Climate variability and prediction

Investigating processes affecting European climate and its variability and analysing the potential for prediction.


Climate prediction on seasonal to decadal time scales is still in its advent. Detailed understanding of the mechanisms governing the natural climate variability and improvement of the prediction systems are of uppermost importance to assess the predictability of the climate system and to improve predictions.

European climate variability is strongly governed by processes in the North Atlantic region. Ocean processes and related air-sea interactions affect moisture and heat transport towards Europe and atmospheric dynamics. Deep water formation in the North Atlantic Ocean is influencing the large scale ocean circulation and has been linked to variations of the Atlantic Meridional Overturning Circulation (AMOC). An important modulator of this deep water formation is the export of freshwater and sea ice from the Arctic. Furthermore, Arctic sea ice variations and recent reductions have been linked to European winter climate conditions and extremes.

Existing predictability studies show the largest interannual to decadal predictability in the North Atlantic and surroundings and the AMOC and related ocean dynamics have been suggested being the main driver for this predictability.  


The main focus area is the European - North Atlantic – Arctic area. However, climate teleconnections are global, which means that relevant processes in other regions, as e.g. the Pacific variability or tropical processes are not excluded. We analyse different time scales, from seasons to decades. Variations of recent past - present day - near future climate are in the centre of the investigations since the ultimate goal of climate prediction is to provide more reliable information for the upcoming years and decades.

Researh questions

The overarching goal is to investigate processes affecting European climate, its variability and analyse the potential for prediction.

More specific research questions are:

  • What are the main drivers of European climate variability at different time-scales?     
  • How does Arctic sea ice trends and variations affect lower latitude climate and the NAO?
  • What are the main drivers of the AMOC and how predictable are they?
  • How large is the uncertainty of regional decadal scale trends due to natural variability?
  • How does an improved representation of the North Atlantic ocean current system affect air-sea exchanges, storm tracks and European weather and climate conditions?
  • How do improved model parameterizations and higher resolution affect the simulated variability and predictability in our regional and global models?
  • How can we improve the initialization of our prediction system?


  • Reanalyses data and longer observational data sets.
  • The global coupled climate model EC-Earth and its component models at different resolutions.
  • The regional models HARMONIE and RCA in different resolution and over different regions as Europe and the Arctic.
  • Existing model simulations (e.g. CMIP5/ CMIP6, CORDEX, GREENICE).

Recent publications

  • Karami, Pasha et al., 2020: West Asian climate during the last millennium according to the EC-Earth model. Canadian journal of earth sciences, Vol. 57, no 1, 102-113 p.
  • Koenigk, Torben et al., 2019: Towards normal Siberian winter temperatures? International Journal of Climatology, 39 (11), 4567-4574.
  • Thomas, Manu et al. 2019: Snowfall distribution and its response to the Arctic Oscillation: an evaluation of HighResMIP models in the Arctic using CPR/CloudSat observations. Geoscientific Model Development, 12 (8), 3759-3772.
  • Fuentes Franco, Ramon et al., 2019: Sensitivity of the Arctic freshwater content and transport to model resolution. Climate Dynamics, 53 (3-4),  1765-1781.
  • Koenigk, Torben et al., 2019: Impact of Arctic sea ice variations on winter temperature anomalies in northern hemispheric land areas.  Climate Dynamics, 52 (5-6), 3111-3137.
  • Koenigk, T. and L. Brodeau, 2017: Arctic climate and its interaction with lower latitudes under different levels of anthropogenic warming in a global coupled climate model. Clim Dyn 49, 471–492.


Figure 1. Shown are winter temperature anomalies over Northern Europe between 1982-2013 in ERA-interim reanalysis data, the ensemble mean from 6 different global atmosphere models and the multi-model ensemble mean of these 6 models. In EXP1 the atmosphere models were forced with observed varying sea surface temperature and sea ice conditions over 1982-2013. In EXP2, the models are forced with observed varying sea ice conditions but with climatological sea surface temperatures. The correlation between observed and simulated Northern Europe temperature is almost the same for EXP1 and EXP2 (0.79 and 0.71), which indicates that Arctic sea ice variations are an important driver of the winter temperature variations over Northern Europe.