The Arctic and Tibetan Plateau have warmed 2–3 times faster than the global warming rate since the late twentieth century. COMBINED aims to better understand interactions between the rapid warming of these two poles, as well as other remote and regional feedback processes. These complex interactions are poorly understood but potentially have profound climate impacts. In addition, COMBINED will exploit the benefits of the Arctic sea ice initialisation and Tibetan Plateau land-surface initialization to improve the prediction of Eurasian climate from weeks, seasons, to a decade in advance. The COMBINED project makes use of available observed datasets and international multi-model databases. We will also perform targeted “pacemaker” experiments and make use of two national prediction systems from Norway (the Norwegian Climate Prediction Model) and China (the Beijing Climate Center Climate System Model).
Our first objective is to distinguish the causes of the synchronous accelerated Arctic and Tibetan Plateau warming. We have found that sea surface temperature (SST) in the extratropical North Pacific in spring have a significant influence on the Arctic summer atmospheric circulation [He et al., 2023]. While we have shown that tropical SST could have been one of the main drivers of Arctic winter tropospheric warming between 1979 and 2013 [Suo, 2023]. Tropical SST variations associated with El Niño events can also enhance poleward heat transport into the Arctic [King et al., 2023]. We have revealed revealing two leading patterns of Arctic and the Tibetan Plateau variability and identified their underlying dynamical linkages [Gao et al., 2025]. Amplified Arctic winter warming can further feed back on Eurasian climate, leading to episodes of Eurasian cooling despite overall global warming [Yin et al., 2025].
In addition, we have reconstructed Arctic sea ice concentration and Tibetan Plateau glaciers back to 1900. These provide exciting insights for a period when instrumental observations were poor in these regions. They show a remarkably large and previously unobserved decrease in sea ice from the 1920 to 1940 concurrent with the Early 20th Century Warming of the Arctic [Semenov et al., 2024]. We have also analysed future projections of Arctic sea ice and produced more reliable projections of newly formed sea ice [Zhao et al., 2024]. These indicate that newly formed winter Arctic sea ice is likely to increase dramatically until the middle of this century regardless of the emissions scenario.
Our second objective is to disentangle the individual and combined effects of the accelerated Arctic and Tibetan Plateau warming on Eurasian climate. We have shown that the Pacific oceanic front substantially increases the Euro-Atlantic blocking frequency [Cheung et al., 2023]. In another study, we found that March surface air temperature variations in Eurasia are influenced by a dipole pattern linked to North Atlantic SST anomalies and Barents Sea ice concentration changes [Yuan et al., 2024]. We have also used a regional atmospheric model to improve projections of spring consecutive rainfall events in the Three Gorges Reservoir area, showing an increase in the intensity and duration of these events in central-west TGR [Zheng et al., 2024].
Our third objective is to enhance understanding of Eurasian climate predictability. First, we have demonstrated that incorporating soil moisture data into the Norwegian climate prediction model significantly improves subseasonal-to-seasonal predictions of soil moisture, precipitation, and temperature, particularly in regions with strong land-atmosphere interactions [Nair et al., 2024]. Secondly, we have shown that a strong western pole of the positive Indian Ocean Dipole accelerates the decay of El Niño by enhancing convection and triggering oceanic upwelling, with important implications for seasonal prediction [Wu et al., 2024]. Furthermore, we contribute have performed pacemaker experiments to better understand how tropical ocean basins interact and affect global teleconnections and predictability [Richter et al., 2025]. In addition, we have developed a new coupled reanalysis dataset (CoRea1860+) extending back to 1860, providing unprecedented opportunities to evaluate mechanisms and improve initialization for predictions [Wang et al., 2025].
Finally, COMBINED strengthens China–Norway collaboration through research mobility and joint events. We co-organized major workshops in Bergen (2022, 2024) with over 100 participants, and in 2025 expanded activities with a Nanjing University summer school in Bergen (6–14 August) and the Climate Neutral Summer School in Rosendal (17–22 August) involving more than 40 international participants. These efforts further deepened collaboration and broadened the COMBINED network.
The Arctic and Tibetan Plateau (TP) have warmed 2–3 times faster than the global warming rate since the late twentieth century. They act as two emerging heat engines, driving large-scale atmospheric circulation anomalies. COMBINED addresses two key open questions: What are the roles of internal and external climate variability and various physical processes in driving synchronous and asynchronous climate variations at the two poles, on sub-seasonal and longer timescales? What is the combined and likely non-linear impact of the warming of these two poles on Eurasian and global climate?
In COMBINED we will take an important step to distinguish the causes of the synchronous accelerated Arctic and TP warming and departures from it on S2S, S2D, and longer timescales. We will disentangle the global impacts of TP amplification and how it acts in concert with that of the Arctic to nonlinearly influence Eurasian climate; and we will assess how these impacts are modulated by interactions with the ocean and land-surface boundary conditions via midlatitude teleconnections. We will finally assess the emerging sources of predictability from the Arctic and TP for Eurasian climate on S2S and S2D timescales. To this end, in addition to advanced statistical approaches, we will perform a suite of pacemaker experiments where conditions over the Arctic, TP, and over the Pacific and Atlantic Oceans are constrained to follow historical observations. Through these we will quantify the relative roles of the various factors in driving TP and Arctic changes, and their combined effects. Parallel S2S-S2D predictions using two national prediction systems (NorCPM and BCC_CSM) with the same accurate initialisation methods for the Arctic and TP cryosphere will be conducted to provide deep insight into the mechanisms and predictability associated with the two-pole interactions.
