Knowledge-driven Development of Causal Model for Predicting Fire Hotspot Escalation in Central Kalimantan Peatlands

Peatland wildfires frequently occurring in Central Kalimantan, Indonesia is a damaging environmental problem on a global scale. To alleviate the occurrence of peatland wildfires, it is important to ensure that hotspots do not escalate into wildfires in the first place. However, insufficient historical data has become a challenge in building a data-driven model of fire escalation in peatlands. This thesis proposes workflow to quantify the causal relationship amongst factors contributing to the peatland fire escalation. The result shows that automated predictive hotspot escalation model can be developed purely using a knowledge driven approach by engaging experts through workshops and surveys.