Mapping Wetland Habitat Changes Using Remote Sensing in Mai Po and Inner Deep Bay, Hong Kong from 1986 to 2020

Wetlands are productive environments under threat by anthropogenic activities. Existing habitat maps in the Mai Po and Inner Deep Bay area in Hong Kong, a Wetland of International Importance designated under the Ramsar Convention, were produced with varying objectives and methods, hindering the comparison of results over time. This study aimed to map five habitats including (i) mangal (native), (ii) mangal (exotic Sonneratia spp.), (iii) intertidal mudflat, (iv) reedbed marsh, and (v) fishpond in the Ramsar Site and Deep Bay in 1986–2020 using remote sensing technology and a standardised analytical framework. It also aimed to explore the use of airborne LiDAR data to identify water-land boundaries at mean sea level for robust delineation of the open intertidal mudflat. High-resolution aerial photos and satellite images were identified and classified to produce initial maps, followed by rectification based on secondary data sources. For intertidal mudflat, since the LiDAR data could not directly indicate the desired coverage due to the acquisition time, satellite images that satisfied the mean sea level requirement were adopted to delineate the areas, which were validated by field measurements and a LiDAR-based simulation. The mapping results showed expansions of the native mangal, Sonneratia, and reedbed marsh coverages in the study period, while the areas were reduced for intertidal mudflats and fishponds. The spatial patterns and locations of changes were also analysed. The methodological framework and mapping results can improve the knowledge of wetland changes to support scientific conservation management.