The LANDFIRE Limited Disturbance (LDist)23 product is a new product introduced with the LF 2023 Update (LF 2023). LDist23 is an early "draft" of the LANDFIRE Annual Disturbance product and includes disturbance events captured through October 31, 2023. LDist23 is the first of three Annual Disturbance products releasing in the LF 2023 Update. LDist23 releases in January 2024, then Preliminary Disturbance (PDist)23 will release mid-year 2024. PDist23 will be the second "draft" of Annual Disturbance for the LF 2023 update. Finally, in the fall of 2024 the final "draft" of Annual Disturbance (Dist23) will be released.

LANDFIRE's (LF) 2023 update (LF 2023) Existing Vegetation Type (EVT) represents the current distribution of the terrestrial ecological systems classification developed by NatureServe for the western hemisphere. In this context, a terrestrial ecological system is defined as a group of plant community types that tend to co-occur within landscapes with similar ecological processes, substrates, and/or environmental gradients. See the EVT product page (https://landfire.gov/vegetation/evt) for more information about ecological systems and NVC classifications. EVT is mapped using decision tree models, field data, Landsat imagery, topography, and biophysical gradient data. Decision tree models are developed separately for tree, shrub, and herbaceous lifeforms which are then used to produce a lifeform specific EVT product. These models are generated for each Environmental Protection Agency (EPA) Level III Ecoregion (https://www.epa.gov/eco-research/ecoregions). Riparian, alpine, sparse, and other site-specific EVTs are constrained by predetermined masks. In LF 2023 Conterminous United States (CONUS) extent, LF will map the lifeform, cover, and height of existing vegetation in areas that were mapped as disturbed over the last twenty years (see LF Annual Disturbance products) using machine learning methods. These disturbed areas were the focus because they are the areas that have changed the most since LF 2016 Remap. To learn more about this new methodology for LF EVC, EVH, and Existing Vegetation Type (EVT) go to https://www.landfire.gov/data/lf2023.

LANDFIRE's (LF) 2022 update (LF 2022) Existing Vegetation Type (EVT) represents the current distribution of the terrestrial ecological systems classification developed by NatureServe for the western hemisphere. In this context, a terrestrial ecological system is defined as a group of plant community types that tend to co-occur within landscapes with similar ecological processes, substrates, and/or environmental gradients. EVT also includes ruderal or semi-natural vegetation types within the U.S. National Vegetation Classification [(NVC) https://usnvc.org/]. See the EVT product page (https://www.landfire.gov/evt.php) for more information about ecological systems and NVC classifications. EVT is mapped using decision tree models, field data, Landsat imagery, topography, and biophysical gradient data. Decision tree models are developed separately for tree, shrub, and herbaceous lifeforms which are then used to produce a lifeform specific EVT product. These models are generated for each Environmental Protection Agency (EPA) Level III Ecoregion (https://www.epa.gov/eco-research/ecoregions). Riparian, alpine, sparse, and other site-specific EVTs are constrained by predetermined masks. Urban and developed areas are derived from the National Land Cover Database (NLCD), and the latest Microsoft Building Footprint dataset. Agricultural lands originate from the 2022 Cropland Data Layer (CDL) and the 2019 California Statewide Crop Mapping layer. Burnable developed classes are identified from building footprint dataset thresholds. LF 2022 retains circa 2016 EVT labels except where shifts in urban, agriculture, and developed classes occur. While Existing Vegetation Cover (EVC) and Height (EVH) are updated using transition rulesets with ST-Sim to account for disturbances, EVT remains unchanged, therefore EVT lifeform is not synchronized to the EVC/EVH lifeform as in some previous versions. LF uses EVT as an input for LF 2022 Fuel Vegetation Type (FVT).

LANDFIRE’s (LF) 2023 Vegetation Departure (VDep) product categorizes departure between current vegetation condition and reference vegetation condition, according to the methods outlined in the Interagency Fire Regime Condition Class Guidebook [FRCC Guidebook (Hann et al 2010)]. VDep differs from the FRCC Guidebook, however, because it is based on the departure of current vegetation condition only, whereas the FRCC Guidebook approach includes departure of current fire regimes for the reference period. For VDep, summary units are defined as a BioPhysical Setting (BpS) with identical reference condition values regardless of map zone. For example, when a BpS is present in map zone 1, 2, 4, 5, 6 and 8, the reference conditions for this BpS are identical in map zones 1, 2, 4, 5, and 8, those map zones become a summary unit for VDep computation. Since reference conditions are unique for this BpS in map zone 6, it is a separate summary unit for calculating VDep. Within each BpS summary unit, we compare the reference percentage of each Succession Class (SClass) to the current percentage, then the smaller of the two is summed to determine the similarity index for the BpS. This value is then subtracted from 100 to determine the departure value, VDep value is always between 0 and 100, with 100 representing maximum departure. Reference conditions are derived from quantitative vegetation dynamics models that mimic native, pre-European colonization disturbance regimes. The current conditions are derived from the corresponding LF 2023 SClass data for each BpS. The proportion of the landscape occupied by each SClass, in each BpS unit, within each summary unit represents current condition of that SClass in VDep calculation. VDep is based entirely on the remaining area of each BpS unit that is occupied by valid SClasses. Each pixel in a BpS within a summary unit has the same VDep value.

LANDFIRE’s (LF) Annual Disturbance products provide temporal and spatial information related to landscape change. Annual Disturbance depicts areas of 4.5 hectares (11 acres) or larger that have experienced a natural or anthropogenic landscape change (or treatment) within a given year. For the creation of the Annual Disturbance product, information sources include national fire mapping programs such as Monitoring Trends in Burn Severity (MTBS), Burned Area Reflectance Classification (BARC) and Rapid Assessment of Vegetation Condition after Wildfire (RAVG), 18 types of agency-contributed "event" perimeters (see LF Public Events Geodatabase), and remotely sensed Landsat imagery. To create the LF Annual Disturbance products, individual Landsat scenes are stacked and made into composites representing the 50th percentile of all stacked pixels (band-by-band) to reduce data gaps caused by clouds or other anomalies. Composite imagery from the specified mapping year, the two prior years, and the following year serve as the base data from which change products such as the Normalized Differenced Vegetation Index (dNDVI), the Normalized Burn Ratio (dNBR), and the Multi-Index Integrated Change Algorithm (MIICA) (Jin et al. 2013) are derived. Image analysts collectively use these datasets (separately or in combination) to isolate the true change from false change (commission errors). False changes can be attributed to many anomalies but are mostly caused by differences in annual or seasonal phenology, and/or artifacts in the image composites. Fire-caused disturbances sourced from MTBS may contain data gaps where clouds obscure the full burn scar from being mapped. Models trained from pre-fire and post-fire Landsat data are used to fill these gaps. The result is gap-free continuous severity and extent information for all MTBS fire disturbances. MTBS pixels derived from gap filling techniques, such as modeling, are noted as such in the Annual Disturbance attribute table. Smaller fires that do not meet the size criteria set forth by MTBS may be attributed using Burned Area (BA), informed from Landsat Level-3 science products and only available in the lower 48 states. Causality and severity information assigned to a disturbance are prioritized by source, with the highest priorities reserved for fire mapping programs (MTBS, BARC, and RAVG) followed by user-contributed events contained in the LF Events Geodatabase, and lastly, Landsat image-based change.

LANDFIRE’s (LF) 2022 Vegetation Departure (VDep) product categorizes departure between current vegetation condition and reference vegetation condition, according to the methods outlined in the Interagency Fire Regime Condition Class Guidebook (FRCC Guidebook (Hann et al 2010)). VDep differs from the FRCC Guidebook, however, because it is based on the departure of current vegetation condition only, whereas the FRCC Guidebook approach includes departure of current fire regimes for the reference period. For VDep, summary units are defined as a BioPhysical Setting (BpS) with identical reference condition values regardless of map zone. For example, when a BpS is present in map zone 1, 2, 4, 5, 6 and 8, the reference conditions for this BpS are identical in map zones 1, 2, 4, 5, and 8, those map zones become a summary unit for VDep computation. Since reference conditions are unique for this BpS in map zone 6, it is a separate summary unit for calculating VDep. Within each BpS summary unit, we compare the reference percentage of each Succession Class (SClass) to the current percentage, then the smaller of the two is summed to determine the similarity index for the BpS. This value is then subtracted from 100 to determine the departure value, VDep value is always between 0 and 100, with 100 representing maximum departure. Reference conditions are derived from quantitative vegetation dynamics models that mimic native, pre-European colonization disturbance regimes. The current conditions are derived from the corresponding LF 2022 SClass data for each BpS. The proportion of the landscape occupied by each SClass, in each BpS unit, within each summary unit represents current condition of that SClass in VDep calculation. VDep is based entirely on the remaining area of each BpS unit that is occupied by valid SClasses. Each pixel in a BpS within a summary unit has the same VDep value.

LANDFIRE's (LF) 2022 update (LF 2022) Existing Vegetation Cover (EVC) represents the vertically projected percent cover of the live canopy for a 30-m cell. EVC is produced separately for tree, shrub, and herbaceous lifeforms. Training data depicting percentages of canopy cover are obtained from plot-level ground-based visual assessments and lidar observations. These are combined with Landsat imagery (from multiple seasons), to inform models built independently for each lifeform. Tree, shrub, and herbaceous lifeforms each have a potential range from 10% to 100% (cover values less than 10% are binned into the 10% value). The three independent lifeform datasets are merged into a single product based on the dominant lifeform of each pixel. The EVC product is then reconciled through QA/QC measures to ensure lifeform is synchronized with Existing Vegetation Height (EVH). Urban and developed areas are derived from the National Land Cover Database (NLCD), and the latest available Microsoft Building Footprint dataset. Agricultural lands originate from the 2022 Cropland Data Layer (CDL) and the 2019 California Statewide Crop Mapping layer. Disturbance events after 2016 are accounted for by incorporating transition rulesets using LF 2022 Fuel Disturbance (FDist). LF uses EVC as an input for LF 2022 Fuel Vegetation Cover (FVC).

The LANDFIRE Preliminary Annual Disturbance (PDist)23 product is a new product introduced with the LF 2023 Update (LF 2023). While LDist23 is a first "draft", PDist23 is akin to a second “draft” of the LANDFIRE Annual Disturbance product and includes disturbance events captured through October 31, 2023. PDist23 is releasing for these extents throughout the calendar year 2024: Conterminous United States (CONUS), Alaska (AK), Hawaii (HI), and Puerto Rico and the US Virgin Islands (PRVI). See the LF 2023 page for more information as this page will be added to as more details are available. https://landfire.gov/data/lf2023 Preliminary Annual Disturbance (PDist23) was created with Landsat imagery, submitted disturbance events from the data call, and online fire program data collected on or before October 31, 2023. Unlike the Limited Annual Disturbance (LDist23) product released earlier in calendar year 2024, PDist23 includes remotely sensed severity for submitted events and unknown disturbances where change was detected.

LANDFIRE’s (LF) Annual Disturbance products provide temporal and spatial information related to landscape change. Annual Disturbance depicts areas of 4.5 hectares (11 acres) or larger that have experienced a natural or anthropogenic landscape change (or treatment) within a given year. For the creation of the Annual Disturbance product, information sources include national fire mapping programs such as Monitoring Trends in Burn Severity (MTBS), Burned Area Reflectance Classification (BARC) and Rapid Assessment of Vegetation Condition after Wildfire (RAVG), 18 types of agency-contributed "event" perimeters (see LF Public Events Geodatabase), and remotely sensed Landsat imagery. To create the LF Annual Disturbance products, individual Landsat scenes are stacked and made into composites representing the 50th percentile of all stacked pixels (band-by-band) to reduce data gaps caused by clouds or other anomalies. Composite imagery from the specified mapping year, the two prior years, and the following year serve as the base data from which change products such as the Normalized Differenced Vegetation Index (dNDVI), the Normalized Burn Ratio (dNBR), and the Multi-Index Integrated Change Algorithm (MIICA) (Jin et al. 2013) are derived. Image analysts collectively use these datasets (separately or in combination) to isolate the true change from false change (commission errors). False changes can be attributed to many anomalies but are mostly caused by differences in annual or seasonal phenology, and/or artifacts in the image composites. Fire-caused disturbances sourced from MTBS may contain data gaps where clouds obscure the full burn scar from being mapped. Models trained from pre-fire and post-fire Landsat data are used to fill these gaps. The result is gap-free continuous severity and extent information for all MTBS fire disturbances. MTBS pixels derived from gap filling techniques, such as modeling, are noted as such in the Annual Disturbance attribute table. Smaller fires that do not meet the size criteria set forth by MTBS may be attributed using Burned Area (BA), informed from Landsat Level-3 science products and only available in the lower 48 states. Causality and severity information assigned to a disturbance are prioritized by source, with the highest priorities reserved for fire mapping programs (MTBS, BARC, and RAVG) followed by user-contributed events contained in the LF Events Geodatabase, and lastly, Landsat image-based change.

LANDFIRE's Annual Disturbance products track how landscapes change across space and time on an annual basis. The Annual Disturbance (Dist) product identifies satellite-detected areas larger than 4.5 hectares (11 acres) that underwent natural or human-caused changes within a specific year (for Dist23, October 1, 2022 – September 30, 2023), or represent fire activity/field treatments as small as 80 square meters. While creating the Annual Disturbance product a variety of data sources are leveraged. 1) National fire mapping programs: This includes information from Monitoring Trends in Burn Severity (MTBS), Burned Area Reflectance Classification (BARC), and Rapid Assessment of Vegetation Condition after Wildfire (RAVG), which offer severity information for fire-caused disturbances. 2) Agency-reported events: There are 18 designated classes for contributed polygon "Event" types such as disease, insects, development, harvest, etc. that are reported by government agencies for inclusion into the disturbance product. 3) Remotely sensed imagery: Harmonized Landsat Sentinel (HLS) satellite images offer a comprehensive-uninterrupted view of the landscape covering all lands, public and private, to fill in the gaps inherent in the previous data sources. These data are reviewed and edited by a team of image analysts to ensure and maintain high quality standards. To create the LF Annual Disturbance product, individual Landsat scenes are stacked and made into composites representing the 15th, 50th, and 90th percentiles of all stacked pixels (band-by-band) to reduce data gaps caused by clouds or other anomalies. Composite imagery from the specified mapping year and the two prior years serves as the base data from which change products such as the Normalized Differenced Vegetation Index (dNDVI), the Normalized Burn Ratio (dNBR), and the Multi-Index Integrated Change Algorithm (MIICA) (Jin et al. 2013) are derived. Image analysts collectively use these datasets (separately or in combination) to isolate the true change from false change (commission errors). False changes can be attributed to many anomalies but are most commonly caused by differences in annual or seasonal phenology, artifacts in the image composites, or difficult to map classes such as wetlands and grasses. Fire-caused disturbances sourced from MTBS may contain data gaps where clouds obscure the full burn scar from being mapped. Models trained from pre-fire and post-fire Landsat data are used to fill these gaps. The result is gap-free continuous severity and extent information for all MTBS fire disturbances. MTBS pixels derived from modeling are noted as such in the Annual Disturbance attribute table. Smaller fires that do not meet the size criteria set forth by MTBS may be attributed as fire by using Burned Area (BA) Level-3 science products derived from Landsat 8 and 9. BA data is only available in the lower 48 states (CONUS). Causality information assigned to annual disturbance products are prioritized by source, with the highest priorities reserved for fire mapping program data (MTBS, BARC, and RAVG) followed by user-contributed events contained in the LF Events Geodatabase, and lastly, satellite image-based change. Severity is assigned directly from fire program data. For events and satellite-detected change, severity is derived from pre- and post-burn standard deviation values of the differenced Normalized Burn Ratio (dNBR). When mapping the LF Annual Disturbance product, the start date is utilized for disturbances from fire program data whereas all other disturbances utilize the end date.