Combining multisource satellite data to estimate lake basin bathymetry for seasonal water bodies
Seasonal water bodies are an important component of global surface water and play an indispensable role in regional flood storage and local biodiversity maintenance.Obtaining high-precision bathymetric information is key to supporting effectively the estimation of water storage and carbon flux in seasonal water bodies,which is helpful in understanding the regional hydrological processes and material-energy balance and other issues.On the one hand,only a few existing studies have focused on seasonal water bodies because of their complex subsurface conditions,particularly small water bodies,which make it even harder to evaluate bathymetry.On the other hand,using traditional bathymetric techniques may encounter great difficulties,where a single sensor cannot balance cost,efficiency,and accuracy.To this end,this study proposes a quantitative estimation method of underwater topography for seasonal water bodies,combined with active LiDAR and passive optical sensor data.The LiDAR data are obtained from ICESat-2/ATLAS global geolocated photon data(ATL03)product,which provides high-precision photons'vertical profile information of the lake basin.Meanwhile,optical sensor data can be derived from Sentinel-2 MSI datasets based on the Google Earth Engine(GEE)cloud platform,where massive commonly used datasets can be accessed,and then the regional Inundation Frequency(IF)distribution is generated.Each photon along ICESat-2 ground tracks is time tagged and geolocated;thus,each photon's height(Hgt)and IF value can be obtained by geographical intersection.In the same lake basin,every point's height and inundation frequency are correlated in theory;thus,we can build an"Hgt-IF"model to fit this correlation.Then,this model is applied so that regional IF distribution can be translated into lake floor elevation over the lake basin.As a typical seasonal lake composed of some dished lakes,Poyang Lake is taken as the research object in this study,and systematical evaluation is performed to evaluate the estimation accuracy and applicability of the method.Results show that the quantitative estimation method is feasible based on the photon elevation of ICESat-2/ATLAS profiles and the inundation frequency information obtained from Sentinel-2 MSI to achieve the"point-to-surface"topography ofseasonal water bodies.As for most dished lakes selected,the R2 values between the predicted and measured yield are greater than 0.7,and the root mean square errors are controlled within 1.0 m.However,the simulation accuracy of dished lakes in different areas also varies due to the combined effects of various factors,such as lake area,subsurface conditions,inundation frequency range,and photon track distribution.In summary,the proposed method can realize the quantitative estimation of underwater topography for seasonal water bodies in general terms.Combining active and passive remote sensing data can make up for the shortcomings of a single sensor,especially when it comes to a large-scale,low-cost,and long time-series situation.This method is also expected to provide ideas and directions for the development of bathymetric retrieval models for seasonal water bodies at the global scale.
seasonal water bodieslake bathymetrysatellite-based bathymetryICESat-2Sentinel-2inundation frequency
万方数据
维普
