Investigation of estuarine mangrove ecosystem changes using unmanned aerial vehicle images: Case study in Xuan Thuy National Park (Vietnam)

This paper introduces an optimized approach for the integration of digital surface models (DSM), orthophotos, and point clouds constructed by unmanned aerial vehicle (UAV) photogrammetry for mapping topography and canopy height of mangrove forests. Object-oriented classification has been applied to orthophoto to accurately determine open mudflats. Elevation points of the point cloud that fall within open mudflats are used to linearly interpolate a digital terrain model (DTM) of mangroves, which is generally unavailable due to the great challenge for both remote sensing and field measurement methods. A canopy height model (CHM) of mangroves is generated by subtracting the DTM from the DSM. Consequently, the constructed CHM has higher accuracy compared to previous studies using only one fixed ground elevation or the water mean level to represent the irregular topography of mangrove forests. The optimized approach is applied to UAV images acquired in 2014 and 2020 to investigate topographical changes and vertical growth of mangroves in a 4 km 2 estuarine area of Xuan Thuy national park in Vietnam. Results show that the study area is vertically eroded with the rate of -0.015 m year −1 due to the groundwater flows, while the vertical growths of the two dominant mangrove species Kandelia obovata and the Sonneratia caseolaris are 0.242 m year −1 and 0.565 m year −1, respectively. A very weak correlation between topographical changes and vertical growth (R = -0.286) is the empirical evidence for the insignificant influence of erosion and/or deposition on the mangrove growth. This study used sole UAV-SfM method to provide comprehensive information on both the topography and canopy structure of fragmented mangrove forests. A combination of LiDAR data and UAV photogrammetry is recommended for dense mangrove inventory.