Daily tropical cyclone intensity response to solar ultraviolet radiation
J. B. Elsner, T. H. Jagger, and R. E. Hodges
GEOPHYSICAL RESEARCH LETTERS, VOL. 37, L09701, doi:10.1029/2010GL043091, 2010
Abstract: An inverse relationship between hurricane activity over the Caribbean and the number of sunspots has recently been identified. Here we investigate this relationship using daily observations and find support for the hypothesis that changes in ultraviolet (UV) radiation rather than changes in other concomitant solar and cosmic variations are the cause.
The relationship is statistically significant after accounting for annual variation in ocean heat and the El Niño cycle. A warming response in the upper troposphere to increased solar UV forcing as measured by the Mg II index (core‐to‐wing ratio) decreases the atmosphere’s convective available potential energy leading to a weaker cyclone. The response amplitude at a cyclone intensity of 44 m s−1 is 6.7 ± 2.56 m s−1 per 0.01 Mg II units (s.d.), which compares with 4.6 m s−1 estimated from the heatengine theory using a temperature trend derived from observations. The increasing hurricane response sensitivity with increasing strength is found in the observations and in an application of the theory. Citation: Elsner, J. B., T. H. Jagger, and R. E. Hodges (2010), Daily tropical cyclone intensity response to solar ultraviolet radiation, Geophys. Res. Lett., 37,
L09701, doi:10.1029/2010GL043091.
Figure 2. Upper air temperature and tropical cyclone intensity response to variations in solar UV radiation. (a) Change in 50 hPa temperature for a change in Mg II index given the Mg II index exceeds the given percentile. The first point to the left is the ordinary least
squares regression coefficient of temperature on Mg II index using all but the lowest 10% of the Mg II values. The next point is the regression coefficient after removing the lowest
20% of the values, and so on. The point‐wise one standarderror band is shown in grey and is computed using a sandwich estimator to account for the autocorrelation in the daily values. (b) Observed and theoretical response of tropical cyclone intensity to variations in solar UV radiation (Mg II index). The observed response is a change in a percentile of tropical cyclone wind speed for all values of Mg II index. The theoretical response is the change in a percentile of wind speed for a set of temperature responses to Mg II index values exceeding a given Mg II index percentile. The solid curve (circles) and the 90% confidence band is based on a bootstrap resampling of the daily data. The dotted curve (squares) is based on equation (2) with the temperature response estimated from NCEP reanalysis data.
Conclusions
Here we show compelling evidence that the relationship between hurricane intensity and solar activity on the daily time scale is physically linked to changes in atmospheric temperature near the top of the cyclone induced by UV radiation. This new finding sheds light on the problem of forecasting hurricane intensification. The overall greater
sensitivity of the response found in the tropical cyclone wind data compared with the heat‐engine theory and temperature data might result from the tropical cyclones themselves warming the temperature aloft and thus dampening the temperature‐UV relationship [Swanson, 2008]. It is noted the theoretical results reflect a change in the maximum potential intensity of a particular tropical cyclone while the observational results reflect a change in the daily maximum wind speed over all tropical cyclones in the
region. Since a tropical cyclone plays a role in moistening the stratosphere [Romps and Kuang, 2009] and since the dissipation of the cyclone’s energy occurs through ocean
mixing and atmospheric transport, a tropical cyclone can act to amplify the effect on the Earth’s climate of a relatively small change in solar output. On longer time scales it is
noted that a portion of the variation in tropical SST’s (0.08 ± 0.2 K) lags the Schwabe cycle by 1 to 3 years, which is roughly equal to the time required for the upper 100mlayer of the ocean to reach radiative equilibrium [White et al., 1997].
=============================
Leif Svalgaard who brought this paper to my attention, has a full copy available for review here
He also mentions:
But see also: http://www.leif.org/research/MgII%20Calibration.pdf
Their MgII proxy for UV [see their Figure 1 is not correct. It is unclear how much that influences the result. Also, I have alerted the lead author. He said he would ‘look into it’.
We’re all gonna die?
DirkH
We are all doomed, at least statistically. The real question is: Shall we survive statistics?
This may be coincidence instead of causality. I would think that any temperature changes, no matter what the cause, would have a lag time of a couple years or so. Looking at the system adiabatically, one would think a temperature gradient of any kind would seek its own level. In a large atmosphere, with all types of sinks operating, that should take time before equilibrium is reached.
I have often thought that it is the switch from one set of conditions to another, and the rate of such change that produces the greatest disequilibrium. Looking at the geologic record, since cold cycles seem to develop more rapidly compared to warm ones, I would expect weather extremes would be most severe at the onset of a cooling cycle.
So the ability of a cyclone to dump the earths heat into the stratosphere and outer space is subject to the random chance of dampening by high levls of UV?
Sound like a random walk, along the big heating cooling highway that the earth has.
Interesting theory, the variable UV’s effect on the formation of O3 and consequent O3 stratospheric “greenhouse” effect seems to me to have the potential for a greater climate effect than the cosmic ray hypothesis. With a very measurable effect during this last solar minimum. And any research into this area will be interesting….
Because of the tilting of this planet is in movement, ONLY one point on this planet recieves DIRECT sunlight. The rest of the planet the sunlight angles and travels further through the atmosphere. At the poles, the angle of sunlight to the surface is thousands of more miles both in mileage and through the atmosphere.
NOW, let us add cloud cover. Some clouds to the pole on an angle from the sun is now hundreds of miles long. More cooling, the denser the cloud cover towards the poles.
I have forgotten the rotation and angles of sunlight reflection when rotated.
“An inverse relationship between hurricane activity over the Caribbean and the number of sunspots has recently been identified”
Not the Gulf of Mexico, not the Atlantic, not the Pacific, just the Caribbean.
The amount of variance of tropical cyclone intensity explained by solar activity has to be necessarily small. Since many of the convective bursts which intensity hurricanes dramatically occur at night, it is difficult to extricate the cause and effect of solar activity changes.
Skeptical.
REPLY: If you are talking about straight linear energy transfer, I’d agree. However, I’ve often suspected there is some sort of transistor effect going on between the solar TSI (true TSI, all energy sources, not just visible) and earth’s atmosphere. -A
Tom, just like Lockwood’s solar influences winters in the UK, only.
Back 1000 years sailors thought they may sail off the Earth as they thought it was flat.
SCIENCE HAS MADE THE SAME ERROR!
This planet is NOT flat and it is NOT a cyclinder.
This is a fast rotating orb that deflects sunlight at angles. Not like in pictures straight down and straight back up.
Cyclones start from the atmosphere down to the surface. This changes the DENSITY of the air as it is being comprssed by speed. Over water the energy increases and pulls more moisture into it.
Anthony, you need to change the format of your blog to allow in-line comments.
The new Elsner paper is a statistical study, and the theory or meteorology content is rather light. Indeed, this is the case with many tropical climatology papers recently, not to pick on just Elsner’s work. This does not mean that the findings are incorrect, just that they are not fleshed out with appropriate additional testing such as modeling. This is one of the disadvantages of GRL journal. Stuff gets published quickly that is rather incomplete, but I guess that’s its mandate.
The upper-level temperature data is a bit sketchy as well considering it is the NCEP-Reanalysis. I have done a simple comparison with the ERA-Interim reanalysis for stratospheric temperatures and found the RMS difference between the two datasets on a monthly time scale quite large — larger than the signal that Elsner is correlating to. It is a cottage industry for folks to use the NCEP reanalysis datasets willy-nilly.
REPLY: Sorry, I have no control over making inline comments available to users, it’s an administrator privilege. Yeah I agree on the upper level NCEP issue. -A
Ryan N. Maue says:
May 5, 2010 at 4:03 pm
REPLY: If you are talking about straight linear energy transfer, I’d agree. However, I’ve often suspected there is some sort of transistor effect going on between the solar TSI (true TSI, all energy sources, not just visible) and earth’s atmosphere. -A
Leif once responded to a question on TSI that it was TOTAL solar irrradience – all energy is absorbed by the sensor and causes heating which is measured. Not just visible colors but all wavelengths.
It doesnot of course measure magnetic waves, gravity waves or other esoteric energies
/harry
Harry Lu says:
Leif once responded to a question on TSI that it was TOTAL solar irrradience – all energy is absorbed by the sensor and causes heating which is measured. Not just visible colors but all wavelengths.
It doesnot of course measure magnetic waves, gravity waves or other esoteric energies
These ‘other’ energies are so low that they are unmeasurable anyway…
Well I don’t like that word “relationship.” What are the SI units of “relationship” ?
I’d rather hear about some real observed Physical mechanism for solar UV (not some other solar phenomenon) causing more intense hurricanes.
A “relationship” is a couple of guys holding hands in a bar (a San Francisco bar anyway); it is not a physical mechanism for some observation.
Some have observed a “relationship” between the stock market behavior, and the results of some election or other; doesn’t mean one causes the other.
Isn’t it about time that we stopped assigning physical meaning to statistical correlation; and pay more attention to process or mechanism instead.
Harry Lu
One obvious difference with UV is that it has the energy to breakup O2 to atomic O, which facilitates O3, which also absorbs uv… And from what i have read on the subject recently it should effect stratospheric temperatures… (and by extension global climate.) http://journals.ametsoc.org/doi/pdf/10.1175/1520-0469(1979)036%3C1084:TROSOI%3E2.0.CO;2
And this nasa article below also raises the question if reduced UV effects upper atmospheric temperatures. And points to some evidence to back it up.
http://www.nasa.gov/topics/earth/features/coolingthermosphere.html
I have no idea as to the effect on the troposphere, but the second law o thermodynamics suggest it should have some effect… just an area i have recently found interesting(thanks to science o doom)
REPLY: If you are talking about straight linear energy transfer, I’d agree. However, I’ve often suspected there is some sort of transistor effect going on between the solar TSI (true TSI, all energy sources, not just visible) and earth’s atmosphere. -A
Ryan N. Maue says: “Leif once responded to a question on TSI that it was TOTAL solar irrradience – all energy is absorbed by the sensor and causes heating which is measured. Not just visible colors but all wavelengths.
It doesnot of course measure magnetic waves, gravity waves or other esoteric energies”
There are more things in heaven and earth, Horatio, Than are dreamt of in your philosophy.
There are NINE data points in this supposed connection between UV and Caribbean hurricanes.
What ever happened to out-of-sample testing? Or common sense?
Ryan N. Maue says: May 5, 2010 at 4:03 pm
REPLY: … However, I’ve often suspected there is some sort of transistor effect going on between the solar TSI – and earth’s atmosphere. -A
I assume you are talking about “amplification” a small change in TSI causes a large change in hurricanes/temperature.
However, a FFT of temperature shows no significant TSI influence.
SSN FFT
http://www.leif.org/research/FFT-SSN-Monthly-1755-2007.png
Hadcru FFT
http://www.leif.org/research/HADCRU-FFT.png
No real sign of the SSN frequencies in the temp plot
What is being amplified to modulate hurricane activities.
UV?
But then lots of this gets filtered (converted to heat) by ozone and it is only a small part of TSI at TOA.
jorgekafkazar says:
May 5, 2010 at 6:09 pm
There are more things in heaven and earth, Horatio, Than are dreamt of in your philosophy.
Is plain nonsense. If they have effects, then they are ‘dreamt of’ [show themselves], and if they don’t have effects then they don’t matter…
Let’s chuck the modeling, m’kay? All modeling does is test the assumptions fed into the model.
Yeah but flip it on its head….
“Man made cyclones cause sun spots”.
We’d need a new branch of environmentalism for that, we’d end up with Gen Y kids painting their faces yellow, in support of the nearby star we’re polluting.
The WWF would come out and condemn us all for killing the ‘Lesser Spotted Sol’, the dark patches on its otherwise perfect orb show the impact of SUV’s, Air conditioners and Fast Food culture.
Millions would march on the worlds capitals, demanding we do something to save the Sun, and then someone would come up with a really expensive solution….hmmm.
/End Post Normal thought experiment.
Hurricane intensity depends on two parameters namely the sea surface temperatures and the temperature at the tropopause.
The bigger the temperature difference then all other things being equal the more intense hurricanes (indeed all convective weather systems) can become.
Pointing to uv levels (or overall solar activity) alone is only half of the equation and may simply be coincidental on the basis that uv levels are usually higher when the sun is more active. It has been pointed out many times on this site that in absolute terms the energy value of uv and other solar variations is pretty small.
However (if one ignores CFCs) the temperature of the stratosphere does appear to fall when the sun is more active which weakens the inversion at the troposphere and would allow an intensification of global convective activity as was observed during the late 20th century though some observers suggest that overall hurricane intensity varies little over time. There are considerable problems measuring total intensity levels but for the moment I am inclined to accept that hurricanes and convective activity in general was greater during the late 20th century than now and in the mid 20th century.
Now that the sun is less active the stratosphere appears to be warming slightly thus increasing the strength of the inversion at the tropopause and convective activity does seem to have reduced with weaker and less numerous hurricanes.
It seems to be the case that when the sun is more active the polar oscillations are more positive which reduces ozone in the stratosphere (despite inreased uv levels) causing the observed stratospheric cooling. I know that Leif thinks that the explanation is human CFC’s but I am inclined to disagree.
Thus the possible chain of causation would be as follows:
i) More active sun.
ii) Faster energy flux upwards from the stratosphere
iii) Cooling stratosphere.
iv) More positive polar oscillations.
v) Weaker inversion at the tropopause.
vi) Greater temperature differential from surface to tropopause.
vii) Enhanced cyclonic activity globally.
Bear in mind that that is only half the story. A warmer ocean surface will further enhance the process. A cooler ocean surface will suppress it.
The longer the current solar quietude continues in parallel with the more negative polar oscillations the more likely to be true this proposition becomes.
If we keep a quiet sun but nevertheless observe a long spell of more positive polar oscillations then that will serve as a falsification.
I know that the proposition seems to be falsified already by reconstructions of past polar oscillations and solar activity levels but I suspect that those reconstructions are sufficiently imperfect in timing and scale to obscure the correlations. I place far more credence in continuing observation.
http://www.realclimate.org scenario http://news.minnesota.publicradio.org
Stephen Wilde says:
May 5, 2010 at 11:31 pm
However (if one ignores CFCs) the temperature of the stratosphere does appear to fall when the sun is more active
There is general agreement that the stratosphere warms when the sun is more active [more UV], that the recent decades cooling is due to increasing CO2, and that the recovery of the ozone due to decreasing CFCs has stopped or diminished the cooling trend, so your ideas don’t work in face of the observational evidence.