Research Of Shallow Bathymetry And Coastal Zone Classification Using ICESat-2 And Sentinel-2

Coastal zones and reefs are sensitive and significant areas for global climate change response.The study of coastal zone changes can help assess the response of climate change hydrological processes.The geographic conditions of coastal zones and reefs are too complex to obtain the ground observation data.Remote sensing technology based on satellite-based platform has unique advantages in monitoring shoreline changes and inversion of shallow sea water depth in a large scale.The new satellite ICESat-2(The Ice,Cloud,and Land Elevation Satellite-2),which uses a 532 nm bluegreen laser beam with a 12 m spot diameter and 0.7m sampling interval in along-track can acquire detailed information of shore zone.The specific research contents and results of the paper are as follows:The large-scale and high-precision seafloor topography is the basis of scientific research on the shape and structure of the earth,as well as important guarantee for marine activities such as marine transportation,fishery,mineral resources development and military struggle.The traditional bathymetric survey is carried out by sonar equipment.Although high-precision depth data can be obtained,it is inefficient.With the maturity of altimetry satellite technology,many sea surface height data can be obtained,which provides an important data source for the study of marine gravity anomalies.According to the isostatic theory,there is a close relationship between sea surface gravity anomaly and seafloor topography,so we can derive a large range of seafloor topography by altimetry technology.In this paper,the inversion of gravity anomalies by altimetry data and the inversion of seafloor topography by gravity anomalies are studied.The main work and innovations are summarized as follows:(1)An accurate refraction correction method was proposed.ICESat-2 does not take into account the refraction occurring at the air-water interface and the corresponding change in the speed of light,resulting in the underwater signal photons geolocation bias and measurements deeper than true.The refraction correction model proposed in this paper integrates the effects of laser beam refraction,sea surface undulation,and sea surface slope variation to obtain more accurate photon coordinates.The final experiment shows that the proposed refraction correction method can effectively correct the bathymetric error,and the error in ICESat-2 direct bathymetry is reduced from the original 2.1m to 0.7m.(2)The regional bathymetry maps were obtained by the inversion of Satellite Derived Bathymetry(SDB)method.The corrected ICESat-2 photons were combined with the pre-processed Sentinel-2 data,and two global bathymetry maps of Yongle Atoll region in South China Sea and Great Barrier Reef region in Australia were obtained by using the Band-Ratio Model inversion.The final experimental results showed that the satellite remote sensing bathymetry inversion was a feasible method to obtain the regional bathymetry,and the median error of bathymetry was 1.5m for the Yongle Atoll region and 1.2m for the Great Barrier Reef region.(3)A method combining ICESat-2 and Sentinel-2 to classify coastal zones was proposed.First,the time-serial images of Sentinel-2 were synthesized using GEE(Google Earth Engine),and different features were calculated for this synthesized image.Second,statistical indicators reflecting the contour characteristics of the coastal zone were extracted from ICESat-2.Finally,a random forest algorithm was used to measure the feature importance and perform the coastal zone classification.The comparison with the measured data showed that the proposed method can effectively classify the coastal zone: the accuracy is 83.61% and the kappa coefficient is 0.81.