基于CART决策树的双尺度流域单元地貌分类研究——以北回归线(云南段)地区为例

针对当前单一地貌划分单元造成的分类结果破碎或漏分问题,该文引入双尺度流域单元划分方法,即采用两种不同大小流域单元的组合作为地貌划分基本单元,以提高地貌划分的细分性和完整性。以30 m ASTER GDEM数据为数据源,基于最佳地形因子组合(高程、地势起伏度、坡度、坡度变率、光照模拟值)、双尺度流域单元、CART决策树算法,实现了北回归线(云南段)地区平原(2类)和山地(7类)共9类地貌的划分,双尺度流域单元划分的最佳流量阈值分别为500、2000。通过平均值、标准差、Moran′s I和人工判读结果对分类结果进行检验,发现基于CART决策树的双尺度流域单元地貌分类方法在北回归线(云南段)地区总体精度可达82.1%,Kappa系数为0.793,总体能够准确识别出研究区的地貌类型空间分布特征,是地貌类型划分的一种可行方法。

Study on Landform Classification of Dual-Scale Watershed Units Based on CART:A Case Study of the Tropic of Cancer (Yunnan Section)

In view of the problem that the classification results are broken or missing due to the landform classification based on a single geomorphic division unit in the current research,we construct a classification method of dual-scale watershed units to improve the subdivision and completeness of landform classification.Based on ASTER GDEM data(30 m resolution),using the best combination of terrain factors,dual-scale watershed unit classification method and CART algorithm,the landforms in the county area along the Tropic of Cancer(Yunnan section)are classified into nine geomorphic types,including plains(2 types)and mountains(7 types).The mean value,standard deviation,Moran′s I and manual interpretation results are used to test the classification results.The results are as follows:1)The best combination of terrain factors in the study area are elevation,relief amplitude,slope,slope variability and hill shade.2)The optimum flow thresholds for dual-scale watershed unit classification in the study area are 500 and 2000,respectively.3)It is found that the overall accuracy of the study area can reach 82.1%,with Kappa coefficient of 0.793.In summary,the proposed method can accurately identify the spatial distribution characteristics of geomorphic types in the study area,which is a feasible method for landform classification.