Southern OceaN optimal Approach To Assess the carbon state, variability and climatic drivers (SONATA)

南大洋（SO）是世界海洋中最令人兴奋和最极端的区域，有着最强的风，最冷的温度和最强烈的风暴。它也被认为是大气中二氧化碳最大的“汇”之一，约占全球海洋吸收二氧化碳的三分之一，平均每年占全球二氧化碳排放量的近十分之一。因此，SO碳汇的演变有可能改变气候变化的速度和程度。尽管它很重要，但我们不知道当代SO碳汇的状态、可变性或气候驱动因素，而且对其最近的演变存在很多争议。近几十年来，南太平洋的气候一直在发生变化：特别是，风加强了（部分归因于平流层臭氧的消耗，部分归因于气候变化引起的温度梯度的增加），海洋酸化正在发生，磷虾种群数量长期下降。这些影响发生在巨大的自然变率和量化不足的气候趋势之上。SONATA将通过提供新的数据和新的见解，整合来自海洋、大气和模型结果的观测结果，实现我们对当代SO碳汇的理解的一步转变。我们将开发三个互补的研究流，即“海洋”、“大气”和“过程与驱动因素”观点，并将使用先进的数学框架将它们结合在一起，以提供具有多个约束和减少不确定性的单一评估。海洋视角将利用现有的和新的海洋碳观测数据。我们将进行一项新的校准实验，以更好地评估目前由美国SOCCOM计划引入的约200个复杂剖面浮标所进行的大量pH测量。这些有可能大大增加可用于计算大气-海洋二氧化碳通量的观测数量，但前提是要进行适当的校准。此外，我们将开发和使用一种新技术，利用受现有水文和碳系统观测约束的上层水柱模型，构建海气通量的季节和时间演变估计。“大气观测”项目将在包括哈雷站（75S）、福克兰群岛（51S）和詹姆斯·克拉克·罗斯号科学考察船在内的偏远地区收集新的大气二氧化碳数据；新的大气O2数据将来自每8周重复一次SO样带的船舶轨迹，以及南极洲沿海的哈雷站。利用这些数据和反向框架方法，索纳塔将提供一个独立的SO碳汇评估，特别是关于变化的地理分布，O2数据有助于告知驾驶员。过程和驱动因素视图将使用两个气候尺度碳模型和一系列后推模拟来确定(a)大气CO2浓度、(b)自然气候变率、(c)气候变化和(d)平流层臭氧消耗对近期SO碳趋势和变率的相对贡献。海洋和大气观测，包括来自SONATA和SOCCOM的新数据，将用于优化模型并验证结果。利用气候模式理想化的强迫将提供理解观测到的变化模式所需的气候驱动因素的“指纹”。最后，这三个研究流将使用贝叶斯融合数学方法进行整合，该方法考虑了每个信息流的优缺点，并将联合不确定性最小化。SO海洋碳汇将以这种方式每年进行评估。然后，我们将测试未来纳入新观测流的附加价值，包括来自浮子、滑翔机、漂流器、自动水面车辆、额外的地面观测和卫星二氧化碳数据。

The Southern Ocean (SO) is the most exciting and extreme region of the world ocean, with the strongest winds, coldest temperatures, and most intense storms. It is believed also to be among the largest 'sink' for atmospheric CO2, accounting for about one third of the uptake of CO2 by the global ocean and nearly one tenth of the global emissions of CO2 on average each year. Thus the evolution of the SO carbon sink has the potential to alter the rate and extent of climate change.In spite of its importance, we don't know the state, variability, or climatic drivers of the contemporary SO carbon sink and there is much controversy over its recent evolution. The climate of the SO has been changing over recent decades: in particular, winds have intensified, (attributed in part to the depletion of stratospheric ozone and in part to increasing temperature gradients arising from climate change), ocean acidification is occurring, and there is a long term decline in krill stocks. These effects take place on top of large natural variability and poorly quantified climatic trends.SONATA will achieve a step change in our understanding of the contemporary SO carbon sink by delivering new data and new insights, integrating observations from the ocean, from the atmosphere, and model results. We will develop three complementary streams of research, an 'Oceanic', an 'Atmospheric', and a 'Processes and drivers' view, and will bring them together using advanced mathematical frameworks to provide a single assessment with multiple constraints and reduction of uncertainties.The Oceanic view will use existing and new observations of ocean carbon. We will undertake a new calibration experiment to better assess the large number of pH measurements now being made by about 200 sophisticated profiling floats introduced by the US SOCCOM programme. These have the potential to greatly increase the number of observations that can be used to calculate air-sea CO2 fluxes, but only if adequately calibrated. In addition we will develop and use a new technique to construct estimates of the seasonal and temporal evolution of the air-sea flux, using a model of the upper water column constrained with available hydrographic and carbon-system observations. The Atmospheric view will collect new atmospheric CO2 data in remote SO locations comprising Halley Station (75S), the Falkland Islands (51S), and aboard the BAS research ship James Clark Ross; new atmospheric O2 data will come from a ship track that repeats a SO transect every 8 weeks, as well as from Halley Station in coastal Antarctica. Using these data and an inverse framework approach, SONATA will provide an independent assessment of the SO carbon sink, which will deliver particularly on the geographic distribution of the changes, with O2 data helping to inform the drivers.The Processes and drivers view will use two climate-scale carbon models and a series of hindcast simulations to identify the relative contributions of (a) atmospheric CO2 concentration, (b) natural climate variability, (c) climate change, and (d) stratospheric ozone depletion to recent SO carbon trends and variability. Ocean and atmosphere observations, including new data from SONATA and SOCCOM, will be used to optimise the model and validate the results. Idealised forcing with climate models will provide the 'fingerprints' of climatic drivers that are needed to understand the observed patterns of change.Finally the three streams of research will be integrated using a Bayesian fusion mathematical approach that considers the strengths and weaknesses of each stream of information and minimises the joint uncertainty. The SO ocean carbon sink will be assessed annually in this way. We will then test the added value of including new streams of observations in the future, including from floats, gliders, drifters, Autonomous Surface Vehicles, additional ground-based observations and satellite CO2 data.