Claim of solar influence is on thin ice: are 11-year cycle solar minima associated with severe winters in Europe?

The following text has been sent in by Geert Jan van Oldenborgh, Jos de Laat, Juerg Luterbacher, William Ingram and Tim Osborn:

Claim of solar influence is on thin ice: are 11-year cycle solar minima associated with severe winters in Europe?

Eight months year ago, news articles claimed that scientists had discovered a strong connection between severe winters in central Europe and the 11-year sunspot cycle. They were based on an article by Sirocko et al, ‘Solar influence on winter severity in central Europe’ (Geophys. Res. Lett. 39 (2012) L16704), which we shall call SBP. Fifty years ago, Lorenz showed that a large part of the variability of winter weather is due to deterministic chaos, i.e., unpredictable fluctuations in atmospheric circulation, particularly in the westerly flow. The role of external forcings (including solar activity) in determining the warmth of individual winters is expected, therefore, to be rather minor – or even negligible if the forcing changes are weak. This has indeed been found in many review articles. SBP was therefore at odds with the current scientific consensus that the role of the sunspot cycle in the climate is small. We investigated why SBP had such unexpected results and came up with a number of fundamental issues with this paper Our results have just been published in Environmental Research Letters, http://iopscience.iop.org/1748-9326/8/2/024014 (open access) 

Below, we provide a brief summary of our findings. SBP's definition of a “severe winter” is one in which the river Rhine was frozen over a particular stretch in Germany. They obtained these years using a simple Google search, their supplementary information gives an incomplete list of references, which include an oil painting and a local newspaper clipping from a few years ago that claims the Rhine was frozen 70 years earlier (and mentions a few other years that are not in SBP's list). A little internet searching by us gave many other years, showing that SBP's list is not reliable.

For half of the period considered by SBP (1780-1963) this Rhine-freezing database is not necessary, as daily observations of minimum and maximum temperatures are available from Frankfurt am Main starting in 1870 (http://www.ead.eu), covering a large part of the series. SBP made no attempt to evaluate and/or calibrate their series against these observations, as is easily done and is standard practice in this field. Comparing the series (Fig. 1a) showed that SBP had included two very mild winters (1914 and 1954, both at low sunspot numbers), and had missed quite a few severe winters, most at high sunspot numbers). The instrumental data show no clear connection with the 11-yr sunspot cycle.

Figure 1. Winter (December–February) temperatures in the Netherlands (Van Engelen et al 2000) and Switzerland (Pfister 1984) compared with each other and the SBP series over the instrumental period (a) (together with winter mean temperature at Frankfurt am Main) and (b) the pre-instrumental period.

Going further back, there are well-calibrated reconstructions of winter temperature in Switzerland (Pfister, 1984, based on the freezing of lakes in Switzerland) and the Netherlands (van Engelen, 2000, partly based on the freezing dates of a canal), both freely available (e.g. from climexp.knmi.nl). They are on opposite sides of the area considered by SBP and are well-correlated with each other, i.e. winter cold often covers a large part of Europe. Comparing with them again shows that SBP had missed about half the severe winters. And again, neither of these carefully constructed and well documented reconstructions shows any clear connection with the 11-yr sunspot cycle.

Finally, SBP put forward explanation for the cold winters in terms of changes in wind direction. However, they use a very different definition of “sunspot minimum” for this analysis. With their original definition their “frozen Rhine” time series would show no connection with the 11-yr cycle. Conversely, using the definition of the first half of the article shows no signal in the analysis of atmospheric circulation.

We concluded that SBP show no real evidence for a strong influence of the 11-yr sunspot cycle on the winter weather of Europe. Our next step was to publish this in the scientific literature to set the record straight, and preferably start a dialogue with the authors on the differences between their and our findings. This proved to be more problematic than we expected. Geophysical Research Letters does not allow comments on other articles, but demands that these are part of a stand-alone scientific article. We wrote what we thought was such an article, but this was rejected without review by the editor-in-chief. Fortunately the editor of another journal, Stefan Rahmstorf of Environmental Research Letters, offered to consider it for peer-review. It was published there on 25 April with some clarifications demanded by two reviewers. This article documents our scientific arguments that the sunspot cycle has no detectable effect on European winter weather, consistent with other work.

About the authors

Geert Jan van Oldenborgh is senior researcher at the Royal Netherlands Meteorological Institute (KNMI). He started there in El Niño research and seasonal forecasting, but as most of the skill in these forecasts in Europe is due to the trends he shifted his interest to verification of decadal forecasts and climate model trends. An extension of this is event attribution, estimating whether and how climate change and other factors affected extreme weather and climate events. He is known for his web site, the KNMI Climate Explorer (http://climexp.knmi.nl), which allows everyone to analyse a large collection of climate data and to check hypotheses in observations and model output.

Jos de Laat works for the Royal Netherlands Meteorological Institute in the department of Earth Observation and Climate. His current focus is on satellite measurements of atmospheric composition - including long term ozone variability and ozone hole dynamics - but he has published on various other aspects of climate including analyses of regional temperature measurements.

Prof. Jürg Luterbacher is deputy Director of the Department of Geography at the University of Giessen, Germany and Chair for Climatology, Climate Dynamics and Climate Change. He is Visiting Professor with the Chinese Academy of Science in China and Guest Professor at the University of Bern, Switzerland. He received his PhD in climate sciences from the University of Bern. He is Lead Author in the chapter 'Information from Paleoclimate Archives' in the IPCC 5th Assessment report.

William Ingram works for the Met Office Hadley Centre & the climateprediction.net project at the Department of Physics, Oxford University.  His technical experience is in General Circulation Models (the most detailed & physically-based models of the climate system), particularly the representation of radiation.  The water vapour feedback on climate change is his major area of research, but he has published on other aspects of climate sensitivity, as well as the rigorous detection & attribution of climate change.

Tim Osborn has worked at the Climatic Research Unit (CRU), part of the School of Environmental Sciences at the University of East Anglia (UEA), since 1990.  A major area of his research is the study of natural climate variability (particularly that driven by changes in the thermohaline circulation of the ocean, related to the North Atlantic Oscillation, or forced by natural external factors) and how this has been observed, modelled and recorded in climate proxies (particularly tree-rings).  Tim's research also include precipitation, its climatology of daily variability and extremes, and possible changes in precipitation variability under future climate scenarios and the resulting impact on drought characteristics.  He is an author of the forthcoming fifth assessment report of the IPCC.