Form Submission: Participation Entry

How can big data from remote sensing generate new insights on urban land use? In this project, we apply CCDC (Continuous Change Detection and Classification), a novel remote sensing algorithm that can analyze a large panel dataset of Landsat images consisting of more than 122 million pixels over a thirty-year period. To identify the pattern of the land use trajectories, this study calculates within-cluster variabilities of 1% of the trajectories and then applies a clustering algorithm to all urban change trajectories. The results include the information on when urban land use changes occur as well as special signatures of land change classes. We find that cropland is the primary land source of urban settlements in Nepal, and there is an oscillating urbanization cycle with intensive land use change. The algorithm provides a unique ability to extract land use patterns which cannot be discovered by conventional remote sensing methods.