Quantifying uncertainty in future carbon cycle feedbacks

在我们对气候系统将如何应对人类持续排放的温室气体的理解上，预测未来的气候变化受制于层层的不确定性。这个谜题的一个重要组成部分是地球系统的自然循环如何调节大气温室气体的水平。特别是，全球碳循环目前以自然碳汇的形式吸收了大约一半的人类二氧化碳排放，但对这些碳汇如何在下个世纪持续存在的估计存在很大的不确定性。这种不确定性在陆地碳汇方面特别大，并直接传播到对未来气候变化和气候影响的预估中的不确定性。为了预测下个世纪气候变暖的程度，我们必须了解目前在植被和土壤中吸收和储存的碳将如何对未来的温度和降水变化做出反应。面对持续的气候变暖，陆地上的碳汇还会继续存在吗？全球模型对当前土地碳吸收的模拟效果如何？限制生长的养分（如氮和磷）的可用性在未来如何变化并影响碳吸收？人类土地利用的持续变化将如何影响大气和生物圈之间的碳流？我们能否理解模式之间的结构差异如何影响未来的预测，以及哪些模式过程对于包括在未来气候变化的模拟中至关重要？我们能否利用现有的数据集，从概率上约束与当前观测一致的未来预测范围？在这项研究中，我计划使用一种方法来解决这些问题，该方法将全球地球系统建模与最近关于全球气候和生态系统过程的观测和实验信息相结合。通过对碳循环不确定性的分析和量化，我希望既能提高我们对碳如何在自然气候系统中流动的理解，又能提高对气候系统如何应对人类持续温室气体排放的预测。

Predicting future climate change is subject to layers of uncertainty in our understanding of how the climate system will respond to continued human emissions of greenhouse gases.  An important component of the puzzle is how natural cycles in the earth system regulate levels of atmospheric greenhouse gases.  In particular, the global carbon cycle currently absorbs about half of human carbon dioxide emissions in natural carbon sinks, but there is substantial uncertainty in estimates of how these carbon sinks will persist over the next century. This uncertainty is particularly large with respect to land carbon sinks, and propagates directly to uncertainty in projected future climate changes and climate impacts.In order to predict how much climate warming will occur over the next century, it is essential that we understand how carbon that is currently absorbed and stored in vegetation and soils will respond to future changes in temperature and precipitation.  Will land carbon sinks persist in the face of continued climate warming?  How well do global models simulate current land carbon uptake?  How might the availability of growth-limiting nutrients (e.g. nitrogen and phosphorus) change in the future and affect carbon uptake?  How will continued changes in human land-use affect the flows of carbon between the atmosphere and the biosphere?  Can we understand how structural differences between models affect future projections, and which model processes are critical to include in simulations of future climate change?  Can we use available data-sets to constrain probabilistically the range of future predictions that are consistent with current observations?  In this research, I plan to address each of these questions using an approach which combines global earth system modelling with recent observational and experimental information about global climate and ecosystem processes.  Through analysis and quantification of carbon cycle uncertainties, I hope to both improve our understanding of how carbon flows through the natural climate system, and sharpen predictions of how the climate system will respond to continued human greenhouse gas emissions.