Initialization

This section describes the main data sources employed to initialize the spatial state variables when we applied the MEDFIRE model to Catalonia (north-eastern Spain). State variables in the MEDFIRE model are spatial variables that describe the landscape context and conditions. Spatial variables are represented in raster format and cover the full extent of the study area at 100 m resolution. The state variables whose values change as a result of spatial dynamic processes are Land cover type (LCT) and Time since last fire (TSF). Other spatial state variables describe additional landscape features but are static: Elevation, Aspect, Ignition probability, Eco-region, Homogeneous fire regime zones and Main wind direction.

An extra sub-section details the initialization of Fire and Vegetation sub-models parameters.

Land cover type (LCT)

LCT describes land cover types and dominant tree species in forest areas of the region. Two LCT have been initialized: in the year 1989 and in year 2000. Information of non-forest categories was obtained from the second and third Land Cover Maps of Catalonia (LCMC) describing land cover from aerial photos (http://www.creaf.uab.es/mcsc/). Dominant tree species per cell were extracted from the Spanish Forest Map (http://www.magrama.gob.es/es/biodiversidad/temas/ecosistemas-y-conectividad/mapa-forestal-de-espana/) and National Forest Inventory (NFI) data available for Catalonia (http://www.magrama.gob.es/es/biodiversidad/temas/inventarios-nacionales/inventario-forestal-nacional/default.aspx). The spatial distribution of each dominant tree species or species groups was derived from the third NFI inventory data by applying krigging interpolation techniques allowing the generation of a continuous layer of information from data measured in NFI forest plots.

LCT is a categorical variable whose states are divided into those land covers that can be affected by fire disturbance: forests, shrubs, croplands and grasslands; and those that cannot: urban, water, rock. Forested cells include information of the dominant tree species in the canopy. The tree species considered are: Pinus halepensis, Pinus nigra, Pinus pinea, Pinus sylvestris, Quercus ilex and Quercus suber. Three additional forest categories are considered to complete the classification of forests in the study area: Other conifers, Other Quercus and Other trees.

(For a detailed description of the GIS procedures carried out to define the LCT state variable see the file below LCTinitialization.pdf )

Time since last fire (TSF)

TSF is a discrete count variable derived from fire historical statistics and accounts for the number of years since the most recent wildfire at a site-specific cell level. The wildfire data used consist of fire perimeters larger than 50 ha occurred in Catalonia between 1975 and 1999 mpaped at 1:50.000. The data set included 319 fires with a total area of approximately 296.000 ha. Non burned areas were assigned a default value of 200 years. Elevation, Aspect, Ignition probability, Eco-region, Homogeneous fire regime zones and Main wind direction

Elevation and Aspect variables values at 100 m resolution are derived from a 30 x 30 meters elevation model available at ICC.

Ignition probability is based on a layer describing the probability of fire ignition in Catalonia (basic fire risk) from anthropogenic and biophysical environmental variables. The base layer was developed by the regional government (http://www20.gencat.cat/portal/site/mediambient). The probabilities assigned to the basic fire risk categories were based on the frequency of fire occurrence between 1975 and 1988. In the following figure the legend for the ignition probability is: high (red), moderate (orange), and low (green).

Eco-regions are bio-climatic defined sub-regions of the study area (Vallecillo et al., 2009) that allow to introduce regionalization in post-fire transition probabilities as well as analyze the model outcomes by Eco-region.

Homogeneous fire regime zones are defined by Castellnou et al., (2009) and are used in the model to identify where is more possible a relief-driven fire happens in contrast with a wind-driven fire spread pattern. The layer has associated the percentage of predominant fire spread pattern type.

Wind main direction accounts for the wind direction in degrees from the north and is derived from historical meteorological data (http://www.meteo.cat/servmet/index.html). Dominant monthly wind direction data in summer (June to September) from 194 meteorological stations of a 10-year period was processed to identify the prevailing wind direction in each station that was spatially interpolated to the whole area.

Initialization of sub-model parameters

Fire sub-model requires as initial variables and parameters: (1) distributions describing the fire regime, (2) the variables modulating ignition, fire spread and burning processes, and (3) the pre-specific thresholds allowing fire suppression to be successful. See the file below FireSubmodelInitialization.pdf for the full description.

The ecological processes implemented in Vegetation sub-model are aimed to describe vegetation regeneration following fire disturbance and natural succession from shrubland to forest. Sub-model parameters are related with those processes and the initialization is specified in VegetationSubmodelInitialization.pdf .

References

Castellnou, M., Pagés, J., Miralles, M., & Pique, M. (2009). Tipificación de los incendios forestales de Cataluña. Elaboración del mapa de incendios de diseño como herramienta para la gestión forestal. Proceedings of the 5th Congreso Forestal Español. Ávila, Spain.

Vallecillo, S., Brotons, L., & Thuiller, W. (2009). Dangers of predicting bird species distributions in response to land-cover changes. Ecological Applications, 19(2), 538–549. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19323209

Files