With the start yesterday of the Cabin Fire in the Angeles National Forest, and the requisite fear that it will become another Station Fire, I decided to take a quick look at the history of wildfires in Southern California. The California Department of Forestry and Fire Protection’s Fire and Resource Assessment Program (FRAP) generates a lot of geospatial data that it disseminates for free to the general public. One of these is a layer containing the boundaries of all known fires 10 acres or greater since 1878. That data set is provided as an ESRI File Geodatabase and is shown below.
Using QGIS 2.10, I selected fires in southern California, roughly defined as Point Concepcion south to the border with Mexico.
I then export the data to an ESRI Shapefile.
From here on out, I’ll leave the GIS world and manipulate the data table from the southern California layer in Excel. The image below shows the data fields in the shapefile’s DBF table. I’ll save the DBF as an Excel file.
I then need to convert the ALARM_DATE (date fire started) and CONT_DATE (fire containment date) to date formats that are usable in Excel. I create fields next to them, and modify the date fields using the formula:
Throwing out the handful of rows with bad date values, the result shows that there have been approximately 6,329 wildfires in southern California since 1902.
I first want to look at how many acres have burned per year. To do that, I create a pivot table, using the Year column and then the sum of the GIS_ACRES field. I then create a chart based on that table. I include a trend line on the chart, which seems to indicate a definite increase in fire frequency.
How about looking at the acreage burned per month?
If I only include data since January 1, 2000, it looks a little different:
There have been 447 fires in the Angeles National Forest between 1910 and 2014.
The obvious peak in August is clearly because of the 2009 Station Fire. The total historical burn acreage for the ANF from all fires is 655,602 acres. The Station Fire was 160,833 acres. Removing the Station Fire from the data paints a very different picture and shows a fairly steady fire risk through the middle and late summer and early fall.
I may play around with the data some more when I get the time.