Initially CDF, and the USDA Forest Service jointly developed a comprehensive fire perimeter GIS layer for public and private lands throughout California. The data covered the period 1950 to 2001 and included USFS wildland fires 10 acres and greater, and CDF fires 300 acres and greater. BLM and NPS joined the effort in 2002 collecting fires 10 acres and greater. Also in 2002 CDF’s criteria expanded to include timber fires 10 acres and greater in size, brush fires 50 acres and greater in size, grass fires 300 acres and greater in size, wildland fires destroying three or more structures, and wildland fires causing $300,000 or more in damage. This project provides for annual updates. Whenever possible, CDF, the USFS, BLM, and NPS solicit additional fire perimeter data from other federal agencies (e.g. BIA, DOD) and local agencies, for incorporation into the fire perimeters mapping project.

The project produces an annual statewide fire history GIS data layer through four steps:

1. Standardize and combine existing digitized fire perimeters into a statewide GIS layer,
2. Identify and remove duplicate fires;
3. Fill in the gaps in data; and
4. Perform annual updates of the GIS layer.

In 1989, CDF units were requested to fill in gaps in their fire perimeter data as part of the California Fire Plan. FRAP provided each unit with a preliminary map of 1950-89 fire perimeters. Unit personnel also verified the pre-1989 perimeter maps to determine if any fires were missing or should be re-mapped. Each CDF Unit then generated a list of 300+ acre fires that started since 1989 using the CDF Emergency Activity Reporting System (EARS). The CDF personnel used this list to gather post-1989 perimeter maps for digitizing. The final product is a statewide GIS layer spanning the period 1950-1999. Annual updates have been made thereafter.

The current fire perimeter layer developed by BLM, CDF, NPS and USFS is the most complete digital record of fire perimeters in California. However it is still incomplete in many respects. Fires may be missing altogether or have missing or incorrect attribute data. Some fires may be missing because historical records were lost or damaged, fires were too small for the minimum cutoffs, documentation was inadequate, or fire perimeters have not yet been incorporated into the database. Agencies are at different stages of participation. CDF and the USFS have completed inventory for the majority of their historical perimeters back to 1950, while only 2002 - 2003 fires are currently present for BLM. So far only Sequoia-Kings Canyon and Yosemite National Park are actively turning in data from NPS. Furthermore, until the data capture process moves to the local level, the most recent fires will only be uploaded to the database once annually, during the month of February.

Other errors with the fire perimeter database include duplicate fires and over-generalization. While the data capture process attempts to identify duplicate fires resulting from multiple data sources (i.e. the USFS and CDF both captured and submitted the fire perimeter), some duplicates may still exist. Additionally, over-generalization, particularly with large old fires may show unburned "islands" within the final perimeter as burned. Users of the fire perimeter database must exercise caution in application of the data. Careful use of the fire perimeter database will prevent users from drawing inaccurate or erroneous conclusions from the data.

In some cases, different agencies record differing fire perimeters for the same fire. Above you can see two instances of duplicate fires with different perimeters, the White Deer fire and the Pierce fire. In the Pierce fire the most accurate fire perimeter is preserved along with the best metadata, such as date, cause, etc. (see tables A and B below). *Note: only sampling of fires and not spatially accurate

When comparing multiple data sets duplicates often arise and are managed by selecting the most accurate and complete record. Table “A” and “B” illustrate how attribute data from different agency records can be merged to produce the most accurate information for each fire record in the database. Cells containing * in table “A” and “B” represent data, which are preserved in the final database. In many cases only one agency recorded the fire perimeter, its cause, and date; therefore only one record exists.

Fire perimeter data from the statewide database are constantly under development. The data are maintained as an Arc/Info REGIONS coverage, which provides effective tools for handling overlapping polygons (for instance areas that burn more than once). Data contained in this coverage come from many sources. Some sources are more accurate or complete than others. The "AGENCY" field identifies the data source. Not all sources report all information, thus some records will have fields with no values.

FRAP produces the annual statewide fire perimeter GIS data layer by combining digitized fire perimeters from BLM, CDF, NPS, and USFS. The long-range goal for maintenance of fire perimeter data will eventually decentralize the data capture process to the individual fire station level. Immediately following a fire event local fire station personnel will map fires into the database. This process will be facilitated by a user friendly computer application that connects the local field user to a client-server database in Sacramento. CDF implemented a similar process, Emergency Activity Reporting System (EARS), for non-spatial data in 1989.