Climate reconstructions have contributed to our understanding of climate variability and trends for many decades. Some reconstructions also address atmospheric dynamics such as circulation variability modes, the monsoon, and jet streams. We recently generated a global, 3-dimensional, monthly reconstruction for the last 600 years [1] and analysed Atlantic-European jet stream variability [2]. However, a seasonal or monthly resolution is insufficient for analyzing dynamical processes, which unfold on a daily timescale, as do weather extremes. The Twentieth Century Reanalysis (20CRv3) [3] provides 3-dimensional, 6-hourly global atmospheric fields back to 1836, even 1806. We can calculate jet stream metric from these fields. Can we go even further back? If we find pressure series from locations that capture the relevant atmospheric dynamics of the eastern North Atlantic–European sector, then a small number of series might suffice to capture jet characteristics. This could eventually lead us back to the late 17th century and would enable us to study the Late Maunder Minimum, a period of low solar activity and high volcanic activity, as well as extreme weather events such as the January 1709 cold spell.
Fig. 1: Blocking frequency 1806-2015 in 20CRv3 (contour interval 10%) and position of the polar front jet (scaled to maximum occupancy in the field) 1941-2024 in ERA5 [6] for extremes in the jet indices in the cold season.
We are not quite there yet. However, in our new paper [4] we construct daily indices of jet stream characteristics (strength, tilt and latitude) 300 years back, based purely on observations. Three suitable locations were available from previous studies: Uppsala and Padua, which capture the North-South gradient, and London, which captures Western European pressure variations. To capture the West-East gradient and central European ridge patterns, we need a fourth series located further east. So, off to the archives we go! After several years of work, we present a daily pressure series for Berlin back to 1721. It is based on many segments, and some gaps had to be filled using data from other stations. Ultimately, using these four locations, we were able to generate the three index series with no missing days. Calculating the same indices in ERA5 reanalysis data [5] and comparing them to the position of the upper-level polar front jet [6] (Fig. 1) shows that these simple surface-pressure indices capture important characteristics of the upper-level jet. They explain a large proportion of the day-to-day surface temperature variability in Western and Central Europe throughout the year, and are closely related to extreme weather events such as cold and warm spells, heavy precipitation and storms due to their close relationship with atmospheric blocking.
For example, 26 December 1836 was the day with the most negative jet strength index within the 20CRv3 period. It shows a strong easterly jet at 50° N, flanked by two westerly jets near 35° and 70° N in a diffluent blocking pattern. This led to heavy snowfall and drifting in southern England, resulting in the country's deadliest avalanche disaster a day later, killing eight (Fig. 2).
Fig. 2. (left) Fields of surface air temperature, 500 hPa GPH (contours, gpm) and 300 hPa wind in 20CRv3 at 12 UTC on 26 Dec 1836. Light green contour: 50% blocking frequency across the ensemble. (right) Lewes avalanche, oil painting by Thomas Henwood, with kind permission, The Sussex Archaeological Society.
The same indices can be calculated in climate model simulations of the past and future. We used our own atmospheric simulations ModE-Sim, and for the 1850–2100 period, the CMIP6 ensemble, and analysed changes in the mean and in extremes (Fig. 3). We find a recent strengthening of the jet in the cold season that is stronger in the observations than in the ensemble of model simulations. Interestingly, positive extremes have not increased, while negative extremes have clearly become much rarer. Positive and negative extremes are also projected to become rarer in the CMIP6 ensemble.
Fig. 3. (left) Time series of seasonal mean jet strength in observations, ModE-Sim and CMIP6 (historical+SSP2-4.5) smoothed with a spline fit. Densities of daily values in (middle) CMIP6 simulations in 1851-1880 and 2071-2100 and (right) in observations in ten 30-yr periods between 1725 and 2024 in observations and (bottom). Insets with coloured dots show the means of the 2nd, 50th and 98th percentiles.
The daily resolution enables us to address not only trends due to climate change or variability caused by the effects of volcanic eruptions or the El Niño/Southern Oscillation phenomenon, but also driving factors on short timescales. We found clear imprints of Sudden Stratospheric Warmings. Most amazing of all, we could detect the imprint of solar storms. As suggested in the literature based on satellite data and model simulations, we observe a significant strengthening, which peaks around 10 weeks after the solar storm. This analysis is possible because astronomical observers recorded unusual events, such as low-latitude aurorae, far back in time. We have the exact day of over 100 solar storms during this 300-year period, even when focusing only on the 'active' season (i.e. the cold season). Eventually, our work links archival sources of pressure measurements with archival sources of solar storms, which I find particularly satisfying.
References
1 Valler, V., et al. ModE-RA - a global monthly paleo-reanalysis of the modern era (1421-2008). Sci. Data 11, 36 (2024).
2 Brönnimann, S. et al. Past hydroclimate extremes in Europe driven by Atlantic Jet and recurrent weather patterns. Nat. Geosci. 18, 246–253 (2025), https://doi.org/10.1038/s41561-025-01654-y.
3 Slivinski, L. C. et al. Towards a more reliable historical reanalysis: Improvements to the Twentieth Century Reanalysis system. Q. J. Roy. Meteorol. Soc. 145, 2876-2908 (2019), https://doi.org/10.1002/qj.3598.
4 Brönnimann, S., Y. Brugnara and P. Kallabis (2026) . A strengthening yet less extreme Atlantic European jet during winter. Nature Geoscience, https://www.nature.com/articles/s41561-026-02069-z.
5 Hersbach, H. et al. The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146, 1999–2049 (2020), https://doi.org/10.1002/qj.3803
6 Rivoire, L., Curbelo, J. & Linz, M. Tracking jet streams as Lagrangian objects. Commun. Earth Environ. (2026), https://doi.org/10.1038/s43247-026-03262-z