In mountain areas, an altitudinal vegetation belt ( AVB ) is the most stable vegetation cover according to the climatic conditions as altitude increases, and it represents the climatic climax community. Investigation and I-dentification of AVBs is significant in ecological and geographical studies thanks to extremely complex environment and diverse vegetation types in mountains. Traditionally, field survey only produced discrete data, and AVBs and their vertical combination are mainly expressed as hand-drawn diagrams. This seriously restricted quantitative analysis of AVBs and their spatial patterns. We present a method for predictive vegetation mapping in the Daqing Mountains of central Inner Mongolia of China. Firstly, we obtain the representative vegetation samples using 1: 250 000 DEM, 1: 1 000 000 vegetation map, 1: 100 000 land-use map and related references. According to the relationship of the vegetation samples and the environmental variables affecting them, we simulate the potential distribution of vegetations by using the Bayes algorithm under the software of ERDAS. Tested by the selected 3 000 grids, the whole accuracy of the mapping is 74. 53 % and the Kappa coefficient is 0.69. The research of potential climax vegetation provides a comparative data source, and the comparison of existing and potential altitudinal patterns is also an interesting issue in geographical and ecological studies. The method makes identifying and comparing AVBs considerably easier by using multi-source data, and could potentially be a solid basis for in-depth analysis of AVBs with their environmental factors.%山地植被带的提取及分析是地学和生态学研究的基础问题之一.利用野外点/线调查和历史文献资料,可以对局域尺度山体的植被带进行归纳和描绘,而在区域乃至全球尺度上更多依赖于学者的经验和知识.利用内蒙古大青山地区1∶100万植被图、1∶10万土地利用图、1∶25万DEM等,设计逻辑判别规则,提取典型的山地植被带斑块；然后基于贝叶斯识别算法,利用地形、水热和太阳辐射等因子对区域尺度山地植被带进行空间分布模拟,最终提取的植被带具有较高的精度,总体精度为74.53％,Kappa系数为0.69.研究表明,利用多源数据可以提取和模拟区域尺度山地植被带连续分布格局,中小比例尺空间数据的集成应用使得该方法的推广具有较大的可行性,为进一步获取大陆及全球尺度的山地植被带数据奠定了基础.
展开▼
