Sea Ice Thickness Simulation And Shortwave Radiation Parameter Sensitivity Analysis Based On ICEPAC

The sea ice column model ICEPACK is a one-dimensional submodule of the sea ice community model CICE(Community Ice Cod E),which is commonly used in climate models.ICEPACK has many empirical parameters,which bring great uncertainty to the simulation and require adjustment or optimization.The present research uses atmospheric observations provided by the Multidisciplinary drifting Observatory for the Study of Arctic Climate(MOSAiC)project to force ICEPACK,aiming to address the following scientific questiosns: 1.Can the ICEPACK model simulate the observed sea ice thickness and its changes based on the MOSAiC atmospheric forcing fields? 2.What are the possible causes of ICEPACK simulation error? 3.Which shortwave radiation parameters are more sensitive in the ICEPACK sea ice albedo simulation?The present research conduct simulations during winter/spring and summer using ICEPACK and the complete atmospheric forcing fields from MOSAiC project.Then we compared the sea ice thickness from the model with the sea ice thickness data from buoy2019T66.We further analyzed the possible causes of errors in ICEPACK simulation.The results show that the simulation error of sea ice albedo plays a major role.Then we developed an adjoint model for the Delta-Eddington shortwave radiation parameterization scheme.The adjoint model is used to analyze the sensitivity of albedo to some parameters in the radiation scheme.The main conclusions are as follows:(1)ICEPACK can simulate the sea ice thickness and its increasing trend in winter and the next year’s spring.The main reason of the error in sea ice thickness is due to excessive accumulation of snow-sea ice conversion in ICEPACK.The sea ice albedo simulated by ICEPACK is bigger than that of MOSAiC.(2)ICEPACK overestimated the surface melting during summer.The sea ice albedo simulated by ICEPACK is also smaller than that of MOSAiC.(3)The albedo of snow is affected by the tuning parameter for snow(R＿snw),the temperature change accompanying snow grain radius change(d T＿mlt),and the maximum melting snow grain radius(rsnw＿mlt).In spring and autumn,the albedo is most sensitive to R＿snw.In the summer melting period,the albedo is most sensitive to rsnw＿mlt.(4)The sensitivity of total albedo to sea ice albedo tuning parameter(R＿ice)and to pond albedo tuning parameter(R＿pnd)is analyzed using both forward integration method and adjoint model method.The results from the two methods are consistent.During summer time,the total surface albedo is more sensitive to sea ice albedo tuning parameter.In future work,the adjoint model built in present research can be used in the optimization of these parameters.