Comparison Of Arctic Sea Ice Thickness In My Country's CMIP6 Models And Analysis Of Error Source

Arctic sea ice is an important part of the global climate system.The existing observational data show that the extent and thickness of sea ice are decreasing,and one-year ice is dominant.At present,the observed data of sea ice thickness are few and generally have the problem of poor time continuity,which cannot fully reflect the change of sea ice thickness.There are also large uncertainties in the simulation of Arctic sea ice in the global climate system model.In present research,the Arctic sea ice output from climate system models are evaluated to quantify their uncertainties and the analysis is further extended to variables such as radiation and snow depth that can influence the variability of sea ice.Please pay attention to capitalization the sea Ice thickness Assimilation product PIOMAS(Pan-Arctic Ice Ocean Modeling and Assimilation System)developed by the University of Washington was used as reference data.Eight Chinese Earth climate system models from the Coupled Model Intercomparison Project Phase 6(CMIP6)are selected.The Arctic sea ice thickness of these models is compared with the sea ice thickness assimilation product of PIOMAS.The temporal variability and spatial distribution of the Arctic sea ice thickness of the selected model during 1980-2014 is evaluated.The simulation capability of each model was quantified by Taylor score.The results showed that all models and PIOMAS were different no matter in March or September,and the deviation are mainly located in the north of Greenland,near the Barents Sea and the Bering Strait.All models tend to underestimate underestimated the thickness of the central region of Arctic sea ice in March and September,and overestimate the thickness of the marginal region of Arctic sea ice in March.Among them,the model closest to PIOMAS in spatial distribution is FIO-ESM-2-0,followed by FGOALS-f3-L.The long-term trend of sea ice thickness shows that the overall trend of sea ice thickness is declining,and the decreasing rate in September is obviously greater than that in March.The model most similar to PIOMAS in terms of spatial distribution of long term trend is NESM3.The analysis of the interannual trend shows that although the average sea ice thickness fluctuates over the years,the sea ice thickness shows the overall trend of decreasing.Based on the heat flux at the ice-air interface,this paper evaluates other variables that may cause the deviation of sea ice thickness simulation.Due to the large differences between the satellite observations of CERES and the radiation components of the three atmospheric reanalysis data,the radiation component area averages of the CERES and the reanalysis data north of 82°N and the eight CMIP6 models were further compared in this paper.The results show that the upward shortwave radiation of FGOALS-f3-L model in March and CIESM model in September are significantly lower,which may be the reason for the error in model simulation.In addition,there is a certain correlation between the distribution of snow depth and the distribution of sea ice thickness,and snow depth may affect the simulation of sea ice thickness to a certain extent.In conclusion,evaluating the simulation results of sea ice thickness of CMIP6 model and diagnosing and analyzing its errors are of great significance for exploring the variation rule of Arctic sea ice thickness,and also have reference value for the development and improvement of Chinas climate system model.