BCC_CSM对北极海冰的模拟:CMIP5和CMIP6历史试验比较

本文利用北京气候中心气候系统模式(BCC_CSM)在最近两个耦合模式比较计划(CMIP5和CMIP6)的历史试验模拟结果，对北极海冰范围和冰厚的模拟性能进行了比较，结果表明:(1) CMIP6改善了CMIP5模拟海冰范围季节变化过大的问题，总体上更接近观测结果；(2)两个CMIP试验阶段中BCC_CSM模拟的海冰厚度都偏小，但CMIP6试验对夏季海冰厚度过薄问题有所改进。通过对影响海冰生消过程的冰面和冰底热收支的分析，我们探讨了上述模拟偏差以及CMIP6模拟结果改善的成因。分析表明，8?9月海洋热通量、向下短波辐射和反照率对模拟结果的误差影响较大，CMIP6试验在这些方面有较大改善；而12月至翌年2月，CMIP5模拟的北极海冰范围偏大主要是海洋热通量偏低所导致，CMIP6模拟的海洋热通量较CMIP5大，但北大西洋表层海流的改善才是巴芬湾附近海冰外缘线位置改善的主要原因。CMIP试验模拟的夏季海冰厚度偏薄主要是因为6?8月海洋热通量和冰面热收支都偏大，而CMIP6试验模拟的夏季海冰厚度有所改善主要是由于海洋热通量和净短波辐射的改善。海冰模拟结果的改善与CMIP6海冰模块和大气模块参数化的改进有直接和间接的关系，通过改变短波辐射、冰面反照率和海洋热通量，使BCC_CSM模式对北极海冰的模拟性能也得到有效提高。     

CICE海冰模式中融池参数化方案的比较研究

冰面融池的反照率介于海水和海冰之间,获得较准确的融池覆盖率对认识极区气冰海耦合系统的热量收支有重要意义。在数值模式中,融池覆盖率的模拟结果直接影响到冰面反照率计算的准确性,本文对CICE5.0中的3种融池参数化方案进行了较系统的比较分析,结果显示3种方案各有优缺点,模拟结果都存在一些问题。cesm方案中判断融池冻结的条件更为合理。比较而言,融池冻结条件更改后的topo方案模拟的北冰洋区域平均融池覆盖率的年际变化幅度、融池覆盖范围、融池发展盛期持续时间与MODIS数据最接近。通过修改CICE5.0中的代码漏洞,研究了融池水的垂向渗透效应,这一效应会带来一些负面影响,如lvl方案中多年冰上几乎没有融池,说明目前的CICE模式中对于海冰渗透性演化或其他物理机制的处理仍有待改进。最后,着重讨论了topo方案的改进思路。     

渤海海冰季节演变的数值模拟

MM5数值产品作为大气强迫,利用冰-海洋耦合模式模拟2003～2004冬季渤海海冰演变过程。与海冰遥感、海洋站监测资料等实测资料相比,初冰日和终冰日和观测数据比较接近,模拟海冰各个发展阶段主要分布特征和MODIS遥感图像相似,但模拟的海冰厚度与辽东湾93平台、202平台的冰厚观测数据相比存在较大出入,说明该模型还有待改善。总体上看,冰-海洋耦合模式基本具备模拟渤海海冰季节演变过程的能力。     

渤海潮汐对冰作用的数值模拟

为研究渤海潮对海冰运动的作用,将渤海海冰模式与Ｂｌｕｍｂｅｒｇ的ＥＣＯＭ－ｓｉ模式通过动力作用联结,构成一种海冰一海洋动力耦合模式。应用上述耦合模式对渤海冬季潮流对冰的动力作用进行模拟研究,得出了渤海海冰运动的一些显著特征。     

地球系统模式FIO-ESM对北极海冰的模拟和预估

评估了地球系统模式FIO-ESM(First Institute of Oceanography-Earth System Model)基于CMIP5(Coupled Model Intercomparison Project Phase 5)的历史实验对北极海冰的模拟能力,分析了该模式基于CMIP5未来情景实验在不同典型浓度路径(RCPs,Representative Concentration Pathways)下对北极海冰的预估情况。通过与卫星观测的海冰覆盖范围资料相比,该模式能够很好地模拟出多年平均海冰覆盖范围的季节变化特征,模拟的气候态月平均海冰覆盖范围均在卫星观测值±15%范围以内。FIO-ESM能够较好地模拟1979-2005年期间北极海冰的衰减趋势,模拟衰减速度为每年减少2.24×104 km2,但仍小于观测衰减速度(每年减少4.72×104 km2)。特别值得注意的是:不同于其他模式所预估的海冰一直衰减,FIO-ESM对21世纪北极海冰预估在不同情景下呈现不同的变化趋势,在RCP2.6和RCP4.5情景下,北极海冰总体呈增加趋势,在RCP6情景下,北极海冰基本维持不变,而在RCP8.5情景下,北极海冰呈现继续衰减趋势。     

渤海海冰漂移数值研究

本文利用海冰－海洋动力耦合模式对渤海典型天气形势下海冰动力过程作了数值模拟，海冰模式建立在动量和质量守恒基础上，忽略了海冰变化的热力过程。海冰厚度被划分为三类：堆积冰、平整冰和开阔冰，冰的形变由一个厚度的重新分布约束条件确定。海洋模式是一个二维风暴潮模式．同时考虑了潮汐的作用。风场资料来自于沿岸气象观测站每日四次的风观测，计算网格是十分之一经纬度。主要的分析和模拟是在大气和海洋共同作用下海冰的漂流特点。观测结果比较表明，数值模拟结果基本上反映了该海域流冰的漂移性质，同时也可为短期冰情预报提供有益的参考。     

基于卫星观测的2003-2013年北极海冰体积变化估算

北极海冰正处于快速减退时期，北极海冰体积变化是全球气候变化的重要指示因子。本文利用两种卫星高度计数据（ICESat和CryoSat-2）反演得到的海冰厚度数据，结合星载辐射计提取的海冰密集度数据以及海冰年龄数据，估算了近期的北极海冰体积以及一年冰和多年冰体积变化。CryoSat-2观测时段（2011-2013年）与ICESat观测时段（2003-2008年）相比，北极海冰体积在秋季（10-11月）和冬季（2-3月）分别减少了1 426 km3和412 km3。其中，秋季和冬季的一年冰的体积增加了702 km3和2 975 km3。相反，多年冰分别减少了2 108 km3和3 206 km3。多年冰的大量流失是造成北极海冰净储量下降的主要原因。     

一种冰-海洋模式的热力耦合方案

冰与海洋的热力耦合对冰与海洋环流的模拟有极其重要的影响,是冰-海洋相互作用的一个重要方面.对其精确确定需要详细考虑冰-海洋界面附近的湍流过程,这在长时期的模拟特别是气候模拟中,常受到技术条件的限制.过去的研究常常假设冰下海洋混合层的温度为冰点,特别是在单纯冰模式的模拟中,但考虑海冰漂移和冰点变化的效应时,这一假设是不精确的.因此,弱化冰下海洋混合层温度为冰点的约束,不考虑详细的冰-海洋界面和海洋混合层的湍流过程,根据冰-海洋耦合系统的能量收支关系,设计了一个简化的冰-海洋热力耦合方案.对该方案引起的海洋混合层适应、热力结构和海冰发展的影响进行了分析,并将其用于全球冰海洋耦合模式的数值试验,结果表明,在大气热力强迫下该耦合方案既可使冰区混合层海水温度向冰点适应,又使冰边缘带海水温度与冰点保持明显差异,能够较好地反映冰-海洋热力相互作用.利用该耦合方案构造的全球冰-海洋耦合模式模拟的海冰范围及季节变化与实际观测非常接近.     

两种盐度参数化方案在海冰模式中的应用比较

本文详细介绍了SIS海冰模式中引进两种盐度参数化方案即等盐度方案和盐度廓线方案对海冰模拟所存在的差异。利用盐度廓线方案导出的表征盐度与海冰温度间关系的方程比等盐度方案多出一项,将定义为盐度差异项。盐度差异项对海冰厚度的热力作用表现为：在海冰厚度增长季节（11月到次年5月）,盐度差异项通过升高海冰内部温度,抑制海冰增长;在消融的第一阶段（6．8月）,盐度差异项通过升高海冰内部温度加快海冰消融;在消融的第二阶段（9．10月）,盐度差异项通过降低海冰内部的温度抑制海冰消融。但尺度分析表明,盐度差异项要比方程中队海冰温度作用最大项小1．2个量级,如果采用一级近似,可以略去盐度差异项,因此盐度差异项对海冰增长和消融影响很小。同时利用冰洋耦合模式（ModularOceanModel,MOM4）,分别采用两种盐度参数化方案模拟北极海冰厚度和海冰密集度的季节性变化,模拟结果也表明两种方案模拟得到的海冰厚度和海冰密集度的季节性变化相差甚小。     

渤海冰-海洋耦合模式Ⅱ.个例试验

以渤海1998~1999和2000~2001年度冬季的海冰发展过程为例,采用实时气象数值预报场作为大气强迫,利用渤海冰海洋耦合模式模拟渤海整个冬季的海冰生消和演变以及渤海冰季冰气、冰水和气水界面的热收支.模拟结果显示大气和海洋的热力效应对渤海的海冰发展非常重要,特别在海冰的冻结和融化阶段,海洋热通量在热力耦合中起着重要作用.模拟还显示了冰覆盖内部区域和冰外缘线附近不同的热力特征,分析讨论了冰区内和冰边缘两个特征点冰厚分布、界面热量收支和海表水温等.     

一个区域海洋——海冰——大气耦合模式中的北极气候偏差：年结果检验

The Coupling of three model components, WRF/PCE （polar climate extension version of weather research and forecasting model （WRF））, ROMS （regional ocean modeling system）, and CICE （community ice code）, has been implemented, and the regional atmosphere-ocean-sea ice coupled model named WRF/PCE- ROMS-CICE has been validated against ERA-interim reanalysis data sets for 1989. To better understand the reasons that generate model biases, the WRF/PCE-ROMS-CICE results were compared with those of its components, the WRF/PCE and the ROMS-CICE. There are cold biases in surface air temperature （SAT） over the Arctic Ocean, which contribute to the sea ice concentration （SIC） and sea surface temperature （SST） biases in the results of the WRF/PCE-ROMS-CICE. The cold SAT biases also appear in results of the atmo- spheric component with a mild temperature in winter and similar temperature in summer. Compared to results from the WRF/PCE, due to influences of different distributions of the SIC and the SST and inclusion of interactions of air-sea-sea ice in the WRF/PCE-ROMS-CICE, the simulated SAT has new features. These influences also lead to apparent differences at higher levels of the atmosphere, which can be thought as responses to biases in the SST and sea ice extent. There are similar atmospheric responses in feature of distribution to sea ice biases at 700 and 500 hPa, and the strength of responses weakens when the pressure decreases in January. The atmospheric responses in July reach up to 200 hPa. There are surplus sea ice ex- tents in the Greenland Sea, the Barents Sea, the Davis Strait and the Chukchi Sea in winter and in the Beau- fort Sea, the Chukchi Sea, the East Siberian Sea and the Laptev Sea in summer in the ROMS-CICE. These differences in the SIC distribution can all be explained by those in the SST distributions. These features in the simulated SST and SIC from ROMS-CICE also appear in the WRF/PCE-ROMS-CICE. It is shown that the performance of the WRF/PCE-ROMS-CICE is determined to a l     

A global coupled sea ice–ocean model

A new sea ice model, GELATO, was developed at Centre National de Recherches Météorologiques (CNRM) and coupled with OPA global ocean model. The sea ice model includes elastic–viscous–plastic rheology, redistribution of ice floes of different thicknesses, and it also takes into account leads, snow cover and snow ice formation. Climatologies of atmospheric surface parameters are used to perform a 20-year global ocean–sea ice simulation, in order to compute surface heat fluxes from diagnosed sea ice or ocean surface temperature. A surface salinity restoring term is applied only to ocean grid cells with no sea ice to avoid significant surface salinity drifts, but no correction of sea surface temperature is introduced. In the Arctic the use of an ocean model substantially improves the representation of sea ice, and particularly of the ice edge in all seasons, as advection of heat and salt can be more accurately accounted for than in the case of, for example, a sea ice–ocean mixed layer model. In contrast, in the Antarctic, a region where ocean convective processes bear a much stronger influence in shaping sea ice characteristics, a better representation of convection and probably of sea ice (for example, of frazil sea ice, brine rejection) would be needed to improve the simulation of the annual cycle of the sea ice cover. The effect of the inclusion of several ice categories in the sea ice model is assessed by running a sensitivity experiment in which only one category of sea ice is considered, along with leads. In the Arctic, such an experiment clearly shows that a multicategory sea ice model better captures the position of the sea ice edge and yields much more realistic sea ice concentrations in most of the region, which is in agreement with results from Bitz et al. [J. Geophys. Res. 106 (C2) (2001) 2441–2463].     

Sea ice volume and age: Sensitivity to physical parameterizations and thickness resolution in the CICE sea ice model

New dynamics parameterizations in Version 5 of the Los Alamos Sea Ice Model, CICE, feature an anisotropic rheology and variable drag coefficients. This study investigates their effect on Arctic sea ice volume and age simulations, along with the effects of several pre-existing model options: a parameter that represents the mean cumulative area of ice participating in ridging, the resolution of the ice thickness distribution, and the resolution of the vertical temperature and salinity profiles.By increasing shear stress between floes, the anisotropic rheology slows the ice motion, producing a thicker, older ice pack. The inclusion of variable drag coefficients, which depend on modeled roughness elements such as deformed ice and melt pond edges, leads to thinner ice and a more realistic simulation of sea ice age. Several feedback processes act to enhance differences among the runs. Notably, if less open water is produced mechanically through ice deformational processes, the simulated ice thins relative to runs with more mechanically produced open water. Thermodynamic processes can have opposing effects on ice age and volume; for instance, growth of new ice increases the volume while decreasing the age of the pack. Therefore, age data provides additional information useful for differentiating among process parameterization effects and sensitivities to other model parameters.Resolution of thicker ice types is crucial for proper modeling of sea ice volume, because the volume of ice in the thicker ice categories determines the total ice volume. Model thickness categories tend to focus resolution for thinner ice; this paper demonstrates that 5 ice thickness categories are not enough to accurately resolve the ice thickness distribution for simulations of ice volume.     

基于Icepack海冰柱模式的融池反照率模拟研究

基于Icepack一维海冰柱模式,以2014年中国第6次北极科学考察长期冰站ICE06的3个融池的辐射参量和气象参量的连续观测作为大气强迫数据,对融池反照率及相关参量进行了模拟。本文引入观测的融池深度及海冰厚度作为初始条件,通过考虑融池覆盖率的作用,改进了平整冰融池参数化方案中海冰干舷的计算,修正了冰上可允许的最大融池深度,成功实现了对融池参数变化的模拟;同时,还修正了入射辐射分量比例系数与对应反照率分量权重系数不一致的问题。标准试验中,模拟的3个融池的反照率与观测结果之间的平均误差分别为0.01、0.05和0.13;入射辐射比例的敏感性试验结果表明,当可见光辐射比例增大8%时,融池反照率的模拟结果增大了6%～8%;融池表面再冻结试验的结果显示,当再冻结冰层厚度小于2 cm时,模拟冰面反照率的增加不足0.006,由此引起的表面能量收支减少了约1.1 W/m2。本文研究指出,准确的入射辐射比例对于改善北极海冰反照率模拟是必要的;并指出目前模式仍存在融池表面再冻结参数化、热收支计算、表面吹雪效应等有待解决的问题。     

A Coupled Ice-Ocean Model in the Pan-Arctic and North Atlantic Ocean: Simulation of Seasonal Cycles

A coupled ice-ocean model is configured for the pan-Arctic and northern North Atlantic Ocean with a 27.5 km resolution. The model is driven by the daily atmospheric climatology averaged from the 40-year NCEP reanalysis (1958–1997). The ocean model is the Princeton Ocean Model (POM), while the sea ice model is based on a full thermodynamical and dynamical model with plastic-viscous rheology. A sea ice model with multiple categories of thickness is utilized. A systematic model-data comparison was conducted. This model reasonably reproduces seasonal cycles of both the sea ice and the ocean. Climatological sea ice areas derived from historical data are used to validate the ice model performance. The simulated sea ice cover reaches a maximum of 14 × 10⁶ km² in winter and a minimum of 6.7 × 10⁶ km² in summer. This is close to the 95-year climatology with a maximum of 13.3 × 10⁶ km² in winter and a minimum of 7 × 10⁶ km² in summer. The simulated general circulation in the Arctic Ocean, the GIN (Greenland, Iceland, and Norwegian) seas, and northern North Atlantic Ocean are qualitatively consistent with historical mapping. It is found that the low winter salinity or freshwater in the Canada Basin tends to converge due to the strong anticyclonic atmospheric circulation that drives the anticyclonic ocean surface current, while low summer salinity or freshwater tends to spread inside the Arctic and exports out of the Arctic due to the relaxing wind field. It is also found that the warm, saline Atlantic Water has little seasonal variation, based on both simulation and observations. Seasonal cycles of temperature and salinity at several representative locations reveals regional features that characterize different water mass properties.     

A coupled ice-ocean model for the Bohai Sea Ⅱ. Case study

The coupled ice-ocean model for the Bohai Sea is used for simulating the freezing, melting, and variation of ice cover and the heat balance at the sea-ice, air-ice, and air-sea interfaces of the Bohai Sea during the entire winter in 1998～1999 and 2000～2001. The coupled model is forced by real time numerical weather prediction fields. The results show that the thermodynamic effects of atmosphere and ocean are very important for the evolvement of ice in the Bohai Sea, especially in the period of ice freezing and melting. Ocean heat flux plays a key role in the thermodynamic coupling. The simulation also presents the different thermodynamic features in the ice covered region and the marginal ice zone. Ice thickness, heat budget at the interface, and surface sea temperature, etc. between the two representative points are discussed.     

A comparison of two global ocean-ice coupled models with different horizontal resolutions

A global eddy-permitting ocean-ice coupled model with a horizontal resolution of 0.25 by 0.25 is established on the basis of Modular Ocean Model version 4 (MOM4) and Sea Ice Simulator (SIS). Simulation results are compared with those of an intermediate resolution ocean-ice coupled model with a horizontal resolution of about 1 by 1 . The results show that the simulated ocean temperature, ocean current and sea ice concentration from the eddy-permitting model are better than those from the intermediate resolution model. However, both the two models have the common problem of ocean general circulation models (OGCMs) that the majority of the simulated summer sea surface temperature (SST) is too warm while the majority of the simulated subsurface summer temperature is too cold. Further numerical experiments show that this problem can be alleviated by incorporating the non-breaking surface wave-induced vertical mixing into the vertical mixing scheme for both eddy-permitting and intermediate resolution models.