The severe El Niño episode of 2015 led to a major and damaging increase in Indonesian peatland fire, highlighting an urgent need to develop operational systems to forecast potentially severe fire events in order to mitigate the impacts of fire and haze. 10 ASEAN states have formally agreed to control peatland and forest fires and urgently need an early fire warning system: a need that we address in this proposal. An operational 'early warning' system for forecasting dangerous burning conditions is within reach using state-of-the-art modelling tools, such as the ECMWF's System 4 seasonal forecast model, but is currently hampered by insufficient knowledge about the influence of fluctuations in peat moisture on fire, particularly during periods of extreme drought, highlighted by the 1997-98 and 2015 El Niño episodes- the strongest and second strongest on record. The majority of present-day fires in Indonesia result from deliberate burning for land clearance, and this human factor means that burning can be influenced by policy and altered land management practice. The translation of scientific research into evidence-based policy and the official regulation and restriction of burning do not work well in Indonesia and new approaches are needed. We plan to both develop a new scientific forecasting tool for fire danger and to influence policy and fire regulations: a novel combination of urgent science and policy research.
This project will develop a suite of climate-, hydrological- and agent-based modelling to predict the incidence of peat fires based on computations and observations for the period 1997 to 2014, and will use the 2015 El Niño event to benchmark the forecast tools. Our working hypothesis is that the increased fire risk associated with dry peat does not trigger appropriate changes in the management practices adopted by local landowners in their use of fire, if there are no
incentives provided by policy. Our anticipated outputs are:
* An operational model of peatland fire occurrence, based on a tropical peatland hydrology model, an agent-based model,
and seasonal climate data derived from state-of-the-art reanalysis data and seasonal forecasts, and Earth Observation Data.
* An operational early fire warning system for peatlands based on seasonal climate forecasts.
* A more complete understanding of how climate, socio-economic and geographic factors interact to drive peatland fires.
* Evidence-based policy tools for reducing the number of fires and area burned each year in Riau province, Sumatra.
* Evidence-based proposals towards new Indonesian fire reduction management strategies and policy input.
The forecasting system will be web-based and accessible, and will predict the risk of peatland fire occurrence up to three months ahead, enabling sufficient time to spread awareness of the impending risk through the community, and to mobilise fire-fighting resources and other fire prevention measures if required. Consideration of non-climate drivers of peatland fire occurrence is critically important because this will help us capture the spatio-temporal patterns in fire in different regions displaying similar climate regimes. Non-climate driven factors also present the most tractable means to develop mitigation actions. We will combine the results of our work on climate, socio-economic and geographic factors to generate a multifactorial model for peatland fire occurrence. The model system will be developed in close collaboration with Indonesian stakeholders following the operational needs of agencies, municipalities and companies in the area. Key stakeholders include the Indonesian peatland restoration agency, Indonesia ministry, ASEAN Regional Haze Support Unit, local communities, forestry companies, local and international researchers. The model system will be robust and simple enough for in-house daily use by our Indonesian stakeholders, who will take over the system, and run and maintain it on their own servers.