Seasonal evolution of Ku- and Ka-band backscattering horizon over snow on first-year and multiyear sea ice

近四十年来，利用连续的被动微波卫星任务提供的长期数据记录，北极海冰面积已被绘制成地图;显示自1979年以来，冰损失的速度加快。人们对海冰变薄的程度知之甚少，部分原因是缺乏类似的长期和一致的海冰厚度数据记录。雷达高度计，如2010年4月以来在欧洲空间局（欧空局）CryoSat-2（CS2）上飞行的雷达高度计，以及2013年2月作为法国国家空间研究中心（法国空研中心）和印度空间研究组织（印度空研组织）联合使命的一部分发射的SARAL/AltiKa卫星，目前正在提供泛北极（或AltiKa高达81.5 N）厚度观测。然而，使用这些数据的一个关键不确定性是雷达实际穿透覆盖的积雪的距离。一般的假设是，雷达回波在Kuband（CS2）频率下来自雪-冰界面，在Ka波段（AltiKa）频率下来自雪-空气界面。利用这些信息，再加上对上覆积雪深度及其密度的假设，科学家们就可以在假设流体静力平衡的情况下，将雷达回波转换为总冰厚。然而，现场证据使这一一般假设受到质疑，即使是对于均匀的积雪。另一个复杂的问题是缺乏关于积雪有多深及其密度的知识。通常情况下，雪的深度和密度的信息的基础上建立在厚厚的多年冰在20世纪80年代的气候。然而，随着海冰覆盖总面积的减少，北极盆地第一年海冰的比例现在有所增加。第一年冰上的雪往往比多年冰上的雪更咸，因此它有可能对雷达回波产生重大影响。此外，在北极某些地区，秋季和冬季的封冻推迟了几周到几个月，缩短了积雪的持续时间。鉴于目前的这些不确定性，很难准确评估海冰厚度每年和长期的变化。由于海冰是气候变化的重要指标，在北极能源和淡水平衡中起着重要作用，是海洋生态系统的关键组成部分，因此我们必须提高雷达测高厚度反演的准确性。该项目的目的正是通过在Ku和Ka波段频率上对雷达穿透深度进行地面观测，从秋季冻结到冬季积雪变质和夏季融化，整个年度周期都是如此。这些信息，连同详细的积雪特征，将使我们能够评估雪的积累，雪的形态和雪的盐度的变化如何影响Ku和Ka波段的穿透系数。MOSAiC漂移站提供了一个独特的机会，可能是唯一的机会，以获得一个基准数据集，包括连贯的现场，空中和卫星数据。对这些信息的分析将使科学家能够更好地描述积雪的物理特性（在不同类型的冰之上）如何影响Ka和Ku波段雷达的穿透力。重要的是，我们将能够评估积雪在第一年（海冰大于几厘米）和多年海冰的季节性演变。MOSAiC还提供了全年观测积雪深度和密度的机会，这将有助于评估雷达测高厚度反演中通常采用的气候学假设的有效性，并为验证积雪深度产品提供数据。这些活动对于改进北极地区许多冰雪条件下雷达测高的海冰厚度反演至关重要。

Arctic sea ice area has been mapped for nearly four decades using the long-term data record provided by successive passive microwave satellite missions; showing an accelerated pace of ice loss since 1979. Less is known about how much the ice has also thinned, in part because of the lack of a similarly long-term and consistent data record on sea ice thickness. Radar altimeters, such as the one flown on the European Space Agency (ESA)'s CryoSat-2 (CS2) since April 2010, and the SARAL/AltiKa satellite, launched in February 2013 as part of a joint mission by the Centre National d'Etudes Spatiales (CNES) and the Indian Space Research Organization (ISRO), are now providing pan-Arctic (or up to 81.5N for AltiKa) thickness observations. However, one key uncertainty in using these data is how far the radar actually penetrates into the overlying snow cover. The general assumption has been that the radar return is from the snow-ice interface at Kuband (CS2) frequencies, and from the snow-air interface at Ka-band (AltiKa) frequencies. Using this information together with assumptions on the depth of the overlying snow pack and its density, scientists can then convert the radar returns into total ice thickness assuming hydrostatic equilibrium. However, field evidence has put this general assumption into question, even for a homogeneous snowpack. A further complication is the lack of knowledge on how deep the snow pack is and its density. Typically, snow depth and density information based on a climatology constructed over thick multiyear ice in the 1980s have been used. However, as the total area in the sea ice cover has declined, there is now a larger proportion of first-year sea ice in the Arctic Basin. Snow over first-year ice tends to be more saline than over multiyear ice, and as such it has the potential for a significant impact on the radar returns. In addition, autumn and winter freeze-up has been delayed by several weeks to months in certain regions of the Arctic, shortening the duration for accumulation of snow. Given these current uncertainties, it is difficult to accurately assess how sea ice thickness is changing from year to year and over the long-term.Because sea ice is an important indicator of climate change, plays a fundamental role in the Arctic energy and freshwater balance, and is a key component of the marine ecosystem, it is essential that we improve the accuracy of thickness retrievals from radar altimetry. This project aims to do just that by making ground-based observations of the radar penetration depth over a full annual cycle at both Ku- and Ka-band frequencies, from autumn freeze-up, through winter snow metamorphism and summer melt. This information, together with detailed snow pack characteristics, will allow us to assess how changes in snow accumulation, snow morphology and snow salinity impact Ku- and Ka-band penetration factors. The MOSAiC drifting station provides a unique opportunity, possibly the only opportunity, to obtain a benchmark dataset that involves coherent field, airborne and satellite data. Analysis of this information will enable scientists to better characterize how the physical properties of the snow pack (above different ice types) influence the penetration of Ka and Ku band radar. Importantly, we will be able to evaluate the seasonal evolution of the snow pack over first-year (sea ice greater than a few cm) and multiyear sea ice. MOSAiC additionally provides the opportunity for year-round observations of snow depth and density that will allow for assessment of the validity of climatological assumptions typically employed in thickness retrievals from radar altimetry and provide data for validation of snow depth products. These activities are essential in order to improve sea ice thickness retrievals from radar altimetry over the many ice and snow conditions found in the Arctic.