ~ 32 °C and negated PDSI > 2 (severe drought conditions).