FuelMap 2020 and 2022: Imputed map of carbon stored in litter, duff, fine woody debris, and coarse woody debris for CONUS forests circa 2020 and 2022

FuelMap 2020 and 2022 are imputed maps of litter, duff, fine woody debris, and coarse woody debris loadings for the forests of the conterminous United States (CONUS) circa 2020 and 2022. In fire science, these strata are often referred to as “fuel” for wildland fire. FuelMap 2020 and 2022 are largely derived from TreeMap 2020 and 2022, which provide a tree-level model of CONUS forests. To create TreeMap, we assigned forest plot data measured by USDA Forest Service’s Forest Inventory and Analysis (FIA) program to a 30x30 meter (m) grid. Specifically, we used a random forests machine-learning algorithm to impute the most similar forest plots to a set of target rasters provided by Landscape Fire and Resource Management Planning Tools (LANDFIRE: https://landfire.gov) and Daymet (https://daymet.ornl.gov/). Predictor variables for both the forest plots (reference data) and rasters (target data) consisted of percent forest cover, forest height, and vegetation type, as well as topography (slope, elevation, and aspect), location (latitude and longitude), 30-year biophysical variable normals from 1981-2010 (precipitation, maximum temperature, minimum temperature, snow water equivalent, shortwave radiation, vapor pressure, and vapor pressure deficit), and disturbance history (time since disturbance and disturbance type) for the landscape circa 2020 and 2022. FIA records downed woody material (DWM) in litter, duff, fine and coarse woody debris pools at some but not all of their forest plots. Thus, many of the FIA plots imputed (assigned) in TreeMap have DWM measurements attached to them. For pixels in TreeMap where the assigned FIA plot recorded DWM, we used the FIA plot assigned in TreeMap 2020 and 2022 in FuelMap 2020 and 2022. For pixels where FIA plots were assigned that did not have DWM measured, we ran a secondary imputation which included only FIA plots where DWM was measured to identify the most similar plot with DWM measured, and then we assigned that identified plot in the FuelMap 2020 and 2022. The main outputs of this project are rasters at 30x30 m spatial resolution of the imputed FIA plot identifier. The plot identifier corresponds to a unique visit to a plot by an FIA field crew, and is also referred to as the plot control number (CN). Using the CN, we look up the loading in each of the carbon pools (in pounds per acre) in the DWM_COND_CALC table of the FIA DataMart and include a raster for each of the following measurements: 1) litter, 2) duff, 3) fine woody debris in the 1-hour (hr) size class (size less than 0.25 inches in diameter), 4) fine woody debris in the 10-hr size class (size from 0.25-1 inch in diameter), 5) fine woody debris in the 100-hr size class (size from 1-3 inches in diameter), 6) coarse woody debris in the 1000-hr size class (size greater than 3 inches in diameter), and 7) “total carbon” in the DWM strata produced by adding these six strata together. We present these data in GeoTIFF formats. The spatial extent is CONUS for landscape conditions circa 2020 and 2022. The carbon loadings for DWM are drawn from the FIA COND_DWM_CALC tables for the assigned plot CN for litter, duff, fine woody debris and coarse woody debris.The FuelMap provides spatial estimates of forest floor carbon at fine spatial resolution (30x30 m). These spatially contiguous estimates are needed for landscape-level analyses of forest biomass and carbon, as well as estimates of emissions from wildland fire. The FuelMap can be used in combination with the TreeMap to provide estimates of forest carbon that include both the trees and forest floor. The FuelMap is the first imputed dataset to partition forest floor carbon into different strata of which we are aware. FuelMap does not provide estimates of soil carbon or carbon stored in roots of trees, nor of carbon stored in understory shrubs or grasses.See the Entity and Attributes section for details regarding the relationship between the data files included in this publication and the FIA DataMart (https://doi.org/10.2737/RDS-2001-FIADB). For additional information regarding TreeMap 2014, 2020, and 2022 see Riley et al. (2021a, 2021b), Riley et al. (2022), Riley et al. (2019), Zimmer et al. (2025), and Houtman et al. (2025).