Google Earth Engine(GEE)——全球1公里分辨率下12种土地覆盖

全球共识土地覆盖
这些数据集整合了多个全球遥感衍生的土地覆盖产品,并提供了关于1公里分辨率下12种土地覆盖类别流行程度的共识信息。有关整合方法和数据集评价的其他信息。前言 – 床长人工智能教程

数据集详情
目前有两个版本的共识土地覆盖数据集。完整版是整合GlobCover(2005-06;v2.2)、MODIS土地覆盖产品(MCD12Q1;v051)、GLC2000(全球产品;v1.1)和DISCover(GLCC;v2)的数据集。缩减版是仅整合前三个产品的数据集(即不含DISCover)。

每个数据集包含12个数据层,每个数据层提供关于一个土地覆盖类别的普遍性的共识信息。所有数据层都包含无符号的8位数值,有效数值范围为0-100,代表共识的流行率,以百分比表示。所有数据层的空间范围为90ºN-56ºS和180ºW-180ºE,空间分辨率为每像素30角秒(赤道上每像素约1公里)。

方法
使用一个通用的分类方案和一个基于精度的整合方法,我们整合了四个全球土地覆盖产品。我们评估了该产品的性能,与估计30米亚像素分辨率土地覆盖的输入进行了比较。我们还比较了用不同产品建立的演绎式和归纳式物种分布模型的准确性,以模拟六种鸟类栖息地专家的大陆分布情况。

结果
我们的产品提供了全球范围内每个标称的1公里像素内12种土地覆被类别流行程度的准确加权共识信息(南极洲除外)。与四个基础产品相比,它能更好地捕捉到细粒度验证数据中包含的所有类别和大多数单独类别的土地覆盖信息。在检测每个细粒土地覆盖类别的存在方面,它也具有最高的灵敏度和总体准确性。用共识数据集建立的演绎模型和归纳模型在建立鸟类物种分布模型方面都具有最高或第二高的准确性。

主要结论
我们的共识产品整合了四个基础产品,并成功地将精度最大化,减少了遗漏的错误。具体来说,共识产品减少了由错误分类、错误缺失率和现有土地覆盖产品的分类格式造成的限制。因此,它在捕捉亚像素土地覆盖信息的能力和对物种分布建模的效用方面超过了单一的基础产品。所提出的方法和共识产品在生物多样性研究以及全球陆地生态系统的理解和建模方面都有多种应用。

论文引证

Tuanmu, M.-N. and W. Jetz. 2014. A global 1-km consensus land-cover product for biodiversity and ecosystem modeling.
Global Ecology and Biogeography 23(9): 1031-1045.

Earth Engine Snippet

var barren = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/barren");
var cultivated_and_managed_vegetation = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/cultivated_and_managed_vegetation");
var deciduous_broadleaf_trees = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/deciduous_broadleaf_trees");
var evergreen_deciduous_needleleaf_trees = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/evergreen-deciduous_needleleaf_trees");
var evergreen_broadleaf_trees = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/evergreen_broadleaf_trees");
var herbaceous_vegetation = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/herbaceous_vegetation");
var mixed_other_trees = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/mixed-other_trees");
var open_water = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/open_water");
var regularly_flooded_vegetation = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/regularly_flooded_vegetation");
var shrubs = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/shrubs");
var snow_ice = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/snow-ice");
var urban_built_up = ee.Image("projects/sat-io/open-datasets/global_consensus_landcover/urban-built-up");

Sample Code: https://code.earthengine.google.com/?scriptPath=users/sat-io/awesome-gee-catalog-examples:earthenv-bd-ecosystems-clim-layers/GLOBAL-CONSENSUS-LANDCOVER

License

EarthEnv Global 1-km Consensus Land Cover Version 1 by Tuanmu & Jetz is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Permissions beyond the scope of this license may be available at Global 1-km Consensus Land Cover - EarthEnv.

Dataset citation

Tuanmu, M.-N. and W. Jetz. 2014. A global 1-km consensus land-cover product for biodiversity and ecosystem modeling.
Global Ecology and Biogeography 23(9): 1031-1045. Data available on-line at http://www.earthenv.org/.

Project Website: Global 1-km Consensus Land Cover - EarthEnv

App Website: App link here

Curated by: Samapriya Roy

Keywords: Earthenv, barren, cultivated and managed vegetation, deciduous broadleaf trees, evergreen broadleaf trees, mixed other trees, shrubs, urban built up, evergreen deciduous needleleaf trees, mixed other trees

Last updated: 2021-05-09

猜你喜欢

转载自blog.csdn.net/qq_31988139/article/details/130890724