The Mediterranean basin is a region historically affected by forest fires. They consume the vegetation layer affecting the ecosystem, causing biodiversity loss, increasing soil erosion and releasing trace gases and aerosols into the atmosphere. In order to assess the impacts of forest fires, researchers and managers highlighted the need of accurate burned area estimations and the quantification of the level of damage produced on soil and vegetation.

In this project the assessment of the impacts is measured by means of burn severity (BS), defined as the degree of damage on soil and plants when the fire is extinguished. Therefore, BS quantifies changes in the vegetation with respect to the pre-fire scenario. In this context, BS can be used as an estimation of the biomass consumed by the fire, namely burning efficiency (BE). Burning efficiency is one of the parameters needed for the estimation of greenhouse gas (GHG) emissions from fires, following the equation proposed by the Intergovernmental Panel on Climate Change (IPCC) in 2006.

L_fire = A x M_Β x C_ƒ x G_εƒ x 10^-3

Where L_fire is the amount of greenhouse emissions from fire (tons of each GHG), A is the burned area (m²), M_Î’ is the mass of fuel available for combustion (tons/m2), C_Æ’ is the combustion factor or burning efficiency (dimensionless), and G_εƒ is the emission factor of each GHG (g/kg of dry burned matter).

Satellite monitoring of burned areas, burn severity and burning efficiency will enable to obtain a more thorough examination of fires characteristics, to observe the impacts in ecosystems, to select priority areas to be restored, to follow ecosystem dynamics after a fire event, and to estimate gas emissions at global scale.

The first aim of FEMM project is to estimate burn severity in Europe using ESA Envisat MERIS Full Resolution data (spatial resolution of 300 meter). The second objective is focused on the development of a method to assess burning efficiency from burn severity estimation and the pre-fire state of the vegetation. The estimation of BS and BE using MERIS images will enable to cover a wide area and will allow generating a burn severity map covering all the European Mediterranean forest fires at the end of the fire season.

In order to achieve the objectives previously described several phases were planned. The first phase consisted in the application of a novel methodology developed to estimate burn severity with NASA Landsat data based on radiative transfer models (RTM). The physical foundation of the RTM assure this method to be applicable to different regions, what makes this method the most recommended to be used at a regional-global scale. This first phase was carried out in two big fires occurred in Spain in 2009. Burn severity was estimated using one MERIS image dated 25th of September 2009, and was validated with the burn severity estimation computed from two Landsat images (202/032 - 2009/08/30, and 199/032 - 2009/07/24). In order to verify the capability of Landsat as reference data in the validation process, the Landsat burn severity estimation was at the same time validated with field data sampled in both fires. The results were very promising with R2 values of 0.92 and 0.95 resulting from the comparison between MERIS and Landsat burn severity estimation. The next phase involved the connection of two processes: (i) burned area mapping, and (ii) burn severity estimation. Burned area mapping was performed using an original methodology especially designed for MERIS imagery. Besides, two wider regions were chosen, namely the Iberian Peninsula (2009 fire season) and Greece (2007 fire season).

The next phase will be related to the second objective of the project, the estimation of burning efficiency. This parameter is usually considered as a static value depending on vegetation type or fuel model, and it is estimated from laboratory or field experiments. However, several authors suggested that this parameter should be dynamic as the biomass is not completely consumed after a fire. Then, in first place a thorough search of these values will be done. Later, the static value of BE will be converted into a dynamic variable according to the burn severity estimated in a specific region. In this process two potential difficulties arise: the first is related to the BE value selected for each vegetation type, and the second refers to how the results can be validated.