3. Create Labeling Data

Create Labeling Data #

FLAM › Create Labeling Data turns fire perimeter polygons into rasters of observed burned area on your model grid. Calculate Burned Area uses these rasters for calibration.

You don’t need this step if your observations are already gridded (for example GFED). In that case, see Input data.

Inputs #

FLAM Create Label Data window

FieldDescription fills from
Fire DelineationPolygon shapefile, one polygon per fire, in a projected CRS in metresFire Shape
Time ColumnAttribute with the fire date as YYYY-MM-DD. The tool preselects the first column whose name contains “date” or “time”.—
Base RasterDefines the output gridBase Map
Output FolderWhere the subfolders are createdResults

Output types #

Tick Daily and/or Monthly for each type:

TypePixel valueDefaultOutput
Burned Area (BA@centroid)Area of the fire (km²), placed in the pixel that contains the fire’s centroidMonthlyLabel_BA_monthly/BA_YYYYMM.tif
Fire Frequency (Fq@centroid)1 where a fire’s centroid falls in the pixelDaily, MonthlyLabel_Fq_daily/Fq_YYYYMMDD.tif
Fractional Burned Area (FB)Share of the pixel covered by fire polygons (0–1)MonthlyLabel_FB_monthly/FB_YYYYMM.tif

Daily outputs go to the matching *_daily folders, named with YYYYMMDD. Pixels without fire are NoData (0).

Sub-divider for FB (default 33) splits each pixel into an n × n grid to measure the burned share. With 33, there are 1089 sub-cells, so the precision is about 0.1 %. Higher values are more precise but slower.

Label data output folders

Burned Area (monthly) and Fire Frequency (daily) are the outputs that Calculate Burned Area needs.

If several fires in the same period have their centroids in the same pixel, values aren’t added up: the pixel keeps the value of only one fire. At coarse resolution, compare the totals with your source data.