The Effect And Correction Of Snow Thickness,Sea Ice Density,Snow Density On Ku-Band Synthetic Aperture Radar Altimeter Sea Ice Thickness Retrieval

Sea ice thickness is an important performance factor for climate and environmental changes.Accurate estimation of sea ice thickness is of great significance for climate chang e research and polar region navigation support.The CryoSat-2 and Sentinel-3 A synthetic aperture radar altimeters operating at the Ku-band are the most advanced altimeters until now,which can provide higher resolution and higher accuracy of sea ice thickness information than traditional radar altimeters.At present,the retrieval of arctic sea ice thickness based on radar altimeter data has the following three problems:(1)Ku band has a short wavelength,so it is easily affected by snow layer when detecting sea ice,this factor likely affects the main scattering horizon of radar signal,shifting toward the snow layer from sea-ice surface,which will bring an additional uncertainty for ice freeboard retrieval.(2)The accuracy of sea ice thickness retrieval is greatly affected by snow density,sea ice density and sea water density,if the commonly fixed values of snow density,sea ice density and sea water density are used,some errors may be introduced into the sea ice thickness retrieval results.(3)Currently,the temporal resolution of Arctic sea ice thickness product data are not high enough,only have monthly sea ice thickness products.Aiming at the above three problems,this paper will use CryoSat-2 and Sentinel-3A radar altimeter data to carry out high-precision arctic sea ice thickness retrieval method research,and use Operation IceBridge measured data for verification.(1)In terms of Ku band radar signal penetrate snow depth factor and sea ice freeboard correction,an improved sea-ice freeboard retrieval method which accounts for the radar penetration of snow layer is proposed.Firstly,the optimal waveform retracking thresholds of CryoSat-2 and Sentinel-3 A radar altimeters were determined.Then the effect of snow on sea ice freeboard retrieval was analyzed.Finally,the radar signal penetrate snow depth factor was calculated.For CryoSat-2,the radar penetration factor is 0.95 over the FYI and 0.889 over the MYI;and for Sentinel-3A,the radar penetration factor is 0.873 over the FYI and 0.856 over the MYI.Compared with the classic snow correction method and ESA L2I sea ice products,the sea ice freeboard accuracy calculated by the improved method is the highest,the root mean square error of the sea ice freeboard retrieved by CryoSat-2 data is 7.49 cm,and the root mean square error of the sea ice freeboard retrieved by Sentinel-3 A data is 8.10 cm.(2)In terms of the optimum density values of snow density,sea ice density and sea water density,we proposed to use the least squares fitting method to back-calculate the sea ice,sea water,and snow density ratio combinations suitable for arctic sea ice thickness retrieval.Firstly,based on the sea ice freeboard,snow thickness and ice thickness data,the density ratios a=sea_water_density/sea_water_density-sea_ice_density and b=snow_density/sea_water_density-sea_ice_density are calculated.Then,we analyzed the effect of density ratio on the sea ice thickness retrieval.Finally,we recommend that when retrieving arctic sea ice thickness,for first year ice,useing density ratio a=9.01 and density ratio b=3.52;for multi-year ice,using density ratio a=7.20 and density ratio b=2.30.(3)In order to retrieve more accurate and reliable arctic sea ice thickness,we proposed to use Inverse Distance Weighted method to fuse the sea ice freeboard data which retrieved by CryoSat-2 and Sentinel-3A radar altimeter,and retrieve the arctic sea ice thickness fusion data.By comparing four arctic sea ice thickness products(CPOM,CS2SMOS,AWI,and the sea ice thickness fusion products in this paper)with the measured data,we found that the accuracy of four products is similar.The correlation between CS2SMOS sea ice thickness and Operation IceBridge sea ice thickness is the highest,which is 67.69%,and their absolute value difference is the smallest,which is 47.15 cm.The root mean square error between the sea ice thickness fusion products in this paper and Operation IceBridge sea ice thickness is the lowest,which is 62.22 cm.