Study On Annual Changes Of Intertidal Wetlands In Guangxi Based On Morphological Operation And Intensive Time Series Data

Tidal wetlands are located in the transitional zone between marine and terrestrial ecosystems,and they have ecological system services and values such as storm and wave buffering,maintaining biodiversity and protecting coastal ecosystems.However,tidal wetlands are under serious threats from sea-level rise,coastal erosion and human activities.In China,more than 70% of large and medium-sized cities are located in coastal areas,and intertidal wetlands face serious anthropogenic disturbances,resulting in the loss of habitats for rare species,the decrease of biodiversity,and frequent occurrence of natural disasters such as red tides and insect pests.Accurate and continuous tracking of the area and distribution of intertidal wetland resources and timely grasp of their spatio-temporal dynamic characteristics are essential for the sustainable management,conservation and restoration of intertidal wetland resources.As the environmental background of intertidal wetlands is complex and variable,how to reduce the uncertainty of remote sensing interpretation caused by tidal inundation and obtain the real spatial distribution information of tidal flat,herbaceous and woody wetlands during the lowest tide is a challenge for intertidal wetland remote sensing monitoring research.Previous studies on remote sensing monitoring of intertidal wetlands relied on monitoring data of tide stations,actual training samples or expert experience,which posed difficulties of application in large-scale and long time series studies.In this study,considering tidal fluctuations and vegetation phenological features,we proposed a new automatic classification scheme for intertidal wetlands to track the spatio-temporal pattern dynamic characteristics in Guangxi from 2016 to 2020,which used Sentinel-2(S2)dense time-series images based on the Google Earth Engine(GEE)platform,with a view to providing basic data and technical support for the sustainable management of intertidal wetlands in Guangxi.The main research results and conclusions obtained are as follows:(1)Fine extraction of tidal flats overcoming tide uncertaintyBased on the Otsu algorithm and Maximum spectral index composite(MSIC)algorithm,the maximum water coverage range was delineated using the maximum water coverage image(AWEIsh-MSIC),which to some extent eased the impact of nonwater information from inland areas on the precise characterization of the maximum water surface of the intertidal zone.The 8-neighborhood Laplacian(Laplacian8)operator was used to convolve the highest tide image(m NDWI-MSIC),enhancing the feature differences between water pixels and surrounding artificial shorelines.Combined with morphological opening,the influence of scattered inland water on the extraction of tidal flats was effectively suppressed.The lowest tide image(NDVI-MSIC)highlighted the feature differences between tidal flats,vegetation,and water bodies.(2)Intertidal vegetation classification based on differences in vegetation phenologyThe Harmonic Analysis of Time Series(HANTS)was applied to analyze the NDVI time series changes in mangrove and Spartina interniflora.The feature differences between mangrove and Spartina interniflora were significantly enhanced by calculating the standard deviation of the NDVI time series(NDVI-std Dev),providing key phenological information for intertidal vegetation classification.(3)Remote sensing mapping and accuracy evaluation of intertidal wetlandUsing the GEE cloud platform,a new intertidal wetland classification scheme was proposed by constructing a high-quality dense time-series S2 image stacks and combining morphological operations and MSIC-Otsu algorithms.In this study,the average overall accuracy and Kappa coefficient of the intertidal wetland classification were 94.23% and 0.92,respectively.Accuracy of intertidal wetland classification results at different times was relatively stable.The method is highly robust and has certain application potential in large-scale and long time series intertidal wetland remote sensing mapping work.(4)Analysis of the spatio-temporal dynamic characteristics of intertidal wetlands from 2016 to 2020From 2016 to 2020,the intertidal wetland resources in Guangxi showed a trend of continuous increase,increasing from 44013.78 ha to 57938.47 ha,an increase of3481.17 ha /yr.The intertidal wetland showed a trend of sea expansion,among which the expansion of Beihai City section is the most significant,followed by the Fangchenggang City section,and the smallest increase was observed in the Qinzhou City section.Connectivity and area of tidal flat patches increased significantly,while the dynamics of mangroves were relatively slight.The mangrove area increased slightly in the Beihai City section and lost slightly in the Qinzhou City section.The dynamics of Spartina interniflora were the highest,with rapid expansion and increasing dominance,and the concentration level continued to increase.According to the spatiotemporal pattern dynamic characteristics of intertidal wetland and the differences in the utilization and protection of intertidal wetland at the local level,the feasibility suggestions for the protection of intertidal wetland are proposed in order to realize the sustainable management and utilization of intertidal wetland resources.