Can we trust statistical predictions?

This is the overarching question raised in a new paper by Erik Kolstad and James Screen. The backdrop is that many people have studied the relationship between the extent of the sea ice in the Barents and Kara Seas in autumn and the wintertime weather farther south several months later (see for example this paper by Hall and co-authors). In recent decades, lower-than-normal sea ice extent has typically been followed by lower-than-average temperatures in Europe.

On the face of it, these kinds of relationships are great, because they can be used for prediction. Other phenomena like El Niño often lead to quite predictable changes in the weather patterns in many places. The potential value of a similar relationship between sea ice and weather is huge. Yet, a crucial question is: Even if there has been a statistical relationship between two variables in the past, does that guarantee that the same relationship will be valid in the future?

In our paper, we found that the lagged relationship between October sea ice extent in the Barents/Kara Seas and the North Atlantic Oscillation in winter (December to February) has changed a lot over time. To illustrate this, we made this figure:

Fig2
The colours show the correlation between October sea ice cover in the Barents/Kara and sea level pressure in the North Atlantic region in December–February during the periods indicated above each panel.

During the first period starting in 1916, the correlation between sea ice and pressure is positive over the Northeast Atlantic and negative farther south. This means that below-normal sea ice cover in autumn was generally followed by westerly winds and mild and wet winter weather in Europe. But in the period between 1980 and 2010, the more familiar pattern with cold winters following low sea ice was evident. This shows how much the correlation can change over time.

The paper shows that climate models have the same fluctuations, and this led us to conclude that—as the title indicates—the relationship is non-stationary. Which again led us to “caution against indiscriminately using Barents‐Kara sea ice to predict the NAO”. We believe this has important implications for seasonal prediction.

Advertisements

Our first 2019 summer forecast

We are now ready with our first ever in-house forecast. It’s based mainly on dynamical forecast models from the Copernicus Climate Change Service (CS3), but it also includes the Bjerknes Centre’s own Norwegian Climate Prediction Model (NorCPM).

Without further ado, here are our predicted temperature anomalies for June:

fc_t2m_6

It requires a bit of explaining. The colours indicate deviations from the normal temperature, which means the average over the past 20-30 years, depending on the model. The first five panels show the predictions from the individual models, the sixth panel shows (more or less) the average of these, and the bottom panel is our own combination of the model forecasts (Multimodel+). For more details on how we do this, please go to our Methodology page.

In short, the forecast is that it will be ever so slightly warmer than normal, but no large anomalies are forecast. The reason for the not-so-exciting forecast is that the models disagree. For instance, the ECMWF model predicts a warm June, while the UK Met Office predicts a cold June.

A pertinent question is: how skilful are these models? To answer this, we have prepared so-called skill maps. Here’s the one for June:

skill_maps_6new

The short story here is that blue means high skill, whereas red means low skill. Deep blue means a 20% improvement on ‘climatology’ (details). The maps are based on the models’ re-forecast skill. Our Multimodel+ forecast is moderately skilful in June in coastal regions.

Can we use ocean temperatures in the Arctic to predict temperatures in Europe?

This question was investigated in a new paper by Erik Kolstad and Marius Årthun in Journal of Climate. The answer is: yes and no. It worked quite well when we used data from the period starting in 1979 (when important weather satellites were launched). But when we used data from earlier in the twentieth century, the relationships that have been valid in recent decades broke down.

An example that we investigated in some detail was central England temperature (CET), for which the observations stretch back to the 17th century. When we used data starting in 1979, we found the following relationship with sea surface temperatures (SSTs) in the Atlantic part of the Arctic:

cet_map_1_0
This picture shows the correlation between SSTs in the Arctic, averaged from October to December, and CET averaged from January to March during the period from October 1979 to March 2017.

The relationship is clearly strongest in the Barents Sea, which is marked with black lines in the picture. When we averaged the SSTs inside that region and created a time series, we could compare it to the CET time series:

cet_timeseries_1_0
Here the time series for Barents Sea SSTs is shown in blue, along with the time series for CET in orange (the latter time series is multiplied by –1 to make the figure more readable).

The correlation between the two time series is –0.44 for the period starting in 1979. But we were interested in what happened during periods starting earlier than 1979. Then we had to use a different data set for the SSTs (CERA-20C, while we used ERA-Interim for the recent period). What we did was to investigate all possible periods with the same length as 1979–2017, starting with 1901–1939. For each period, we computed the correlation between SSTs in the Barents Sea and CET, and found this:

cet_stationarity_0
The correlation between Barents Sea SST and CET for each possible 38-year period starting with 1901–1939. The correlations in blue are based on CERA-20C, and the orange star shows the correlation for the ERA-Interim data.

The recent negative correlation of –0.44 is unprecedented. For periods starting early in the twentieth century the correlations are even positive. One of our main conclusions is that one should take great care when using Arctic SSTs to predict climate in later months.

There is also a Kudos web page about this article.

Erik Kolstad is currently investigating the relationship between Arctic sea ice and the North Atlantic Oscillation.

How does the ocean influence temperature over land?

Norway is much warmer than most other places at similar latitudes. The picture below shows fishing activity in northern Norway in winter; this could not have been possible without the warm Norwegian Atlantic Current, which brings warm water masses from the south throughout the year.

'Winter_in_Lofoten,_1886'_by_Otto_Sinding,_Bergen_Kunstmuseum
“Vinter i Lofoten” (1886), painted by Otto Sinding

In a new paper, Marius Årthun, Noel Keenlyside and Erik Kolstad from our project group and colleague Tor Eldevik from the Bjerknes Centre for Climate Research investigate the causes of European temperature variations on long time scales. For instance, we know that the temperature in parts of Scandinavia fluctuates with a period of 14 years. The paper shows that this is probably because the sea surface temperatures in the Nordic Seas also fluctuate on the same time scale. The mean westerly winds then transport air masses form over the ocean to over land, and thus the connection between the ocean and land is mediated. We can use this information to predict variations between colder-than-normal and warmer-than-normal, potentially several years ahead.

The title of the paper is Time scales and sources of European temperature variability, and it has now been accepted for publication in Geophysical Research Letters (GRL). Here is the abstract:

Skillful predictions of continental climate would be of great practical benefit for society and stakeholders. It nevertheless remains fundamentally unresolved to what extent climate is predictable, for what features, at what time scales, and by which mechanisms. Here we identify the dominant time scales and sources of European surface air temperature (SAT) variability during the cold season using a coupled climate reanalysis, and a statistical method that estimates SAT variability due to atmospheric circulation anomalies. We find that eastern Europe is dominated by sub-decadal SAT variability associated with the North Atlantic Oscillation, whereas interdecadal and multi-decadal SAT variability over northern and southern Europe are thermodynamically driven by ocean temperature anomalies. Our results provide evidence that temperature anomalies in the North Atlantic Ocean are advected over land by the mean westerly winds, and, hence, provide a mechanism through which ocean temperature controls the variability and provides predictability of European SAT.