It’s been a soaking wet Autumn for much of Europe. Many areas have seen between 150 and 200% of the long-term average rainfall (Figure 1). Parts of Spain have seen even larger surpluses; seasonal anomalies were near 250% for the Cantabria to Navarra provinces in the north, along with Murcia, the eastern half of Castile la Mancha and easternmost Andalusia.
Most of the worst-affected areas are in the western half of Europe. This region has seen many instances of disruptive and destructive flooding during Sep-Nov 2019. For example, Spain was hit by what’s known as a ‘gota fría’ storm in September, leading to what Murcia Today describe as ‘disastrous flooding’ in the Murcia province. There were also severe flooding incidents in France and Italy in November. Floods have also been widespread in England and Wales, where some locations saw their wettest autumn on record.
Figure 1: Map showing total rainfall for autumn (Sep-Nov) 2019 as a percentage of the 1981-2010 average.
Show Me the Culprit
When it comes to weather, attribution of anomalous conditions is never straightforward. There’s a myriad of different possible driving forces to consider.
Just how many? Well, to investigate the wet autumn of 2019, I’ve explored 21 different measures of the atmosphere and ocean!
By explore, I mean that I’ve conducted some simple correlation analysis. After all, this is a blog, not a (fully…) scientific publication!
Below, you can see the plot for Arctic sea ice extent anomaly (ASIE-A) as an example, next to a plot of mean decadal trend in total precipitation since the 1950s. The two are linked by something you may have heard of…
Climate Change – A Notorious Accomplice!
This was inevitably going to be a factor to discuss. Rainfall has shown an overall upward trend since the 1950s – but it’s not consistent between the decades (Figure 2). The 1960s were almost as wet as the 2010s. This may have something to do with North Atlantic sea surface temperature patterns, which are discussed later.
In any case, the trend since the 1980s is large enough that it may affect the correlation analysis on other factors. I believe this is behind ASIE-A having what’s proved to be the strongest correlation with rainfall of all that I investigated. You see, Arctic sea ice has been declining dramatically since the 1980s. All of the lowest ASIE-A values are in the past two decades, each much wetter than the 1980s.
Figure 2: Plots showing, for the western two-thirds of Europe, (left) the decadal means of total precipitation (rain, snow and ice e.g. hail) and (right) a scatter plot of this against Arctic sea ice extent anomaly, including an order-2 polynomial fit (red dots) to illustrate the negative correlation.
Luckily for the investigation, most of the other factors I’ve considered don’t have a steady long-term trend. They oscillate on varying timescales – a good example is the Atlantic Multidecadal Oscillation (AMO).
That greatly limits how much climate change can distort the correlation results.
Of the 20 other measures, 14 oscillate it’s 3 of those which have correlations that stand out to me:
- The Western Hemisphere Warm Pool (WHWP) – averaged near 5.0 in 2019
- The Atlantic Multidecadal Oscillation (AMO) – averaged near 1.5 in 2019
- The Tropical North Atlantic index (TNA) – averaged near 5.0 in 2019
All three of these are based on changes in ocean sea surface temperatures. The oceans are the most major source of moisture for rainfall on Earth, so this makes intuitive sense.
They also share another trait; positive correlation with autumn wetness. West-European autumn rainfall appears to increase with each of them.
They stand out not just for their correlation, but the fact that autumn 2019 has seen high values for each. Especially the WHWP and TNA, which are near record-high!
To see how much this counts for, I’ve mapped (Figure 3) the mean autumn rainfall for the years in which at least two of the months met the following criteria:
- WHWP and TNA at least 3.5
- AMO at least 1.5
Figure 3: Map showing total rainfall for autumn (Sep-Nov) during the years with the most comparable WHWP, AMO and TNA values to those observed in 2019, as a percentage of the 1981-2010 average.
I’ve also included a side-by-side of this and the 2019 map (Figure 4), to save some frantic scrolling.
Clearly, there is a great deal of similarity across the western half of Europe, along with most of Scandinavia. On the other hand, this ceases to be the case when looking at the remainder of Europe or northern Africa.
Figure 4: Figures 1 & 3, side-by-side for easy comparison.
This demonstrates that correlations can be of great value for anticipating seasonal characteristics, but only if carried out on a regional basis. The smaller the region, the more effective this pattern-matching is likely to be.
By providing historical data focused to the hyperlocal scale, MetSwift gives you the potential to achieve maximum effectiveness*.
A Harbinger of Things to Come?
The changing climate will likely bring about more frequent very wet autumns in Europe when other key climate drivers support that.
It just so happens that within the next two decades, we’re likely to see a counter to the climate trend. The AMO exhibits positive and negative phases which each last 2-3 decades. It’s been in a positive phase since 1995, 24 years ago. So, it’d be surprising not to see a negative phase begin within the next dozen years.
At that point, a shift toward drier autumns, perhaps even years, is plausible. It depends how the WHWP and TNA behave. The former has been trending warmer this past decade, while the latter has been very variable – a bit of a wild card!
* We’ve done the hard part at MetSwift gathering and processing and analysing the weather observations. Most of the data for key climate drivers is freely and easily acquirable via ESRL/NOAA.
James Peacock MSc
Head Meteorologist at MetSwift