The LANDFIRE Annual Disturbance products for 1999-2007 are included within the LF 2001 version and are developed through a multistep process. Inputs to this process include (but are not limited to): Landsat imagery derived Normalized Burn Ratio (NBR) data (in CONUS only), polygon data developed by local agencies, fire data obtained from Monitoring Trends in Burn Severity (MTBS), Burned Area Reflectance Classification (BARC), and Rapid Assessment of Vegetation Condition after Wildfire (RAVG) fire mapping efforts, and Protected Area Database (PAD) data. Event polygon data are provided to LANDFIRE by various local, regional, and national agencies and organizations. Disturbance type and year information is included as attributes for each polygon and transferred to the disturbance grids. Severity is determined by using dNBR (difference Normalized Burn Ratio) data classified into high, medium, and low severity levels based on a statistical comparison with the MTBS, BARC, and RAVG fire severity.
Vegetation Tracker (VCT, Huang, et. al. 2008) algorithms are used to identify disturbances outside of LF 2001 Events. VCT data are developed for each year identifying disturbed areas and severity. Since disturbance type (i.e. causality) is not determined in the VCT process, a spatial analysis was done comparing the VCT output to buffered (1 kilometer) LF 2001 Events and PAD GAP land use characteristics. While not providing a precise type of disturbance, this analysis provides information useful for narrowing down the types of disturbance that could or could not typically occur. Each zone has ten disturbance grids, one for each year 1999 to 2007. Each grid is attributed with year, disturbance type (if known, otherwise a description of possible types), severity, and the data sources used to create the data.