Monitoring long-term annual urban expansion (1986–2017) in the largest archipelago of China

Abstract

Abstract Islands that support numerous biodiversity are subject to increasing anthropogenic disturbance with the ever-growing coastal urbanization especially in developing countries. It is essential to monitor the island urban expansion to support sustainable policy making for ecological conservation and environmental management. However, current methods developed for mainland have limitations in capturing the annual urban expansion from land reclamation on islands. Besides, differences of urban expansion among various island development types remain unclear. This study developed an efficient framework on the Google Earth Engine platform for mapping annual urban dynamics in island regions using long-term time series of Landsat, by integrating coastal dynamic mapping approach, random forest classifier, and time-series change detection method. We implemented the developed framework in Zhoushan Archipelago, the largest archipelago in China that contains different island development types. The mapped urban areas and their conversion sources were reliable with overall accuracies over 90%. The overall accuracy of urbanized years was 86% using the one-year tolerance strategy. The total urban area expanded from 97 ± 24 km2 to 438 ± 34 km2 during 1986–2017, at the cost of 148 ± 24 km2 agricultural land, 138 ± 14 km2 water body, 41 ± 13 km2 forest and 14 ± 8 km2 tidal flat. The urban growth accelerated since 2004 driven by a series of government policies, as well as the growth of the population and socio-economy. Moreover, most urban expansion was concentrated in islands with comprehensive development type (65%), followed by the islands with harbor and logistics (15%), coastal tourism (10%), coastal industry (8%), scientific fishery (1%) and marine science and education (1%). The speed and scale of future urban expansion will play an important role for island sustainability. The proposed framework is transferable in other regions for a better understanding of the long-term island urban dynamics at large scales.