Global Pasture Watch - Annual grassland class and extent maps at 30-m spatial resolution (2000—2024)

Sub-dataset: Dominant grassland class, 2015-2017 Description Global annual grassland class and extent for 2000—2024 produced by Parente et al. (2024) within the scope of the Global Pasture Watch initiative. The mapped grassland extent includes any land cover type, which contains at least 30% of dry or wet low vegetation, dominated by grasses and forbs (less than 3 meters) and a: maximum of 50% tree canopy cover (greater than 5 meters), maximum of 70% of other woody vegetation (scrubs and open shrubland), and maximum of 50% active cropland cover in mosaic landscapes of cropland & other vegetation. The grassland extent is classified into two classes: Cultivated grassland: Areas where grasses and other forage plants have been intentionally planted and managed, as well as areas of native grassland-type vegetation where they clearly exhibit active and 'heavy' management for specific human-directed uses, such as directed grazing of livestock. Natural/semi-natural grassland: Relatively undisturbed native grasslands/short-height vegetation, such as steppes and tundra, as well as areas that have experienced varying degrees of human activity in the past, which may contain a mix of native and introduced species due to historical land use and natural processes. In general, they exhibit natural-looking patterns of varied vegetation and clearly ordered hydrological relationships throughout the landscape. Open shrubland (v2-beta): Land on which the vegetation is dominated by low-growing woody plants, characterized by a sparse distribution of shrubs and dominated by woody perennials. Typically covers 50—75% of the area, with significant open ground (with or without herbaceous understory) between them, where shrub canopies are less than 10 meters in diameter, and tree cover is below 10%, meaning they do not form a continuous or semi-continuous canopy. The dataset is organized in 69 global mosaics (25 years for each time series) in COG (Cloud Optimized GeoTIFF) format, WGS84 Coordinate Systems (EPSG:4326) and pixel size equal to 0.00025 degrees, including: Probabilities of cultivated grassland (values range from 0–100), Probabilities of natural/semi-natural grassland (values range from 0–100), and Probabilities of open shrubland (values range from 0–100), and Dominant class (0-other land cover, 1-cultivated grassland and 2-natural/semi-natural grassland, 3-open shrubland). All raster files are in unsigned 8-bit integer format and use 255 as no-data value (pixels ignored by prediction), following an specific naming convention: Project name: Global Pasture Watch (gpw) Class name: cultivated grassland (cultiv.grassland), natural/semi-natural grassland (nat.semi.grassland), open shrubland (open.shrubland) and dominant grassland (grassland) Procedure combination: Random Forest (rf), median filter (med.filt) and balanced threshold (bthr). Variable type: probability (p) and factor class (c) Spatial resolution: 30m Begin of time reference: date of first Landsat composite used by the modeling (20240101) End of time reference: date of last Landsat composite used by the modeling (20241231) Spatial extent: global (go) Coordinate system: World Geodetic System 1984, used in GPS (epsg.4326) Version: v2 Related resources Maps of dominant grassland:2000-2002 2003-2005 2006-2008 2009-2011 2012-2014 2015-2017 2018-2020 2021-2023 2024 Probability maps of cultivated grassland:2000-2024 (All URLs) Probability maps of natural/semi-natural grassland:2000-2024 (All URLs) Grassland reference samples based on VHR imagery (2000–2024):GeoPackage files Global machine learning models (Random Forest):Parquet and joblib python files Reference sampling design derived by FSCV:GeoPackage and raster files Harmonized reference samples based on existing LULC dataset:GeoPackage and raster files Source code for reproducibility:GitHub release Mapping feedback tool:GeoWiki Data catalogues:OpenLandMap STAC Google Earth Engine Support For questions of bugs/inconsistencies related to the dataset raise a GitHub issue in https://github.com/wri/global-pasture-watch子数据集：2015-2017年优势草原类 描述 本数据集由Parente等人（2024）在全球牧场观察（Global Pasture Watch）计划框架下生成，包含2000—2024年的全球年度草原类型与分布范围。 本次映射的草原分布范围涵盖满足以下全部条件的土地覆盖类型：以草本和非禾本草本植物（高度低于3米）为优势种，干湿低矮植被占比不低于30%，且需同时满足： 1. 乔木冠层覆盖占比最高不超过50%（乔木高度高于5米）； 2. 其他木本植被（灌丛与开阔灌丛地）覆盖占比最高不超过70%； 3. 在农田与其他植被镶嵌的景观中，活跃耕地覆盖占比最高不超过50%。 草原分布范围被划分为以下两类： 1. 栽培草原：经人工有意种植与管理的草本及其他饲用植物分布区域，以及原生草原型植被经明确的高强度人工管理以服务于特定人类用途（如定向畜牧放牧）的区域。 2. 天然/半天然草原：相对未受干扰的原生草原/低矮植被区域，如温带草原与冻原，以及历史上经历过不同程度人类活动、因过往土地利用与自然过程同时包含本土与引入物种的区域。此类区域整体呈现自然化的多样植被格局，且景观内水文关系清晰有序。 此外还包含开阔灌丛地（Open shrubland，v2-beta）：植被以低矮木本植物为优势的土地类型，特征为灌木稀疏分布，以多年生木本植物为主。通常覆盖区域的50%—75%，其间存在大面积裸露地面（可伴生草本下层植被）；灌木冠幅直径小于10米，乔木覆盖占比低于10%，无法形成连续或半连续的冠层。 本数据集以云优化GeoTIFF（Cloud Optimized GeoTIFF，COG）格式存储，共包含69幅全球镶嵌影像，每条时间序列涵盖25年数据，采用WGS84坐标系（EPSG:4326），像素分辨率为0.00025度（约30米），包含以下数据图层： - 栽培草原概率（取值范围0–100） - 天然/半天然草原概率（取值范围0–100） - 开阔灌丛地概率（取值范围0–100） - 优势类层（0代表其他土地覆盖，1代表栽培草原，2代表天然/半天然草原，3代表开阔灌丛地） 所有栅格文件均采用无符号8位整型格式，以255作为无数据值（预测时将忽略该像素值），并遵循特定命名规范： - 项目名称：Global Pasture Watch（缩写gpw） - 类名：栽培草原（cultiv.grassland）、天然/半天然草原（nat.semi.grassland）、开阔灌丛地（open.shrubland）与优势草原类（grassland） - 处理组合：随机森林（Random Forest，rf）、中值滤波（median filter，med.filt）与平衡阈值（balanced threshold，bthr） - 变量类型：概率型（p）与因子类（c） - 空间分辨率：30米 - 时间参考起始：建模所用首幅Landsat合成影像的日期（20240101） - 时间参考结束：建模所用末幅Landsat合成影像的日期（20241231） - 空间范围：全球（go） - 坐标系：GPS所用的世界大地测量系统1984（epsg.4326） - 版本：v2 相关资源 - 优势草原类地图：2000-2002、2003-2005、2006-2008、2009-2011、2012-2014、2015-2017、2018-2020、2021-2023、2024 - 栽培草原概率地图：2000-2024（全部链接） - 天然/半天然草原概率地图：2000-2024（全部链接） 基于超高分辨率（Very High Resolution，VHR）影像的草原参考样本（2000–2024）：GeoPackage格式文件 全球机器学习模型（随机森林）：Parquet与joblib格式Python文件 由FSCV衍生的参考采样设计：GeoPackage与栅格格式文件 基于现有土地覆盖与土地利用（Land Use and Land Cover，LULC）数据集的协调参考样本：GeoPackage与栅格格式文件 可复现性源代码：GitHub发布版本 制图反馈工具：GeoWiki 数据目录：OpenLandMap STAC、Google Earth Engine 支持与反馈 若您发现与本数据集相关的错误或不一致问题，请在https://github.com/wri/global-pasture-watch 提交GitHub Issue。