Saturday, September 27, 2008

Atmospheric Roulette – Numerical Models Tank

The views expressed are those of the author and do not necessarily represent those of the National Weather Service.”

Please keep in mind the ESRL/PSD GSDM web link, below, while reading this discussion.

The 91-day signal to noise ratio (snr) anomaly composites are now updated daily, centered on the date shown (see product descriptions). Most map room issues have been resolved.

The spatial patterns of global SSTs have not changed significantly during the past several weeks. The tropical Indian and west central Pacific Oceans remain well above normal with anomalies ~2C and widespread totals roughly 29-30C. The warmest SSTs, shifting south with the seasonal cycle, are from east of the Philippines to New Guinea with totals ~30-31C. Negative anomalies, ~1-2C, continue from the equatorial Dateline into portions of the east Pacific Ocean cold tongue, with totals well under 28C. Respectively, anomalous spatial horseshoes of warmth and coolness extend into the extratropical North and South Pacific Oceans from these regions (negative phase of the PDO understood). We generally consider 29C SSTs as a critical temperature for supporting persistent deep moist convection.

Latest TAO buoy data show a steeper than climatology equatorial Pacific Ocean thermocline leading to roughly minus 4C anomalies (five-day average) ~150m depth/140W. Previously discussed upwelling ocean Kelvin waves have been responsible for the latter.

Overall, there are many similarities to the SST patterns compared to a year ago. This in itself demonstrates the scientific incompleteness of just using Nino 3.4 to define not only ENSO, but the impacts of global tropical SSTs onto the atmospheric circulation. (note the initial projection) (link 18).

Tropical convective forcing is still enhanced across portions of the Tropical Northwest Pacific Ocean (TNWP; including Super Typhoon Jangmi approaching Taiwan) and the Western Hemisphere. In fact, a respectable surface westerly wind event has been occurring north of Indonesia off the equator. Strong suppression continues across the equatorial Indian Ocean into Indonesia. An important monitoring issue heading into boreal winter will be if westerly wind events accompanying any additional coherently eastward shifting tropical convection across the west Pacific Ocean occur on the equator.

Understanding this spatial pattern of tropical forcing is not trivial, and requires intense daily monitoring within the WB (2007) GSDM framework. An important point to stress is that numerical model performance has decreased substantially over the last few weeks. Subseasonal processes (linking model skill as a function of the WB (2008) measure of the Global Wind Oscillation (GWO) is on the list of efforts for WB to quantify at some point), in addition to seasonal transition, have been responsible. I do feel comfortable with what I think are the on-going global dynamical processes, and will do my best to summarize in the following.

Animations of upper tropospheric daily mean vector wind anomalies show generally zonally oriented chains of midlatitude Rossby wave energy dispersions (RWDs) during the last few weeks. This suggests, and is objectively shown by the real-time WB (2008) measure, that the GWO has been dominating. A major component has consisted of a robust global mountain-frictional torque index cycle during about the last couple weeks (see AAM plots). Around 11 September the global mountain torque involving most north-south massifs was greater then plus 20 Hadleys, followed by a negative event of similar magnitude a few days ago. The frictional torque has loosely been in quadrature. The last occurrence of this type of (weaker) behavior was ~ 1 August (see WB (2008) submitted paper for power spectra).

The point is that the variation discussed above, through processes tied to meridional momentum transports, helped to shift anomalous zonal mean westerly wind flow from the equatorial regions into the midlatitudes. As I type, anomalous easterly wind flow (~3-5 m/s at 200mb) is quickly returning to the equatorial and subtropical atmospheres. Additionally, loosely zonal mean subtropical AAM sinks with poleward source regions have become established in both hemispheres. Finally, updated through 25 September global relative AAM is already at least 1 sigma below the R1 data climatology (without the stratospheric QBO). These are all characteristics of the global circulation trying to return to a La-Nina base state, represented by phase 3 of the GWO 250mb psi snr composite anomaly plot (more said below).

The WB (2008) measure of the GWO and WH (2004) measure of the MJO remain out of sync, the former having a strong projection in octant 8 and the latter in octant 6 of their phase spaces. However, global tropical circulation anomalies are much better represented by the MJO phase 3 250mb psi snr plots, rather than phase 6. For instance, there are upper tropospheric twin tropical anticyclones across the Western Hemisphere, not cyclones. Again, this tells us that the WB (2008) GWO (which considers the MJO) is “calling the shots”. Our current Western Hemisphere signal of tropical forcing was initiated by the midlatitude RWDs linked to the positive global mountain torque per above roughly 2 weeks ago.

There is no change in my predictive thoughts from what I typed a week ago. Tropical convective forcing is already festering across the South Indian Ocean. I expect a rapid increase of enhanced rainfall across much of the equatorial Indian Ocean during the next 1-2 weeks. The WH (2004) measure of the MJO is probable to orbit into octants 2-3 over the next few weeks. The WB (2008) GWO is likely to continue its ~10-20 day circuits from phases 4-5 to 8-1, but weakly “spiral” closer to octant 3 of phase space, also during the next few weeks. Hence coupling between the tropics and extratropics could resume by about the middle of October, and the GWO and MJO phase 3 250mb psi snr plots may broadly represent the global circulation.

The current response developing across the PNA sector, shown by the GWO phase 5 250mb psi composite anomaly for the extratropics, is the result of a fast transient RWD forced by the TNWP enhanced convection. I like phases 8-1 of the GWO 250mb psi composites by week 2, meaning some form of troughs coming into to the western states. Afterwards, while GWO variations per above continue, on average (mid-late October?), the phase 3 GWO and MJO 250mb psi and temperature snr composite anomaly plots, typical of a “La-Nina like” low AAM base state, may be most representative. This means a probability shift toward ridging near the west coast and Caribbean, and a central states trough.

There are also seasonal cycle issues not addressed in this long discussion. Overall, an enhancement of the weather typical for October appears most likely across the USA during the next few weeks. Per previous postings and above, a rendition of the low AAM base state from 2007-08 may be returning. Synoptically this may mean western and central USA troughs but with added subtropical westerly wind flow. Hence the probability for high impact weather could increase centered on the Plains, Mississippi and Ohio Valleys going into this upcoming boreal winter.

The tropical cyclone risk should hopefully decrease across the TNWP after week-1, while staying at least climatology for the Western Hemisphere “until further notice”. In addition to cyclogenesis from the deep tropics, I have concerns for hybrids especially around the Caribbean. After weeks of flooding, etc, suppression (dryer weather) is probable to shift northeast into northern India and Southeast Asia ~weeks 2-3. Per above, intense to severe thunderstorm activity is probable to erupt across the equatorial Indian Ocean during the next 1-2 weeks, then propagate east afterward. The risk of Bay of Bengal and Arabian Sea tropical cyclogenesis should be at least climatology during October.

Please see the latest official tropical cyclone forecasts for all basins. I trust the expertise of the appropriate meteorological centers to alert the public of additional weather hazards worldwide.


Links to CPC and PSD ENSO discussions:

The following is a link to information about the stratosphere and other nice monitoring tools:

The following is a link to NCEP model verifications (surf around for lots more) .

The following is a link discussing recent global weather and related events.

These are probabilistic statements. We hope that an opportunity will arise for us (soon) to allow our dedicated web page effort to mature, expediting objectively and accountability. This web page effort will hopefully include an objective predictive scheme for the GWO with hindcasts.

The WB (2007) paper on the GSDM has been published in the February issue of MWR. In addition, the first of a two-part paper, where WB formally introduce the GWO (WB (2008)), has been accepted for publication in MWR. The authors are in the process of resubmitting after a minor revision. A pdf version (before revision) can be downloaded from the following link:

Overlapping seasonally varying subseasonal composites for variables such as surface temperature, precipitation, geopotential height and streamfunction anomalies are planned on being posted on the web site mentioned above and presented in part-2 of our paper. We want to emphasize notions such as global-zonal mean-regional scale linkages as well as forcing-response-feedback (with subsequent interactions) relationships. An important purpose is to provide a dynamical weather-climate linkage framework to evaluate the numerical models in a sophisticated manner as part of a subseasonal (and any time scale) forecast process, in addition to a climate service for all users. Relying on the numerical models alone is a cookbook! I should be able to do a discussion next weekend, ~4-5 October.

Ed Berry


Accudean said...

What if the low AAM base state bottoms out in late December unlike 2007/08? The other question for North America in regards to upcoming winter would be related to the north central Atlantic's SST anomalies trending cooler, Negative NAO(?).

2007-08 featured strong warm anomalies south of Greenland which led to a Positive NAO for much of that winter season. An early bottoming of AAM and a greater frequency of -NAO blocking would shift cold anomalies for the season to the Great Lakes, right?

Thanks again for opening up my worldview of climate.


Ed Berry said...

Hi Dean,

Thank you for reading the discussion and questions. There is, of course, uncertainty in the details. There were multiple periods of low global AAM last cold season including December 2007 and March 2008 (~2 sigma anomalies).

I can link the North Atlantic blocking from the 2007-08 winter to RWDs from enhanced west central Pacific Ocean tropical convection during mid Jan-early February 2008 (for example). Extratropical SST anomalies are generally forced by the atmosphere (as opposed to the tropics). There are numerous modeling studies that support these observations.

Should the SSTs warm near the eq. Dateline due to WWB on the equator later this cold season, then tropical forcing may come well into the Pacific Ocean. RWDs from this region can support a negative projection onto the NAO leading to other wx. ramifications such as possible cold in the Great Lakes.

Ed Berry