Each data file contains 1440 columns and 720 rows, and has a 0.25 degree latitude by a 0.25 degree longitude spatial resolution. Data is available from 1997 through the present. The data is divided into 3 main datasets.

  1. Burned area
  2. Monthly emissions and fractional contributions of different fire types
  3. Daily / 3-hourly fields to scale the monthly emissions to higher temporal resolutions

Please be aware of the differences between GFED4 (without small fires) and GFED4s (with small fires). The mapped burned area is without small fires, this is the GFED4 burned area described in Giglio et al. (2013) and distributed via the UMD server as described below. The emissions fields are based on this dataset but boosted with small fires, GFED4s. If you are interested in the effect of this boost, the emissions fields also contain a layer with the fraction of emissions resulting from this boost.

The GFED4 burned area data set provides global, monthly burned area from 1997 through the present (as well as July 1995-June 1996), and higher temporal resolution daily burned area for a subset of the time series extending back to August 2000. Instructions for downloading the GFED4 burned area data are available here.

Annual hdf5 files with monthly GFED4s burned area (including small fires), emissions, and scalars to distribute the monthly emissions over the days as well as the diurnal cycle

The hdf5 files include fire carbon emissions (g C m-2 month-1) and dry matter emissions (kg DM m-2 month-1). In addition, fractional contributions of different fire types:

  1. Savanna, grassland, and shrubland fires
  2. Boreal forest fires
  3. Temperate forest fires
  4. Deforestation and degradation
  5. Peatland fires
  6. agricultural waste burning
A more extensive readme file https://www.geo.vu.nl/~gwerf/GFED/GFED4/Readme.pdf
Recommended emission factors (g species per kg dry matter burned) to calculate specific trace gas and aerosol emissions in combination with above listed fractional contributions https://www.geo.vu.nl/~gwerf/GFED/GFED4/ancill/
Regional estimates for the various trace gases and aerosol emissions https://www.geo.vu.nl/~gwerf/GFED/GFED4/tables/
Example code (Python and Matlab [the latter provided by Wolfgang Knorr]) on how to compute read the hdf5 files and compute trace gas or aerosol emissions https://www.geo.vu.nl/~gwerf/GFED/GFED4/ancill/code/

Biosphere fluxes: The CASA-GFED biosphere fluxes include NPP, Rh, and fires (BB). Units are in g C m-2 month-1. Our model does not have CO2 and nitrogen fertilization and is spun up to almost neutral conditions. The monthly datafiles are can be found along with the emissions estimates, please see the accompanying readme file for the file structure.


Tables with annual emission estimates for the various trace gas and emissions sources for the globe and the 14 basis regions we identified below. A digital version of this basis region map is included in the annual emissions files.


Below are a number of graphs (clickable for larger versions) about GFED4s and differences with GFED3. The abbreviations refer to the GFED basis regions.

GFED Basis Regions
Relative differences between GFED4 (grey bars) and GFED4s (black bars) versus GFED3, averaged over 1997-2011.

Left panel: the small fire burned area led to increased burned area in most regions where small fires (for example agricultural waste burning) are dominant. 2nd panel: herbaceous fuel consumption was lowered in our model to better match observations. 3th panel: total carbon emissions were somewhat higher on a global scale, but regionally differences were larger depending on the balance between the in general increased burned area and decreased fuel consumption. Right panel: the use of new emission factors also led to changes. For carbon monixide (CO) as shown here, separating extratropical forest into boreal and temperate forest as done in Akagi et al. (2011) led to increased CO:C ratios in the boreal, and decreases in this ratio in temperate forests. For some species the differences in emission factors was substantial, on a global scale CH4 emissions for example were 25% lower than in GFED3 mainly because of the new emission factors used.