25 I feel that someone in the Western Pacific should say something
about storms there, though this is not the area of my expertise.
(Typhoon experts, help!)
I have browsed the papers of Emanuel (2005) and of Chan and Liu
(Though my institution has subscription of "Nature", the file of
Emanuel's paper which I got was an incomplete one --no figures, no
math formulas, no substantial list of reference. But a complete PDF
file is available from the author's site at MIT
), and a PDF of its
supplement is also there.)
Chan and Liu's paper discusses correlation of year-to-year values
between SST and indices of strong TCs. It is a not study of long-term
trends. Chan and his colleagues have already discussed the
relationship between ENSO and Western North Pacific TCs in Wang and
Chan (2002) and other papers. This time they examine whether local
SST is important in addition to ENSO, and their answer is "no" in the
year-to-year time scale. Their story is clearer when variability in
the longer (interdecadal) time scale is excluded. From that part I
infer that the correlation between local SST and TC indices is likely
to be positive in the longer time scale, though it requires a
specific study to establish such a relationship.
I am a little surprised to know Chan's previous finding that Western
Pacific TCs are more active in the El Nino phase of ENSO. My
understanding which had not been updated since 1980s was that El Nino
suppresses cumulus convection in the Western Pacific and thus
suppresses TCs there as well. I still think that the conclusion
depends on the target area and seasons. Chan and Liu took 120 - 180
E, May - November. I think that suppression of TCs by El Nino
prevails in the western part of the Western Pacific, and mainly in
winter and spring.
It is true that the area with SST above 27 deg. C in the Central and
Eastern Pacific is larger in the El Nino phase. Probably the high-SST
area in the whole Pacific is larger then. But El Nino also tends to
suppress TCs in some regions where local SST is high enough. Probably
the correlation between the high-SST area and the total TC activity
is positive, but it is a result of spatial aggregation of complicated
Emanuel (2005) shows the correlation of _smoothed_ time series of SST
and "Power Dissipation Index". I think that the smoothing reduces the
ENSO signal, but that it does not eliminate it. Thus it is difficult
to connect the discussion of the paper with the "ENSO and the rest"
view of Chan and Liu.
Emanuel has made a good effort to compensate for the inhomogeneity of
data quality, as he describes in the supplement. But, it is still a
As Morita and Watanabe (2005) reported, the "best track" data shows
decrease of the frequency of strong TCs in the 1990s in the area 15-
30 N, 120-150 E.
(Morita refers to a data set compiled by Japan Meteorological Agency.
I think it is essentially the same as the Western Pacific part of the
JTWC data set used by Emanuel and by Chan, but I have not confirmed
In some more detail, TCs with central pressure lower than 920 hPa
decreased, those around 950 hPa increased, and those around 980 hPa
decreased. Though it cannot be denied that these are real trends,
Morita suspects that these are artifacts due to changes in observing
practice. Aircraft reconnaissance in the Pacific was phased out in
1987, and since then surface pressure have been determined by
satellite image interpretation (Dvorak method) except occasionally by
island stations or ships. Morita's results suggest that the image
interpretation underestimates very strong TCs (typhoons) but somewhat
overestimates moderate ones (TSs). I am not sure whether this causes
significant bias in Emanuel's PDI. Also, Morita's observation is
about the specific region. Dvorak method may have different
sensitivity in different climatic regions.
When we discuss whether the influence of global warming has appeared
in TCs, there is a fundamental problem that we are not very sure
about theoretically how TCs should react to (greenhouse-gas-induced)
For middle latitudes, extratropical cyclones are the principal actor
in the energy cycle in the atmosphere. Therefore the total power of
cyclones (the number times the intensity) in the whole mid-latitude
zone should somehow correspond to the global forcing. If the
characteristics of transient warming is similar to those of
equilibrium warming, the north-south gradient of temperature will
decease, and therefore the total power of extratropical cyclones will
(Things may not be so simple, however, even in this zone. Increased
water vapor will provide energy by latent heat release and somewhat
compensate for the loss. And perhaps more important role of
moistening is to increase inhomogeneity within individual cyclones.
Surely the maximum rainfall rate will increase. Also the maximum wind
speed may increase despite of decreased average value.)
There is no guarantee for such large-scale determinism for TCs. The
essential feature of the tropical atmosphere is cumulus convection,
whose individual horizontal scale and time scale are of the order of
1 km and of hours. The largest-scale feature is the Hadley
circulation whose upward branch is none other than the collective
activity of cumulus convection. The Hadley circulation requires
cumulus convection, but it does not require TCs. Whether cumuli
organize into TCs, easterly waves, Madden-Julian oscillations or
something else, or maybe remain rather random, is not constrained by
If we assume large-scale determinism, we can have some conclusions
which will be valid as long as the assumption is valid. Emanuel's
previous theoretical work (Emanuel, 1987) assumes that there is a
circular vortex coupled with convection, and discusses how strong it
would be. The GFDL model has more degree of freedom, but their
experiments assume a circular vortex as the initial condition. It is
reasonable that results of these studies are mutually consistent.
Global warming experiments with a "20 km grid" (actually spectral)
GCM of the Meteorological Research Institute (MRI, of Japan) shows
intensification of strong TCs (consistent with the GFDL model study),
and increase of the life time of individual TCs (as Emanuel
suggests), but also decrease of the total number of tropical storms.
(Unfortunately the only on-line information about that study I have
found is a short abstract for the previous AMS meeting (Oouchi et
al., 2005). Off-line and in Japanese, there is a little more
information in the abstract volume of MSJ 2005 Spring Meeting
(presentation Nos. A203 and A204)).
I am not sure whether the collective total power of TCs, or Emanuel's
PDI, increases or decreases as climate warms in that model.
A caveat is that all GCMs as well many TC models (including GFDL's)
that have been used for climate change experiments employ hydrostatic
approximation and "cumulus parameterization". They assume some ways
of self-organization of cumulus convection which may or may not be
true. Non-hydrostatic, cloud-resolving models are promising in
reduction of this kind of uncertainty. But, climate simulation with
these models requires much more computer resources than currently
available. Whether the society can afford the cost is another problem.
Another difficulty for the 21st-century projection of TCs (and also
for the projection of tropical climate in general) is that it is not
certain how ENSO behaves as the mean climate warms. With the same
scenario of greenhouse gas concentration, some coupled GCMs produce
more EN-like climate, and others less EN-like. ENSO may be a kind
of "free" mode which can shift either way.
We can say some small thing relatively confidently for those mid-
latitude areas which are sometimes affected by TCs. Warmer _local_
SST helps maintain TCs which happen to arrive there. Thus, we will
encounter more cases of strong TCs there _unless_ the situation at
the area of TC generation changes much.
* Emanuel K.A., 1987:
The dependence of hurricane intensity on climate.
Nature, 326, 483-485.
* Morita M. and Watanabe S., 2005:
On a problematic issue of the Dvorak method for observations of
typhoons---Have typhoons really become weaker than before? (in
Meteorol. Soc. Japan 2005 Spring Meeting, Presentation No. C201.
* Oouchi K., Yoshimura J., Yoshimura H., Mizuta R. and Noda A., 2005:
Tropical cyclones in a greenhouse-warmed climate: a projection from a
20-km mesh global climate model.
Amer. Meteorol. Soc. 2005 Annual Meeting, Suki Manabe Symposium, P1.1.
* Wang, B. and Chan J.C.L., 2002:
How strong ENSO events affect tropical storm activity over the
western North Pacific?
J. Climate, 15, 1643 - 1658.
Comment by Kooiti Masuda 16 Aug 2005 @ 4:56 pm
25. Kooiti Masuda
I have to write again about event horizons. The global computer
models are crap for more than a lack of computing power. Even if the
computer was as large as the earth itself, it would be GIGO.
If your change your set of assumptions there is another way.
It is essential to discuss event horizons in climate.
For instance the Dane research on cosmic ray flux--with glacial
epochs associated with the movement of the solar system up and down
in the plane of the galaxy--this is a good example of an event
horizon that has long term climate predictability. The spinning of a
hurricane--you cannot even predict mesovortices an hour ahead of time-
-so now you want to use the same event horizon and explain what is
happening now, what is happening 100 years from now? Terrible.
Junkscience! The math shows you that turbulance and viscosity limit
future behavioral prediction. So we have to look at other mechanisms
that we know with certainty have predictable event horizons. But
then the critical comments here are that with warmer climates come
different baratropical behaviors. It's the same problem, stated
differently. The idea of ENSO as a factor is a better approach,
because the timescale of ENSO has a longer event horizon, and there
are correlations of ENSO with tropical storm activity. Then there is
the SSTs, which again is the same problem--as they too haven't
necessarily been directly associated with increases in both intesity
and frequency of tropical storms.
So let me digress a moment and come back to this. The weather
community is dominated by those who study and consider baratropical
behaviors--as well as they practically should. However, the viscosity
issue becomes dominate over longer timescales--where ELECTRICAL
influences on cloud microphysics start to dominate. The turbulance
problem is then like talking about turbulance of the water in a pipe--
all you need to really talk about is the pipe to know where the water
is going to go. And while large scale electrical features hold the
key, none of the so called 'skeptics' have looked at electrical
behaviors as it pertains to climate--as they are not trained as such:
Of course, meteorologists are also lacking in such training. And I
say this with real respect, as my father himself is a meteorologist
and I was born on in the early sixties on an Air Force base where he
gave weather to the SAC pilots flying B-52s. Pure respect for this
profession--but it isn't a profession where there is training in
Finally there are the 'warmers', and they too approach the problem of
cloud dynamics as a direct feedback heating. They tend to, if they
have education on what the GHG physics really are, a corresponding
lacking education in electro magnetic behaviors, and specifically
what electro magnatic properties CO2 has in the oceans.
ENSO itself because it is a longer time scale cloud dynamics behavior
is ELECTRICALLY forced. The SOI is largely about the relative
depressurization and stirring of the oceans either over Darwin or
Tahiti and then one end gets discharged, like a flat beer, per Bates
et al Nature on Hurricane Felix, and the conductivity dynamic changes
and capacitive couplings between ocean and ionosphere which organize
cloud microphysics flips back to the other end of the tropical
Pacific. Where ENSO comes in is the sustained winds start to have
induction meaning with the moving salt spray and surface, and, again,
SSTs factor in about a drop in one percent of resistance per increase
of one degF. For purposes of this discussion, I will leave out cold
upwelling of initially more resistive water, but that same water
contains nutrients for large scale biological conductivity factoring
here, and I will also leave out space weather, and its implications
on the ionosphere and the capacitive couplings involved.
I will only talk about ENSO in terms of CO2 and the general
temperature of the Pacific. This is critical because ENSO really has
only been around in the 2 to 7 year peridicity for only 5,000 years.
You go into the Wisconsonian and El Nino may come every 20 years or
not at all. Then it comes up in still other climates long ago, but
the point is, the oceans warmed, CO2 increased, and ENSO came. If
you think only about the problem thermally, only think about the
oceans warming overall, it doesn't really provide explaination for
what ENSO 'is', or why it came about 5,000 years ago, and not, say,
12,000 years ago or 8,000 years ago during other warm periods
preceeding the Wisconsonian.
But an electrical view reveals mechanism of the climate change. As
the oceans warm up, the induction meaning of the sustained winds
along the tropics starts to become more significant relative to
capacitive couplings of ocean and ionosphere. It has to because
induction has to overcome the relative conductivity changes that
occur from the SOI depressurizing and stirring one end of the Pacific
while the other recharges. Eventually one end gets 'flat' from
carbination coming out of solution and the other end is more prone to
microphysics organizations and hence pressure changes.
When an El Nino event occurs, its induction meaning is so significant
and such a concentration of the energies of the global electrical
circuit that other areas cannot be so organized. That's why
hurricanes won't occur as well during an El Nino.
Now, the idea that there is a different event horizon with and
electrical approach should be more clear. That is, no matter weather
now or 100 years from now, CO2 concentrations will have very specific
conductivity meaning as it comes out of solution from the low
pressure and winds of a tropical low, where it rises to the ocean
surface and then goes back there into solution and rises ion counts--
exactly where capacitive couplings with the ionosphere then occur and
alter cloud microphysics in the more intense field and relatively
organize a tropical storm.