Study On The Comprehensive Application Of High-resolution Satellite Observation And Dynamic Simulation

The estuarine and coastal zone is a dynamic geomorphological unit of the land-sea interface.It is subject to a variety of factors and is undergoing rapid dynamic changes.On the one hand,the tidal river at the estuarine and coastal zone is affected by factors such as tides,and water quality factors change rapidly;On the other hand,the muddy wetland topography is also undergoing rapid changes due to the inflow of water and sand from the river upper reaches,changes in sea level under global climate change,and human activities,which have an important impact on the wetland topography evolution.In order to improve the accuracy of dynamic monitoring,this research combines high-resolution satellite observations with hydrodynamic simulation to promote the monitoring of tidal river water quality and tidal flat topography in the estuary and coastal zone.Satellite images with high spatial resolution can observe small-scale tidal rivers.However,due to the influence of light,weather,and revisit frequency,they are faced with limited continuous distribution of data in time series and cannot directly monitor some water quality parameters of non-photosensitive components.Aiming at the problem of dynamic monitoring of water quality,this study combines the advantages of satellite observation time-space continuous coverage and high time resolution of hydrodynamic water quality model.We analyze the water quality change of Huangpu River tidal reach,and solves the shortage of remote sensing observation of the rapid change of water quality in:(1)Based on the measured spectrum and synchronized suspended particulate matter(SPM)and chlorophyll-a concentration parameters.We used the corrected Landsat8/OLI and Sentinel-2/MSI data based on ACOLITE method to model the SPM and chlorophyll-a concentration.The results were verified by K-fold cross-validation.The fitting accuracy R² of the Landsat8/OLI model are 0.76 and 0.68,respectively,and the Sentinel-2/MSI model are 0.77 and 0.63 respectively.(2)Through the Delft 3D hydrodynamic and water quality model,the dissolved oxygen(DO),ammonia nitrogen(NH₃-N)and potassium permanganate index(COD_Mn)of the Huangpu River were simulated.And it is verified by the measured data of Yangpu water plant.The results show that the mean square root error of simulated and measured water level is 0.22 m;the verified root mean square errors of the water quality parameters DO,NH₃-N,and COD_Mn are respectively 0.53 mg/L,0.16 mg/L,0.27 mg/L.Further research and analysis found that the Huangpu River chlorophyll-a concentration from August 2013 to July 2014 showed a good positive correlation with DO and NH₃-N,and a good negative correlation with COD_Mn.(3)Based on the correlation between the chlorophyll-a concentration and water quality parameters,the Landsat8/OLI satellite was used to retrieve the DO,NH₃-N,and COD_Mn of the Huangpu River.It was found that for the tidal reach,the water quality parameters vary greatly with the seasons.The combination of satellite observations and dynamic models is of great significance to the tidal river water quality monitoring with high dynamic changes.Faced with the rapid and dynamic changes of wetland topography,remote sensing technology cannot achieve high-temporal-spatial monitoring.High spatial resolution satellite images usually have low revisit frequency and are greatly affected by cloud cover.The impact is large,so it is impossible to provide images with high time resolution images.And satellite images with high revisit frequency usually have lower spatial resolution.In this study,through the high-temporal-spatial resolution satellite data fusion technology,and combined with the dynamic model to obtain the tidal level at the moment of image transit,the remote sensing dynamic monitoring of the wetland topography is realized.Through the comprehensive application of high-resolution satellite observation and dynamic simulation,the difficulty of dynamic monitoring of wetland topography is solved,and a new method for dynamic monitoring of muddy wetland topography is established:(1)In this study,by useing the STARFM spatio-temporal data fusion method,GF-1/WFV satellite data(16 meters spatial resolution,4 days revisit frequency)and geostationary satellite GOCI data(500 meters spatial resolution,1 hour revisit frequency)are fused to generate hour-by-hour images with a spatial resolution of 16meters.(2)Using the BP neural network supervised classification method,a series of waterlines are extracted from the time series satellite images,combined with the tidal level information simulated by the dynamic model.The daily digital terrain elevation(DTM)of the intertidal wetland can be constructed.The method was applied to the Chongming North Beach of the Yangtze Estuary.(3)Validated by DTM obtained by UAV LiDAR data during the same period,the root mean square error of the daily DTM obtained by this method is 0.16 m,and the root mean square error of the DTM constructed by the unfused satellite imagery spanning one year is 0.27 m,which shows that the time resolution does have a certain impact on the DTM accuracy.The experiment of extracting the waterlines at the spatial resolution of 500 m and 16 m respectively shows that the error between the two in the tidal flat slope of 0.15°is about 0.35 m,which shows that the spatial resolution has an impact on the DTM accuracy.The application in Chongming Dongtan shows that the method proposed in this study is an effective method to monitor the wetland topography dynamic changes.