LANDFIRE 2016 Remap Includes 90-kilometer Buffer

Submitted by atripp on

The recent release of LANDFIRE (LF) 2016 Remap for Alaska concludes a multi-year effort to map vegetation, fuel, and fire regime for the entire United States (CONUS, Alaska, Hawaii, and Insular Areas). This new product suite will better inform wildland fire decision support systems and facilitate national and regional strategic fire and resource management (forest, habitat, carbon, wildlife) planning efforts.

LF’s modeling and mapping process is complex, requiring many iterative steps to generate a suite of 20+ geospatial rasters. The process involves developing composite imagery, disturbance maps, and vegetation maps such as Existing Vegetation Type (EVT), Existing Vegetation Cover (EVC), and Existing Vegetation Height (EVH).  Data fuels and fire regime products are then derived from these products.

For disturbance mapping, late- and early-season Landsat imagery for the year before and the year after the year of interest were used to create best-pixel composites by season and year. Subsequently, with the aid of the modified Multi-Index Integrated Change Algorithm (MIICA), Normalized Burn Ratio, and Normalized Difference Vegetation Index, analysts compared the composites to identify disturbances on the landscape.

For vegetation mapping, a similar process of creating best pixel composites was used, although it drew on more years and dates within the year to better distinguish vegetation. Utilizing these composites, EVT, EVC, and EVH were mapped to the Nature Serve’s Ecological Systems Classification Vegetation Type and to National Vegetation Classification (NVC) System with a Classification and Regression Tree (CART) model. CART model training utilized plot data from USDA-Forest Inventory and Analysis program among many other sources to acquire dependent variables (such as EVT) and a suite of extracted independent variables (from plot centers, e.g. bands, band wavelengths, and spectral indices) generated from Landsat data.. Other remotely sensed independent variables included digital elevation model data (slope and elevation) and biophysical gradients (precipitation and temperature). Lidar data were acquired and processed from point clouds and used in tandem with cover and height data from plots to model EVC and EVH.

Consistent with the rest of the LF 2016 Remap, project imagery and data were acquired and processed to model and map vegetation for the 90-kilometer buffer along the 1,538-mile (2475-kilometer) border with Canada, which will support analysis of near- and cross-border incidents and land management planning.

Including a 90-kilometer buffer in the LANDFIRE 2016 Remap for Alaska along the international border with Canada will support fire behavior projections and assessments, fuel treatments, risk assessments, firefighter and public safety issues, wildlife habitat, and land management planning that abut international borders.


Author Name
Timothy D. Hatten
Author Email