Quantifying the state of the current carbon cycle: a model-data fusion approach using multiple constraints

我们已经知道，全球碳循环和大气中二氧化碳(CO2)和甲烷(CH4)等温室气体的浓度变化在塑造地球气候方面发挥着重要作用。通过对冰芯的测量，我们可以重建大气中二氧化碳含量的变化，以及对10万年前全球气温的估计。在地球历史上，从冰川时代到暖期的变化伴随着大气二氧化碳浓度大约百万分之100的差异。在过去的10000年里，大气中的二氧化碳水平一直相对稳定，在260ppm至280ppm之间，工业革命之前约为280ppm。我们正在燃烧大量的化石燃料，从而向大气中排放二氧化碳，从而扰乱了自然的碳循环。与海洋-大气-生物圈系统中的自然碳排放(每年约1000亿吨)相比，人为碳排放(1990年代每年约70亿吨碳)很小，但它们会导致地球环境的重大变化。目前大气中的二氧化碳浓度为380ppm，预计到2050年将再增加70ppm至170ppm，这引发了人们对气候变化的担忧。从这个角度来看，迫切需要研究气候和碳循环之间的关系，以及是什么控制了大气中的二氧化碳浓度。这就是我的研究内容，也是最让我兴奋的地方：它解决了全球范围内的“大图景”问题。例如，目前正在从大气中吸收二氧化碳的陆地植被对未来气候变暖有何反应？然而，问题是，这种汇的强度正受到全球变暖的威胁；换句话说，全球变暖将导致更少的二氧化碳吸收，甚至二氧化碳释放，这反过来将导致额外的变暖。这种所谓的气候-碳循环反馈并不是很广为人知，但对这种反馈的理解是未来气候保护措施的必要组成部分。由于我们没有多余的地球副本，一个常用的研究工具来实验和测试关于大气二氧化碳和气候变化的原因的想法是地球系统的计算机表示。这些计算机模型是基于对调节过程的简化，因此不能准确地代表地球系统的功能。然而，我们必须确保这些模型与我们观察地球系统的方式一致。这通常是通过将模型结果与地球系统的这些观测结果(例如大气二氧化碳浓度的测量)进行定性比较来实现的。另一种更严格、更定量的方法来对比和优化这些观测模型，是基于一种数学方法，称为数据同化。我在这个项目中的研究涉及开发一个全面的碳循环数据同化系统。我将利用碳循环的广泛测量来广泛优化我的计算机模型，使计算机模拟的结果与观测结果最吻合。使用广泛的不同观测是一个重要的方面，因为各种观测有助于约束模型的不同部分，例如，大气CO的测量有助于量化野火释放的碳量。但是，这与我们今天和未来的社会有什么关系呢？例如，我的研究将量化直接土地管理对当代土地碳汇的贡献。这一点具有重大的政策意义，因为《京都议定书》只允许各国因直接土地使用管理而获得碳积累的碳信用。

We already know that the global carbon cycle and changes in the concentration in the atmosphere of heat trapping 'greenhouse' gases such as carbon dioxide (CO2) and methane (CH4) play a substantial role in shaping the planet's climate. From measurements of ice cores we can reconstruct the change of the amount of CO2 in the atmosphere as well as an estimate of the global temperature through several 100,000 years back in time. Changes from glacial times to warm phases in Earth's history go along with a difference of approximately 100 parts per million (ppm) in atmospheric CO2 concentrations. During the preceding 10000 years atmospheric CO2 levels have been relatively stable between 260 and 280 ppm and were about 280 ppm before the Industrial Revolution. We are burning large quantities of fossil fuels, thereby emitting carbon dioxide into the atmosphere, and thereby disturbing the natural carbon cycle. Anthropogenic carbon emissions are small (about 7 Gigatons of carbon per year in the 1990s) compared to the natural fluxes (about 100 Gigatons of carbon per year) in the ocean-atmosphere-biosphere system, but they cause significant changes in the environment of the Earth. At the moment atmospheric CO2 concentration is 380 ppm, and is predicted to increase by another 70 to 170 ppm by 2050, which raises concerns about climate change. From this perspective, there is an urgent need to investigate the relationship between climate and carbon cycle and what controls the concentration of CO2 in the atmosphere. This is what my research is about and this is also what excites me most about it: it addresses the 'big picture' questions on a global scale. For instance, how does the land vegetation, which is currently taking up CO2 from the atmosphere, react to a warmer climate in the future? The problem is, however, that the strength of this sink is under threat from the warming; in other words, global warming will lead to less CO2 uptake, or even CO2 release, which in turn will lead to additional warming. This so-called climate-carbon cycle feedback is not very well known, but the understanding of this feedback is a necessary ingredient for future climate protection measures. As we do not have spare copies of our planet, a common research tool to experiment and test ideas about the causes of atmospheric CO2 and climate changes is a computer representation of the Earth system. These computer models are based on simplifications of the regulating processes and therefore cannot exactly represent the functioning of the Earth System. However, we have to make sure that these models are consistent with how we observe the Earth system. This is usually done by qualitatively comparing model results with these observations of the Earth System such as measurements of the atmospheric CO2 concentration. Another, more rigorous and quantitative way of contrasting and optimizing the models against these observations is based on a mathematical approach, which is called data assimilation. My research in this project concerns the development of a comprehensive data assimilation system for the carbon cycle. I will make use of a wide range of measurements of the carbon cycle to extensively optimize my computer models such that the results from the computer simulations are in best agreement with the observations. Using a wide range of different observations is an important aspect as the various observations help to constrain different parts of the models for instance measurements of atmospheric CO help to quantify the amount of carbon released by wildfires. But what relevance does this have to our society today and for the future? For example, my research will quantify the contribution to the contemporary land carbon sink due to direct land management. This has great policy relevance since the Kyoto protocol only allows nations to claim carbon credits for carbon accumulation due to direct land-use management.