Investigation of past and present climate dynamics and impact of climate tipping elements by means of a spatio-temporal analysis of climate data using complex networks

研究气候系统的稳定性及其所经历的变化的影响需要数据分析和建模方法。最近关于气候系统中特别敏感的子系统（引爆要素）的发现，引起了人们的关注，即普通气候模式仍然低估了气候变化。本项目以数据分析部分为重点，利用复杂网络理论的思想，对气候数据进行时空、多变量分析，探讨引爆要素在全球和区域气候与环境变化中的作用和连通性。这种方法将受到严格的审查和测试。为此，将开发和实现推断耦合方向和间接耦合的新方法。引爆要素将通过时空相互作用进行研究；这将有助于理解由于引爆要素的变化和引爆点的交叉而导致的全球气候网络的稳定性或多重稳定性。本研究还通过比较现代时空气候相互作用与全新世期间过去的气候变化，包括开发一种适当的方法来比较稀疏分布数据（古数据）与复杂网络数据（近期气候相互作用网络）。因此，该项目将对气候变化研究（更好地理解引爆因素）和方法发展（时空分析的新方法）做出重大贡献，最终也可用于改进气候模型。

The study of the stability of the climate system and the impacts of the changes it undergoes requires both data analysis and modelling approaches. Recent findings of particularly sensitive sub-systems in the climate system (tipping elements), have risen concerns that common climate models still underestimate climate change. Focusing on the data analysis part, this project aims to investigate the role and connectivity of tipping elements with respect to global and regional climate and environmental change, using ideas from complex network theory for a spatio-temporal, multi-variate analysis of climate data. This approach will be critically reviewed and tested. For this purpose, new methods for inferring coupling directions and indirect couplings will be developed and implemented. Tipping elements will be studied by their spatio-temporal interactions; this will help to understand the stability or multi-stability of the global climate network due to changes of tipping elements and the crossing of tipping points. This study is complemented by comparing modern spatio-temporal climate interactions with past changes of climate during the Holocene, including the development of an appropriate approach to compare sparsely distributed data (palaeo-data) with complex network data (recent climate interaction network). The project will, therefore, significantly contribute to the climate change studies (better understanding tipping elements) and the methodological development (new approaches for spatio-temporal analysis), which could eventually also be used to improve climate models.