GLC_FCS30D - 全球 30 米土地覆被变化数据集(1985-2022 年)
注
本数据集是正在提交的论文的一部分,因此没有引用和 DOI 信息。请在使用本数据集时注意这一点。
GLC_FCS30D 数据集是全球土地覆被监测领域的一项开创性进展,它以 30 米的分辨率全面揭示了从 1985 年到 2022 年期间的土地覆被动态。GLC_FCS30D 采用连续变化探测方法开发,利用了谷歌地球引擎平台中大量的陆地卫星图像档案,包括 35 个土地覆被子类别和 26 个时间步长,2000 年之前每五年更新一次,之后每年更新一次。通过严格的完善过程,包括时空分类和时间一致性优化,该数据集实现了高置信度的准确性,经过 84,000 多个全球样本的验证,总体准确率达到 80.88%。值得注意的是,GLC_FCS30D 阐明了重要趋势,揭示了森林和耕地变化是过去 37 年全球土地覆被变化的主要驱动因素,森林净损失约 250 万平方公里,耕地面积净增加约 130 万平方公里。GLC_FCS30D 具有多样化的分类系统、高空间分辨率和广泛的时间覆盖范围,是气候变化研究和可持续发展分析的宝贵资源。前言 – 人工智能教程
GLC_FCS30D 是首个采用连续变化探测技术的 30 米分辨率全球精细土地覆被动态产品。它采用了一套包含 35 个土地覆被类别的精细分类系统,时间跨度为 1985 年至 2022 年。2000 年之前,更新周期为每 5 年一次,2000 年之后,更新周期为每年一次。具体而言,该系统结合了连续变化检测方法、局部自适应更新模型和来自密集时间序列大地遥感卫星图像的时空优化算法,经过验证,基本分类系统(10 个主要土地覆被类型)的总体精度为 80.88%(±0.27%),LCCS 一级验证系统(17 个 LCCS 土地覆被类型)的总体精度为 73.24%(±0.30%)。
数据链接
土地分类类型
展开以显示土地覆被等级、RGB 值和十六进制代码
LC Id | Classification System | RGB value | Color |
10 | Rainfed cropland | (255,255,100) | |
11 | Herbaceous cover cropland | (255,255,100) | |
12 | Tree or shrub cover (Orchard) cropland | (255,255,0) | |
20 | Irrigated cropland | (170,240,240) | |
51 | Open evergreen broadleaved forest | (76,115,0) | |
52 | Closed evergreen broadleaved forest | (0,100,0) | |
61 | Open deciduous broadleaved forest (0.15<fc<0.4) | (170,200,0) | |
62 | Closed deciduous broadleaved forest (fc>0.4) | (0,160,0) | |
71 | Open evergreen needle-leaved forest (0.15< fc <0.4) | (0,80,0) | |
72 | Closed evergreen needle-leaved forest (fc >0.4) | (0,60,0) | |
81 | Open deciduous needle-leaved forest (0.15< fc <0.4) | (40,100,0) | |
82 | Closed deciduous needle-leaved forest (fc >0.4) | (40,80,0) | |
91 | Open mixed leaf forest (broadleaved and needle-leaved) | (160,180,50) | |
92 | Closed mixed leaf forest (broadleaved and needle-leaved) | (120,130,0) | |
120 | Shrubland | (150,100,0) | |
121 | Evergreen shrubland | (150,75,0) | |
122 | Deciduous shrubland | (150,100,0) | |
130 | Grassland | (255,180,50) | |
140 | Lichens and mosses | (255,220,210) | |
150 | Sparse vegetation (fc<0.15) | (255,235,175) | |
152 | Sparse shrubland (fc<0.15) | (255,210,120) | |
153 | Sparse herbaceous (fc<0.15) | (255,235,175) | |
181 | Swamp | (0,168,132) | |
182 | Marsh | (115,255,223) | |
183 | Flooded flat | (158,187,215) | |
184 | Saline | (130,130,130) | |
185 | Mangrove | (245,122,182) | |
186 | Salt marsh | (102,205,171) | |
187 | Tidal flat | (68,79,137) | |
190 | Impervious surfaces | (195,20,0) | |
200 | Bare areas | (255,245,215) | |
201 | Consolidated bare areas | (220,220,220) | |
202 | Unconsolidated bare areas | (255,245,215) | |
210 | Water body | (0,70,200) | |
220 | Permanent ice and snow | (255,255,255) | |
0, 250 | Filled value | (255,255,255) |
数据集后处理
数据集由约 961 个图层组成,年度图层由约 23 年的图像组成,每个波段代表从 2000 年开始的一年,5 年图层从 1985 年开始,3 个波段代表 5 年的间隔,因此 1985-1990 年为 b1,1990-1995 年为 b2,1990-2000 年为 b3。
论文引用
Zhang, X., Zhao, T., Xu, H., Liu, W., Wang, J., Chen, X., and Liu, L.: GLC_FCS30D: The first global 30-m land-cover dynamic monitoring product with
a fine classification system from 1985 to 2022 using dense time-series Landsat imagery and continuous change-detection method, Earth Syst. Sci. Data
Discuss. [preprint], https://doi.org/10.5194/essd-2023-320, in review, 2023.
数据引用
Liangyun Liu, Xiao Zhang, & Tingting Zhao. (2023). GLC_FCS30D: the first global 30-m land-cover dynamic monitoring product with fine classification
system from 1985 to 2022 [Data set]. Zenodo. https://doi.org/10.5281/zenodo.8239305
数据集代码
var annual = ee.ImageCollection("projects/sat-io/open-datasets/GLC-FCS30D/annual");
var five_year = ee.ImageCollection("projects/sat-io/open-datasets/GLC-FCS30D/five-years-map");
代码链接
License¶
This work is licensed under and freely available to the public under Creative Commons Attribution 4.0 International license.
Created by: Zhang et al. 2023
Curated in GEE by : Samapriya Roy
Keywords: GLC_FCS30D, 1985-2022, Land-cover change, Landsat, change detection, Google Earth Engine
Last updated in GEE: 2024-02-20