Impacts of Topography on the Land Cover Classification in the Qilian Mountains, Northwest China

There are many disagreements and uncertainties among global land use/land cover (LULC) products, which make it unsuitable to apply these products directly to a specific region. In this study, Enhanced Vegetation Index (EVI) time-series data from the Moderate Resolution Imaging Spectroradiometer (MODIS) with 250 m spatial resolution, combining with geographic features, were used for LULC classification in the Qilian Mountains, northwest China. The authors tested the Random Forest (RF), the Classification and Regression Tree (CART), and the Support Vector Machine (SVM) classifiers. Their final classification product was also compared with 3 global LULC maps. The results showed that (1) topographic information could improve the classification results to some extent when they were integrated with spectral information in complex mountain regions; (2) the classification results using an RF classifier reached the highest overall accuracy (88.84%); (3) there was consistency with homogenous classes between our classification result and the 3 LULC products, but inconsistency with heterogeneous classes; and (4) the overall accuracy of this study was improved about 10% compared with the 3 LULC products. Therefore, this classification product is more suitable than global LULC products when considering complex terrain factors in mountain regions.