This dataset represents a high resolution urban land cover classification map across the southern California Air Basin (SoCAB) with a spatial resolution of 60 cm in urban regions and 10 m in non-urban regions. This map was developed to support NASA JPL-based urban biospheric CO2 modeling in Los Angeles, CA. Land cover classification was derived from a novel fusion of Sentinel-2 (10-60 m x 10-60 m) and 2016 NAIP (60 cm x 60 cm) imagery and provides identification of impervious surface, non-photosynthetic vegetation, shrub, tree, grass, pools and lakes.

Land Cover Classes in .tif file: 0: Impervious surface 1: Tree (mixed evergreen/deciduous) 2: Grass (assumed irrigated) 3: Shrub 4: Non-photosynthetic vegetation 5: Water (masked using MNDWI/NDWI)

Google Earth Engine interactive app displaying this map: https://wcoleman.users.earthengine.app/view/socab-irrigated-classification

A portion of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. Support from the Earth Science Division OCO-2 program is acknowledged. Copyright 2020. All rights reserved.

Land cover describes the surface of the earth. Land cover maps are useful in urban planning, resource management, change detection, agriculture, and a variety of other applications in which information related to earth surface is required. Land cover classification is a complex exercise and is hard to capture using traditional means. Deep learning models are highly capable of learning these complex semantics and can produce superior results.Using the modelFollow the guide to use the model. Before using this model, ensure that the supported deep learning libraries are installed. For more details, check Deep Learning Libraries Installer for ArcGIS.Fine-tuning the modelThis model can be fine-tuned using the Train Deep Learning Model tool. Follow the guide to fine-tune this model.Input8-bit, 3-band high-resolution (80 - 100 cm) imagery.OutputClassified raster with the same classes as in the Chesapeake Bay Landcover dataset (2013/2014). By default, the output raster contains 9 classes. A simpler classification with 7 classes can be performed by setting the the 'detailed_classes' model argument to false.Note: The output classified raster will not contain 'Aberdeen Proving Ground' class. Find class descriptions here.Applicable geographiesThis model is applicable in the United States and is expected to produce best results in the Chesapeake Bay Region.Model architectureThis model uses the UNet model architecture implemented in ArcGIS API for Python.Accuracy metricsThis model has an overall accuracy of 86.5% for classification into 9 land cover classes and 87.86% for 7 classes. The table below summarizes the precision, recall and F1-score of the model on the validation dataset, for classification into 9 land cover classes:ClassPrecisionRecallF1 ScoreWater0.936140.930460.93329Wetlands0.816590.759050.78677Tree Canopy0.904770.931430.91791Shrubland0.516250.186430.27394Low Vegetation0.859770.866760.86325Barren0.671650.509220.57927Structures0.80510.848870.82641Impervious Surfaces0.735320.685560.70957Impervious Roads0.762810.812380.78682The table below summarizes the precision, recall and F1-score of the model on the validation dataset, for classification into 7 land cover classes: ClassPrecisionRecallF1 ScoreWater0.950.940.95Tree Canopy and Shrubs0.910.920.92Low Vegetation0.850.850.85Barren0.470.480.47Developed, Medium Intensity0.790.690.74Impervious Surfaces0.840.840.84Impervious Roads0.820.830.82Training dataThis model has been trained on the Chesapeake Bay high-resolution 7 class 2013/2014 NAIP Landcover dataset (produced by Chesapeake Conservancy with their partners University of Vermont Spatial Analysis Lab (UVM SAL), and Worldview Solutions, Inc. (WSI)) and other high resolution imagery. Find more information about the dataset here.Sample resultsHere are a few results from the model.

This dataset provides high-resolution, species-specific land cover maps for the Hawaiian island of Lāna'i based on 2020 WorldView-2 satellite imagery. Machine learning models were trained on extensive ground control polygons and points. The land cover maps capture the distribution and diversity of vegetation with high accuracy to support conservation planning and monitoring. This data release consists of two child items, one containing the field and expert collected ground control data used to train our models, and another consisting of resulting land cover maps for the island of Lāna‘i. The research effort that generated these input data, and products are carefully described in the associated manuscript Berio Fortini et al. 2024. Full citation is listed in the larger work section of this XML file. Inputs: Ground control polygons used for model training and evaluation Ground control points used for independent pixel-level model validation Outputs: Raster 1. Species-specific land cover map for the island of Lāna‘i, based on expert-adjusted class posterior probabilities. Raster 2. Community-specific land cover map for the island of Lāna‘i, based on land cover classification including expert-adjusted class posterior probabilities. Raster 3. Mixed hierarchical land cover map for the island of Lāna‘i, based on land cover classification including expert-adjusted class posterior probabilities. Raster 4 (stack) Individual cover class membership probability maps.

The following dataset provides a high-resolution (1 metre) land cover classification for a portion of the Trail Valley Creek (TVC) Research Watershed, NWT. The areal coverage of this product covers the Siksik Creek, Big Bear Lake, Little Bear Lake, Inuvik-Tuktoyaktuk Highway (ITH) bridge, and TVC valley study areas including coverage of the TVC Main Meteorological (TMM) station. Collectively, throughout this document we refer to this region as the “Greater Siksik Area”. In total, the classification product covers an aerial footprint of 5.4 km2. The classification was created using high-resolution RGB imagery collected using a fixed-wing Unmanned Aerial System (UAS) in the Fall of 2019 at the peak of plant senescence where autumn leaf colour was at its maximum and before leaves had fallen.

We used Digital Soil Mapping (DSM) technologies combined with the real-time collations of soil attribute data from TERN's recently developed Soil Data Federation System, to produce a map of Australian Soil Classification Soil Order classes with quantified estimates of mapping reliability at a 90m resolution.

The map gives an estimate of the spatial distribution of soil types across Australia.

Soil classes are based on The Australian Soil Classification - Second Edition by the National Committee on Soil and Terrain, R Isbell - https://ebooks.publish.csiro.au/content/australian-soil-classification-9781486304646

Detailed information about the Soil and Landscape Grid of Australia can be found at - https://esoil.io/TERNLandscapes/Public/Pages/SLGA/index.html

Period (temporal coverage; approximately): 1950-2021; Spatial resolution: 3 arc seconds (approx 90m); Number of pixels with coverage per layer: 2007M (49200 * 40800); Data license : Creative Commons Attribution 4.0 (CC BY); Format: Cloud Optimised GeoTIFF; Lineage: The map was produced as per methods described at - https://aussoilsdsm.esoil.io/slga-version-2-products/australian-soil-classification-map

Soil classification data was extracted from the SoilDataFederator (SDF) - https://esoil.io/TERNLandscapes/Public/Pages/SoilDataFederator/SoilDataFederator.html)

A total of 195,383 observations with either an Australian Soil Classification (ASC) or a Principal Profile Form (PPF) classification or a Great Soil Group (GSG) classification were extracted (Figure 1). Of these observations 130,570 of them had an ASC directly assigned by a pedologist. The remaining 64,813 observations either had a PPF or an ASC assigned to them by pedologists. The PPF and GSG classification where then transformed to an ASC using these remap tables.

The 90m raster covariate data was obtained from TERNs publicly available raster covariate stack - https://esoil.io/TERNLandscapes/Public/Products/TERN/Covariates/Mosaics .A parsimonious set of these covariates was used in the modelling.

We used the R "Ranger" Random Forest package to implement a machine learning model as per standard Digital Soil Mapping (DSM) methodologies.

The observed geographic locations in the ASC data set were used to extract cell values from the raster covariate stack using the R "raster" package. This data set was then divided into a 90/10% split of training and external validation sets. The training data was then bootstrapped sampled 50 times to create 50 bootstrap training sets. These training sets were then used to generate 50 Random Forest model realisations.

Using the CSIRO Pearcey High Performance Compute (HPC) cluster the Random Forest models were evaluated against the input covariate raster data stack. This was done for each 90m raster cell across the nation for each of the 50 bootstrapped model realisations. The modal ASC value across the 50 realisations for each cell was determined and assigned as the most probable soil type for that cell in the output raster. The ratio of the second most probable soil to the most probable soil was also calculated to generate a model confusion index, an estimate of the structural uncertainty in the Random Forest model.

The Australian Soil Resource Information System (ASRIS) contains a product that is a compilation of all existing polygon mapping conducted by state and federal soil survey agencies across all of Australia. This product is made up of a diverse range of field mapping products at a range of mapping scales. From this product we extracted all polygons that were mapped at a scale of 1:100,000 or finer, as defined in the Guidelines For Surveying Soil And Land Resources (Blue Book). Polygons mapped at this scale are high quality spatial estimates of the distribution of soil attributes. We then rasterised these polygon ASC values and merged these values into our final estimates of ASC, i.e., where an ASRIS 100,000 scale polygon exists it will replace the modelled ASC value.

All processing for the generation of these products was undertaken using the R programming language. R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.

Code - https://github.com/AusSoilsDSM/SLGA Observation data - https://esoil.io/TERNLandscapes/Public/Pages/SoilDataFederator/SoilDataFederator.html Covariate rasters - https://esoil.io/TERNLandscapes/Public/Pages/SLGA/GetData-COGSDataStore.html

Soil map units are the basic geographic unit of the Soil Survey Geographic Database (SSURGO). The SSURGO dataset is a compilation of soils information collected over the last century by the Natural Resources Conservation Service (NRCS). Map units delineate the extent of different soils. Data for each map unit contains descriptions of the soil’s components, productivity, unique properties, and suitability interpretations.Each soil type has a unique combination of physical, chemical, nutrient and moisture properties. Soil type has ramifications for engineering and construction activities, natural hazards such as landslides, agricultural productivity, the distribution of native plant and animal life and hydrologic and other physical processes. Soil types in the context of climate and terrain can be used as a general indicator of engineering constraints, agriculture suitability, biological productivity and the natural distribution of plants and animals.Dataset SummaryPhenomenon Mapped: Soils of the United States and associated territoriesCoordinate System: Web Mercator Auxiliary SphereExtent: The 50 United States, Puerto Rico, Guam, US Virgin Islands, Marshall Islands, Northern Marianas Islands, Palau, Federated States of Micronesia, and American SamoaVisible Scale: 1:144,000 to 1:1,000Source: USDA Natural Resources Conservation ServicePublication Date: November 2023Data from the gSSURGO database was used to create this layer.AttributesKey fields from nine commonly used SSURGO tables were compiled to create the 173 attribute fields in this layer. Some fields were joined directly to the SSURGO Map Unit polygon feature class while others required summarization and other processing to create a 1:1 relationship between the attributes and polygons prior to joining the tables. Attributes of this layer are listed below in their order of occurrence in the attribute table and are organized by the SSURGO table they originated from and the processing methods used on them.Map Unit Polygon Feature Class Attribute TableThe fields in this table are from the attribute table of the Map Unit polygon feature class which provides the geographic extent of the map units.Area SymbolSpatial VersionMap Unit SymbolMap Unit TableThe fields in this table have a 1:1 relationship with the map unit polygons and were joined to the table using the Map Unit Key field.Map Unit NameMap Unit KindFarmland ClassInterpretive FocusIntensity of MappingIowa Corn Suitability RatingLegend TableThis table has 1:1 relationship with the Map Unit table and was joined using the Legend Key field.Project ScaleSurvey Area Catalog TableThe fields in this table have a 1:1 relationship with the polygons and were joined to the Map Unit table using the Survey Area Catalog Key and Legend Key fields.Survey Area VersionTabular VersionMap Unit Aggregated Attribute TableThe fields in this table have a 1:1 relationship with the map unit polygons and were joined to the Map Unit attribute table using the Map Unit Key field.Slope Gradient - Dominant ComponentSlope Gradient - Weighted AverageBedrock Depth - MinimumWater Table Depth - Annual MinimumWater Table Depth - April to June MinimumFlooding Frequency - Dominant ConditionFlooding Frequency - MaximumPonding Frequency - PresenceAvailable Water Storage 0-25 cm - Weighted AverageAvailable Water Storage 0-50 cm - Weighted AverageAvailable Water Storage 0-100 cm - Weighted AverageAvailable Water Storage 0-150 cm - Weighted AverageDrainage Class - Dominant ConditionDrainage Class - WettestHydrologic Group - Dominant ConditionIrrigated Capability Class - Dominant ConditionIrrigated Capability Class - Proportion of Mapunit with Dominant ConditionNon-Irrigated Capability Class - Dominant ConditionNon-Irrigated Capability Class - Proportion of Mapunit with Dominant ConditionRating for Buildings without Basements - Dominant ConditionRating for Buildings with Basements - Dominant ConditionRating for Buildings with Basements - Least LimitingRating for Buildings with Basements - Most LimitingRating for Septic Tank Absorption Fields - Dominant ConditionRating for Septic Tank Absorption Fields - Least LimitingRating for Septic Tank Absorption Fields - Most LimitingRating for Sewage Lagoons - Dominant ConditionRating for Sewage Lagoons - Dominant ComponentRating for Roads and Streets - Dominant ConditionRating for Sand Source - Dominant ConditionRating for Sand Source - Most ProbableRating for Paths and Trails - Dominant ConditionRating for Paths and Trails - Weighted AverageErosion Hazard of Forest Roads and Trails - Dominant ComponentHydric Classification - PresenceRating for Manure and Food Processing Waste - Weighted AverageComponent Table – Dominant ComponentMap units have one or more components. To create a 1:1 join component data must be summarized by map unit. For these fields a custom script was used to select the component with the highest value for the Component Percentage Representative Value field (comppct_r). Ties were broken with the Slope Representative Value field (slope_r). Components with lower average slope were selected as dominant. If both soil order and slope were tied, the first value in the table was selected.Component Percentage - Low ValueComponent Percentage - Representative ValueComponent Percentage - High ValueComponent NameComponent KindOther Criteria Used to Identify ComponentsCriteria Used to Identify Components at the Local LevelRunoff ClassSoil loss tolerance factorWind Erodibility IndexWind Erodibility GroupErosion ClassEarth Cover 1Earth Cover 2Hydric ConditionHydric RatingAspect Range - Counter Clockwise LimitAspect - Representative ValueAspect Range - Clockwise LimitGeomorphic DescriptionNon-Irrigated Capability SubclassNon-Irrigated Unit Capability ClassIrrigated Capability SubclassIrrigated Unit Capability ClassConservation Tree Shrub GroupGrain Wildlife HabitatGrass Wildlife HabitatHerbaceous Wildlife HabitatShrub Wildlife HabitatConifer Wildlife HabitatHardwood Wildlife HabitatWetland Wildlife HabitatShallow Water Wildlife HabitatRangeland Wildlife HabitatOpenland Wildlife HabitatWoodland Wildlife HabitatWetland Wildlife HabitatSoil Slip PotentialSusceptibility to Frost HeavingConcrete CorrosionSteel CorrosionTaxonomic ClassTaxonomic OrderTaxonomic SuborderGreat GroupSubgroupParticle SizeParticle Size ModCation Exchange Activity ClassCarbonate ReactionTemperature ClassMoist SubclassSoil Temperature RegimeEdition of Keys to Soil Taxonomy Used to Classify SoilCalifornia Storie IndexComponent KeyComponent Table – Weighted AverageMap units may have one or more soil components. To create a 1:1 join, data from the Component table must be summarized by map unit. For these fields a custom script was used to calculate an average value for each map unit weighted by the Component Percentage Representative Value field (comppct_r).Slope Gradient - Low ValueSlope Gradient - Representative ValueSlope Gradient - High ValueSlope Length USLE - Low ValueSlope Length USLE - Representative ValueSlope Length USLE - High ValueElevation - Low ValueElevation - Representative ValueElevation - High ValueAlbedo - Low ValueAlbedo - Representative ValueAlbedo - High ValueMean Annual Air Temperature - Low ValueMean Annual Air Temperature - Representative ValueMean Annual Air Temperature - High ValueMean Annual Precipitation - Low ValueMean Annual Precipitation - Representative ValueMean Annual Precipitation - High ValueRelative Effective Annual Precipitation - Low ValueRelative Effective Annual Precipitation - Representative ValueRelative Effective Annual Precipitation - High ValueDays between Last and First Frost - Low ValueDays between Last and First Frost - Representative ValueDays between Last and First Frost - High ValueRange Forage Annual Potential Production - Low ValueRange Forage Annual Potential Production - Representative ValueRange Forage Annual Potential Production - High ValueInitial Subsidence - Low ValueInitial Subsidence - Representative ValueInitial Subsidence - High ValueTotal Subsidence - Low ValueTotal Subsidence - Representative ValueTotal Subsidence - High ValueCrop Productivity IndexEsri SymbologyThis field was created to provide symbology based on the Taxonomic Order field (taxorder). Because some mapunits have a null value for soil order, a custom script was used to populate this field using the Component Name (compname) and Mapunit Name (muname) fields. This field was created using the dominant soil order of each mapunit.Esri SymbologyHorizon TableEach map unit polygon has one or more components and each component has one or more layers known as horizons. To incorporate this field from the Horizon table into the attributes for this layer, a custom script was used to first calculate the mean value weighted by thickness of the horizon for each component and then a mean value of components weighted by the Component Percentage Representative Value field for each map unit. K-Factor Rock FreeEsri Soil OrderThese fields were calculated from the Component table using a model that included the Pivot Table Tool, the Summarize Tool and a custom script. The first 11 fields provide the sum of Component Percentage Representative Value for each soil order for each map unit. The Soil Order Dominant Condition field was calculated by selecting the highest value in the preceding 11 soil order fields. In the case of tied values the component with the lowest average slope value (slope_r) was selected. If both soil order and slope were tied the first value in the table was selected.Percent AlfisolsPercent AndisolsPercent AridisolsPercent EntisolsPercent GelisolsPercent HistosolsPercent InceptisolsPercent MollisolsPercent SpodosolsPercent UltisolsPercent VertisolsSoil Order - Dominant ConditionEsri Popup StringThis field contains a text string calculated by Esri that is used to create a basic pop-up using some of the more popular SSURGO attributes.Map Unit KeyThe Mapunit key field is found throughout SSURGO and

New global maps of the Köppen-Geiger climate classification at an unprecedented 1-km resolution for the present day (1980–2016) and for projected future conditions (2071–2100) under climate change. The maps are stored in GeoTIFF format as unsigned 8-bit integers. We also provide a legend file (legend.txt) linking the numeric values in the maps to the Köppen-Geiger climate symbols.Please cite the following paper when using the maps in any publication: Beck, H.E., N.E. Zimmermann, T.R. McVicar, N. Vergopolan, A. Berg, E.F. Wood: Present and future Köppen-Geiger climate classification maps at 1-km resolution, Scientific Data 5:180214, doi:10.1038/sdata.2018.214 (2018).

This dataset combines the work of several different projects to create a seamless data set for the contiguous United States. Data from four regional Gap Analysis Projects and the LANDFIRE project were combined to make this dataset. In the northwestern United States (Idaho, Oregon, Montana, Washington and Wyoming) data in this map came from the Northwest Gap Analysis Project. In the southwestern United States (Colorado, Arizona, Nevada, New Mexico, and Utah) data used in this map came from the Southwest Gap Analysis Project. The data for Alabama, Florida, Georgia, Kentucky, North Carolina, South Carolina, Mississippi, Tennessee, and Virginia came from the Southeast Gap Analysis Project and the California data was generated by the updated California Gap land cover project. The Hawaii Gap Analysis project provided the data for Hawaii. In areas of the county (central U.S., Northeast, Alaska) that have not yet been covered by a regional Gap Analysis Project, data from the Landfire project was used. Similarities in the methods used by these projects made possible the combining of the data they derived into one seamless coverage. They all used multi-season satellite imagery (Landsat ETM+) from 1999-2001 in conjunction with digital elevation model (DEM) derived datasets (e.g. elevation, landform) to model natural and semi-natural vegetation. Vegetation classes were drawn from NatureServe's Ecological System Classification (Comer et al. 2003) or classes developed by the Hawaii Gap project. Additionally, all of the projects included land use classes that were employed to describe areas where natural vegetation has been altered. In many areas of the country these classes were derived from the National Land Cover Dataset (NLCD). For the majority of classes and, in most areas of the country, a decision tree classifier was used to discriminate ecological system types. In some areas of the country, more manual techniques were used to discriminate small patch systems and systems not distinguishable through topography. The data contains multiple levels of thematic detail. At the most detailed level natural vegetation is represented by NatureServe's Ecological System classification (or in Hawaii the Hawaii GAP classification). These most detailed classifications have been crosswalked to the five highest levels of the National Vegetation Classification (NVC), Class, Subclass, Formation, Division and Macrogroup. This crosswalk allows users to display and analyze the data at different levels of thematic resolution. Developed areas, or areas dominated by introduced species, timber harvest, or water are represented by other classes, collectively refered to as land use classes; these land use classes occur at each of the thematic levels. Raster data in both ArcGIS Grid and ERDAS Imagine format is available for download at http://gis1.usgs.gov/csas/gap/viewer/land_cover/Map.aspx Six layer files are included in the download packages to assist the user in displaying the data at each of the Thematic levels in ArcGIS. In adition to the raster datasets the data is available in Web Mapping Services (WMS) format for each of the six NVC classification levels (Class, Subclass, Formation, Division, Macrogroup, Ecological System) at the following links. http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_NVC_Class_Landuse/MapServer http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_NVC_Subclass_Landuse/MapServer http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_NVC_Formation_Landuse/MapServer http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_NVC_Division_Landuse/MapServer http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_NVC_Macrogroup_Landuse/MapServer http://gis1.usgs.gov/arcgis/rest/services/gap/GAP_Land_Cover_Ecological_Systems_Landuse/MapServer

This dataset consists of the 1km raster, dominant aggregate class version of the Land Cover Map 2015 (LCM2015) for Great Britain. The 1km dominant coverage product is based on the 1km percentage product and reports the aggregated habitat class with the highest percentage cover for each 1km pixel. The 10 aggregate classes are groupings of 21 target classes, which are based on the Joint Nature Conservation Committee (JNCC) Broad Habitats, which encompass the entire range of UK habitats. The aggregate classes group some of the more specialised classes into more general categories. For example, the five coastal classes in the target class are grouped into a single aggregate coastal class. This dataset is derived from the vector version of the Land Cover Map, which contains individual parcels of land cover and is the highest available spatial resolution. LCM2015 is a land cover map of the UK which was produced at the Centre for Ecology & Hydrology by classifying satellite images from 2014 and 2015 into 21 Broad Habitat-based classes. LCM2015 consists of a range of raster and vector products and users should familiarise themselves with the full range (see related records, the CEH web site and the LCM2015 Dataset documentation) to select the product most suited to their needs. LCM2015 was produced at the Centre for Ecology & Hydrology by classifying satellite images from 2014 and 2015 into 21 Broad Habitat-based classes. It is one of a series of land cover maps, produced by UKCEH since 1990. They include versions in 1990, 2000, 2007, 2015, 2017, 2018 and 2019. Full details about this dataset can be found at https://doi.org/10.5285/711c8dc1-0f4e-42ad-a703-8b5d19c92247

This dataset contains the Soil map of the Grand-Duchy of Luxembourg at scale of 1:25.000. It is made up of soil typological units (SMU) containing information on texture, stoniness, nature of coarse fragments, drainage, depth, substrate and simplified pedogenetic classification. A simplified map (texture, nature of coarse fragments) has been produced from this dataset to reduce complexity while meeting most user needs. The current dataset has been produced between 1964 and 2020.

This land cover map of Dakar (Senegal) was created from a Pléiades very-high resolution imagery with a spatial resolution of 0.5 meter. The methodology followed a open-source semi-automated framework [1] that rely on GRASS GIS using a local unsupervised optimization approach for the segmentation part [2-3].

Description of the files:

• "Landcover.zip" : The direct output from the supervised classification using the Random Forest classifier.
• "Landcover_Postclassif_Level8_Splitbuildings.zip" : Post-processed version of the previous map ("Landcover"), with reduced misclassifications from the original classification (rule-based used to reclassify the errors, with a focus on built-up classes).
• "Landcover_Postclassif_Level8_modalfilter3.zip" : Smoothed version of the previous product (modal filter with window 3x3 applied on the "Landcover_Postclassif_Level8_Splitbuildings").
• "Landcover_Postclassif_Level9_Shadowsback.zip" : Corresponds to the "level8_Splitbuildings" with shadows coming from the original classification.
• "Dakar_legend_colors.txt" : Text file providing the correspondance between the value of the pixels and the legend labels and a proposition of color to be used.

References:

[1] Grippa, Taïs, Moritz Lennert, Benjamin Beaumont, Sabine Vanhuysse, Nathalie Stephenne, and Eléonore Wolff. 2017. “An Open-Source Semi-Automated Processing Chain for Urban Object-Based Classification.” Remote Sensing 9 (4): 358. https://doi.org/10.3390/rs9040358.

[2] Grippa, Tais, Stefanos Georganos, Sabine G. Vanhuysse, Moritz Lennert, and Eléonore Wolff. 2017. “A Local Segmentation Parameter Optimization Approach for Mapping Heterogeneous Urban Environments Using VHR Imagery.” In Proceedings Volume 10431, Remote Sensing Technologies and Applications in Urban Environments II., edited by Wieke Heldens, Nektarios Chrysoulakis, Thilo Erbertseder, and Ying Zhang, 20. SPIE. https://doi.org/10.1117/12.2278422.

[3] Georganos, Stefanos, Taïs Grippa, Moritz Lennert, Sabine Vanhuysse, and Eleonore Wolff. 2017. “SPUSPO: Spatially Partitioned Unsupervised Segmentation Parameter Optimization for Efficiently Segmenting Large Heterogeneous Areas.” In Proceedings of the 2017 Conference on Big Data from Space (BiDS’17).

Founding:

This dataset was produced in the frame of two research project : MAUPP (http://maupp.ulb.ac.be) and REACT (http://react.ulb.be), funded by the Belgian Federal Science Policy Office (BELSPO).

While there have been many maps produced that depict vegetation for the state of Hawai‘i only a few of these display land cover for all of the main Hawaiian Islands, and most of those that were created before the year 2000 have very generalized units or are somewhat inaccurate as a result of more recent land use changes or due to poor resolution (both spatial and spectral) in the imagery that was used to produce the map. Some of the more detailed and accurate maps include the Hawai‘i GAP Analysis (HI-GAP) Land Cover map (Gon et al. 2006), the NOAA C-CAP Land Cover map (NOAA National Ocean Service Coastal Services Center 2012), and the more recently released Hawai‘i LANDFIRE EVT Land Cover map (U.S. Geological Survey 2009). However, all of these maps as originally produced were not considered to be detailed enough, current enough, or had other classification issues that would not allow them to be used as the primary base for the Hawai‘i Carbon Assessment. For the Hawai‘i Carbon Assessment we integrated components from several of these previously mentioned land cover and land use mapping efforts and combined them into a single new land cover map (CAH Land Cover) that was further updated using very-high-resolution imagery. The hierarchical classification system of the CAH Land Cover map allows for grouping the mapped units into different configurations, ranging from very detailed plant communities reflecting current conditions to very generalized major land cover units and biomes that represent land use and potential vegetation zones, respectively. The CAH Land Cover classification is hierarchical with forty-eight CAH Detailed Land Cover units which can be grouped into twenty-seven CAH General Land Cover units, thirteen CAH Biome units, and seven CAH Major Land Cover units (Appendix 1). The CAH Detailed Land Cover units generally correspond to the rUSNVC Association level, the CAH General Land Cover units are related to the rUSNVC Group level, and the CAH Biome units connect to the rUSNVC Subclass level.

The 2023 Wildland Hazard Potential (WHP) represents areas of significant fuel buildup. The WHP is a raster geospatial product produced by the USFS Fire Modeling Institute in the Fire, Fuel, and Smoke Program. Higher values have a higher probability of high-intensity fire, with torching, crowning, and other forms of extreme fire behavior. Data used represent areas classified as “high” or “very high” risk. The LANDFIRE Fire Regime Groups (FRG) raster dataset (LC16_BPS_200) was used to identify moderate fire regimes (Regime groups I and II). Fire regime groups represent expected historical fire regimes based on interactions between vegetation dynamics, fire spread, fire effects, and spatial context. The definitions of the regimes are outlined in the Interagency Fire Regime Condition Class Guidebook. The data for this indicator represent areas identified as having “moderate”, “high”, or “very high” wildland hazard potential within fire regime groups I and II. Wildfire hazard potential, version 2023, 4th Edition. https://www.firelab.org/project/wildfire-hazard-potential (October 2, 2023)

The Cooperative Land Cover Map is a project to develop an improved statewide land cover map from existing sources and expert review of aerial photography. The project is directly tied to a goal of Florida's State Wildlife Action Plan (SWAP) to represent Florida's diverse habitats in a spatially-explicit manner. The Cooperative Land Cover Map integrates 3 primary data types: 1) 6 million acres are derived from local or site-specific data sources, primarily on existing conservation lands. Most of these sources have a ground-truth or local knowledge component. We collected land cover and vegetation data from 37 existing sources. Each dataset was evaluated for consistency and quality and assigned a confidence category that determined how it was integrated into the final land cover map. 2) 1.4 million acres are derived from areas that FNAI ecologists reviewed with high resolution aerial photography. These areas were reviewed because other data indicated some potential for the presence of a focal community: scrub, scrubby flatwoods, sandhill, dry prairie, pine rockland, rockland hammock, upland pine or mesic flatwoods. 3) 3.2 million acres are represented by Florida Land Use Land Cover data from the FL Department of Environmental Protection and Water Management Districts (FLUCCS). The Cooperative Land Cover Map integrates data from the following years: NWFWMD: 2006 - 07 SRWMD: 2005 - 08 SJRWMD: 2004 SFWMD: 2004 SWFWMD: 2008 All data were crosswalked into the Florida Land Cover Classification System. This project was funded by a grant from FWC/Florida's Wildlife Legacy Initiative (Project 08009) to Florida Natural Areas Inventory. The current dataset is provided in 10m raster grid format.Changes from Version 1.1 to Version 2.3:CLC v2.3 includes updated Florida Land Use Land Cover for four water management districts as described above: NWFWMD, SJRWMD, SFWMD, SWFWMDCLC v2.3 incorporates major revisions to natural coastal land cover and natural communities potentially affected by sea level rise. These revisions were undertaken by FNAI as part of two projects: Re-evaluating Florida's Ecological Conservation Priorities in the Face of Sea Level Rise (funded by the Yale Mapping Framework for Biodiversity Conservation and Climate Adaptation) and Predicting and Mitigating the Effects of Sea-Level Rise and Land Use Changes on Imperiled Species and Natural communities in Florida (funded by an FWC State Wildlife Grant and The Kresge Foundation). FNAI also opportunistically revised natural communities as needed in the course of species habitat mapping work funded by the Florida Department of Environmental Protection. CLC v2.3 also includes several new site specific data sources: New or revised FNAI natural community maps for 13 conservation lands and 9 Florida Forever proposals; new Florida Park Service maps for 10 parks; Sarasota County Preserves Habitat Maps (with FNAI review); Sarasota County HCP Florida Scrub-Jay Habitat (with FNAI Review); Southwest Florida Scrub Working Group scrub polygons. Several corrections to the crosswalk of FLUCCS to FLCS were made, including review and reclassification of interior sand beaches that were originally crosswalked to beach dune, and reclassification of upland hardwood forest south of Lake Okeechobee to mesic hammock. Representation of state waters was expanded to include the NOAA Submerged Lands Act data for Florida.Changes from Version 2.3 to 3.0: All land classes underwent revisions to correct boundaries, mislabeled classes, and hard edges between classes. Vector data was compared against high resolution Digital Ortho Quarter Quads (DOQQ) and Google Earth imagery. Individual land cover classes were converted to .KML format for use in Google Earth. Errors identified through visual review were manually corrected. Statewide medium resolution (spatial resolution of 10 m) SPOT 5 images were available for remote sensing classification with the following spectral bands: near infrared, red, green and short wave infrared. The acquisition dates of SPOT images ranged between October, 2005 and October, 2010. Remote sensing classification was performed in Idrisi Taiga and ERDAS Imagine. Supervised and unsupervised classifications of each SPOT image were performed with the corrected polygon data as a guide. Further visual inspections of classified areas were conducted for consistency, errors, and edge matching between image footprints. CLC v3.0 now includes state wide Florida NAVTEQ transportation data. CLC v3.0 incorporates extensive revisions to scrub, scrubby flatwoods, mesic flatwoods, and upland pine classes. An additional class, scrub mangrove – 5252, was added to the crosswalk. Mangrove swamp was reviewed and reclassified to include areas of scrub mangrove. CLC v3.0 also includes additional revisions to sand beach, riverine sand bar, and beach dune previously misclassified as high intensity urban or extractive. CLC v3.0 excludes the Dry Tortugas and does not include some of the small keys between Key West and Marquesas.Changes from Version 3.0 to Version 3.1: CLC v3.1 includes several new site specific data sources: Revised FNAI natural community maps for 31 WMAs, and 6 Florida Forever areas or proposals. This data was either extracted from v2.3, or from more recent mapping efforts. Domains have been removed from the attribute table, and a class name field has been added for SITE and STATE level classes. The Dry Tortugas have been reincorporated. The geographic extent has been revised for the Coastal Upland and Dry Prairie classes. Rural Open and the Extractive classes underwent a more thorough reviewChanges from Version 3.1 to Version 3.2:CLC v3.2 includes several new site specific data sources: Revised FNAI natural community maps for 43 Florida Park Service lands, and 9 Florida Forever areas or proposals. This data is from 2014 - 2016 mapping efforts. SITE level class review: Wet Coniferous plantation (2450) from v2.3 has been included in v3.2. Non-Vegetated Wetland (2300), Urban Open Land (18211), Cropland/Pasture (18331), and High Pine and Scrub (1200) have undergone thorough review and reclassification where appropriate. Other classification errors were opportunistically corrected as found or as reported by users to landcovermap@myfwc.com.Changes from Version 3.2.5 to Version 3.3: The CLC v3.3 includes several new site specific data sources: Revised FNAI natural community maps for 14 FWC managed or co-managed lands, including 7 WMA and 7 WEA, 1 State Forest, 3 Hillsboro County managed areas, and 1 Florida Forever proposal. This data is from the 2017 – 2018 mapping efforts. Select sites and classes were included from the 2016 – 2017 NWFWMD (FLUCCS) dataset. M.C. Davis Conservation areas, 18331x agricultural classes underwent a thorough review and reclassification where appropriate. Prairie Mesic Hammock (1122) was reclassified to Prairie Hydric Hammock (22322) in the Everglades. All SITE level Tree Plantations (18333) were reclassified to Coniferous Plantations (183332). The addition of FWC Oyster Bar (5230) features. Other classification errors were opportunistically corrected as found or as reported by users to landcovermap@myfwc.com, including classification corrections to sites in T.M. Goodwin and Ocala National Forest. CLC v3.3 utilizes the updated The Florida Land Cover Classification System (2018), altering the following class names and numbers: Irrigated Row Crops (1833111), Wet Coniferous Plantations (1833321) (formerly 2450), Major Springs (4131) (formerly 3118). Mixed Hardwood-Coniferous Swamps (2240) (formerly Other Wetland Forested Mixed).Changes from Version 3.4 to Version 3.5: The CLC v3.5 includes several new site specific data sources: Revised FNAI natural community maps for 16 managed areas, and 10 Florida Forever Board of Trustees Projects (FFBOT) sites. This data is from the 2019 – 2020 mapping efforts. Other classification errors were opportunistically corrected as found or as reported by users to landcovermap@myfwc.com. This version of the CLC is also the first to include land identified as Salt Flats (5241).Changes from Version 3.5 to 3.6: The CLC v3.6 includes several new site specific data sources: Revised FNAI natural community maps for 11 managed areas, and 24 Florida Forever Board of Trustees Projects (FFBOT) sites. This data is from the 2018 – 2022 mapping efforts. Other classification errors were opportunistically corrected as found or as reported by users to landcovermap@myfwc.com.Changes from Version 3.6 to 3.7: The CLC 3.7 includes several new site specific data sources: Revised FNAI natural community maps for 5 managed areas (2022-2023). Revised Palm Beach County Natural Areas data for Pine Glades Natural Area (2023). Other classification errors were opportunistically corrected as found or as reported by users to landcovermap@myfwc.com. In this version a few SITE level classifications are reclassified for the STATE level classification system. Mesic Flatwoods and Scrubby Flatwoods are classified as Dry Flatwoods at the STATE level. Upland Glade is classified as Barren, Sinkhole, and Outcrop Communities at the STATE level. Lastly Upland Pine is classified as High Pine and Scrub at the STATE level.

This dataset provides global maps describing the land surface into 22 classes, which have been defined using the United Nations Food and Agriculture Organization’s (UN FAO) Land Cover Classification System (LCCS). In addition to the land cover (LC) maps, four quality flags are produced to document the reliability of the classification and change detection. In order to ensure continuity, these land cover maps are consistent with the series of global annual LC maps from the 1990s to 2015 produced by the European Space Agency (ESA) Climate Change Initiative (CCI), which are also available on the ESA CCI LC viewer. To produce this dataset, the entire Medium Resolution Imaging Spectrometer (MERIS) Full and Reduced Resolution archive from 2003 to 2012 was first classified into a unique 10-year baseline LC map. This is then back- and up-dated using change detected from (i) Advanced Very-High-Resolution Radiometer (AVHRR) time series from 1992 to 1999, (ii) SPOT-Vegetation (SPOT-VGT) time series from 1998 to 2012 and (iii) PROBA-Vegetation (PROBA-V), Sentinel-3 OLCI (S3 OLCI) and Sentinel-3 SLSTR (S3 SLSTR) time series from 2013. Beyond the climate-modelling communities, this dataset’s long-term consistency, yearly updates, and high thematic detail on a global scale have made it attractive for a multitude of applications such as land accounting, forest monitoring and desertification, in addition to scientific research.

Region 5 Forest Health Treatment Priority MappingThe number of acres of forests burning at high severity in recent years, combined with the recent drought-induced tree mortality event of 2015-2016 have more than ever highlighted unsustainable forest health conditions in California. Urgency for implementing preventative landscape-level tree density and fuels reduction treatments to restore and maintain forest resiliency to wildfires and drought (bark beetles) has now become an emergency. To accomplish meaningful landscape level treatments, land managers must be able to prioritize areas of highest risk that are conducive to project implementation. Forest Health Protection has analyzed a variety of readily available corporate GIS data sets to identify areas that are considered most at risk to high levels of bark beetle-caused tree mortality, have a high likelihood of experiencing stand replacing wildfire and are accessible and appropriate for mechanical thinning. This product has been used on several R5 National Forests for 5-year planning, identifying cross collaboration, all lands opportunities, and guiding layout of new projects using the Farm Bill insect and disease treatment Categorical Exclusion authority under NEPA. This webmap illustrates areas deemed at high risk of tree mortality, due to bark beetles, on all lands throughout the state. These same areas should also be considered at a risk to high-severity wildfire due to overstocked conditions and generally high fuel loading from past tree mortality. The webmap is suitable for landscape-level planning, rather than stand-level planning, as the data used to identify priority treatment areas are not sufficiently detailed for use at the stand level. Ground verification of areas identified in the map as priorities for treatment is highly recommended. Areas mapped outside of USDA National Forest System lands may not reflect recent management activities. Basic consideration for classification as high priority for treatment required that areas:Have not suffered moderate or high severity wildfire since at least 1998;Have not been thinned by the USDA Forest Service since at least 2005;Have not experienced stand-replacing disturbance, owing to clear-cut or natural mortality, since at least 2005;Contain stands with 60% or higher relative stand density;Are dominated by trees with diameter at breast height (DBH) of 11” or more.Lands that met the basic conditions were then classified as high priority for treatment based on the species composition and density of the stands that they contain.Highest priority was assigned to locations with stands that contain:Pines principally, and have stand density index (SDI) of 220 or higher; OR Fir-dominated mixed conifer and white fir, have SDI 270 or higher, and historically contained mostly pines; OR Pine-dominated mixed conifers, and have SDI 270 or higher.Pine-dominated stands are typically associated with drier sites and often experience higher levels of tree mortality associated with high stand density, bark beetles, and drought.Second priority was assigned to locations with stands that:Contain fir-dominated mixed conifer and white fir, have SDI 330 or higher;Were not classified as highest priority.Fir-dominated stands found on more mesic sites can also experience elevated tree mortality associated with high stand density, bark beetles, and drought, though generally at a lower level than pine-dominated stands or fir-dominated stands growing on historically pine-dominated sites.Download the thinning priority layers displayed in this WebMap. In addition to what is displayed on this webmap, the download also includesThird priority including smaller DBH of 6" - 11" 50% relative stand density (dependent on dominant species)Regional Dominance Type for each priority pixel

This is a 10-metre pixel data set representing the land surface, classified into 21 UKCEH land cover classes, based upon Biodiversity Action Plan broad habitats. The pixel product is given as a two-band raster in geoTiff format. The first band gives the most likely land cover type; the second band gives the probability associated with this land cover. The probability layer is an indicator of uncertainty (0 to 100). Low values correspond to low certainty (higher uncertainty). This is the first 10m resolution land cover map produced by UKCEH. It succeeds 20m resolution classified pixel products from 2017, 2018 and 2019. A full description of this and all UKCEH LCM2020 products are available from the product documentation accompanying this data.

The choropleth map is a device used for the display of socioeconomic data associated with an areal partition of geographic space. Cartographers emphasize the need to standardize any raw count data by an area-based total before displaying the data in a choropleth map. The standardization process converts the raw data from an absolute measure into a relative measure. However, there is recognition that the standardizing process does not enable the map reader to distinguish between low–low and high–high numerator/denominator differences. This research uses concentration-based classification schemes using Lorenz curves to address some of these issues. A test data set of nonwhite birth rate by county in North Carolina is used to demonstrate how this approach differs from traditional mean–variance-based systems such as the Jenks’ optimal classification scheme.

This data release includes a file geodatabase with land cover maps for a 313-kilometer segment along the mainstem Klamath River corridor downstream from Iron Gate Dam, CA. The maps were derived from high-resolution (15cm) imagery and topobathymetric elevation data, collected by NV5 Geospatial (formerly QSI, Inc) and published on Open Topography (https://doi.org/10.5069/G9DN436N). Acquisition dates spanned 06/01/2018 to 06/14/2018. The area of interest was discretized into three zones and land cover maps were generated using geoprocessing tools, implemented in ArcMap 10.8.1, and image classification tools, implemented in Ecognition 10.2. The maps include six land cover classes: 1) trees, 2) shrubs, 3) grass, 4) bare ground, 5) rock outcrop and 6) water. Water was defined by a lidar hydrobreak raster product. Trees, shrubs and grass were classified using a lidar-derived canopy height model and threshold values in vegetation indices. The bare ground and rock outcrop classes were further classified using a random forest classifier. A total of 639 independent reference points were used in an accuracy assessment. The overall accuracy of the Level II classification (all 6 classes) is 87%. The overall accuracy of the Level I classification [ 6) water, 7) vegetation (trees, shrubs, grass) and 8) unvegetated (bare ground, rock outcrop) ] is 95.1%. See attached Klamath_AccuracyAssessment.csv for more details. The file geodatabase includes land cover maps for three zones. Zone A covers Iron Gate Dam to Fort Goff. Zone B covers Fort Goff to Somes Bar. Zone C covers Somes Bar to the estuary. The USGS Integrated Water Availability Assessments (IWAAs) Program and the USGS National Land Imaging Program funded generation of these high-resolution digital products.

This data publication contains multiple maps of Puerto Rico scanned at 600 dots per inch: full map scans, scans clipped to mapped areas only, and georeferenced scans of 1:10,000-scale land-use maps from 1950-1951 that were produced by the Rural Land Classification Program of Puerto Rico, a project led by Dr. Clarence F. Jones of Northwestern University. These historical maps classified land use and land cover into 20 different classes, including 13 different types of crops, two classes of forests, four classes of grasslands and other areas, which is a general class for non-rural areas. This package includes maps from 76 out of the 78 municipalities of Puerto Rico, covering 422 quadrangles of a 443-quadrangle grid for mainland Puerto Rico. It excludes the island municipalities of Vieques and Culebra, Mona Island and minor outlying islands.The Rural Land Classification Program of Puerto Rico produced 430 1:10,000-scale maps. That program also produced one island-wide land-use map with more generalized delineations of land use. Previously, Kennaway and Helmer (2007) scanned and georeferenced the island-wide map, and they converted it to vector and raster formats with embedded georeferencing and classification. This data publication contains the higher-resolution maps, which will provide more precise historical context for forests. It will better inform management efforts for the sustainable use of forest lands and to build resilience and resistance to various future disturbances for these and other tropical forest landscapes.

The maps were scanned and georeferenced to help with the planning and application process for the USDA Forest Service (USDA) Forest Legacy Program, a competition-based program administered by the USDA Forest Service in partnership with State agencies to encourage the protection of privately owned forest lands through conservation easements or land purchases. Geospatial products and maps will also be used by personnel at the Department of Natural and Environmental Resources and partners in Non-Governmental Organizations working with the Forest Stewardship Program. This latter program provides technical assistance and forest management plans to private landowners for the conservation and effective management of private forests across the US. The information will provide local historical context on forest change patterns that will enhance the recommendations of forest management practices for private forest landowners. These data will also be useful for urban forest professionals to understand the land legacies as a basis for planning green infrastructure interventions.

Data depict the rural areas of Puerto Rico around 1951 and how they were classified by geographers then. Having it georeferenced allows managers, teachers, students, the public and scientists to compare how these classifications have changed throughout the years. It will allow more precise identification and mapping of the past land use of present forests, forest stand age, and the past juxtaposition of different land uses relative to each other. These factors can affect forest species composition, biodiversity and ecosystem services. Forest stand age, past land-use type and past disturbance type, forest example, help gauge current forest structure, carbon storage, or rates of carbon accumulation. Another example of how the maps are important is for understanding how watersheds have changed through time, which helps assess how forest ecosystem services related to hydrology evolve. These maps will also help gauge how the forests of Puerto Rico are responding to recent disturbances, and how past disturbances over a range of scales relate to these responses.For more information on the Rural Land Classification Program of Puerto Rico, generated maps, and the island-wide land-use map, please see Jones (1952), Jones and Berrios (1956), as well as Kennaway and Helmer (2007).