<![CDATA[Wildfires in Indonesia have a widespread impact on health, the environment, society, and the economy. The number of hotspots, detected through satellite imagery, is a key indicator in monitoring the severity of fires. Because hotspot data is seasonal and prone to spikes due to extraordinary events such as El Niño, an adaptive forecasting method is needed. The SARIMA model is effective for capturing seasonal patterns, but it is less responsive to extreme spikes. Therefore, intervention analysis with pulse functions is used as an alternative to model sudden and temporary changes in time series data. This study aims to compare the performance of the SARIMA model and an intervention model using a pulse function in forecasting the number of hotspots in Indonesia. The data used in this study were obtained from the Ministry of Environment and Forestry through the SiPongi platform, consisting of monthly data from January 2014 to December 2022. The modeling results show that the SARIMA model produced a MAPE value of 36.93%, an RMSE of 66.27, and an MAE of 47.83. In contrast, the intervention model with a pulse function at order b=0, s=0, and r=1 specifically SARIMAachieved a MAPE of 8.06%, an RMSE of 8.45, and an MAE of 6.67, substantially outperforming the SARIMA model across all metrics. These findings indicate that the intervention model provides much more accurate forecasts of hotspot occurrences in Indonesia. Furthermore, forecasts up to 2025 indicate a declining trend in the number of hotspots over time. However, seasonal patterns remain evident, with expected increases in hotspot activity during the months of February, August, and October. These results are expected to contribute valuable insights for developing more effective forest fire mitigation strategies in Indonesia ]]>
