Zhao Jing-zhu, Sun Yi, Bai Guang-xin, Wu Gang, Shao Guo-fan
Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China.
J Environ Sci (China). 2003 Jul;15(4):520-4.
Vegetation or land cover maps have been made directly or indirectly available for the entire territory of China. The certainties and uncertainties of land cover statistics were analyzed by comparing three data sources: China's Vegetation Map, IGPB DISCover, and University of Maryland Product. Great similarities in the statistics of 7 aggregated land cover types were found among the three data sources,particularly between the two global land cover maps. The per-pixel agreement between any 2 of 3 maps was between 38.0%-51.4%; the per-pixel agreement among all three maps was only 27.1%. Certainties were found in regions where vegetation types are typical and human land use practice is relatively homogenous; the uncertainties occurred to either vegetation transition zones or regions where land cover types and land use practice are relatively diversified. Systematic and multidisciplinary efforts are necessary to promote accurate mapping of nationwide land cover types in China. Intensive efforts should be made in regions where uncertainties of land cover information are found.
中国全境已直接或间接地获取了植被或土地覆盖图。通过比较三个数据源(中国植被图、国际地圈-生物圈计划(IGBP)的全球土地覆盖数据产品DISCover和美国马里兰大学的土地覆盖数据产品),分析了土地覆盖统计数据的确定性和不确定性。在7种汇总的土地覆盖类型统计数据中,发现这三个数据源之间存在很大的相似性,特别是在两张全球土地覆盖图之间。三张图中任意两张之间的逐像素一致性在38.0%-51.4%之间;三张图之间的逐像素一致性仅为27.1%。在植被类型典型且人类土地利用方式相对单一的地区,数据具有确定性;而在植被过渡带或土地覆盖类型和土地利用方式相对多样的地区,则存在不确定性。需要系统的多学科努力,以促进中国全国土地覆盖类型的准确制图。应在发现土地覆盖信息存在不确定性的地区加大力度。