Back to: Climate change article.
Back to: Evert's homepage.
Back to: Menu
Hurricane formation and climate change
When "Andrew" made its hit on Florida in 1992 local residents and authorities received a rather rude wake-up call on how vulnerable they were. The city of Homestead was ravaged. The total damage done was $ 30 billion, of which $ 17 billion was insured. A number of insurance companies went bankrupt.
The last really violent hurricanes which hit the area in the vicinity of Miami were
forgotten by most people. In 1926 a cat. IV (* see below) made a straight hit
on Miami destroying much of the - in those days - smaller city.
In 1928 the Lake Okeechobee storm took the lives of 2000 people, and in 1935
the Labor Day Storm, the most violent and vicious hurricane ever making landfall
in the USA (a cat. V), made its onslaught onto the Florida Keys, killing 400 on
the islands and making an end to Flaglers' Railway. The train on it perished in
The year 2005 provided a second wake-up call. The name was 'Katrina'. The final toll probably will end up to be over ±1500. The total damage (including all long term economic damage caused by 'Katrina' and 'Rita' – offshore oil & natural gas installations) may well end up in the order of $ 150 billion.
Hurricanes & global warming
This also raises the question whether people living in vulnerable areas should prepare for more violent hurricanes in the event of a warming climate.
In three articles (ref. 1, 2 and 3) the conclusion is reached that in case of a higher CO2 world the Sea Surface Temp.'s in the NW Pacific would rise by some 2°C and the intensity on the hurricanes would increase by 5 - 11% in terms of maximum wind velocity. Their central pressure would decrease by 7 - 24 mbar, their radius with hurricane force winds would increase by 2 - 3% and, probably most important in many cases, their near storm precipitation would increase with some 30% according to the model studies of ref. 2. On a qualitative basis these conclusions would also hold for other hurricane areas in the world like in the Atlantic and Caribbean.
Another result (from ref. 3) is that the occurrence of hurricanes will probably not increase. The availability of warm surface water with a temperature of over 26°C is a prerequisite for the formation of hurricanes. But also the amount of wind shear has to be low. They can only form under barotropic conditions. If warm water masses expand further North or South, they will reach areas with strong winds aloft. In these areas the formation of hurricanes is prevented. This point is illustrated by the Caribbean: Water temperatures are sufficient for the formation of hurricanes all year, but only in late summer and autumn low shear conditions prevail enabling the formation of hurricanes.
More recent publications point in the same direction. They are summarized on a website of the Geophysics Fluid Dynamics Laboratory. In an article by Thomas R. Knutson and Robert E. Tuleya (ref. 4) much more robust and comprehensive model study was made which resulted in an expected increase of 6% in maximum wind velocity, a increase of 14% in central pressure fall and an increase of 18% in rainfall near the core of hurricanes.
On the other hand: Up to now no real trend in hurricane strength and occurrence has been observed in The Atlantic and the Caribbean. The same holds for the Pacific. An apparent trend upwards in the earlier records can be explained by under-reporting in the satellite-less era. Some of the smaller storms were missed altogether and the moment of maximum intensity (that is how they are rated) was also missed in many cases. In the Atlantic Ocean and Caribbean the records from the start of air reconnessance (1944) are reasonably reliable. In the Pacific this is the case from about 1965 - 1970 onwards.
Summarizing: Up to now no upwards trend in strength and/or occurrence in hurricane activity has been observed. However: Based on model studies quoted above the intensity of hurricanes is expected to increase. On the other hand: Their occurrence is not expected to increase.
A recent publication by Holland and Webster (ref. 5) points in the direction of an increase in hurricane activity in the Atlantic basin. It should however be stressed that this results - up to now - only holds for this part of the tropics.
Damage done by hurricanes
One final remark on this subject: The escalation of damage done by hurricanes can be attributed to the fact that much more people live on dangerous ground today. About 20 million people live in Florida. In 1961 a hurricane (Hattie, like Mitch in 1998 a cat. V storm!) hit Belize and Honduras with torrential rains. It was, however, no disaster like in 1998. In Honduras it was the change in land use which turned a major inconvenience in 1961 into an all out disaster in 1998.
1: Thomas R. Knutson, Robert E. Tuleya and Yashio Kurihara
2: T.R. Knutson and R.E. Tuleya
3: A. HendersonSellers, H. Zhang, G. Berz, K. Emanuel, W. Gray,
C. Landsea, G. Holland, J. Lighthill, S.L. Shieh, P. Webster and
4: T.R. Knutson and R.E. Tuleya
5: Greg J. Holland and Peter J. Webster
"Heightened tropical cyclone activity in the North Atlantic:natural variability or climate trend?"
Atlantic Hurricane Data by Year
A graph of some Atlantic hurricane data of the period 1944-2014. Below the Saffir-Simpson scale:
Type Category Pressure Winds Surge mb knts mph ft Depression TD ------- < 34 < 39 ----- Tropical Storm TS ------- 34- 63 39- 73 ----- Hurricane 1 > 980 64- 82 74- 95 4- 5 Hurricane 2 965-980 83- 95 96-110 6- 8 Hurricane 3 945-965 96-112 111-130 9-12 Hurricane 4 920-945 113-134 131-155 13-18 Hurricane 5 < 920 > 134 > 155 > 18
From this the conclusion can be drawn that, up to now, no trend in hurricane
actvity, in terms of occurence, can be observed. It is however striking that
from the second half of the 90-ties on there has been quite some activity
in the Atlantic and Caribbean. Whether this is the harbinger of more intense
hurricanes due to global warming, or caused by a multidecadal cycle is still
an open question. (See below.)
The year 2005 might have been a once in a lifetime event. It was an incredible season, which ended with 'Zeta' at the end of the year (30/12!).
Another way of looking at this subject is considering the number of major (catagory 3 or above) landfalling hurricanes in the US. These data do not suffer from the drawback of missing smaller storms or under rating due to missing of the moment of maximum intensity (for the older data). Their strength at the moment of landfall is precisely known.
This gives rise to the following overview:
period Number of major landfalling hurricanes Cat. 3 Cat. 4 Cat.5 Total 1891 - 1900 5 1 6 1901 - 1910 5 1 6 1911 - 1920 3 2 5 1921 - 1930 3 2 5 1931 - 1940 6 1 1 8 1941 - 1950 9 1 10 1951 - 1960 5 3 8 1961 - 1970 4 1 1 6 1971 - 1980 4 4 1981 - 1990 3 1 4 1991 - 2000 3 1 4 2001 - 2010 4 3 7No clear upward trend can be observed.
One more graph including some older data is shown below. However, it should be borne in mind that these data (definitely before the 1920-ties) are (much) less accurate in terms of hurricane strength.
How much energy is transferred by Hurricanes?
An interesting question is: How much energy is transferred by Hurricanes?
It is well known that hurricanes mix warm surface waters with cooler water within the
thermocline, leaving colder wakes that over a period of weeks are restored to normal
conditions by mixing and reheating by the sun. This restoration is associated with net
heating of the oceans to some depth. This implies a net export of heat from the regions
affected by the storms.
The conclusion from this with respect to the ideas of "taming" hurricanes is that caution should be exercised when considering this. 1.4 1015 Watt is substantial if one bears in mind that the net solar energy flow towards the Earth is 120 1015 Watt!
Ref: K Emanuel,
Some notes on the Atlantic Multidecadal Oscillation
When considering the question whether there is a link between global warming and hurricane formation the Atlantic Multidecadal Oscillation (AMO) is one of the complicating factors.
How does it work?
The tropical part of the Atlantic Ocean and the Caribbean basin are enjoying high sea surface temperatures. Due to this - in average - the tropical storm activity is high. It also results in ample rainfall, making the ocean water less salty. When this less salty water reaches the Arctic it leads to a weakening of the gulfstream. In its turn this causes some cooling, less tropical storm activity, and less rain. Due to this the ocean becomes more salty and finally the gulfstream becomes stronger again. This leads to higher temperatures. The circle is closed.
The picture below shows this oscillation:
The increased hurricane activity after 1995 fits quite nicely in this overall picture.
However, one may raise some doubt to this story when one considers the hurricane statistics more closely. In the table below it is shown. 'TS' stands for the number of tropical storms in the considered decade, '1' for the number of category 1 hurricanes, etc. up to and including 5. 'Katrina', 'Rita' and 'Wilma' were the most recent category 5 hurricanes.
The jump after 1994 is now much more striking. The question arises whether this is still part of the AMO pattern, or that it is the harbinger of a change in the tropical storm climate in the Atlantic and Caribbean basin. This isn't yet completely resolved, as the hurricane activity and intensity in the Atlantic and Caribbean basin increased after 1994, but didn't, up to now, do so in other parts of the world.
Ref: Richard Kerr, 'Atlantic Climate Pacemaker for Millennia Past, Decades Hence?',
Some on links between warmer sea water and more intense tropical storms
Various publications on the intensification of tropical storms due to higher sea surface temperatures have appeared in the open literature. Also a publication by M.E. Mann en K.A. Emanuel (Eos, 87, (24), 13 June 2006, 233-244) was made in which doubt was cast on the role of the Atlantic Multidecadal Oscillation. (see the first 2 references below)
However, lately I decided to take a closer look at the developments after 2010. It is striking that - except in case of the Atlantic and Caribbean basin - there still isn't a clear increase in hurricane intensities visible. First one can take a look at the numbers sorted by category.
There is no clear trend visible in the Pacific, while for the Atlantic and Caribbean basin from about 1970 onwards there is. However, one may say the counting tropical storms by category is rather rude. A small storm is counted on an equal basis as a much bigger and langer lasting one of the same category.
One can also "count" them by means of the so called accumulated cyclone energy index, in which the square of the velocity (in knots) is counted each 6 hours during which a storm has tropical storm force or higher. This leads to the following results.
The picture remains the same: No clear trend visible in the Pacific; from about 1970 onwards an increase in
numbers and intensity in the Atlantic and Caribbean basin.
Finally on can consider a measure introduced by Kerry Emanuel, the so called power dissipation index in which the wind velocities cubed are counted. This is reasonable, as the power dissipation is proportional to the power 3. Think of "Betz' Law" for windmills.
He fitted the sea surface temperatures against the power dissipation index and obtained a quite handsome fit which is shown on the left side of the figure below.
As I couldn't find data on power dissipation indexes on the internet (and generating them myself would be too
elaborate) I decided to use a fit berween the accumulated cyclone energy index and the power dissipation index
given in the picture below (see reference 3):
Using the 1.36 power "law" I then attempted to reconstruct the PDI figures of Emanuel. The result is given in the right side of the figure given above the ACE-PDI fit. The resemblance is quite reasonable. Using the 1.36 power "law" I then made a figure showing the power dissipation index from the accumulated cyclone energy index data. These are shown below.
When I compare these with publications I found on the internet (on ACE
and PDI) the similaritity is quite reasonable. As the number of
hurricanes is quite constant (see also reference 4) this leads to
the conclusion that - except the Atlantic and Caribbean basin - there is no clear increase in tropical storm activity.
One cannot draw from this the conclusion that there is no link between a warming climate and the intensity of tropical storms. There are good thermodynamic reasons to expect such a link to be present. However, there are probably other factors - e.g.: think of wind shear and the thickness of the warm water layer in the tropical oceans - at play which currently prevent a rise in intensity.
1. "Atlantic Hurricane Trends Linked to Climate Change",
Back to: Climate change article.
Back to: Evert's homepage.
Back to: Menu