黑河流域1公里土地覆盖格网数据集（2000）

黑河流域1公里土地覆盖图是冉有华等（2009；2011）发展的融合了多源本地信息的中国1公里土地覆盖图（MICLCover）的子集。MICLCover土地覆盖图采用IGBP土地覆盖分类系统，基于证据理论，融合了2000年中国1：10万土地利用数据、中国植被图集（1：100万）的植被型、中国1：10万冰川分布图、中国1：100万沼泽湿地图和MODIS 2001年土地覆盖产品（MOD12Q1）。MICLCover的验证结果表明，其与中国土地利用图在7类水平上的总体一致性达到88.84%，其中，耕地、城市、湿地和水体类型的一致性达到95%以上；通过与MODIS2001年土地覆盖数据产品和IGBPDISCover土地覆盖图在三个典型地区的视觉比较，MICLCover在保持了中国土地利用图的总体精度，增加了中国植被图叶属性和叶型信息的同时，所反映的局部土地覆盖细节更加详细。采用国家森林资源调查数据，在甘肃省、云南省、浙江省、黑龙江和吉林省的验证结果表明，MICLCover的森林类型的精度相对于MODIS土地覆盖产品有大幅度的提高；采用甘肃省祁连山国家自然保护区管理局的一类森林资源调查数据对MICLCover的森林类型进行验证，结果表明，MICLCover的森林类型在该地区的精度为82.94%。
总之，MICLCover土地覆盖图在保持了中国土地利用数据的总体精度条件下，补充了中国植被图中对植被类型及植被季相的信息，更新的中国湿地图、中国冰川图最新信息，使得中国土地覆盖数据的精度得到大大提高，分类系统更加通用，该数据可为陆面过程模型提供更高精度的土地覆盖信息。

The 1-kilometer land cover map of the Heihe River Basin is a subset of the 1-kilometer China land cover map (MICLCover) developed by Ran Youhua et al. (2009; 2011) by fusing multi-source local data. The MICLCover land cover map adopts the IGBP land cover classification system, and is constructed based on evidence theory by fusing the 2000 Chinese 1:100,000 land use data, vegetation types from the Vegetation Atlas of China (1:1,000,000), the Chinese 1:100,000 glacier distribution map, the Chinese 1:1,000,000 swamp and wetland map, and the MODIS 2001 land cover product (MOD12Q1).

Validation results of MICLCover indicate that its overall consistency with the Chinese land use map reaches 88.84% at the 7-category level, with the consistency of cropland, urban, wetland and water body types exceeding 95%. Through visual comparison with the MODIS 2001 land cover product and the IGBP DISCover land cover map in three typical regions, MICLCover maintains the overall accuracy of the Chinese land use map, adds information on leaf attributes and leaf types from the Chinese vegetation atlas, and presents more detailed local land cover features. Using national forest inventory data, validation conducted in Gansu Province, Yunnan Province, Zhejiang Province, Heilongjiang Province and Jilin Province shows that the accuracy of forest types in MICLCover has been greatly improved compared with the MODIS land cover product. Validation of the forest types in MICLCover using the first-class forest inventory data from the Administration of Qilian Mountain National Nature Reserve in Gansu Province shows that the accuracy of forest types in this region reaches 82.94%.

In summary, while retaining the overall accuracy of Chinese land use data, the MICLCover land cover map supplements information on vegetation types and vegetation phenology from the Chinese vegetation atlas, and incorporates updated data from the Chinese wetland map and Chinese glacier map, thereby substantially improving the accuracy of Chinese land cover data and enhancing the generalizability of the classification system. This dataset can provide land cover information with higher precision for land surface process models.