Methodology

Data used in Spain in Flames, both in the fires map and the journalistic investigations, comes from the General Statistic of Forest Fires (EGIF, in Spanish), elaborated by the Coordination Centre of National Information about Forest Fires (CCINIF, in Spanish). The information is collected by each autonomous community in Spain from the fire record filled in situ in each fire. This record contains more than 150 text entries. When data from the autonomous communities arrives to CCINIF, it is processed and consolidated into the EGIF. The information is then sent to the European Commission. This long, complicated process make the data available only two years afterwards.

We used the Freedom of Information Act in 2012 to access to the data of fires forest from 2001 to 2010. Since then, we have asked for the EGIF each year, being 2015 the most recent database available. In order to process and analyze the data, we converted the database from its original format MS Access to MySQL by using Cynergi’s conversion utility. We also converted the location coordinates from UTM to WGS 84 and the control and extinction times to minutes.

In the database analysis, a series of shortcomings were identified in the homogenization of the data:

• The cause of each forest fire was a supposed cause and not a certain cause in more than half of the cases.

• The geolocation of the fire was not specified in almost 18% of the cases. For these scenarios, we used the location of the town to which the fire is linked for its display in the map. There are 24 fires without geolocation nor town linked to the fire, so we decided to left them out of the map. We also detected 441 fires with a wrong geolocation since the coordinates were out of the region in which it happened. In these cases, we used the town coordinate as well.

• Economic losses information is limited, which make almost impossible the idea of analyzing the real cost of the fires in Spain. More than 30% of the fire records don’t include the costs of extinction and almost 9% don’t include economic losses.

Plus, official data could contain mistakes. Civio can’t corroborate all and each one of the data registers of each fire, so we have to work with the official information available. This is the reason why we decided to locate the fires with the coordinates given by the official records -except for the cases mentioned above- even if they locate fires in the Mediterranean Sea.

In order to offer a good performance in the map, only fires that burned 1 ha or more are displayed. These are 82.583 fires, a 35% of the total, but they represent almost 98% of the burned surface between 2001 and 2015. To display this data, we use the location of the fire as a center point and then we expand the circle area in proportion with the surface burned. This doesn’t mean that the fire propagated in a circular figure, but it has the informative purpose of showing a visual idea of the real extension of the burned surface.

Here is available to download the .csv file with all fires that burned 1 ha or more displayed on the map.