Gedday,

Since reading  recently the paper by Dickinson and Molinari:  Mixed Rossby–Gravity Waves and Western Pacific Tropical Cyclogenesis. Part I: Synoptic Evolution. Journal of the Atmospheric Sciences: 2002, Vol. 59, No. 14, pp. 2183–2196) I occasionally go to the NOAA CDC web-site and construct a hovmoller of meridional wind along the equator.  I have placed on my web-page (Follow links to synoptic discussion/today's date) such Hovmollers for the past 1 1/2 months at 700 hPa and at 850 hPa.

                             850 hPa                                                                                        700 hPa

Looking particularly at the 700 hPa Hovmoller and following the logic in Dickinson and Molinari, you can see we have had a sequence of westward propagating mixed-Rossby gravity waves between about 140W and 100 E over the past month.  Interestingly one of these seems to have been the precursor to tropical storm Haishen, currently at about 14 N, 136E.  Looking at the meridional wind Hovmoller, the MTRG originated at about 120W on 11 November, but first appeared at low-levels on the dateline on about  16 November.  On my web-page I have placed a sequence of one-per-day 850 hPa charts from TLAPS for the period 0000 UTC 17 November to 0000 21 November.

On tehf irst image (17 Nov) The MRG (if thats what it is) appears as an isolated single cell on the equator at about 163 E.  It moves westward over the following days, and over the last couple of days of the sequence drifts northward off the equator to transform into a TD, much as described in Michael and John's paper.

The resultant Tropical Storm, Haishen,  is impressive because the satellite signature covers quite a large area; so I have also put up a recent satellite image.

cheers

John McB

Matt Wheeler

Good example of the MRG wave (and its transitioning to a TD-type disturbance and then a TC) John. It shows up consistently in the 2-d antisymmetric OLR filtering as well (as attached from http://www.bom.gov.au/bmrc/clfor/cfstaff/matw/maproom/OLR_modes/f.1.mrg.html).
(Admittedly, the contour interval could be reduced to make it clearer).

And on the satellite image you show (21 Nov) you can defintely see the antisymmetric nature of the convection in the WPacific. At the longitudes of 140E-150E it is all in the NH. At the longitudes of 160E-170E it is in the SH, and from 180W-160W it is all in the NH. This suggests a nominal zonal wavelength of about 40-50 degrees for this example.

-Matt.

Matt Wheeler
Folks,

Attached is a better plot of the antisymmetric OLR filtering and the corresponding GMS infrared image from John's web-page. Just one more exaple of large-scale tropical weather being produced by an equatorially-trapped wave! And to think that Matsuno (1966) derived all these solutions before any such observations as these!

-Matt.

Paul Roundy

Thanks for pointing this out!  It appears that in order for the MRG mode to become involved in cyclogenesis, the wavetrain must interact with a broad region of convection that is enhanced by some other process.  In the Dickinson and Molinari case, it appeared to be the MJO.  The MJO does appear to be approaching the West Pac region right now, but in this case,
it appears as though the westward-propagating modes were more important.

Matt Wheeler's page, giving the ER n=1 mode does indeed show a region of enhanced convection in about the right place.
http://www.bom.gov.au/bmrc/clfor/cfstaff/matw/maproom/OLR_modes/f.3.n1ER.html (saved copy here)
This indicates that multiple modes, simultaneously enhancing local convection, work together to provide an environment suitable for cyclogenesis.

My broader filter for westward-propagating intraseasonal modes gives a much larger region of enhanced convection west of the date line than appears in Matt's ER n=1 filter result. Long satellite image sequences confirm that the broad region of enhanced convection that the storm developed from originated south west of Hawaii. I have been warning about that region for several weeks now.  If the old forecast from November 7 continues to verify, the active region near the date line has already begun its gradual weakening period.

Michael Dickinson
John, Matt, Paul... et al.

Thank you for pointing out these cases. It is amazing how these events can jump out in the raw (unfiltered) data.

In looking at the phase of the MJO from Matt Wheeler's OLR maps  http://www.bom.gov.au/bmrc/clfor/cfstaff/matw/maproom/OLR_modes/f.5.MJO.html (saved copy here)
for example), the MJO appears to be present in the western Pacific during this time and *may account* for the apparent MRG to TD-type disturbance transition.  Paul Roundy points out another possibility. He suggests that multiple modes,
simultaneously enhancing local convection, can work together to provide an environment suitable for cyclogenesis. Whatever the case, the presence of the broad region of convection along the equator appears to be a necessary
ingredient for the MRG to TD-type transition.

In our case, the MRG wave packet appeared and amplified within the convectively active envelope of the MJO, but the exact process by which this occurred was uncertain. We suggested wave accumulation (e.g., Sobel and Bretherton 1999) as
one possible mechanism. We did not investigate the contribution of the other modes in this process.

Regards

Mike Dickinson