Forest cover is rapidly changing at the global scale as a result of land-use change (principally deforestation in many tropical regions and afforestation in many temperate regions) and climate change. However, a detailed map of global forest gain is still lacking at fine spatial and temporal resolutions. In this study, we developed a new automatic framework to map annual forest gain across the globe, based on Landsat time series, the LandTrendr algorithm and the Google Earth Engine (GEE) platform. First, samples of stable forest collected based on the Global Forest Change product (GFC) were used to determine annual Normalized Burn Ratio (NBR) thresholds for forest gain detection. Secondly, with the NBR time-series from 1982 to 2020 and LandTrendr algorithm, we produced dataset of global forest gain year from 1984 to 2020 based on a set of decision rules. Our results reveal that large areas of forest gain occurred in China, Russia, Brazil and North America, and the vast majority of the global forest gain has occurred since 2000. The new dataset was consistent in both spatial extent and years of forest gain with data from field inventories and alternative remote sensing products. Our dataset is valuable for policy-relevant research on the net impact of forest cover change on the global carbon cycle and provides an efficient and transferable approach for monitoring other types of land cover dynamics.

Global Forest Resources Assessment has been monitoring the world forests at 5 to 10 year intervals since 1946. The Global Forest Resources Assessments (FRA) are now produced every five years in an attempt to provide a consistent approach to describing the world\u2019s forests and how they are changing. The Assessment is based on two primary sources of data: Country Reports prepared by National Correspondents and remote sensing that is conducted by FAO together with national focal points and regional partners. The scope of the FRA has changed regularly since the first assessment published in 1948. These assessments make an interesting history of global forest interests, both in terms of their substantive content, but also in their changing scope. Land spanning more than 0.5 hectares with trees higher than 5 meters and a canopy cover of more than 10 percent, or trees able to reach these thresholds in situ. It does not include land that is predominantly under agricultural or urban land use.

The data included in Data360 is a subset of the data available from the source. Please refer to the source for complete data and methodology details.

This collection includes only a subset of indicators from the source dataset.

To advance our understanding of forest cover changes, given the discrepancies, this work providesan original analysis by assessing five available remote sensing datasets (ALOS PALSAR forest and non-forest data, ESA CCI Land Cover, MODIS IGBP, Hansen/GFW on global tree cover loss, and Terra-I) toestimate the likely extent of current forests (circa 2018) and forest cover loss from 2001-2018, forwhich data was available. This assumes that no single approach or data source can capture majortrends everywhere; therefore, an all-available data approach is needed to overcome shortcomings ofindividual datasets. The main shortcomings of this approach, however, are that it does not account for forest gains, tends tounderestimate the conversion in dry forests ecosystems and lacks explicit assessment ofuncertainties across the different datasets.“Forest cover loss” in the all-available data analysis consists of observations (pixels) changing fromforest to non-forest at any time during 2000 to 2018. The spatial resolution chosen was 250m giventhe original resolutions of the datasets incorporated and on the understanding that forest areasshould be a minimum of 250 x250m (6.25 ha) to contain the functional attributes of a forest (e.g.species distribution, ecology, ecosystem services), rather than depicting individual trees or groups oftrees.According to our analysis, about 20% of total forest cover loss takes place in core forest, which welabel “primary forest loss”, while the remaining 80% results from the conversion of edge andpatched forests, which is labelled as “secondary forest loss”. Two thirds of total forest cover loss inthe period from 2000-2018 occurred in the tropics and subtropics, followed by boreal and temperateforests. A portion of the loss in temperate and boreal forests will not be permanent and might referto other types of natural forest disturbances produced by insects, fire, and severe weather, as wellas by felling of plantations or semi-natural forests as part of forest management.Much tropical forest cover loss is in South America and Asia, while subtropical forest cover loss ismainly in South America and Africa. When looking at countries by income levels, as defined by theWorld Bank, much of deforestation takes place in upper middle and lower middle-income countries.To the risk of simplifying, this suggests an increasing pressure on forests in the transition that occurswhen countries increase economic development. In the tropics, upper-middle income countriesdominate forest cover loss in South America, due to the influence of Brazil, and lower middle-income countries in Asia, due to the influence of Indonesia. Forest cover loss in the subtropics occursmainly in Brazil and Argentina in South America, many lower-middle income countries in SouthAmerica, and lower-income countries in sub-Saharan Africa. Most temperate and boreal forest coverloss, likely not all permanent, occurs in high-income countries (Russia), and North America (UnitedStates and Canada) Unfortunately, this data does not identify changes over time or land use interactions amongcountries. Reduced forest cover loss in some mainly high-income countries, except North America, isassociated with forest cover loss, particularly in lower- and upper-middle countries in the tropics. Interactions are informed by the “forest transition” effect. Forest transition dynamics occur whennet forest restoration replaces net forest cover loss in some specific place. The countries thatunderwent a forest transition that reduced forest loss and encouraged regrowth may have placedadditional pressure on forests outside their borders, thus displacing deforestation. The debate onforest transitions and leakage is quite controversial given its policy implications.Recent analysis, based on a land-balance model that quantifies deforestation due to global trade atcountry level in the tropics and sub-tropics, linked to a country-to-country trade model, found thatfrom 2005-2013, 62% of forest loss was caused by commercial agriculture, pasture and plantations.About 26% of total deforestation was attributed to international demand, 87% of which wasexported to countries with decreasing deforestation or increasing forest cover in Europe and Asia(i.e. China, India). Some of this displacement pressure may be reduced by land intensification. Global patterns of forest fragmentationIn this analysis we consider forest degradation alongside forest cover loss. Degradation is a multi-factorial phenomenon that includes amongst others loss of native species, appearance of invasivespecies, pollution damage, structural changes, selective timber removal and many more. Here weuse fragmentation as a proxy that can be detected through remote sensing; this is a critical aspect offorest degradation but does not capture all aspects. The change in spatial pattern and structure byfragmentation of forest into smaller patches or “islands” damages forest ecosystem services such ascarbon storage and climate mitigation, regulation, water provision, and habitat for biodiversity. These impacts are created by changes at forest edges, which include increased exposure to differentclimate, fire, wind, mortality, and human access. The increasing isolation of forest patchescontributes to long-term changes in biodiversity, including species richness and productivity,creating fundamental changes in forest ecosystems.We evaluated the fragmentation of forests using morphological spatial pattern analysis (MSPA)assessed on the two all-available data global forest cover maps corresponding to 2000 and 2018, todetermine forest cover transitions between different type of fragmentation classes (i.e. stable core,inner edges, outer edges, and patches). Changes between fragmentation classes over time aredefined as primary and secondary degradation based on their initial state, in contrast to forestswhich remain in the same fragmentation class as stable core, inner edge, outer edge, and patch. Inthis definition, primary degradation is a result of the fragmentation of core forests into forest withmore edges, reducing the area of continuous forest extent, and resulting in greater losses of carbonand associated ecosystem services such as biodiversity present in intact forests. Secondarydegradation is the conversion of edge forests into more fragmented classes, occurring in secondaryforests which may already be degraded and are more accessible and easier to deforest

The "Forest Proximate People" (FPP) dataset is one of the data layers contributing to the development of indicator #13, “number of forest-dependent people in extreme poverty,” of the Collaborative Partnership on Forests (CPF) Global Core Set of forest-related indicators (GCS). The FPP dataset provides an estimate of the number of people living in or within 5 kilometers of forests (forest-proximate people) for the year 2019 with a spatial resolution of 100 meters at a global level.

For more detail, such as the theory behind this indicator and the definition of parameters, and to cite this data, see: Newton, P., Castle, S.E., Kinzer, A.T., Miller, D.C., Oldekop, J.A., Linhares-Juvenal, T., Pina, L. Madrid, M., & de Lamo, J. 2022. The number of forest- and tree-proximate people: A new methodology and global estimates. Background Paper to The State of the World’s Forests 2022 report. Rome, FAO.

Contact points:

Maintainer: Leticia Pina

Maintainer: Sarah E., Castle

Data lineage:

The FPP data are generated using Google Earth Engine. Forests are defined by the Copernicus Global Land Cover (CGLC) (Buchhorn et al. 2020) classification system’s definition of forests: tree cover ranging from 15-100%, with or without understory of shrubs and grassland, and including both open and closed forests. Any area classified as forest sized ≥ 1 ha in 2019 was included in this definition. Population density was defined by the WorldPop global population data for 2019 (WorldPop 2018). High density urban populations were excluded from the analysis. High density urban areas were defined as any contiguous area with a total population (using 2019 WorldPop data for population) of at least 50,000 people and comprised of pixels all of which met at least one of two criteria: either the pixel a) had at least 1,500 people per square km, or b) was classified as “built-up” land use by the CGLC dataset (where “built-up” was defined as land covered by buildings and other manmade structures) (Dijkstra et al. 2020). Using these datasets, any rural people living in or within 5 kilometers of forests in 2019 were classified as forest proximate people. Euclidean distance was used as the measure to create a 5-kilometer buffer zone around each forest cover pixel. The scripts for generating the forest-proximate people and the rural-urban datasets using different parameters or for different years are published and available to users. For more detail, such as the theory behind this indicator and the definition of parameters, and to cite this data, see: Newton, P., Castle, S.E., Kinzer, A.T., Miller, D.C., Oldekop, J.A., Linhares-Juvenal, T., Pina, L., Madrid, M., & de Lamo, J. 2022. The number of forest- and tree-proximate people: a new methodology and global estimates. Background Paper to The State of the World’s Forests 2022. Rome, FAO.

References:

Buchhorn, M., Smets, B., Bertels, L., De Roo, B., Lesiv, M., Tsendbazar, N.E., Herold, M., Fritz, S., 2020. Copernicus Global Land Service: Land Cover 100m: collection 3 epoch 2019. Globe.

Dijkstra, L., Florczyk, A.J., Freire, S., Kemper, T., Melchiorri, M., Pesaresi, M. and Schiavina, M., 2020. Applying the degree of urbanisation to the globe: A new harmonised definition reveals a different picture of global urbanisation. Journal of Urban Economics, p.103312.

WorldPop (www.worldpop.org - School of Geography and Environmental Science, University of Southampton; Department of Geography and Geosciences, University of Louisville; Departement de Geographie, Universite de Namur) and Center for International Earth Science Information Network (CIESIN), Columbia University, 2018. Global High Resolution Population Denominators Project - Funded by The Bill and Melinda Gates Foundation (OPP1134076). https://dx.doi.org/10.5258/SOTON/WP00645

Online resources:

GEE asset for "Forest proximate people - 5km cutoff distance"

This data set was created by Transparent World, with the support of Global Forest Watch. Many studies depicting forest cover and forest change cannot distinguish between natural forests and plantations. This data set attempts to distinguish tree plantations from natural forest for seven key countries: Brazil, Cambodia, Colombia, Indonesia, Liberia, Malaysia, and Peru.Given the variability of plantations and their spectral similarity to natural forests, this study used visual interpretations of satellite imagery, primarily Landsat, supplemented by high resolution imagery (Google Maps, Bing Maps, or Digital Globe), where available, to locate plantations. Analysts hand-digitized plantation boundaries based on several key visual criteria, including texture, shape, color, and size.Each polygon is labelled with the plantation type and when possible, the species. A “gr” in front of the species name indicates a group of species, such as pines or fruit, where the individual species was not identifiable. The percentage of plantation coverage indicates a rough estimate of the prevalence of plantation within apolygon (as in the case of a mosaic). Types are defined as follows:* Large industrial plantation: single plantation units larger than 100 hectares* Mosaic of medium-sized plantations: mosaic of plantation units < 100 hectares embedded within patches of other land use* Mosaic of small-sized plantations: mosaic of plantation units < 10 hectares embedded within patches of other land use.* Clearing/ very young plantation: bare ground with contextual clues suggesting it will become a plantations (shape or pattern of clearing, proximity to other plantations, distinctive road network, etc)For more information on this data set and how it was produced, see the forthcoming WRI Technical Note associated with this project.

Global Forest Watch (GFW) is an online platform that provides data and tools for monitoring forests. By harnessing cutting-edge technology, GFW allows anyone to access near real-time information about where and how forests are changing around the world. Data on tree cover gain and loss have been obtained from Global Forest Watch. Tree cover gain data is from the Global Land Analysis & Discovery (GLAD) lab at the University of Maryland and "measures areas of tree cover gain from the year 2000 to 2020 across the globe at 30x30 meter resolution" (Global Forest Watch). This data was aggregated to country/economy level. Tree cover loss data is a collaboration of the University of Maryland, Google, USGS, and NASA, and uses Landsat satellite images to map annual tree cover loss at a 30 × 30 meter resolution. This collection includes only a subset of indicators from the source dataset.

AbstractForests of Australia (2023) is a continental spatial dataset of forest extent, by national forest categories and types, assembled for Australia's State of the Forests Report. It was developed from multiple forest, vegetation and land cover data inputs, including contributions from Australian, state and territory government agencies and external sources.A forest is defined in this dataset as "An area, incorporating all living and non-living components, that is dominated by trees having usually a single stem and a mature or potentially mature stand height exceeding two metres and with existing or potential crown cover of overstorey strata about equal to or greater than 20 per cent. This includes Australia's diverse native forests and plantations, regardless of age. It is also sufficiently broad to encompass areas of trees that are sometimes described as woodlands".The dataset was compiled by the Australian Bureau of Agricultural and Resource Economics and Sciences (ABARES) for the National Forest Inventory (NFI), a collaborative partnership between the Australian and state and territory governments. The role of the NFI is to collate, integrate and communicate information on Australia's forests. State and territory government agencies collect forest data using independent methods and at varying scales or resolutions. The NFI applies a national classification to state and territory data to allow seamless integration of these datasets. Multiple independent sources of external data are used to fill data gaps and improve the quality of the final dataset.The NFI classifies forests into three national forest categories (Native Forest, Commercial plantation, and other forest) and then into various forest types. Commercial plantations presented in this dataset were sourced from the National Plantation Inventory (NPI) spatial dataset (2021), also produced by ABARES.Another dataset produced by ABARES, the Catchment scale land use of Australia CLUM dataset (2020), was used to identify and mask out land uses that are inappropriate to map as forest.The Forests of Australia (2023) dataset is produced to fulfil requirements of Australia's National Forest Policy Statement and the Regional Forests Agreement Act 2002 (Cwth) and is used by the Australian Government for domestic and international reporting.Previous versions of this dataset are available on the Forests Australia website spatial data page and the Australian Government open government data portaldata.gov.au.CurrencyDate modified: 30 November 2023Modification frequency: Every 5 yearsData extentSpatial extentNorth: -8.2°South: -44.4°East: 157.2°West: 109.5°Source informationData, Metadata, Maps and Interactive views are available from ABARES website.Forests of Australia (2023) – Descriptive metadata.The data was obtained from Department of Agriculture, Fisheries and Forestry - Australian Bureau of Agricultural and Resource Economics and Sciences (ABARES). ABARES is providing this data to the public under a Creative Commons Attribution 4.0 license.Lineage statementPresented on this page is a summarised lineage on the development of state and territory datasets for Forests of Australia (2023). The dataset has been produced using the Multiple Lines of Evidence (MLE) method for publication in the Australia’s State of the Forests Report – 2023 update. Detailed lineage information can be found here.Forests of Australia (2023) is a continental spatial dataset of forest extent, by national forest categories and types, assembled for Australia's State of the Forests Report – 2023 update. It was developed from multiple forest, vegetation and land cover data inputs, including contributions from Australian, state and territory government agencies and external sources.For each state or territory, except for the ACT where there was no new data, intersection of the Forests of Australia (2018) dataset with a forest cover dataset supplied by the jurisdiction, and with other available and appropriate independent forest cover datasets, identified:High confidence areas – areas where all the examined datasets agreed with the Forests of Australia (2018) dataset that the areas were forest or non-forest. No further assessment was required for these areas.Moderate confidence areas – areas where the Forests of Australia (2018) dataset agreed with the forest cover dataset supplied by state or territory, and with external or independent datasets, that the areas were forest or non-forest. These areas were identified as potential errors and needed further analysis in order to determine the correct allocation (forest or non-forest). The required analyses and validation were conducted by ABARES, in consultation with relevant state and territory agencies, using various ancillary data including high-resolution imagery such as World Imagery by ESRI, Bing Maps and Google Earth Pro.Low confidence areas – areas where the Forests of Australia (2018) dataset disagreed with the forest cover dataset supplied by state or territory, and with external or independent datasets, that the areas were forest or non-forest. All such areas were identified as potential errors and needed further analysis in order to determine the correct allocation (forest or non-forest). The required analyses and validation were conducted by ABARES, in consultation with relevant state and territory agencies, using various ancillary data including high-resolution imagery such as World Imagery by ESRI, Bing Maps and Google Earth Pro.External or independent datasets used include:H_Woody_Fuzzy_2_Class dataset is based on the NGGI dataset produced by DCCEEW from Landsat data and was developed to support New South Wales Natural Resources Commission’s (NRC) Monitoring, Evaluation and Reporting Program. NRC applied Fuzzy Logic and Probability modelling to the NGGI dataset to derive annual layers distinguishing between forest and non-forest at 25 m raster resolution. Each of five annual layers, 2015 to 2019, was resampled to a 100 m raster by classifying as forest the 100 m pixels that had more than half their area as forest as determined from 25 m pixels. The five annual layers were combined and every pixel in the combination that had been classified as forest in any year during 2015-2019 period was allocated as forest (and the balance non-forest). This approach was taken to prevent areas where the crown cover had reduced temporarily below 20%, through events such as fire, harvesting, drought or disease, from being incorrectly classified as non-forest.State-wide Land and Tree Study (SLATS) dataset is based on data collected by the Landsat satellite. This dataset was available for Queensland only. Foliage Projective Cover (FPC) values of 11 or greater (equivalent to crown cover 20% or greater) were considered as forest candidates in this SLATS dataset. The National Vegetation Information System (NVIS) version 6.0 dataset was used to identify areas in this SLATS dataset that met the height requirements of the forest definition used by the National Forest Inventory.The National Greenhouse Gas Inventory (NGGI) dataset is produced from Landsat satellite Thematic Mapper™, Enhanced Thematic Mapper Plus (ETM+) and Operational Land Image (OLI) images for the Australian Government Department of the Climate Change, Energy, the Environment and Water (DCCEEW), and identifies woody vegetation of height or potential height greater than 2 metres, crown cover greater than 20%, and with a minimum patch size of 0.2 hectares (DISER, 2021a) . The dataset is compiled using time-series data since 1972 and is produced at a 25 m × 25 m resolution. The NGGI dataset used was developed from the five annual layers (2016-2020, inclusive) from the ‘National Forest and sparse woody vegetation data (Version 5.0) spatial dataset produced using the algorithms for land-use change allocation developed for the National Inventory Reports (DISER, 2021b). Each layer of the original 25 m resolution, three-class (forest, sparse woody and non-forest) dataset was resampled to a binary (forest and non-forest) 100 m raster by classifying as forest the 100 m pixels that had more than half their area as forest; the sparse woody and non-forest classes were combined into a non-forest class. The five annual layers were then combined and every pixel in the combination that had been classified as forest in any year during 2016-2020 period was allocated as forest (and the balance non-forest). This approach was taken to prevent areas where the crown cover had reduced temporarily below 20%, through events such as fire, harvesting, drought or disease, from being incorrectly classified as non-forest.All input datasets were converted to 100m rasters (ESRI GRID format), aligning with relevant standard NFI state or territory masks (also known as NFI SNAP grids), in Albers projection. Where the input dataset was in polygon format, the Polygon to Raster tool was used to convert the polygon dataset to raster format, using the Maximum_Combined_Area option.Validation assessment results were incorporated to give improved and high-confidence forest cover datasets for each state or territory.Look-up tables translating the state or territory forest cover data to NFI forest types were used where provided. Where this information was not provided, it was derived by ABARES from translating Levels 5 and 6 of the National Vegetation Information System (NVIS) version 6.0 attribute information to NFI forest types.This dataset has been converted from GeoTIFF to Multidimensional Cloud Raster Format (CRF) to facilitate publishing to the Digital Atlas of Australia (DAA).Date of extraction: February 2024.Data dictionaryAttribute nameDescriptionVALUEIdentifier of every unique combination of the following attributes: STATE, FOR_SOURCE, FOR_CODE, FOR_TYPE, FOR_CAT, HEIGHT and COVER.COUNTNumber of cells that belong to a particular VALUE. For this dataset, in which cell resolution is 100 by 100 metres.

The data in this repository is available under the Open Database License: http://opendatacommons.org/licenses/odbl/1.0/. Any rights in individual contents of the database are licensed under the Database Contents License: http://opendatacommons.org/licenses/dbcl/1.0/This repository includes two datasets. The first is a collection of polygons covering mines globally and the associated forest cover loss from 2000 to 2019. The polygons were derived by merging the "global-scale mining polygons version 2" (Maus et al., 2022) and mining and quarry polygon features extracted from the OpenStreetMap database (OpenStreetMap contributors, 2017). To remove double counting of areas the overlaps between the datasets were resolved by uniting intersecting features into single polygon features, i.e. keeping only the external borders of intersecting features. A random visual check was conducted, and a few small manual editing of polygons was performed where errors were identified.

The resulting dataset is encoded as a Geopackage in the file 'global_mining_polygons.gpkg'. The GeoPackage includes a single layer with 192,584 entries called 'mining_polygons' with the following attributes:

id unique feature identifier

isoa3 ISO 3166-1 alpha-3 country codes

country country names

area area of the polygon in squared kilometres

geom the geometry of the features in geographical coordinates WGS84

The second dataset provides annual time series of global tree cover loss within mines from 2000 to 2019, covering all polygons in the above dataset. The area of tree cover loss for each polygon was calculated from the Global Forest Change database (Hansen et al., 2013). Each polygon also has additional string attributes with biomes derived from Ecoregions 2017 © Resolve (Dinerstein et al., 2017) and the level of protection derived from The World Database on Protected Areas (UNEP-WCMC and IUCN, 2022).

This dataset is encoded in CSV format in the file 'global_mining_forest_loss.csv', which includes 416,412 entries and 53 variables, such that:

id unique feature identifier

id_hcluster unique feature identifier

list_of_commodities a comma-separated list of commodities

isoa3 ISO 3166-1 alpha-3 country codes

country country names

ecoregion ecoregion name

biome biome name

year the year

area_forest_loss_XXX_YYY the area of forest cover loss within a polygon per year in squared kilometres.

The values of tree cover loss are disaggregated per initial percentage of tree cover (XXX) and per protection level (YYY).

XXX can take one of:

000: total tree cover loss independently from the initial tree cover

025: tree cover loss on pixels with initial tree cover between 0 and 25%

050: tree cover loss on pixels with initial tree cover between 25 and 50%

075: tree cover loss on pixels with initial tree cover between 50 and 75%

100: tree cover loss on pixels with initial tree cover between 75 and 100%

YYY can take one of:

la: tree cover loss within strict nature reserve

Ib: tree cover loss within wilderness area

II: tree cover loss within national park

III: tree cover loss within natural monument or feature

IV: tree cover loss within habitat/species management area

V: tree cover loss within protected landscape/seascape

VI: tree cover loss within PA with sustainable use of natural resources

p: tree cover loss within any type of protection, including not applicable, not assigned, or not reported

none: when YYY is omitted, total tree cover loss within the polygon

For details about the protection levels definition see the UNEP-WCMC and IUCN (2022). The id can be used to link polygons to forest loss data.

Crowther_Nature_Files.zip This description pertains to the original download. Details on revised (newer) versions of the datasets are listed below. When more than one version of a file exists in Figshare, the original DOI will take users to the latest version, though each version technically has its own DOI. -- Two global maps (raster files) of tree density. These maps highlight how the number of trees varies across the world. One map was generated using biome-level models of tree density, and applied at the biome scale. The other map was generated using ecoregion-level models of tree density, and applied at the ecoregion scale. For this reason, transitions between biomes or between ecoregions may be unrealistically harsh, but large-scale estimates are robust (see Crowther et al 2015 and Glick et al 2016). At the outset, this study was intended to generate reliable estimates at broad spatial scales, which inherently comes at the cost of fine-scale precision. For this reason, country-scale (or larger) estimates are generally more robust than individual pixel-level estimates. Additionally, due to data limitations, estimates for Mangroves and Tropical coniferous forest (as identified by WWF and TNC) were generated using models constructed from Topical moist broadleaf forest data and Temperate coniferous forest data, respectively. Because we used ecological analogy, the estimates for these two biomes should be considered less reliable than those of other biomes . These two maps initially appeared in Crowther et al (2015), with the biome map being featured more prominently. Explicit publication of the data is associated with Glick et al (2016). As they are produced, updated versions of these datasets, as well as alternative formats, will be made available under Additional Versions (see below).

Methods: We collected over 420,000 ground-sources estimates of tree density from around the world. We then constructed linear regression models using vegetative, climatic, topographic, and anthropogenic variables to produce forest tree density estimates for all locations globally. All modeling was done in R. Mapping was done using R and ArcGIS 10.1.

Viewing Instructions: Load the files into an appropriate geographic information system (GIS). For the original download (ArcGIS geodatabase files), load the files into ArcGIS to view or export the data to other formats. Because these datasets are large and have a unique coordinate system that is not read by many GIS, we suggest loading them into an ArcGIS dataframe whose coordinate system matches that of the data (see File Format). For GeoTiff files (see Additional Versions), load them into any compatible GIS or image management program.

Comments: The original download provides a zipped folder that contains (1) an ArcGIS File Geodatabase (.gdb) containing one raster file for each of the two global models of tree density – one based on biomes and one based on ecoregions; (2) a layer file (.lyr) for each of the global models with the symbology used for each respective model in Crowther et al (2015); and an ArcGIS Map Document (.mxd) that contains the layers and symbology for each map in the paper. The data is delivered in the Goode homolosine interrupted projected coordinate system that was used to compute biome, ecoregion, and global estimates of the number and density of trees presented in Crowther et al (2015). To obtain maps like those presented in the official publication, raster files will need to be reprojected to the Eckert III projected coordinate system. Details on subsequent revisions and alternative file formats are list below under Additional Versions.----------

Additional Versions: Crowther_Nature_Files_Revision_01.zip contains tree density predictions for small islands that are not included in the data available in the original dataset. These predictions were not taken into consideration in production of maps and figures presented in Crowther et al (2015), with the exception of the values presented in Supplemental Table 2. The file structure follows that of the original data and includes both biome- and ecoregion-level models.

Crowther_Nature_Files_Revision_01_WGS84_GeoTiff.zip contains Revision_01 of the biome-level model, but stored in WGS84 and GeoTiff format. This file was produced by reprojecting the original Goode homolosine files to WGS84 using nearest neighbor resampling in ArcMap. All areal computations presented in the manuscript were computed using the Goode homolosine projection. This means that comparable computations made with projected versions of this WGS84 data are likely to differ (substantially at greater latitudes) as a product of the resampling. Included in this .zip file are the primary .tif and its visualization support files.

References:

Crowther, T. W., Glick, H. B., Covey, K. R., Bettigole, C., Maynard, D. S., Thomas, S. M., Smith, J. R., Hintler, G., Duguid, M. C., Amatulli, G., Tuanmu, M. N., Jetz, W., Salas, C., Stam, C., Piotto, D., Tavani, R., Green, S., Bruce, G., Williams, S. J., Wiser, S. K., Huber, M. O., Hengeveld, G. M., Nabuurs, G. J., Tikhonova, E., Borchardt, P., Li, C. F., Powrie, L. W., Fischer, M., Hemp, A., Homeier, J., Cho, P., Vibrans, A. C., Umunay, P. M., Piao, S. L., Rowe, C. W., Ashton, M. S., Crane, P. R., and Bradford, M. A. 2015. Mapping tree density at a global scale. Nature, 525(7568): 201-205. DOI: http://doi.org/10.1038/nature14967Glick, H. B., Bettigole, C. B., Maynard, D. S., Covey, K. R., Smith, J. R., and Crowther, T. W. 2016. Spatially explicit models of global tree density. Scientific Data, 3(160069), doi:10.1038/sdata.2016.69.

Global Forest Watch (GFW) is an online platform that provides data and tools for monitoring forests, allowing for near real-time information about where and how forests are changing around the world.

The World Resources Institute and Google DeepMind created a global map of the dominant drivers of tree cover loss from 2001 to 2022 at 1 km spatial resolution. We used a deep learning model to classify seven driver classes: permanent agriculture, hard commodities, shifting cultivation, forest management, wildfires, settlements & infrastructure, and other natural disturbances. As part of the study, we collected a set of training samples through interpretation of very high resolution satellite imagery and developed a single world-wide customized residual convolutional neural network model (ResNet) using satellite data (Landsat and Sentinel-2) and ancillary biophysical and population data. In addition, we collected a stratified random sample of validation plots through interpretation of very high resolution satellite imagery to estimate the accuracy of the final classification map.

In this repository, we make the following available:

• The training and validation data, available in two separate files. Both datasets were collected by a team of image interpreters and assessed for quality by two additional interpreters. Note that while for the validation data the quality of the primary driver was rigorously assessed, the secondary driver wasn’t subject to the same level of quality control.

• The global drivers of forest loss raster (drivers_forest_loss_1km.tif), including the discrete classification and probabilities for each class.

Creator & Contact

Created by World Resources Institute and Google DeepMind

Contact: Michelle Sims: Michelle.Sims@wri.org

Definitions

A driver is defined as the direct cause of tree cover loss, and can include both temporary disturbances (natural or anthropogenic) or permanent loss of tree cover due to a change to a non-forest land use (e.g., deforestation). The dominant driver is defined as the direct driver that caused the majority of tree cover loss within each 1 km cell over the time period.

Classes are defined as follows:

This occurrence dataset provides primary data on repeated tree measurement of two inventories on the permanent sampling plot (8.8 ha) established in the old-growth polydominant broadleaved forest stand in the “Kaluzhskie Zaseki” State Nature Reserve (center of the European part of Russian Federation). The time span between the inventories was 30 years, and a total of more than 11 000 stems were included in the study (11 tree species and 3 genera). During the measurements, the tree species (for some trees only genus was determined), stem diameter at breast height of 1.3 m (DBH), and life status were recorded for every individual stem, and some additional attributes were determined for some trees. Field data were digitized and compiled into the PostgreSQL database. Deep data cleaning and validation (with documentation of changes) has been performed before data standardization according to the Darwin Core standard.

Представлены первичные данные двух перечетов деревьев, выполненных на постоянной пробной площади (8.8 га), заложенной в старовозрастном полидоминантном широколиственном лесу в заповеднике “Калужские засеки”. Перечеты выполнены с разницей в 30 лет, всего исследовано более 11 000 учетных единиц (деревья 11-ти видов и 3-х родов). Для каждой учетной единицы определяли вид, диаметр на высоте 1.3 м и статус, для части деревьев также измеряли дополнительные характеристики. Все полевые данные были оцифрованы и организованы в базу данных в среде PostgreSQL. Перед стандартизацией данных в соответствии с Darwin Core выполнена их тщательная проверка, все внесенные изменения документированы.

1. Tropical forests are subject to diverse deforestation pressures while their conservation is essential to achieve global climate goals. Predicting the location of deforestation is challenging due to the complexity of the natural and human systems involved but accurate and timely forecasts could enable effective planning and on-the-ground enforcement practices to curb deforestation rates. New computer vision technologies based on deep learning can be applied to the increasing volume of Earth observation data to generate novel insights and make predictions with unprecedented accuracy.
2. Here, we demonstrate the ability of deep convolutional neural networks (CNNs) to learn spatiotemporal patterns of deforestation from a limited set of freely available global data layers, including multispectral satellite imagery, the Hansen maps of annual forest change (2001-2020) and the ALOS PALSAR digital surface model, to forecast deforestation (2021). We designed four model architectures, based on 2D CNNs, 3D CNNs, and Convolutional Long Short-Term Memory (ConvLSTM) Recurrent Neural Networks (RNNs), to produce spatial maps that indicate the risk to each forested pixel (~30 m) in the landscape of becoming deforested within the next year. They were trained and tested on data from two ~80,000 km2 tropical forest regions in the Southern Peruvian Amazon.
3. The networks could predict the location of future forest loss to a high degree of accuracy (F1 = 0.58-0.71). Our best performing model (3D CNN) had the highest pixel-wise accuracy (F1 = 0.71) when validated on 2020 forest loss (2014-2019 training). Visual interpretation of the mapped forecasts indicated that the network could automatically discern the drivers of forest loss from the input data. For example, pixels around new access routes (e.g. roads) were assigned high risk whereas this was not the case for recent, concentrated natural loss events (e.g. remote landslides).
4. CNNs can harness limited time-series data to predict near-future deforestation patterns, an important step in harnessing the growing volume of satellite remote sensing data to curb global deforestation. The modelling framework can be readily applied to any tropical forest location and used by governments and conservation organisations to prevent deforestation and plan protected areas. Methods Input raster data all freely available online Original input raster data from:

Global Forest Change - https://glad.earthengine.app/view/global-forest-change ALOS JAXA - https://www.eorc.jaxa.jp/ALOS/en/dataset/aw3d30/aw3d30_e.htm

Processed with code at https://github.com/PatBall1/DeepForestcast Dataset includes:

Input shapefiles for each study site. Input geotiff files (.tif) for each study site. Input PyTorch tensors (.pt) for each study site. Model weights (.pt) for trained networks (for testing and forecasting). Output deforestation forecasts for each study site as geotiffs (.tif).

OverviewThe tropical tree cover data maps tree extent at the ten-meter scale and tree cover at the half hectare scale to enable accurate monitoring of trees in urban areas, agricultural lands, and in open canopy and dry forest ecosystems. The data extends over 4.3 billion hectares of the global tropics. The data is derived from multi-temporal convolutional neural network models applied to Sentinel optical and radar imagery. The 10-meter dataset is a binary tree extent layer that is similar to a land cover map, while the tree cover data represents fractional cover at a half-hectare scale. More details on the methodology and analyses can be found on the GitHub page.Resolution: 0.5 haGeographic Coverage: 4.3 billion hectares of the tropics (-23.44 to 23.44 latitude)Frequency of Updates: Annual change detection maps starting in 2017 are planned for 2024 releaseDate of Content: 2020CautionsThis dataset uses a different definition of a tree and a different definition of tree cover than does Hansen et al. (2013). This dataset defines a tree according to both the height and crown diameter. Woody vegetation higher than 5 meters regardless of crown diameter, or between 3 and 5 meters with a minimum crown diameter of 5 meters is considered a tree. This definition is different from Hansen et al. (2013) which defines a tree as any vegetation at least 5 meters in height. The tropical tree cover dataset does not disambiguate plantation trees from non-plantation trees.Analyses or statistics derived for shapefiles smaller than 0.5ha may not be accurate

Important Note: This item is in mature support as of May 2025 and will retire in December 2025.This map features forests for the world, which are provided here as an excerpted subset of the World Land Cover 30m BaseVue 2013 layer. Separating forests is useful as a cartographic layer on environmentally oriented maps and analytically as a basis for ecosystem and habitat definition.This web map uses the vector Terrain with Labels as its basemap.Dataset SummaryBaseVue 2013 is a commercial global, land use / land cover (LULC) product developed by MDA. BaseVue covers the Earth’s entire land area, excluding Antarctica. BaseVue is independently derived from roughly 9,200 Landsat 8 images and is the highest spatial resolution (30m), most current LULC product available. The capture dates for the Landsat 8 imagery range from April 11, 2013 to June 29, 2014. The following 4 classes of forest are featured in this layer: Deciduous Forest: Trees > 3 meters in height, canopy closure >35% (<25% inter-mixture with evergreen species) that seasonally lose their leaves, except Larch.Evergreen Forest: Trees >3 meters in height, canopy closure >35% (<25% inter-mixture with deciduous species), of species that do not lose leaves. (will include coniferous Larch regardless of deciduous nature).Woody Wetlands: Areas where forest or shrubland vegetation accounts for greater than 20% of vegetative cover and the soil or substrate periodically is saturated with, or covered by water. Only used within the continental U.S.Mixed Forest: Areas dominated by trees generally greater than 5 meters tall, and greater than 20% of total vegetation cover. Neither deciduous nor evergreen species are greater than 75% of total tree cover. Only used within the continental U.S.What can you do with this layer?This layer has query, identify, and export image services available. The layer is restricted to an 16,000 x 16,000 pixel limit, which represents an area of nearly 300 miles on a side. This layer is part of a larger collection of landscape layers that you can use to perform a wide variety of mapping and analysis tasks.Important Note: This layer is available for users with an ArcGIS Organizational subscription. To access this layer, you'll need to sign in with an account that is a member of an organizational subscription. If you don't have an organizational subscription, you can create a new account and then sign up for a 30-day trial of ArcGIS Online. For more information, see the Landscape Layers group on ArcGIS Online.

The World Database on Protected Areas (WDPA) is the most comprehensive global database of marine and terrestrial protected areas, updated on a monthly basis, and is one of the key global biodiversity data sets being widely used by scientists, businesses, governments, International secretariats and others to inform planning, policy decisions and management. The WDPA is a joint project between UN Environment and the International Union for Conservation of Nature (IUCN). The compilation and management of the WDPA is carried out by UN Environment World Conservation Monitoring Centre (UNEP-WCMC), in collaboration with governments, non-governmental organisations, academia and industry. There are monthly updates of the data which are made available online through the Protected Planet website where the data is both viewable and downloadable. Data and information on the world's protected areas compiled in the WDPA are used for reporting to the Convention on Biological Diversity on progress towards reaching the Aichi Biodiversity Targets (particularly Target 11), to the UN to track progress towards the 2030 Sustainable Development Goals, to some of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) core indicators, and other international assessments and reports including the Global Biodiversity Outlook, as well as for the publication of the United Nations List of Protected Areas. Every two years, UNEP-WCMC releases the Protected Planet Report on the status of the world's protected areas and recommendations on how to meet international goals and targets. Many platforms are incorporating the WDPA to provide integrated information to diverse users, including businesses and governments, in a range of sectors including mining, oil and gas, and finance. For example, the WDPA is included in the Integrated Biodiversity Assessment Tool, an innovative decision support tool that gives users easy access to up-to-date information that allows them to identify biodiversity risks and opportunities within a project boundary. The reach of the WDPA is further enhanced in services developed by other parties, such as the Global Forest Watch and the Digital Observatory for Protected Areas, which provide decision makers with access to monitoring and alert systems that allow whole landscapes to be managed better. Together, these applications of the WDPA demonstrate the growing value and significance of the Protected Planet initiative.

Private Forest Wind Damage Assessment Spatial Database - May 2025. Published by Department of Agriculture, Food and the Marine. Available under the license Licence Not Specified (notspecified).Following Storm Darragh and Storm Éowyn during the winter of 2024/2025, and noting that many forests have been windblown around the country, Minister for Agriculture, Food and the Marine, Martin Heydon, and Minister of State for Forestry, Horticulture and Farm Safety, Michael Healy-Rae, invited key stakeholders to join department officials on a taskforce to ensure that storm-damaged forests were managed safely and appropriately. A Forestry Windblow Taskforce was set up to quantify forest damage and to identify approaches to facilitate the mobilisation of wind damaged timber.

Part of the Taskforce’s work involved initiating a detailed mapping assessment using high-resolution satellite imagery to provide information at a local or forest stand level scale. The detailed assessment of windblow damage was undertaken using high resolution satellite imagery from SkySat, and supplemented with pre and post storm Sentinel-2 and PlanetScope satellite data. In addition, drone imagery was also acquired for a number of specific locations.

The mapping exercise relied on the tasking of SkySat imagery during cloud-free weather conditions to acquire the necessary imagery data. The mapping was conducted largely between early February and the beginning of April 2025. The mapping effort focused on a target area of interest where the damage was deemed most likely to have occurred. These target areas were forests stands that were predominantly coniferous species and at least 15 years of age. These age and species criteria were used to filter both Coillte’s sub-compartment database and DAFM’s private forest dataset to confine the wind damage mapping exercise to the most relevant forests.

The windblow mapping exercise utilised a range of available EO datasets of varying spatial and temporal resolution which included: SkySat: 75% (c. 0.50 m resolution), Sentinel-2/PlanetScope: 20% (c. 10 m resolution/c. 3 m resolution), and drone imagery: 5% ( c. 0.2 m resolution).

The national estimate of private wind damage area (11,414 ha) as included in the private forest wind damage spatial database is within approximately +/- 500 hectares of the actual windblown private forest area. This uncertainty is due in part to the fact that for some parts of the country, SkySat satellite imagery has not yet been acquired. It is expected that there will be an ongoing refinement of the private forest windblow area estimate when new SkySat or other Earth Observation data becomes available over the coming months.

As part of this mapping exercise, “older” windblown areas, i.e. windblown forest areas that are more than 4 years old, were also identified and mapped. It is estimated these damage forest areas represent between 750 and 1,000 hectares of the total national area estimate of private wind damaged forest.

The area of wind damage in broadleaf stands may be greater than identified in the private forest wind damage database given the focus in the mapping exercise on coniferous species that were at least 15 years of age. This is also due in part to the fact that the identification of windblow in broadleaf stands is more challenging, particularly if the damage impacts individual trees.

The output from the mapping assessment is an ESRI Shapefile polygon database of wind damaged, privately owned forest areas greater than or equal to 0.1 hectares. The Shapefile is provided in the Irish Transverse Mercator geographic coordinate system. The main attribute included the spatial database table is area in hectares for each wind damaged forest area delineated.

These data are provisional in that they are a record of DAFM data holding in relation to private wind damaged forest at this time (May 2025). They are not published as legal definitions of the current actuality with regard to their geographic extent. They may contain errors and omissions and it should also be noted that the data cannot be taken as being absolutely current. Therefore they should be treated as indicative of the actual geographic situation. The Department of Agriculture, Food and the Marine will accept no liability for any loss or damage suffered by those using this data for any purpose whatsoever....

This data set, a collaboration between the GLAD (Global Land Analysis & Discovery) lab at the University of Maryland, Google, USGS, and NASA, measures areas of tree cover loss across all global land (except Antarctica and other Arctic islands) at approximately 30 × 30 meter resolution. The data were generated using multispectral satellite imagery from the Landsat 5 thematic mapper (TM), the Landsat 7 thematic mapper plus (ETM+), and the Landsat 8 Operational Land Imager (OLI) sensors. Over 1 million satellite images were processed and analyzed, including over 600,000 Landsat 7 images for the 2000-2012 interval, and more than 400,000 Landsat 5, 7, and 8 images for updates for the 2011-2022 interval. The clear land surface observations in the satellite images were assembled and a supervised learning algorithm was applied to identify per pixel tree cover loss.In this data set, “tree cover” is defined as all vegetation greater than 5 meters in height, and may take the form of natural forests or plantations across a range of canopy densities. Tree cover loss is defined as “stand replacement disturbance,” or the complete removal of tree cover canopy at the Landsat pixel scale. Tree cover loss may be the result of human activities, including forestry practices such as timber harvesting or deforestation (the conversion of natural forest to other land uses), as well as natural causes such as disease or storm damage. Fire is another widespread cause of tree cover loss, and can be either natural or human-induced.This data set has been updated five times since its creation, and now includes loss up to 2022 (Version 1.10). The analysis method has been modified in numerous ways, including new data for the target year, re-processed data for previous years (2011 and 2012 for the Version 1.1 update, 2012 and 2013 for the Version 1.2 update, and 2014 for the Version 1.3 update), and improved modelling and calibration. These modifications improve change detection for 2011-2022, including better detection of boreal loss due to fire, smallholder rotation agriculture in tropical forests, selective losing, and short cycle plantations. Eventually, a future “Version 2.0” will include reprocessing for 2000-2010 data, but in the meantime integrated use of the original data and Version 1.7 should be performed with caution. Read more about the Version 1.7 update here.When zoomed out (< zoom level 13), pixels of loss are shaded according to the density of loss at the 30 x 30 meter scale. Pixels with darker shading represent areas with a higher concentration of tree cover loss, whereas pixels with lighter shading indicate a lower concentration of tree cover loss. There is no variation in pixel shading when the data is at full resolution (≥ zoom level 13).The tree cover canopy density of the displayed data varies according to the selection - use the legend on the map to change the minimum tree cover canopy density threshold.

To improve scientific understanding of the extent and distribution of mangrove forests of the world the status and distribution of global mangroves were mapped using recently available Global Land Survey (GLS) data and the Landsat archive.The project interpreted approximately 1000 Landsat scenes using hybrid supervised and unsupervised digital image classification techniques. Results were validated using existing GIS data and the published literature to map ‘true mangroves’.The total area of mangroves in the year 2000 was 137,760 km2 in 118 countries and territories in the tropical and subtropical regions of the world. Approximately 75% of world's mangroves are found in just 15 countries, and only 6.9% are protected under the existing protected areas network (IUCN I-IV). Our study confirms earlier findings that the biogeographic distribution of mangroves is generally confined to the tropical and subtropical regions and the largest percentage of mangroves is found between 5° N and 5° S latitude.The remaining area of mangrove forest in the world is less than previously thought; the estimate provided in this study is 12.3% smaller than the most recent estimate by the Food and Agriculture Organization (FAO) of the United Nations. This data set presents the most comprehensive, globally consistent and highest resolution (30 m) global mangrove database ever created

This global gridded dataset depicts vegetation biomass carbon stocks at the native processing resolution of 0.0089 decimal degrees (~1km by ~1km). We used the mean aggregate ArcInfo command to resample this dataset to 5 and 10 minute spatial resolution (please note that the 10 minute data includes the extent of Antarctica, while the others do not). The 1km data is expressed in 0.01 tons of biomass carbon per hectare, while the 5 and 10 minute data are expressed in tons of biomass carbon per hectare; soil carbon stocks are not included. Each map is geo-referenced to the WGS1984 coordinate system, and in geographic projection.

The vegetation biomass carbon database was created in two main steps: 1) estimate
carbon stocks, and 2) map values using a range of spatiallyexplicit climate and vegetation datasets. Creators followed the IPCC GPG Tier1 method for estimating vegetation carbon stocks using the globally consistent default values provided for aboveground biomass (IPCC 2006). They added belowground biomass (root) carbon stocks using the IPCC root to shoot ratios for each vegetation type, and then converted total living vegetation biomass to carbon stocks using the carbon fraction for each vegetation type (varies between forests, shrublands and grasslands). All estimates and conversions were specific to each continent, ecoregion and vegetation type (stratified by age of forest). Thus, we compiled a total of 124 carbon zones
or regions with unique carbon stock values based on the IPCC Tier1 methods. Please refer to Tables 1a-i (http://cdiac.ornl.gov/epubs/ndp/global_carbon/carbon_documentation.html)
to review the details associated with each of these carbon zones. A small number
of carbon zones were not included in the original IPCC default data but were in the land
cover map such as mixed and burnt forest and natural vegetation/cropland mosaic categories.

The continental regions, ecofloristic zones, and frontier forest shapefiles were combined to
determine the spatial distribution of global carbon_zones. These data were then gridded and
combined with the GLC2000 data. An ESRI ArcInfo script was used to apply the associated
carbon values to each pixel within a carbon zone. Specifically, we clipped out the carbon
zone boundaries from the GLC2000 gridded land cover data and then used a series of carbon
remap tables, created from the values listed in tables 1a1i, to assign carbon values to the gridded data. These clipped GLC2000 carbon zone grids were then
merged back together to form a single contiguous global dataset at 1 kilometer by 1 kilometer
resolution.

This spatial database is likely the best available, globally consistent
map depicting vegetation carbon stocks, circa 2000, and follows the widely accepted IPCC methods for estimating carbon stocks at the national level. However, the methods employed here are not directly linked to groundbased measures of carbon stocks and have not been validated with field data. We essentially applied a sophisticated paint-by-numbers
approach, which consequently masks variations within classes and may lead to unnatural, abrupt gradients between vegetation classes as defined by the GLC 2000 and FAO ecoregions (Gibbs et al. 2007).