Wednesday, September 04, 2013

Global warming and hurricanes


The North Atlantic hurricane season has begun, the time when hurricanes are more likely to form up. It is important to understand what is happening with hurricanes.

It is a mistake to attribute all hurricanes to man-made global warming, and it is an even bigger mistake to deny that our emissions of greenhouse gases can have any effect on hurricanes.

Experts are very cautious about confirming that hurricanes are getting worse. (Experts are always cautious, which is why it is so remarkable that all respected scientific societies have warned about global warming, and even more remarkable that media personalities such as Andrew Neil and politicians such as Owen Paterson and George Osborne should ignore the experts in this one case. But I digress.) 

One reason for this expert caution is that hurricanes have a quantifiable economic cost, and if it is reasonably certain that, say, 10% of a hurricane's cost is down to carbon emissions, the fossil fuel industry might find itself defending against a claim for 10% of the economic cost of a hurricane. And that would never do.

So it is important to find out the truth about hurricanes as far as we can.

First, the physics:
Hurricanes arise when certain factors coincide, mainly:
  1. Sea temperatures greater than 26C
  2. Low vertical wind shear i.e. little change of wind direction and strength with altitude
  3. High air humidity 
  4. Locations 20 degrees from the Equator so that the Coriolis force can set the air spinning
Global warming undoubtedly produces higher sea surface temperatures (SST) and therefore it will tend to produce more hurricanes. However, models predict that vertical wind shear is also increased by increasing temperature, so that tendency will be damped down somewhat.

Is there any evidence of increased hurricane strength with the small amount of heating we have experienced to date? 

Emanuel 2005 finds a convincing relationship between sea surface temperatures (SST) and hurricane energy measured as Power Dissipation Index, PDI



Click the graphs to make them bigger.  Major hurricanes are clearly getting more frequent globally, though southern hemisphere tropical storms (storms that may or may not develop into hurricanes) are becoming less frequent.

This map via Flemlose gives a good idea of where hurricanes are to be found:


Interestingly, Judith Curry, who always tends to lean towards the no-worries view of climate, is a co-worker in a paper (Webster, Holly, Curry Chang, Changes in Tropical Cyclone Number, Duration, and Intensity in a Warming Environment, Science 2005)  that concludes yes, hurricanes are intensifying. 

Now look at this graph :

Fig 4

Apologies for the quality. I made it by using MSPaint to superimpose a Wood for Trees graph  of Northern Hemisphere SST onto Ryan Maue's PDI graph. PDI, Power Dissipation Index, is a measure of hurricane activity. I chose Northern Hemisphere (NH) SSTs because most hurricanes happen in the NH as can be seen in Fig 3.

The composite shows a sketchy relationship between SST and hurricane activity. The correlation is not perfect, as is to be expected, since SST is not the only factor in hurricane formation. There is a pretty good relationship between the 26 year trend from 1972 and 1998, with hurricane activity rising in step with rising NH sea temperatures, but after 1998 the hurricane activity falls away, although sea temperatures do not fall. Note however that the hurricane activity is still greater in 2009 than it was in 1975.

However, if we match hurricane PDI with the Pacific Decadal Oscillation (PDO) a measure of El Nino/La Nina activity in the Pacific, this is what we see:

Fig 5

I obtained the trace of the PDO from the excellent Wood for Trees site.

There is a better match between declining PDO and declining hurricane activity.

The match is by no means perfect. There is a lag between rising hurricane activity 1990-93, although PDO leads hurricanes thereafter. There is also a mismatch in 1982-90, and another in 1970-74.

Let us add global sea surface temperatures in to the picture.

Fig 6

The green line is of global SST, again from Wood for Trees, but coloured in by hand in order to distinguish it from the PDO trace. The global SST supplies heat in 1972-4 where the PDO could not account for the hurricane energy. It supplies heat to start the hurricane spike in 1990, which is then continued by the rising PDO.

In short, we can say that between them PDO and global sea temperatures do relate to observed hurricane activity, though why PDO should predominate at some times and Global SST at others is an open question.

Of course, correlation is not causation, and we would need to integrate atmospheric humidity and wind shear to get more accurate correlation.

If we assume that the PDO is a strong component of hurricane activity, what can we expect in coming years?

Fig 7

Fig 7 shows a view of the PDO going back to 1900. The period 1920-2010 suggests that there is a 60-year cycle to the PDO. It is true that one and one half apparent cycles is not enough to establish the presence of a cycle. But equally, it may be taken as suggestive evidence for a 60 year cycle. If there is a 60 year cycle, we have had 10-15 years of cooling influence from the PDO (as evidenced in the hiatus in warming of surface temperatures), so we may be due for another 20 years of negative, cooling PDO, and therefore another 20 years of relatively quiet hurricane activity.

This however is not certain. Two additional factors come into play. First, the deep ocean has taken in a lot of heat since 1980, and this heat is going to have an effect. It will inevitably emerge at some point in the future, and it may act to disrupt any regular cycle that may exist. So we may not get 20 years of relatively quiet hurricanes.

Furthermore, there is another higher frequency cycle in the ENSO pattern.

ENSO is the El Nino Southern Oscillation, and refers to the ocean surface temperatures in the Pacific, rather than the whole Pacific region climate, which is what the PDO refers to.

Fig 8

Fig 8 shows the ENSO going back to 1880. In the 130 years covered, there have been 13 warm El Nino events and 14 cool La Nina events, giving a cycle length of 9-10 years. On this cycle, we are due to move into a decade of predominantly warming El Nino events in 2014 or 2015 that will last until 2024-5.

A 5 year period of cooler Pacific temperatures may follow, and then around the year 2030, a strong positive PDO will begin, with warmer global sea surface temperatures to delivering frequent, record breaking major hurricanes which will have significant economic impact.

Anyone who is under 60 years old and who lives in a hurricane-prone part of the world should take note of this projection.

More reading: Arthur Smith predicts warming in 2013
More on this blog: Atlantic Hurricanes relate to AMO

No comments: