Tuesday, September 19, 2006


The following is a link to our recently accepted paper by MWR which discusses the GSDM (Weickmann and Berry 2006).


From taking into consideration the interactions of 4 different subseasonal time scales, a sequence of maps depicting a coherent set of repeatable events has been derived for the Northern Hemisphere cold season from November-March. This set is broken up into 4 stages, referred to as GSDM (for Global Synoptic-Dynamic Model) Stages 1-4 in the text of my Blog. Figure 13 in our paper presents a schematic of the GSDM. Ideally it would be advantageous to post our weather-climate discussions with greater frequency to provide additional detail while having a more complete weather-climate record of attribution and prediction. In these discussions I adapt the GSDM for the warm season. Our list of work includes a seasonally adjusted rendition of the GSDM.

Our latest weather-climate discussion dated August 18th, 2006 (and updated September 9th), has been posted on the ESRL/PSD MJO web site at


Please see past postings for web site links. I am going to discontinue inserting most of them in an effort for brevity. I also need to do the same with these postings.

Global tropical SSTs remain above average across the Western Hemisphere particularly along the equatorial Pacific cold tongue. Weekly mean anomalies (September 10-16) were ~plus 1-2C with values ~plus 3C west of South America. At depths of roughly 100m magnitudes ~plus 3-4C were present around 130W along the equator per latest TAO buoy data (September 18th). The latter are the result of weak eastward propagating oceanic Kelvin waves initiated by past westerly wind events on the equator across the west central Pacific. Again, the seasonal cycle of SSTs needs to be kept in mind with these anomalies. Climatologically SSTs along the equatorial cold tongue are cooling this time of year. Actual cold tongue SSTs vary from around 22C near the South American coast to in excess of 30C just west of the equatorial date line with the 29C isotherm at about 165W. Recall we use the 29C SST as a threshold for supporting persistent deep moist tropical convection.

SSTs across the equatorial Indian Ocean have cooled as a response to recent clouds and rainfall in that region, while substantial negative anomalies (~ minus 1-2C) remain around Indonesia south of the equator. Finally, respectable positive anomalies of SSTs also extend from the date line to the coast of Southeast Asia, with large “pools” of 29C and greater values. As I have stated in earlier postings, cool Indonesian SSTs with warm SSTs around the date line and east are typical of warm ENSOs in their mature phase which usually occurs during boreal winter. Literally only time will tell us how our current weak warm event evolves (models understood). I personally feel this is unclear (see past postings), also understanding equatorial Pacific SST anomalies this time of year tend to persist through winter. Please see the appropriate web sites for global SST details.

The following are links to ENSO discussions.



Please also see the following CPC link (and others therein) for further ENSO, etc., insights, and remember that official USA information on anything related to ENSO comes from CPC.


The MJO signal is once again very weak, assuming we actually did have one a week to 10 days ago. This matter brings up an important related point. In general, with out getting into the level of detail I would use to scientifically defend this issue, the MJO signal has been generally weak since about 2002. I have raised this point in past writings, and may be related to a longer term global change signal such as global warming. However, the point is even when there are strong MJOs, on average these convectively coupled modes on explain ~20% of the tropical convective forcing. Additionally, no two MJOs are the same. If anyone is going to attempt to use MJO information for predictive purposes, detailed disciplined daily monitoring within a dynamical framework is required (understanding the importance of averaging to extract low frequency signals)! This is one of the motivations for deriving the GSDM. Rules of thumb based on statistical techniques, indices, etc., have their usefulness; however, there is no cookbook (insert picture of angry dog showing fangs)!!!

Taking the above one step further, we have been dealing with a weather-climate situation where signals have been weak since at least April 2006 (2002?). This does include tropical convective forcing arising from highly complicated nonlinear dynamical interactions with the extratropics (with forcing and responses then switching and “going all over the place” like any white noise process would) which project onto anything but MJOs. Again, only disciplined detailed daily monitoring within a dynamical framework linking weather and climate and research support (to help with numerical modeling work) has any hope of catching these things, particularly for efforts such as improving predictions of extreme weather events for days 3-14. And please remember, such kind of information must be expressed probabilistically for needed periods of time, and that numerical models do very poorly predicting tropical convective flare-ups after about day 5.

During the last 3-5 days a consolidation of tropical convective forcing has occurred centered around 10-15N/120E. Full-disk satellite imagery shows a nice band of enhanced tropical convection extending from the northern Arabian Sea east-southeast across northern Indonesia (keep in mind the SSTs) into the South Pacific islands. Frontal activity from the Southern Hemisphere is contributing to the extension into the SPCZ (which is by definition). Per coherent modes Hovmollers an eastward propagating convectively coupled Kelvin wave from the Indian Ocean merging with a Rossby mode from the west Pacific may have contributed to the merger. Three-day averaged OLR anomalies are ~minus 50-70 W/m**2. The anomalous convection that was across the equatorial Indian Ocean roughly 2 weeks ago has shifted to about 20N leaving suppression in its wake (as expected from the seasonal cycle), leading to the band of thunderstorm clusters (see links for possible regions of tropical cyclogenesis). Weaker tropical forcing remains across the northern equatorial date line region, and from the Americas into the North Atlantic and Africa.

My own thoughts, with very low confidence, are there will be a fairly rapid eastward propagating convectively coupled mode (dynamical signal) from the location of the consolidation into the Tropical Northwest Pacific and date line regions. Frontal activity from the Southern Hemisphere may also excite tropical thunderstorm activity across the warm SSTs around the date line. Thus by late week 1 into week 2 we may observe positive anomalies of tropical convection from around the date line into the Philippines and perhaps Southeast Asia. Other areas of tropical forcing may remain around the Americas and North Atlantic, and perhaps re-emerge across the Indian Ocean just south of the equator (where SSTs are still slightly above average). In fact, frontal activity from the Southern Hemisphere may already be starting the latter. The forcing across the Western Hemisphere may be a local response linked to the Indian Ocean flare-up 2 weeks ago and perhaps even our weak warm ENSO.

Based on monitoring during the past several months (and boreal falls from several past years) I would not be surprised to see two regions of tropical convective forcing across both the equatorial Indian Ocean and west central Pacific re-appear during the next 1-3 weeks. My defense for that notion is still the loose presence of these 20-30 day modes of tropical convective variability which may (and not defensible at this time) be linked to submonthly modes in the extratropics.

Global AAM signals are still not clear. Global mountain torque, based on the reanalysis data through September 16th and its 1968-1997 climatology, is still at about plus 10 Hadleys with most of that coming from East Asia. The global frictional torque is roughly minus 10 Hadleys, with most of that coming from intense storm track activity across the Southern Hemisphere extratropics (probably also a contribution from the Coriolis torque). The former (along with the Earth component) is contributing to a positive global tendency of around 15 Hadleys while the latter has contributed to a global tropospheric relative AAM of about minus 1 sigma. I think I can see where a lot of this is coming from; however, it is not trivial for me to explain that without additional presentation resources. The following is a brief attempt.

If a time-latitude section of 200mb zonal mean zonal wind anomalies is done (care understood of just using one level), there is a signal of poleward propagation of zonal mean westerly wind anomalies (~ 5m/s) from the equator to presently around 30N and 30S starting early August 2006 linked to a central Pacific tropical convective flare-up. Given the seasonal cycle, these zonal mean westerlies may have re-invigorated what was already an active storm track season for the Southern Hemisphere extratropics. That may contribute a negative tendency to both the global frictional and Coriolis torques. Meanwhile, flare-ups across the warm SSTs of the west central Pacific may have contributed to (or be a response to) positive East Asian mountain torques. All of these behaviors may be warm ENSO signals, in the presence of other “things going on”.

Currently, there are 200mb zonal mean easterly wind anomalies throughout the tropical and subtropical atmospheres, with ~5m/s magnitudes. The more up to date ESRL/PSD operational global AAM plot (1979-1998 climatology) still has anomalies slightly less than minus 1 sigma, with strong signal of vertically averaged zonal mean Southern Hemisphere subtropical easterly winds (with some symmetry to the Northern Hemisphere). The 120E consolidation had led to the formation of twin subtropical anticyclones in that region, and these are linked to Rossby wave energy dispersions (RWDs) into both extratropics (per animations of 150mb and 250mb daily mean vector wind anomalies). These RWDs have recently contributed to strong anticyclones across the high latitudes of both hemispheres, and possibly an increase in the flux convergence of AAM transport at 30S and perhaps starting to appear at 30N on September 16th.

Turning to the Asia-Pacific-North American sector, not only is there anomalous outflow of upper tropospheric southwesterly flow from the 120E convection, there is another source starting near the date line (see plots and animations on the ESRL/PSD web site). The former is loosely from many modes of subseasonal variability while the latter is a warm ENSO signal (also contributing to short time scale processes). These outflows are contributing to subtropical jets (STJs). The outflow tied to the 120E consolidation linked up with a baroclinic wave packet moving through northern Asia a couple of days ago. It is this interaction which may lead to a second trough across the western USA by this weekend after its predecessor. Recalling the local Western Hemisphere response from above, there have also been weak twin subtropical anticyclones from the East Pacific into the Atlantic.

To summarize, I think we have 1) a weak warm event whose future evolution is unclear, 2) the consolidation of tropical convective forcing around 10-15N/120E which may lead to an eastward propagating dynamical signal, 3) other arguably random regions of tropical convective forcing including the Western Hemisphere, 4) a sub-monthly component that has contributed to mixed global AAM signals especially for the East Asian Mountain torque, and 5) lots and lots of white noise. Of course, the global circulation is well on its way to boreal autumn, only adding more uncertainty. I do think that an intermediate state between GSDM Stages 4 and 1 best describe the current weather-climate situation. Uncertainty remains very high for any predictions.

Week 1 (20-26 September 2006): GSDM Stage 1 is most probable, with STJ interaction. Like nearly all the models show, 1-2 strong mobile synoptic baroclinic troughs digging into the western part of the country then heading east is probable. The details are unclear, especially for the second trough. My own feeling has been to favor, with very low confidence, the models depicting a slower and deeper solution for this second Rocky Mountain trough, as was shown by models such as the ECMWF 4-5 days ago. We are in a base state that favors troughs to break anticyclonically into the western USA states (which is why there would be a NE-SW tilt; there are also AAM transport considerations not discussed). During this week, portions of the Northern and Central Rockies are likely to experience heavy snowfall while severe local storms are probable from the Plains into the south central and southeast USA.

While temperatures colder than normal may persist over the western states, much of the Deep South will keep summertime. The Tropical North Atlantic will need to be monitored for any additional cyclone formation. Right now conditions for tropical cyclogenesis there remain favorable (from a weather-climate linkage viewpoint). With troughs digging into the western USA, the East Coast and Gulf States may become more vulnerable with time, particularly from hybrid systems. Please see http://www.nhc.noaa.gov/ for the latest tropical cyclone information.

Week 2 (27 September – 03 October 2006): A transition to GSDM Stage 2 may be most probable, meaning an eastward shift of the week 1 pattern across the USA. The trough position may be become established ~95-100W with the ridge just off the west coast into Alaska. Synoptic features would be expected to modulate this pattern. However, I could see this situation being transient, with perhaps GSDM Stage 3 by the end of this week or during week 3. This is time of year we see the North Pacific Jet “outrun” the tropical convective forcing.

Week 3 (04-10 October 2006): Unclear. However, similar to past boreal autumns, a GSDM Stage 4-1 may return at some point as our “war of the oceans” resumes. This may return us to a synoptic pattern similar to the one now present across the USA.

I think opportunities for rainfall will exist for Southwest Kansas later this week for system # 1 and hopefully system #2 for the Friday-Saturday time frame. Per above, I have been weakly favoring the model solutions showing a slower closed Rocky Mountain cyclone for the latter. I hope we do not have problems with the dry intrusion for the second storm. I think another trough will follow for later next week into week 2; however, its maximum amplitude may occur across the central portion of the country meaning another chance for rainfall followed by another surge of a chilly airmass. Afterwards, right now I would have to favor odds toward near-above average rainfall with overall near normal temperatures (with likely large variations) possibly through most of October. I think there will be more strong western USA troughs for at least the next several weeks.

I may not be able to do another posting for the next 2-3 weeks due to so many commitments. Please keep checking.

Ed Berry

No comments: