Mangrove monitoring and change analysis with landsat images: A case study in pearl river estuary (china)

Abstract

Under the synergistic influence of global climate change and rapid urbanization, the conservation of mangrove wetlands is facing great challenges. Accordingly, mangrove wetland management and protection requires investigating the covered area, distribution, and dynamic changes of mangrove wetlands during different periods. In this study, we detected changes of mangroves and analyzed their driving-factors in Pearl River Estuary (Shenzhen Bay and Qi’ao Island) from Landsat images and field work during the period of 1990–2020 using long short-term memory neural network (LSTM) method. The combination of spectral information and texture information yielded high-precision mangrove interpretation results when extracting mangrove areas from remote sensing data, and the LSTM method has the better interpretation effect compared with the random forest (RF) method. Furthermore, based on the interpretation results for the mangroves, this study analyzed dynamic changes in mangrove area and its distribution in Shenzhen Bay and Qi’ao Island from 1990 to 2020 and showed that the overall mangrove area displayed an increasing trend in the past 30 years, with increasing landscape dominance and strong cohesion, mainly due to the ecological restoration projects, and the rapid growth of mangrove area from 2010 to 2015 may be related to rising temperature. However, the cumulative effect of natural factors (precipitation and sea level rise) may cause the growth rate of mangrove area to tend to decrease.

Keywords

• Mangrove
• Landsat images
• Extraction
• Change monitoring
• Driving factors