Thursday, April 5, 2012
I wanted to draw attention to a new climate science paper Evidence linking Arctic amplification to extreme weather in mid-latitudes by Francis and Vavrus. The paper is behind a paywall, alas, but I will try to summarize the main idea here, since I think this line of research is potentially very important if it turns out to be correct.
To work our way into the paper in easy stages, we need to first refresh our memory of the atmospheric circulation from high school geography (at least that's where I first heard it, so long ago that the memory is translucent with age). The Wikipedia has a picture with the main points:
This picture shows a stylized cross-section of the circulation in the troposphere (the lower portion of earth's atmosphere in which temperature decreases with height and the atmosphere is well mixed). At the equator, it's hot and so the air rises, and as it does so it cools, which lowers the amount of water it can hold below the amount that is actually in it, causing rain, thunderstorms, etc. The resulting drier air eventually returns to the earth's surface around 30o north and south of the equator - this is the Hadley cell and the dry descending air is the cause of deserts at around those latitudes.
Similarly at the poles it is very cold, causing sinking air, and driving a similar polar cell which requires air to be lifted up at around 60o north and south of the equator. In between there is a (somewhat amorphous and loosely structured) Ferrell cell which moves air northward at the surface and southward high up. All three of these cells essentially serve to move heat from the equator (where the sun shines the most) to the poles (where it shines the least).
High up at the boundaries of the cells, air of quite different origins with large pressure and temperature differences meets. These differences drive strong winds, which you'd think would flow in a north south direction, but instead the rotation of the earth (Coriolis force) bends them around so that you end up with very rapid zonal (latitudinal) winds - the jet streams - occurring at high altitude.
Note that the diagram above shows the atmosphere sloping down to the left. This is because the warmer atmosphere at the equator is expanded (at lower density) relative to the cold air at the poles. These height gradients can be thought of as equivalent to the temperature differences that drive the whole atmospheric circulation.
The above description is of course incredibly simplified. And in particular, the jet streams do not pursue a perfectly latitudinal course at their respective positions, but instead meander about. You'll get almost as much insight as a chapter full of equations would give you by watching this little video (also from the Wiki) which shows the jet streams in the northern hemisphere as the red regions:
If you click on my screenshot above you'll get the Wiki video in a separate window and I highly recommend going through it a couple of times. You'll notice that the jet streams have big chaotic-looking meanders in - the one in the screenshot above is about the size of North America, and that the meanders progress in a generally westward direction. These are called Rossby waves (also known as planetary waves).
These things have a massive effect on the weather. If you think of the jet stream as being at the boundary of the polar cell and the Ferrel cell, it should be clear that when a big loop is south over your location, it's likely to be cold, whereas when you are in a northward tongue with the jet stream well north of you, it's going to be warmer. Similarly, since it's the mid latitudes where the regular pattern of cyclonic storms dominate the weather, the Rossby waves have a lot to do with where the storm tracks go and thus where it's wet and where it's dry.
As the Rossby waves are slow - often taking a week to cross the US - they have a lot to do with the persistence of the weather - the fact that things are often nice for a few days or a week, and then wet and miserable for a few days at a time. Indeed, anomalies in the jet streams can cause much longer patterns - sometimes a meander will become detached from the main course of the jet stream and just kind of hang around (one kind of what is referred to as atmospheric blocking) leading to weather that may not change for weeks or longer.
For example, in 2007, the jet stream over Europe got stuck in an unusual southerly configuration for a long period of time, and this caused extended raininess over the UK that eventually led to unprecedented flooding there:
So if the jet stream gets stuck - the Rossby waves stop or slow down - it can lead to weather extremes - one place may have flooding as rain continues for weeks or months, while another place may have drought as it stays sunny and dry there for an extended period.
With that background, let us now turn to the new paper where we will be focussing on the northern hemisphere polar jet stream (which mainly influences weather in northern Europe and the US). My readers will be well aware the the Arctic has been warming significantly faster than the rest of the planet, and in particular this has led to a fairly abrupt loss of summer/fall sea ice that occurred much faster than climate scientists were expecting:
This in turn means that the Arctic ocean absorbs a lot more sunlight in the late summer and fall, which it then gives back to the atmosphere in the winter. Because the atmosphere is then warmer than it used to be as a result, it is thicker, reducing the height gradient I talked about a few figures back. In particular, the paper finds these changes in the thickness of the atmosphere (to the 500hPa level):
Here the red regions shows a thicker atmosphere than historically present (and the white asterisks represent statistically significant changes). The upper left figure is for OND=October, November, and December where the atmosphere is much warmer and extends quite a bit higher than it used to. The anomalies persist into January, February, and March (especially over the north Atlantic) and then are smaller (but still present) in spring and early summer.
It is expected on theoretical grounds that if the height and temperature gradients between the mid latitudes and the poles are smaller, then the jet stream winds (which result from those gradients) will slow down a bit, and this in turn will cause the Rossby waves to move more slowly and become bigger and loopier. For example, the paper presents evidence of the northern most latitude of the Rossby wave gradually moving north during the fall over the last few decades (blue curve):
Similarly the fall and winter average speed of high altitude northern hemisphere polar jet stream winds has been dropping, particularly in recent years since the major collapse in sea-ice in 2007:
Still, I think the weather of the last few years has certainly caught everyone's attention in the northern hemisphere with extremes of snowiness in some places, extraordinarily mild winters in others, and then killer droughts in the southern US and unprecedented floods in Europe. This story - jet streams slowing down and meandering more slowly and loopily, causing weather to change more slowly and thus cause more extremes certainly has the potential to make sense of a lot of different phenomena. It also fits with the Hansen paper we discussed a while back which showed that not only are summers getting warmer on average they are also becoming more variable in their warmth:
If the weather is becoming slower changing, we would expect that to increase the likelihood of both unusually cool damp summers, and that of long hot summers, relative to whatever the current average is. And this applies even more forcefully to winters, when the Arctic amplification of global warming is most pronounced. This would make sense of the extraordinarily mild winter in much of the US this year, following on a particularly harsh winter the previous year.
So I think this is definitely something to follow. It's also worth making this point: if this turns out to be right, it will be the latest in a series of climate developments which were not well anticipated by climate science. For example in the relevant section of the IPCC AR4 (from 2007) there is no discussion of the possibility that Rossby waves might slow down and cause more weather extremes. This will be like the largely unanticipated early collapse in Arctic sea ice, or the discovery that the Greenland and Antarctic ice sheets were degrading at the edges much faster than expected, and through an unanticipated mechanism (basal lubrication).
In general, it appears to me that the climate system is more complex than climate science is currently able to fully model, and so it's constantly throwing up surprises in the way it reacts to the changes humanity is making in the atmosphere. This should not be reassuring.