北极冰海耦合模式对两种不同大气再分析资料响应的分析

本文中我们比较了Climate Forecast System Reanalysis(CFSR)高分辨率的再分析数据集和低分辨率的Japanese 25-year Reanalysis Project(JRA25)再分析数据集在向下短波辐射、向下长波辐射、10m风场、近地面气温、降水、湿度上的不同,发现二者差异最大的为降水数据,其次为向下短波辐射数据、向下长波辐射数据。用这两个数据集驱动同一冰海耦合模式,CFSR强迫的海冰、北冰洋中层水和加拿大海盆温盐结构与实测相比有很大差距,等密度面上的地转流速在加拿大海盆和欧亚海盆比JRA25强迫的结果高20%,同时等密度面的深度偏深、位温偏高,在弗拉姆海峡的流通量也比海洋再分析数据Simple Ocean Data Assimilation(SODA)偏多。CFSR的向下辐射数据更加接近实测,采用此数据的敏感性实验模拟结果与实测符合的更好。对于海冰的模拟,云量起着至关重要的作用,降水带来的淡水通量通过影响大西洋入流水携带的热量进而影响到冰区。此外,CFSR过量的降水也是二者对于北冰洋温盐结构、弗拉姆海峡流通量以及地转流强度模拟产生偏差的主要原因。尽管风场的分辨率不同,在海盆尺度上对于海冰和海水温盐结构的影响并不大。

Analysis on the response of an Arctic ice-ocean coupled model to two different atmospheric reanalysis datasets

The downward radiative fluxes, wind speed, near surface temperature, precipitation, humidity got from Climate Forecast System Reanalysis (CFSR) and the Japanese 25-year Reanalysis Project (JRA25) are compared in this article. We find that most significant biases between the two datasets are precipitation, downward fluxes for both shortwave and longwave radiation. Driving by these two datasets, model results forced by CFSR shows big differences on sea ice, Atlantic water and thermohaline structure in Canada basin compared to in situ observations, with the simulated geostrophic current on isopycnal surface 20% higher than that forced by JRA25 and a larger volume fluxes than that derived from SODA data. Sensitivity experiment forced by downward radiative fluxes from CFSR, which have been evaluated to be close to observed values, demonstrates comparable results to observational results. The cloud data plays a key role in modeling sea ice while freshwater flux brought by precipitation can change the heat transport of Atlantic inflow prominently and carry a further effect on sea ice in the Arctic. The overestimated precipitation in CFSR is the major for large biases of volume flux through Fram Strait, geostropic current on isopycnal surface and thermohaline structure in central Arctic. Although reanalysis wind have different resolution for the two datasets, our results indicates that it carries an ignorable effect on modeling sea ice and thermohaline structure on basin scale.