Speaker: Xinyue Chang
Abstract: The information of urban dynamics at fine spatiotemporal resolutions is crucial to urban growth modeling and sustainable city development. And knowing when the urbanization starts or ends would be helpful to understand the urbanization process and investigate further environmental impact. Motivated by mapping annual urban dynamics in nationwide locations, we develop a sparse multivariate functional approach of detecting change and estimating change-year in urbanization process by using reflected energy data from land surface. Based on FPCA techniques and CUSUM, the proposed methodology outperforms the existing regression method in both simulation and real data applications.