Using hydrological modelling to improve the Fire Weather Index system over tropical peatlands of peninsular Malaysia, Sumatra and Borneo
- Rafal Tekreeti
- Apr 7
- 3 min read
By: J. Mortelmans , S. Apers , G. J. M. De Lannoy, S. Veraverbeke, R. D. Field, N. Andela, S. E. Page and M. Bechtold
Tropical peatland fires are a major source of carbon emissions and air pollution, with significant environmental and societal consequences. The tropical peatlands in Southeast Asia store approximately 68 Gt of Carbon (C), roughly 77% of the global C stock in tropical peatlands. Under natural, wet conditions, these peatlands act as net C sinks, storing atmospheric C into the soil by inhibiting the decomposition of organic plant matter. However, when disturbed, e.g. through drainage for agriculture or road construction, the microbial decomposition is fostered, gradually converting these C sinks into sources leading to long-term effects.
By drying out the top peat layers, these disturbances also increase the risk of fires in these ecosystems. Wildfires turn these peatlands very abruptly into a C source by releasing large quantities of carbon dioxide (CO₂), methane (CH₄), and toxic aerosols over prolonged periods. Major fire events, such as those in Indonesia during the 1997–1998 El Niño and the 2015 fire crisis, have demonstrated the severity of peatland fires, emitting carbon at levels comparable to global fossil fuel emissions over short timeframes.
The VSC's advanced computational resources made handling the large datasets necessary to do this research possible.
To assess the risk of a potential fire, fire danger rating systems, such as the Canadian Fire Weather Index (FWI) system are widely used. However, the FWI, which assesses fire danger based on meteorological variables, was originally designed for boreal forests and may not fully capture the hydrological characteristics of tropical peatlands. The FWI’s moisture codes, which estimate fuel dryness based on temperature, humidity, precipitation, and wind speed, were calibrated for mineral soils rather than organic peat. As a result, its ability to predict fire danger in peatlands remains uncertain.
This study evaluates the effectiveness of a peatland-specific version of the FWI (FWIpeat) that incorporates hydrological modeling to improve fire danger assessment in the tropical peatlands of peninsular Malaysia, Sumatra, and Borneo. We modified the FWI by replacing its meteorological-based moisture codes with groundwater table estimates from a peatland-specific land surface model (PEATCLSM; Figure 1). The performance of FWIpeat, the standard FWI (FWIref), and the Drought Code (DC) was assessed against satellite-based fire occurrence data from 2002 to 2018.
Our results show that DC performs best in predicting fire occurrence over undrained peatlands, outperforming both FWIref and FWIpeat. Over drained peatlands, DC and FWIpeat perform similarly, both providing a better fire danger assessment than FWIref. These findings highlight the importance of incorporating deeper soil moisture information when evaluating fire risk in tropical peatlands. However, the improvements from FWIpeat are smaller than those observed in boreal peatlands, likely due to the challenges of accurately modeling hydrology in drained areas and the dominant role of human ignitions in the region.
This study underscores the potential for integrating peatland-specific hydrology into fire danger assessment but also identifies key areas for improvement. Refining hydrological input data and incorporating additional constraints from Earth observation datasets and human management practices could enhance the predictive capability of fire danger models in tropical peatlands, supporting more effective fire management and mitigation strategies.
The VSC's advanced computational resources made handling the large datasets necessary to do this research possible.
Read the full publication in the international journal of Wildland Fire here