Northwest Europe can expect a couple of winters with relatively frosty conditions, as one key driver of the North Atlantic Oscillation (NAO) is set to favour blockades of westerlies, allowing periods dominated by a supply of cold and relatively dry polar or continental air to flow in from the East or the North.
Current sunspot observations show (the relatively weak) solar cycle 24 has come to a close, entering a new – and possibly prolonged solar minimum. In this article we discuss ramifications for pending European winters.
Based on the developing closure of the current solar cycle (number 24), a relatively long minimum and taking into account a documented time lag these effects may be observed for a relatively large cluster of winters, starting in 2019 and possibly lasting as long as 2026. Do to increased likelihood of a negative phase in the NAO for Northwest Europe these winters may be relatively cool – but not extremely cold:
Again, no ‘horror winter(s)’ – Astrology is no base for weather forecasts
First, sadly, an important disclaimer to again distance ourselves from “horror winter” forecasts. Presuming a ‘horror winter’ defines as an extremely cold winter lasting for at least three months (1963 would qualify for Northwest Europe) the main reason is quite simple: there isn’t enough cold air (or water) around.
Superimposed on annual winter fluctuation (that depends on many factors apart from the 11-year solar cycle) is a clear warming trend (for all European seasons, including winter) due to increased atmospheric heat absorption, which implies sun cycle-driven changes in air circulation can only lead to redistribution of a limited and decreasing amount of cold air on the entire northern hemisphere. (Relatively cool winters in one region therefore correspond to (even) warmer winters elsewhere on the northern hemisphere: for instance relatively cold European winters are linked to Arctic warm winters.) Annually forecasted “horror winters” (presumed prolonged and extremely cold winters) have an extremely poor track record, since they first appeared in European media: these forecasts have never been proven right, and always lacked scientific background or even any argumentation as to the presumed mechanisms.
The horror of horror winter forecasts in British (tabloid) news media, for each subsequent year between 2012 and 2018. Image made by Dutch weather forecaster WeerPlaza in their critical Dutch-language commentary of this (annual!) phenomenon, that in fact dates back to well before 2012. The WeerPlaza meteorologist notes these forecasts always lack credible foundation and have always turned out to be wrong, ‘producing’ in fact very mild winters (indeed since 2013). This year though the horror winter forecast has also again turned up in Dutch news media – where a weather forecaster claimed to know exact temperatures beforehand (“minus 20 degrees”) without offering any reason “why and how” – to simplify the driving questions of science for a broader audience. So, we’ll offer a helping hand there:
But yes, climate science does offer (1) reason to expect a group of ‘relatively frosty European winters’ ahead
Contrary to what some argue annual, decadal and centennial fluctuations in solar activity only have a very small climatic effect – when it comes to global temperature this stays within a very small margin of just 0.07 degrees Celsius, exerting btw a net cooling effect, as the solar activity trend over the last decades shows a steady decline (therefore slightly masking the warming effect of simultaneously increasing heat absorption by greenhouse gases).
On a timescale of years to centuries, the influence of natural fluctuations in solar activity on global temperatures is almost negligible, as clearly illustrated by this powerful interactive temperature graph made by Bloomberg. If we were to distill a net influence, over the last couple of decades (since ±1960) the very slight net effect of the solar fluctuations on global temperatures is (much like aerosols) negative, promoting cooling, not warming. As the global temperature trend over that same period shows an accelerated rise, solar observations further strengthen the understanding that increasing greenhouse gas concentration are the cause of the current warming – not the Sun(!). The solar cycle does exert an influence on regional temperature anomalies though, particularly the European (and North-American) winter.
The ‘Solar Influences on Climate’ are neatly summarised in a 2010 publication under that same title in the journal Reviews of Geophysics by an international team of 15 climatologists. These researchers have also looked at regional effects, and what is of more interest there, is an approximately 11-year oscillation on top of the above-mentioned (declining) trend – the so-called sunspot cycle, a natural sinus wave of slowly increasing and decreasing solar activity, expressed in the number of observed sunspots.
This solar cycle is comprised of a cluster of years that are defined as a (numbered) solar maximum, when sunspots are above 50, followed by a group of years that are defined as a solar minimum, during which visible sunspot activity can drop to zero – and for instance Aurora Borealis and Aurora Australis are much weaker.
Having one mechanism in action that would promote a negative phase in the North Atlantic Oscillation does not mean you’ll have a cold winter straight away. In fact temperate zone winter variability is very high and European winters are notoriously difficult to capture in seasonal forecasts. One of the available models at hand, NOAA NCEP, does not forecast below average temperatures for Europe – to the contrary, as the above panels show (the cold air would be for eastern Canada). For Europe this forecast does not pass the model’s own skill filter, meaning a very low confidence level. Please note our article is about the influence of the solar cycle (that will be exerted between 2019-2026), not a full weather forecast.
Evidence for a negative NAO during solar minima
Quick explanation of the relevance: during a negative NAO phase high-pressure blockades form over Northern Europe and Atlantic depressions are forced on a more southern route. In Northwest Europe the effects on winter weather of a negative NAO are large, especially when it stays for several weeks and Siberian cold air is carried in.
Interestingly, as the researchers have found (by filtering out the signal of volcanic aerosol peaks in the dataset, that are also of influence [favouring positive NAO after 1-2 year time lag]), during a solar minimum the North Atlantic Oscillation – an air pressure distribution that in its normal positive phase is expressed as high pressure around the Azores and low pressure around Iceland (carrying Atlantic depressions and mild winter weather to Northern Europe) – is more likely to favours its opposite state, the negative phase.
Possible mechanism(s) and climate models
Most likely the mechanism has to do with an interaction between the troposphere (the area of the atmosphere that includes our pressure systems) and the overlying stratosphere, with the latter being more directly influenced both by fluctuations in total solar irradiance and more importantly changes in ultraviolet radiation – as uv of course also influences ozone concentrations (a greenhouse gas inside the stratosphere, the concentration of which is therefore important for the temperature inside this high-altitude atmospheric layer).
[It should be noted that there are also alternative mechanisms of influence between the solar cycle and NAO suggested. The most remarkable we found is by a Swedish geophysicist who in 2010 in the journal Global and Plantery Change suggested strong solar minima influence atmospheric (and ocean) currents by speeding up Earth’s rotation, due to a weakening interaction of solar wind with Earth’s magnetosphere(!) The same researcher in 2015 in Natural Science also predicted a new Grand Solar Minimum by 2030-2040. We'll get to GSMs and such forecasts below.]
Also climate models are now able to reproduce the negative NAO during solar minima, for instance in this 2011 publication in Nature Geoscience by a research team led by Sarah Ineson of the UK Met Office.
And if we remain in the UK for a bit longer, by detrending British temperatures from the overall northern hemisphere temperatures, a University of Reading-led team found a clear correlation between solar minima and the British winters specifically, they wrote in Environmental Research Letters in 2010. [The used method a good way to illustrate that such relatively cold winters originate in redistribution of cold air – there is no net northern hemisphere cooling during negative NAO (while net warming due to rising greenhouse gases continues). In fact, the Arctic is often (a lot) warmer, as negative NAO often coincides with negative Arctic Oscillation (AO) – carrying warm air from lower latitudes to the high North.]
Other climate researchers express skepticism of the strength of the correlation. For instance a 2013 publication in Environmental Research Letters (aptly titled ‘Claim of solar influence is on thin ice: are 11-year cycle solar minima associated with severe winters in Europe?’) led by Geert Jan van Oldenborgh of the Royal Netherlands Meteorological Insitute (KNMI) [an authority on the influence of multi-annual climate variations like ENSO and the Atlantic Multidecadal Oscillation (that does influence local air pressure patterns)].
These researchers could not replicate earlier claims by yet another research group (not discussed here) about the connection between the solar cycle and severity of Central-European winters – criticising poor methodology, and finding no significant connection:
“We show that there is no significant connection beyond common trends between sunspot numbers and instrumental series of the North Atlantic Oscillation and Frankfurt am Main temperature, nor with reconstructions of winter weather in the Netherlands and Switzerland.”
Interestingly several more recent studies have suggested that the link between the solar cycle and NAO may be stronger if you allow for a delay. For instance this 2016 publication in the Quarterly Journal of the Royal Meteorological Society and this publication, from February 2018 in Environmental Research Letters, describe a time lag of 2-4 or 3-4 years of observed negative NAO following a solar minimum for early winter months (December and January), where ocean coupling is presumed, while the solar cycle effects for February are quicker (a time lag of 0-2 years), suggesting a direct atmospheric route.
Current ocean temperature anomalies according to NOAA. Although scientific literature shows that for late winter (February) the solar cycle exerts a direct atmospheric influence on the North Atlantic Oscillation (NAO), the more pronounced effects are delayed. This time lag is created through atmosphere-ocean interactions, that can sometimes keep an anomaly in place, strengthening the (combined) signal. Image shows the current regional deviations from mean ocean temperature anomalies (blue: below average, yellow and red: above). Note specifically the pattern for the North Atlantic Ocean. High temperature anomalies off the US coast favour (extra strong) formation of Atlantic depressions, the course of which is influenced by the NAO air pressure pattern. Currently relatively cold water around Iceland and relatively warm water around the Azores, could (also) favour a negative NAO, promoting a southern course for Atlantic depressions, over southern Europe, enabling higher pressure and colder weather up north.
So, what can we expect for upcoming European winters? If we stick to the science at hand, chances of cold February months are highest (including February 2019) – as the solar-NAO correlation for this month is immediate, and the solar minimum is already in place.
Then what about December 2018 and January 2019? Well, there it becomes a bit of a definition game – of whether or not the solar minimum has already started some two years ago:
Comparison with winter conditions during previous solar minimum
If we look at NOAA’s curve more closely we can see that the last (very weak) cycle has on average already been below 20 sunspots for about two years (beginning 2017) – so just within the documented time lag.
If we look back at the end of the previous solar cycle, that started producing negative NAOs by December 2009, we see a similar pattern, where sunspots were also below ±20 for about two years beforehand (beginning 2007).
So yes, a bit quick and dirty, it is possible, for December 2018 and January 2019 too – but more so for February 2019. The stronger signal however should be expected for the winters from 2019-2020.
For the upcoming cluster of winters indeed the winters influenced by the previous solar minimum offer an indication of what to expect. Winters with regular negative NAO and therefore frosty periods of two or three weeks, but with overall temperatures that would classify as ‘normal’ (for the 20th century climate average) – not as ‘cold’ or ‘extremely cold’. Examples are the winters of 2009-2010, 2010-2011 and 2011-2012.
It is probably true to state though that the upcoming solar minimum will be stronger (therefore longer) than the previous. The declining trend leads some to state we’re in for something stronger, something resembling ‘Little Ice Age’ conditions. So what does the science actually say about that?
What would be the effects of a new Grand Solar Minimum? And will we have one?
With a maximum sunspot number of 68.9 (in August 2013) the last solar maximum (Cycle 24) was the weakest in over a century – after Cycle 14 peaked at a maximum of 64.2 in February 1906. This prompts a discussion of what would happen if the trend of gradually declining solar activity would continue into a so-called Grand Solar Minimum, of which paleoclimatologists have described several over the course of the late Holocene, with the most famous one the Maunder Minimum – a prolonged sunspot minimum between 1645 and 1715 during which sunspots could hardly be observed and that partially overlapped with a period of cold European winters known as the Little Ice Age – that however started earlier (after the Medieval Warm Period, unclear starting border between 1300-1500) and lasted longer (approximately until 1850). Although this overlap is rather poor, it should be noted that general sunspot activity was relatively low for a longer period of time, indeed between ±1300 and ±1850, including smaller ‘Grand Solar Minima’ like the Wolf Minimum and the Spörer Minimum, preceding the Maunder Minimum, and the later Dalton Minimum.
[Also important to note that not all paleoclimatologists agree Grand Solar Minima caused the Little Ice Age. Other suggested mechanisms include large fluctuations in human population (driving land use change), changes in the AMOC/Gulf Stream and volcanoes. It's a complex story, even superscale volcanic eruptions can be weak coolers and most probably it's all factors combined that exert a climate-forcing signal.]
Possibly driving larger timescale solar changes (and leading to sets of Grand Solar Minima), scientists have also discovered other (weaker) solar cycles, on much longer timescales than the clearly observable 11-year sunspot cycle. Interestingly this includes a ±2.400-year long cycle, evidence of which has even been found in Permian sediments, dating to more than 250 million years ago, as described by Roger Anderson in the Journal of Geophysical Research in 1982.
It’s commonly referred to as the Hallstatt Cycle, named after the Central European (Celtic) Hallstatt culture of 800 to 600 BC – a relatively cool period of the Holocene, during which mountain glaciers extended, and that coincided with the Homeric Grand Solar Minimum.
For the Holocene record this cycle was described by an international research team in a 2016 paper in Astronomy & Astrophysics – illustrating how over the past 9,000 years this ±2,400-year Hallstatt cycle seems to correlate with the occurance of (clusters of) grand solar minima:
It requires a bit of a trained eye, but from this 9,000-year record the ±2,400-year Hallstatt cycle is linked to the occurence of Holocene Grand Solar Minima (shown as white dots in the graph).
This of course leads to speculate whether a ‘Hallstatt minimum’ occurred during the Little Ice Age (that would fit the cycle best!). If so, perhaps a return to similar conditions becomes less likely – as we are some 2,800 years separated from the Homeric Minimum and indeed already almost 400 years beyond the start of the Maunder Minimum – meaning we would have to wait about two thousand years for a next grand minima to manifest.
So why is the solar activity trending down then? Well, that is common too. There are more cycles – it goes up and down all the time (remember it was even lower at the start of the 20th century than it is now, but that did not classify as a Grand Minimum).
Not straight into a Grand Minimum, but the set of relatively cold European winters can last until 2026(!)
Mike Lockwood, Professor of ‘Space Environment Physics’ of the University of Reading in this 2009 publication in Proceedings of the Royal Society A considers it an 8 percent chance we’ll enter a new Grand Solar Minimum in the next 50 years. A later study by UK’s Met Office Hadley Centre and the National Centre for Atmospheric Science under lead author Sarah Ineston, taking into account the lower-than-expected Cycle 24, and published in 2015 in Nature Communications suggests this chance has since somewhat increased – but still places it at ‘just’ 15 to 20 percent.
And indeed a study led by NASA specialist David Hathaway and published two years ago in the Journal of Geophysical Research suggests we won’t be rushing into a Grand Solar Minimum. These experts expect after the relatively small Solar Cycle 24, we’ll have Cycle 25 that will be equally small or perhaps even slightly less active – but that sunspot activity will not fall silent:
NASA specialist forecast shows after Cycle 24 we’ll ‘just’ have Solar Cycle 25, albeit a weak one – and with an extra long solar minimum in between.
What is important to note though is that these less-intense solar cycles are characterised by somewhat more prolonged solar minima. And that’s relevant to our winter forecast! The current minimum is forecast to reach the deepest point after 2020 (most likely 2021) – and then stay low (sunspots <20) for at least another year. This means that given the time lag of 2 to 4 years in the correlation with NAO the current minimum can influence European winters up to 2024-2026!
Frosty weather is normal, ‘horror winters’ don’t exist (anymore)
Even in this situation and contrary to the usual hyped statements there is little reason to assume extremely cold winters. (In fact, as stated above, it should be noted that in the history of the annually predicted ‘horror winter’ such very cold winter conditions have never come to pass.)
The reason is the much stronger (and global) influence of anthropogenic climate change. Not only is there a clear Northwest-European trend of rising average winter temperatures and a trend of decreasing cumulative winter frost – but what is of more importance is the fact that climate change is strongly pronounced in the regions that would supply cold air to Northwest Europe: Siberia, and more specifically the Arctic.
Temperature rise from anthropogenic climate change follows a strongly skewed pattern, with the northern hemisphere (and especially the Arctic) warming much faster than the global average. Temperature graph from this 2017 study in the journal Weather.
If we look at the expression of anthropogenic climate change as a result of rising concentrations of greenhouse gases we see a strongly skewed pattern: on average the northern hemisphere (where +2 degrees has been reached!) is warming about 4 times as fast as the southern hemisphere, and over the Arctic the effects are even further amplified, with up to a full degree of warming per decade (and warming of four to six times the combined hemispheric average).
As warm air is geographically fixed by upper-ocean layer interaction and for instance decreasing sea ice, it has large ramifications for the chance of prolonged cold spells. Short-duration cold temperature extremes are still possible locally (especially at lower latitudes) due to changing patterns of air circulation (decreasing strength of the Polar Cell, a consequence of rapid Arctic warming) – but for such conditions to spread out over large areas and to remain for multiple winter months many decades of rapid heat absorption would have to be compensated. The climate system does not have an alternative natural route to achieve such required cooling, not on our human timescales – and definitely not in the winter of 2018-2019 or subsequent European winters ahead.
© Rolf Schuttenhelm | www.bitsofscience.org