As I pointed out a few days ago, we had a nice pair of Equatorially trapped Rossby Waves (or Rossby-Smossby waves... or, indeed, Edson Waves), one in the Western Pacific and one in the Central Indian Ocean: each wave having an equatorial westerly jet and a pair of symmetric vortices.

Well... something interesting has happened over the past three days:  For both pairs, the whole pattern has moved northward such that we still have a well-defined cyclonic vortex in the Northern Hemisphere; but the Southern vortex has moved northward with it, and now sits on the equator.......Anyone have any insight on this?

The relevant images (large scale 850 hPa analysis, shading representing vorticity) are on my web-page.

John McB

Three days ago:

Two Days ago

One Day ago

Latest analysis with the Southern Vortex, in each Ocean, now sitting squarely across the equator

John Molinari
John,

         You have raised an important question.  Mike Dickinson and I in our recent JAS article found similar behavior in a different equatorial mode, the mixed Rossby-gravity wave.  The waves seem to remember their  origins (i.e., keep the structure they had at the equator) for a while even after they move away from the equator.  Michael Reeder suggested to us that the waves are refracted off the equator by a spatially varying background  flow.  The latter can be caused by the MJO or some other broad convective forcing.  To my knowledge, this issue is still open.  I think we need some help from theoreticians to resolve this...

         Cheers,
         John

Pat Harr

John,

The large-scale events associated with the monsoon depression/TS 26W have certainly undergone an interesting evolution.  The formation of the depression and subsequent organziation into 26w seems to be clearly related to the Eq.
Rossby-wave activity.  Now, if one had looked at the maps for the first time at 1200 UTC 9 October it would seem that 26W was associated with the transformation of a MRG wave to TD-waves as discussed by Takayuba and Nitta, Mike Dickenson and John Molinari, and Hung-Chi Kuo and collaborators. As John M. mentioned, the transformation from MRG to TD wave has been associated with the varying background flow in that part of the basin.  Although the MJO signal has not been strong, the Eq. westerly winds associated with the Eq. Rossby waves surely must have contributed to low-level convergence that has been associated with the wave transformation.

An additional mechanism to be considered is the influence of beta-induced wave train from 26W as the very large circulation moves northward.  In terms of the circulation pattern as of late (9-10 October and most likely coming days) the question would be is the signal a MRG wave-TD wave transformation or dispersion from the TC to the southeast?  The large-scale flow pattern probably contains some contribution from all of these factors.

This seems to be a period when some careful decomposition of the flow over the past week using tools such as those defined by M. Wheeler to be based on specific wave criteria might yield some insight as how the flow evolved from
the Eq. Rossby wave pattern to the MRG-TD wave pattern or wave dispersion from the TS circulation patterns.

Pat.

John Knaff

Folks,

I'm going to make speculation on John M's observation.  I apologize for not showing up with a reference for such an idea, but the northward movement of this complex may be explained by the northward drift of the dynamical equator,
which by my recollection basically divides Southern and Northern Hemisphere air masses (inertial flows).  Since these equatorially trapped systems move along the dynamical equator, if the dynamical equator moves north then so does the
equatorially trapped system.  I think there is sufficient cross equatorial flow associated with 26W to have this affect.

John Molinari

John, John, and Pat,

        I believe the dynamic equator implies that the vorticity = 0 contour moves  away from the equator.  This could occur, for instance, if negative  vorticity occurred at very low latitudes in the Northern Hemisphere, which  would make absolute vorticity = 0 off the equator.  If this is correct, then it would actually most likely be -du/dy < 0 that would move the
dynamic equator rather than strong v components.  [But if the strong south  winds were unbalanced, they would be deflected to the east in the Northern Hemisphere and potentially create the needed negative vorticity.]  We have found in the past that very rarely does absolute vorticity = 0 get very far from the equator, i.e., not as far as 5N.  Still, it is possible this is
playing a role in some fashion.  As you can tell, this is all very speculative, especially since I am writing it at 9:15 pm on a Friday after going to a local pub with some of our graduate students!
        These tropical waves are very interesting, and slowly but surely I am moving some of my efforts from hurricanes to the more general tropics...
 Cheers,
        John M.

Mark Lander

Hi Pat,

    Like Roger Edson, I am a bit skeptical of the easy identification of all these equatorial wave modes running around out here.  The N=1 ER Rossby is probably the easiest to recognize (it's just the twin TC pattern/with a west wind burst in the middle).  I have become a bit adept even at "seeing" the eastward moving Kelvin waves that show up as such on
Matt's page.   I have not yet, however, been able to identify the MRG in the cloud pattern (not even sure what it would look like); and, even on Matt's OLR modes page, this mode flip-flops day-by-day all over the place.
    TS 26W (Bavi) started long before the 9th.  On Oct 4 th, the SLP dropped to 1002 mb at Majuro (6N 172E).  The westerly monsoonal flow had already commenced at Kosrae (6N 162E) into the southern Marshalls.  If one wants to trace the structrue of this disturbance, one must go way back to the end of September/early October.
    It might be of help that the numerical guidance during the last week of September indicated the development of the large monsoon depression beginning from a disturbance/monsoon depression in Marshall Islands.  I'm sure that the model wind fields can readily be decomposed into their wave modes, and any transformations that occurred during the "model" evolution
of TS Bavi determined.  Now, what Nature did might actually have been a bit different.
     In any case, however Bavi got going, the ingredients were there at the end of September, and numercial guidance (actually I only looked at NOGAPS/AVN/and MRF) made a very good forecast (a week out) of TC genesis.

    Regards,  Mark Lander

John Molinari
Hi Mark,

         The MRG wave behavior on Matt Wheeler's page is realistic.  Those waves tend to collect in wave packets (could we say they run in packets?).  The Dickinson and Molinari JAS paper from this year shows them pretty well.  The waves grow until they reach the middle of the packet, then die at the other end.  Because the packet motion (group velocity) is
opposite to the wave motion (phase velocity), a given low, for instance, will grow, then quickly die while a new low grows one wavelength behind it.  The weird-looking behavior in the animations reflects that behavior.
         Having said that, I agree that it is very difficult to identify these waves in real time on satellite pictures, much more difficult than equatorial Rossby waves.  Dickinson and I have maps with winds and OLR filtered for the MRG waves, and even then the signature is somewhat subtle and can be hidden on satellite images by the MJO.  This is the conundrum
for tropical forecasting: I think these waves do matter a lot, but they are hard to identify, because, as Matt Wheeler's page shows, they cross one another and interfere with the signatures of one another.  I would argue (and probably many of you already agree) that tropical precipitation forecasting is even more difficult than summertime convective precipitation
forecasts in the middle latitudes, where the organizing disturbances (short wave troughs, MCCs) are easier to identify.

         Cheers,
         John

John McBride

Thanks John and Pat for noting that as the Southern vortex moved over the equator, it undertook the transformation into a mixed Rossby-Gravity wave.  As Mark and John have noted, these wave (the Mixed R-G's) can be hard to see on unfiltered and un-processed wind plots.  The one over the weekend, however, was beautiful.

On my web-page, I have put the Matsuno plot of the structure of the mixed wave. (My plot having been lifted from Hendon, H. H. and B. Liebmann, 1991: Structure and annual variation of antisymmetric fluctuations of tropical convection and their relationship to      Rossby-gravity waves. J. Atmos. Sci., 48, 2127-2140. ).  Immediately below it I have placed a subsection from the raw GASP tropical analyses at 850 hPa for this weekend (12UTC on the 11th and 12th October) showing the structure of the Southern member of the Rossby-pair over the Pacific, once it had moved northward and sat over the equator.  The resemblance in structure to the classsical mixed wave is striking.

One further comment re diagnosing the waves in real-time.  I just remind everyone that Matt (the Boy Wonder)'s diagnostic are based entirely on the OLR signature, whereas in looking at the wind-field analyses, I am generally going totally by the wind-structure.  Matt and I have a few ideas as to how to better see the waves in the wind-fields on real-time...we'll do some work on it and get back to you (this will be with the help of a few talented real-time types like Noel Davidson and Gordon Jackson, though they don't necessarily know it yet)

cheers

John McB

850 hPa 02_10_11-1200 UTC                                               850 hPa 02_10_12-1200 UTC

 

Paul Roundy
Greetings from Germany!  These events are very interesting.  Matt Wheeler's website does a good job isolating the MRG mode, especially during the northern fall.  During other seasons (and occasionally during the northern fall as well), the modes that appear to be antisymmetric across the equator in the filtered data are actually produced by waves
that are asymmetric across the equator (dominated by one hemisphere only, like a TD-type disturbance or easterly wave). The antisymmetric filter forces the asymmetric anomaly to produce an anomaly in the opposite hemisphere of opposite sign.  That anomaly is bogus when applied in a diagnostic analysis for an antisymmetric wave-type.  One can discern such
a case by examining the unfiltered anomalies.  Perhaps movies of horizontal maps of the raw anomalies could be posted on the web along with the filtered anomalies, to help users discern which waves are most like MRG waves and which are probaby asymmetric.

I have no problem with the idea that Rossby waves, TD-type disturbances, and MRG waves can undergo some type of metamorphosis from one type into another.  A careful analysis of the OLR spectrum (without applying symmetry filtering) reveals that the peaks associated with each of  these wave-types overlap each other. This is consistent with the idea that one of these waves could evolve by some kind of refraction, convective process, or wind interaction into another wave
type.

Paul

John McBride

Thanks Paul..... I am starting to become astounded by these waves.

I have known about them since I was in graduate school when Wayne Schubert had us reading Longuet Higgins and Matsuno.  I used to know Adrian Gill quite well and spent many a white board session with him theorising about  the waves in my office..... Still... it never occurred to me I would ever  see them on the daily weather maps.

The funny thing is that I don't even look for them: they just jump out at me.  I arrive at work each  day, turn on the computer, bring up the satellite image and the tropical  analysis, and there they are... they now stand out to me like a cold front or low pressure system does on a mid-latitude map.

I believe there may be a new era in tropical forecating coming out of this.... Currently the tropical NWP modelling systems go to great lengths to initalise the intial moisture and divergence fields (with obvious good cause).... possibly in the future the initalisation/assimilation phase will pay equal attention to diagnosing and retaining the location, phase and amplitude of the various tropical waves that exist in the initial state.

Michael Dickinson
I thought would take the opportunity to make a few comments.

For those of you unfamiliar with the JAS article John Molinari was referring to,  I maintain a web site that shows many of the results from this study (a copy of  the paper can also be downloaded from there). The page contains a series of  animations that show the transition of a MRG wave into an off-equatorial  "TD-type" disturbance that occurred during July 1987. Three tropical cyclones formed on the edges of three consecutive cyclonic gyres as they moved  off the equator.

http://www.atmos.albany.edu/student/mjd/dm1987.html

Note the transition of MRG waves to off-equatorial TD-type disturbances was  first shown by Takayabu and Nitta (1993, their fig. 20 I believe). I also have a series of animations from this case as well.
(http://www.atmos.albany.edu/student/mjd/takayabu1985.html)

What I found interesting in the maps that John McBride shows (thanks John) is the development of another cyclonic (using Northern Hemisphere terminology) gyre near the dateline by 1200 UTC 10 Oct. This was certainly not present earlier in
the map sequence. This suggests that this additional gyre may be a result of dispersion from the strong cyclonic circulation near 20N (I have not looked east of the dateline to see if additional "alternating" gyres are present along the equator).

Sobel and Bretherton (1999) make the comment that the source of TD-type disturbances may be from tropical cyclones themselves through dispersion. This may be an example of that.

The current case is different from the situation that John Molinari and I discuss. In the present case, the equatorial waves (appearing on 10 Oct) appear to be a result of dispersion. In our case (July 1987), we argue that the equatorial wave packet development/growth occurred within a large region of convection during the active phase of the MJO. So, one could say this is an example of "in situ" development.

Regards

Mike Dickinson