Rapid change in the Arctic – North Atlantic climate: using a dynamical systems framework to explore local stability, observability and teleconnections

Abstract

There is growing evidence for rapid Arctic climate change, including the intrusion of warm and saline Atlantic Water into the Arctic, seasonal sea ice loss and shifts in circulation pathways. These rapid Arctic changes can modulate subpolar North Atlantic processes, the connecting link between the Arctic and the North Atlantic, with potentially far-reaching and profound implications for the Northern Hemisphere weather and climate.  Across observations and model datasets, untangling this complex set of multiscale phenomena in the Arctic and their potential teleconnections is a major challenge. Yet intercomparison of different data products is necessary for understanding the mechanisms underlying the rapid Arctic changes.  In this study, we seek to understand recent changes and their underlying physical processes so that we can better observe, model and predict future climate changes. The spatiotemporal resolution of data needed to cover a sufficient range of multiscale processes tends to be high, which leads to huge volumes of data to analyse.     Fortunately, however, there have been new advances in data-driven methods. These methods abstract the problem to a dynamical systems framework of Reduced Order Modelling with the aim of discovering the key physical processes buried within the complexity.Therefore, this study simplifies the complexity of the multiscale climate variability , starting with analysis of the Arctic Subpolar Gyre sTate Estimate (ASTE), spanning 2002-2017.   We use sparse sampling methods to robustly approximate the underlying dynamics efficiently, even on huge datasets.  An advantage is that we can explore periods of rapid changes in the system as a whole, i.e. not solely via one measure or variable.   This provides a natural way to study changes to the dynamical balance across all state variables and to extract teleconnection patterns.   By taking advantage of these new, data-driven approaches, the findings of our study will inform the detection and attribution of rapid changes to the Arctic – North Atlantic climate system, especially in regions where the relevant state variables are hard to observe or estimate accurately.