Tuesday, July 24, 2007

Dog Days (Dobermans from the Devil) are a Real Drag!!!

Sea surface temperatures anomalies (SSTA) throughout much of the global tropics are generally less than 2C with warmth across the Indian Ocean and west central Pacific Ocean and coolness along the East Pacific equatorial cold tongue. Totals across the former are ~29-30C. At depth negative anomalies less than minus 3C remain ~140W/100m per latest TAO buoy data. Interestingly, the largest anomalies and warmest SSTs as of 23 July per CPC data are present across much of the Tropical Northwest Pacific (TNWP) from west of the date line to the South China Sea. Anomalies are in excess of 2C with totals 30-31C. The latter is part of the horseshoe spatial distribution of warm SSTAs across the west Pacific Ocean into the extratropics.

Most numerical and statistical forecast tools continue to suggest La-Nina conditions (ocean forcing-circulation response, etc.) to be probable through at least boreal fall. As I stated a week ago, my own feelings are the coolest anomalies will remain across the East Pacific and this will not be a basin wide situation. The positive SST anomalies across the west Pacific warm pool are deep and have been persistent for at least the past 5-7 years.

Full disk satellite imagery have presented a nice signal of expanding tropical convection centered ~80-90E on the equator for the past several days. Latest 3-day averaged OLRA are ~minus 50-90W/m**2. This region of very intense tropical rainfall has been surging northward into India and southern China (tied to both the monsoon and severe thunderstorm frontal band activity) during the last few days. Meanwhile, strong suppression has been present from the west central Pacific into the Philippines, allowing the SSTs to warm across those regions. Satellite imagery as of this writing shows tropical forcing expanding along the equator to at least 150E, with a contribution coming from a convectively coupled Kelvin wave. These satellite, OLR and signals from numerous other diagnostic tools (a few discussed below) tell us a MJO is in the early stages of developing. In fact, this may be the strongest MJO signal since the December 2006-January 2007 event.

Over the last week this enlarging region of Eastern Hemisphere tropical convective forcing has started to slowly move east along the equator. Accounting for the Kelvin wave, I computed a phase speed of roughly 5m/s (4 deg long/day). However, there have been too many events of recent tropical forcing that shift east quickly, then stall, etc. How this latest situation will behave is still unclear (more said below). Statistical tools such as the multivariate Wheeler and Hendon (2004) technique currently suggest a strong MJO projection of ~2 standard deviations. While not unreasonable, these and other similar techniques from various global operational weather centers tend to both lose the signal and eastward motion after about a week.

There is some recent research (Jiang et. al. 2007) which suggests a “predictability barrier” exists for MJO evolution as the moist convection emerges from Africa into the Indian Ocean. It is possible we are currently seeing this impact. However, my own thoughts to all of this are: 1) predictability of this MJO should start to improve; and 2) I think it is probable for very intense convective forcing to shift into Southeast Asia while extending through the TNWP during the next few weeks.

The global circulation is dynamically responding. Animations of fields such as operational data upper tropospheric daily mean vector wind anomalies suggests a reversal in wind patterns across the tropical atmosphere during the last week or so. Twin anticyclones are developing across the western Indian Ocean ~60E with downstream lows ~120E. Additionally, the twin tropical/subtropical anticylones that were ~140E roughly 10 days ago have been displaced poleward into the central Pacific Ocean of both hemispheres. I hope this will be a “good thing” for the western USA during the next few weeks. However, feedback processes from the extratropics are still supporting Rossby wave energy dispersions across the North Pacific that is maintaining the wicked western USA ridge.

Zonal mean easterly wind flow anomalies are strengthening throughout the subtropics particularly around 15N, where magnitudes of ~5-10m/s at 200mb are now present. The equatorward propagation of zonal mean zonal wind anomaly bands discussed last week appears to have ceased. In the lower levels there is a trade wind surge along the equator with anomalies ~5-10m/s west of the date line as a response. How far east this trade surge spreads along the equator may be important for additional La-Nina development.

ESRL/PSD R1 data AAM plots updated through 21 July do not (yet) present strong torque and transport signals. The global frictional and mountain torques are slightly negative with the largest contribution to the latter coming from East Asia (~minus 10 Hadleys). A negative East Asian mountain torque, meaning below normal mean sea level pressures along the east slopes of the topography, is probable as MJO convection organizes across the Indian Ocean. The global AAM tendency is close to minus 20 Hadleys with actual AAM (including the QBO) ~minus 1.5 standard deviations. Updated through 15 July (5-day average), the Global Wind Oscillation (GWO) clearly suggests GSDM Stage 4-1.

Before going into the forecast, I want to make some comments about this summer’s western USA ridge from Hell. First, I have done a very poor job (in my view) with predicting the magnitude and extent of the anomalies (unprecedented excessive heat, drought, etc.) associated with it. During the last several weeks I have offered this ridge would shift off the USA west coast, and that has only occurred maybe once or twice for very brief periods. While this option may have been the "right" forecast, the atmosphere choose to take a different path given its stochastic nature. One of the premises of numerical ensemble weather prediction is to offer these alternate trajectories through phase space since it is a dynamical systems approach. An interesting study in itself would be to see how many, if any, ensemble members captured this alternate evolution.

In retrospect, I am not surprised by the magnitude of the ridge. In fact, there were comments made in past Blog postings during the late spring expressing concerns over a western USA ridge this summer. However, I did not think the situation would get as bad as it has. A message from all of this is the need for us to express subseasonal predictive information making use of probability shifts while having a complete forecast process including the GSDM and the numerical models. Any lesser effort is not scientifically and practically fair.

There has been past modeling work that suggests summer western USA drought patterns with a warm and convectively active Indian Ocean (~GSDM Stage 1). The trend for anomalously intense ridges to be shifted northwest from climatology across western North America during summer has been an unfortunate occurrence during roughly last 5 years. Again, to make better use of, and intelligent evaluation, of any numerical model output a diagnostic weather-climate linkage framework such as the GSDM is needed to allow attribution efforts as presented here. The information learned will translate to better subseasonal forecasts.

Uncertainty remains very high and my confidence is in the toilet for any week 1-4 prediction. That said, I am not changing my forecast reasoning from a week ago. It still looks on track, which is “something”. As discussed above, I think we have a MJO dynamical signal, and tropical forcing spreading into the TNWP weeks 2-3 is probable. The South Asian monsoon system still represents a stationary component forcing the global circulation. The GWO should undergo a slow decent orbit through GSDM Stage 1 and possibly Stage 2 by week 3. Of course, other faster variations are likely to contribute to smaller quasi-phase space circuits, and the GWO amplitude due to the MJO is unclear. However, similar to large variations seen from March-May 2007, the orbits are probable to be tilted toward GSDM Stage 1. For the PNA sector I think this will translate to a “net” retrogression of the western North American ridge to possibly ~140W by week 3. After all, I am bound to get this “right” eventually. Deliberately continuing to beat a dead bull, the truth of the matter is this effort needs to be properly supported to allow expression of these predictions probabilistically, with appropriate and fair verification.

International ramifications include locations from the Indian Ocean into Southeast Asia and southern China getting repeatedly pounded with severe thunderstorms and flooding rainfall. Areas that have been hit hard during much of this summer are likely not to see much relief for at least weeks 1-2. In fact, I would not rule out tropical cyclone activity for the Arabian Sea week 1 and Bay of Bengal week 2. Rainfall should slowly increase from west to east across Indonesia weeks 1-2. The current suppression across the TNWP should give way to active convection weeks 2-3. This includes tropical cyclone activity perhaps threatening the Philippines and Japan. Finally, the environment for tropical cyclone development across the Atlantic Ocean basin looks to be more favorable later week 1 and week 2.

Impacts across the global extratropics include the continuation of the parade of intense storm systems hammering much of Europe and more strong troughs for western and southwestern South America, weeks 1 and 2. The latter is the Southern Hemisphere analog to western North American troughs during boreal winter GSDM Stage 1 situations.

For the USA, we still have to deal with predictability issues caused by the shorter wavelengths, existing zonal mean and circulation anomalies from past “events”, and seasonal cycle matters going into August. Week-two ensemble means from various centers are starting to exhibit more disagreement; no surprise given the above. The ESRL/PSD ensemble mean actually shows a spatial distribution of 500mb geopotential height anomalies consistent with the proposed retrogression scenario favoring a central USA trough. Other models do not support this notion very well. Quoting from my 17 July posting, “In fact, more consistent with GSDM Stage 1 and possibly Stage 2 working with the seasonal cycle heading into August, the ridge may shift into the east Pacific Ocean (~140W) leading to trough development across the Rockies and western Plains with a Southeast states ridge. This may allow some reversal of the temperature and precipitation anomalies seen for the past several weeks”. Hey, at least I am being consistent!


An experimental quasi-phase space plot of the GSDM utilizing a time series of normalized relative AAM tendency anomaly (Y-axis) and normalized relative AAM anomaly time series (X-axis) can be found at



We call the behavior of this plot the Global Wind Oscillation (GWO), and one of its purposes is to extend thinking beyond the MJO.

These are probabilistic statements, and work is ongoing to quantify in future posts. We hope that an opportunity will arise for us to have a dedicated web page effort to expedite more objectively, with rigor, thoroughness and verification. The WB (2007) paper on the GSDM has been published in the February issue of MWR. I will try to at least briefly update this Blog on Friday (27 July). In general, due to covering shifts and travel, my postings will be irregular through at least August.

Ed Berry

1 comment:

beaudodson said...


Posted your last thoughts on easternuswx