Global fire danger forecast in support to the Global Wildfire Information System

Ongoing research project |
2013 - 2016

Project Objective

Tender JRC/IPR/2013/H.3/0014/OC

This successful collaboration between ECMWF  and the Joint Research Centre (JRC) ran from February 2013 to July 2016. It achieved its key objectives of producing fire danger weather forecast based on ECMWF ensemble prediction  system

In particular, making the first step towards the creation of an operational early warning system for fire  went far beyond the original feasibility  plans. This collaboration was the first step toward what  lead ECMWF to bid successfully to become the computation centre for the Copernicus Emergency Management Service-fire


In 2013 it was clear that the European Forest Fire Information System (EFFIS), the EC focal point of information on forest fires, was planned  to be gradually expanded into a Global Wildfire Information Systems, supporting global initiatives such as GOFC/GOLD and GEOSS. The FORESTMOD Action is member of the GOFC-GOLD (Global Observation of Forest Cover-Global Observation of Land Dynamics) Fire Implementation Team which aims at the coordination of fire related activities at global level, including fire danger assessment. Within these activities, the Joint Research Centre (JRC), next to the Canadian Forest Service, was leading  the development of a global fire early warning system. The prototype Global Wildfire Information System was foreseen as the basis for a cooperative approach towards an integrated system that provides access to all fire related available information on a global scale.

In 2011 scientific developments had already  shown that in addition to the well established fire danger forecasts with lead times of a few days, skilful predictions of fire occurrence is possible up to several months ahead. For example, Chen et al. 2011 showed the good predictability of fire occurrence in the Amazon with 3-5 month lead time based on predictions of two sea-surface-temperature-based climate indices. Full seasonal weather forecasts could  provide even more information.

As these evidences became more and more convincing, it  seemed appropriate to extend fire danger forecasts with global coverage from several days lead time to cover all time scales up to seasonal predictions using meteorological input from the best available weather forecasting system, including an ocean model. Such forecasts were expected to aid the appropriate development and distribution of emergency response resources in and among different regions, amongst other applications.

This Project had as a amjor objective  to enhance and extend to forecast horizon of the European Forest Fire Information System (EFFIS) fire danger forecast capability for Europe, which, at that time, were limited to 6 days lead time with 25 km spatial resolution. Longer lead times in the predicting the Fire Weather Index  were achieved by using ECMWF  weather forecast. Moreover, as single weather forecasts are restricted by the numerical representation of physical processes, the resolution of the simulated atmospheric dynamics and the sensitivity of the solutions to the pattern of initial conditions, Ensemble Prediction Systems (EPS) that take account of these uncertainties showed to be more skill full in medium range (3-15 days) weather predictions. It was therefore important to utilize this performance advantage of EPS in order to predict the Fire Weather Index (FWI) beyond three days.

This projects for three years  delivered medium range Fire hazard predictions which were used  for the EFFIS system (


(Chen, Y., Randerson, J. T., Morton, D. C., DeFries, R. S., Collatz, G. J., Kasibhatla, P. S., Giglio, L., Jin, Y., and Marlier, M. E. (2011). Forecasting fire season severity in south america using sea surface temperature anomalies. Science, 334:787–791.)

JRC logo

Francesca Di Guiseppe

Florian Pappenberger