Edward Berry, NWS and Klaus Weickmann, CDC

At the time of our last update (22 December 2004) the MJO signal was still very weak. The circulation  in the Pacific North American region had just transitioned to a positive PNA pattern in association with a strong flareup of convection west of the dateline.  Since that time convection has redeveloped further west over Indonesia and the PNA region circulation anomalies have again retrograded, producing a strong ridge near 160W.  A weak-moderate MJO also seems in the process of moving over the west Pacific region although the confidence in this continuing is below average.

The mid-latitude westerly flow continues to be weaker than normal with a pronounced "split flow" over the oceans.  There are indications of a shift toward increasing the westerly flow near the equator and possibly in the mid-latitudes.  A strong stratospheric vortex is in place consistent with a positive phase of the North Atlantic Oscillation.

For information on the status of El Nino and the MJO see the following links:

Latest CPC ENSO Advisory

Latest CPC MJO Discussion
 

Part 1.   Weather-Climate Overview

Figure 1 (below) shows Hovmoller plots of outgoing longwave radiation anomalies (OLRA) for the  equatorial and southern tropics.  OLR anomalies are used as a proxy for deep tropical convection anomalies.  The contours isolate three coherent OLR modes, including the Madden Julian Oscillation (MJO), the Kelvin wave and the equatorial Rossby wave.  The red shading (negative OLR anomalies) shows convective flareups have continued over the Indian Ocean and the west Pacific Ocean (see previous discussions).  Convection anomalies have been more persistent over the equatorial Indian Ocean while they are more transient over the west Pacific with a recurrence interval of 20-30 days.  This time scale should continue to be important.

The eastward shift of negative OLR anomalies at the bottom of both panels projects onto a weak MJO (~10 W/m² anomaly in OLR) according to the coherent OLR mode contours.   This is supported by the phase space diagram of Wheeler (click here), which shows a weak to moderate MJO starting over the western Indian Ocean on about 25 December.  It is currently over the western Pacific Ocean and weakening (< 1 sigma anomaly).  Recent satellite images show a tropical depression at 10N, 140E suggesting MJO convective forcing has matured over the western Pacific Ocean.  Convection also continues to stay active over Africa and the west Indian Ocean.  For the latest satellite imagery see: Latest Indian Ocean Satellite Picture;  Latest Western Pacific Satellite Picture
 
 Figure 1  latest images
 

Part 2. Technical assessment of recent subseasonal variability
 
In this section we take a long term perspective on recent subseasonal variations of the atmosphere and make a scientific assessment of forcing-response relationships among various physical processes. This technical analysis may be too detailed for some readers.  Those desiring a discussion of the forecast insights only can skip to Part 3 of this discussion.  However, the important forecast issues are brought into perspective by the analysis in this section.

Fig. 2 (below) shows the time mean anomalies in northern hemisphere 500 hPa geopotential height and tropical global vector wind for 1 September 2004 through 10 January 2005, a 132 day average.  This period was averaged because an anomalous meridional transport of westerly momentum transport out of mid-latitudes started in early September 2004 (shown later).  Since then, episodic zonal momentum transports out of mid-latitudes have recurred and have given rise to the large amplitude split flow pattern seen in the top panel of Fig. 2. Midlatitude anticyclones over the oceans accompany the split flow and are largest at the exit regions of the east Asian and North American jets.  In the tropics Fig. 2 (panel 2) a pattern of twin anticyclones over most regions of the globe and twin cyclones in the Indonesia region are linked with the midlatitude split flow.

Figure 2

A preliminary assessment of the forcing for the tropical wind anomalies is obtained by viewing Fig. 3 (below). The figure shows the four stages of our synoptic model for the MJO portion only (see webpage discussion for more information).  The heavy shading highlights negative OLR anomalies and shows the typical eastward shift that occurs during a MJO.  The blue and red shading represent the subtropical twin cyclones and anticyclones respectively that accompany the convection.  In Stage 4 convection is more or less active at three locations: near the dateline (barely), over South America and over Africa/the Indian Ocean.  Suppressed convection is seen over Indonesia and the west Pacific Ocean. Despite being used for the MJO these pictures also provide clues about the circulation's "response" when tropical convective forcing is more persistent than a MJO, as occurred during the recent 132-day period.  We have already referred to the persistent African/Indian Ocean convection, the transient flareups over the west Pacific and the suppressed conditions over Indonesia when discussing Fig. 1.  Such forcing would favor a circulation regime somewhere between Stage 4 to Stage 1.  The latter even has a weak signal of split flow in the northern mid-latitudes.  The persistent convection over the Indian Ocean may be partially related to weak positive SST anomalies there and the seasonal suppression of convection over the larger positive SST anomalies near the dateline (especially seen in Nino 4, the farther west index). Alternatively, stochastic forcing of the circulation due to wave mean flow interaction, baroclinic life cycles and meridional momentum transports is another important mechanism.

Figure 3

Figure 4 (below) illustrates time history of the zonal mean split flow regime.  The negative zonal wind anomalies (easterly flow anomaly) that define the split flow are first seen in early September 2004 and can be loosely organized into three distinct episodes: 1) throughout most of September and early October 2004, 2) during the first 20 days of November 2004 and 3) from about December 15 to the present (January 12, 2005).  In the intervening periods the split flow weakens or changes sign.

Figure 4

The zonal mean picture in Fig. 4 does not show which regions are contributing to the split flow or how it is shifting around with time.  A regional picture of the mid-latitude's (30-60N) split flow evolution is presented in Fig. 5 (see below).  The annotations are extensive but the figure shows a 130-day period that includes ~35 synoptic systems, ~15 baroclinic life cycles and the 20-30 day variations seen in equatorial convection.  The highlighted features have been confirmed as coherent using daily animations of various fields.  Recall that the time mean split flow is located at the exit regions of the mid-latitude jet streams. In Fig. 5 this is represented by a H, L pattern in the Pacific North American region and a similar H, L over the Atlantic Ocean/Europe.  If we concentrate on the PNA region, three amplified split flow periods have been highlighted. The first and the third are quasi-stationary in space while the second is linked with a westward propagation across the date line.  In all three cases the split flow is alternating with a combined flow (L, H pattern).  The former projects onto a negative PNA teleconnection pattern whereas the latter onto the positive phase. This alternation is later linked to transient convective flareups over the western Pacific described in Fig. 1 and the persistent forcing of convection over the Indian Ocean.

Finally, Fig. 5 also has highlighted and labelled: a) a  prominent September 15th baroclinic wave packet, b) a succession of six synotic-scale anticyclones in the last month and c) an October 15th isolated wave breaking event.  The recent anticyclonic amplifications over the Atlantic Ocean contribute strongly to the 130-day time mean although prominent anticyclones also developed there during late September and mid-November 2004.  The North Atlantic Oscillation is generally negative during this time and becomes positive on about 15 December in association with the first of six amplifying anticyclones over the Atlantic Ocean.  The central Pacific Ocean anticyclone develops later on 22 December after retrograding off the USA west coast.

 Figure 5

Tropical convection can be persistent (i.e., it lasts longer than a synoptic life cycle) and can thus force a Rossby wave response such that the convection leads the circulation anomaly.  Feedbacks from other processes can obscure the relationship between the tropical convection and the circulation at any time. This leads to noise in a "signal" that for MJO purposes is shown in the synoptic model (Fig. 3).  Figure 6 (see below) addresses the question of the relationship between tropical convection and the circulation anomalies over the PNA region. Three pairs of contemporaneous 7-day averaged 150 mb vector wind and OLR anomaly patterns are shown in Fig. 6. 

Figure 6

The three represent cases where there is an alternation from a cyclone anomaly east of the date line (left side column) to an anticyclone in the same region (right side column).  The left column leads the right one in time.  In Fig. 6 we number the bottom row (1) to the top row (3).  These numbers (1-3) are marked on Fig. 1 for convection/OLR and on Fig. 5 for the 250 mb geopotential height. To understand fully the simultaneous OLR/convection anomalies requires a detailed analysis but some broad obervations are possible.  In the bottom panel the convection shifts from the west Pacific to the Indian Ocean as the circulation changes from cyclone to anticyclone.  In the middle panel, northwestward propagation of convection occurs and the Pacific Ocean circulation anomalies also retrograde.  In the top panel, the strongest OLR anomalies lead the circulation.  The simultaneous pictures in this case show weakening convection over the west Pacific (top, left) and eastward shifted convection over Indonesia (top, right).  The split flow in this case takes on a large zonal scale suggesting a dynamical interaction with the climatological flow is initiated.  This can lead to more persistence of the anomaly.
 
 3.   Predictive Insights

In the 22 December 2004 discussion, we suggested the on-going retrogression and increasing tropical convection across Indonesia and southcentral Africa  could lead to Stage 1 of the SDM  ( link to discussion with SDM ) by week 2 (the period of 30 December to 05 January 2005). This favors a trough along the USA west coast and a negative phase of the PNA (see PNA index)).  The pattern would suggest a stormy situation for at least Rockies and Plains, with colder (warmer) than normal temperatures for the west (southeast).  Confidence was low due to a lack of a MJO signal at that time.

Convection did continue to increase and consolidate around 5S/120E, and move east as a weak to moderate MJO.  The weather was very active for most of  country during this outlook period, particularly from California into the central states.  Very heavy and devastating precipitation occurred in California, and a strong winter storm system affected the Plains and Midwest states with heavy snow and rain.  Temperatures were generally below (above) normal across much of the west (southeast).  While our outlook had reasonable success, we did not predict the extreme precipitation event for California nor the evolution of the tropical convection into a MJO.  Also, the circulation continued to display characteristics of an intermediate structure between Stages 1 and 4 of the SDM.

Currently, the MJO signal has once again become weak.  In fact, the centriod of the tropical convection associated with the MJO has actually shifted back to the west over the last couple of days, and is around 140E on the equator.  Furthermore, this area of convection is splitting.  At this time (14 January 2005) there is a weak tropical cyclone developing across the northwest Pacific Ocean and thunderstorm activity is increasing along the South Pacific Convergence Zone (SPCZ).  Tropical convection is also again increasing across South Africa and the South Indian Ocean.

As of 13 January 2005, animations of 150mb vector wind anomalies (link to latest 30-day animation) show a western Pacific wave train, typical of Stage 2 of the SDM, tied to the western Pacific convection.  This has led to a cold regime for particularly the central USA.  Additionally, zonal mean westerly flow has increased throughout the tropics and subtropics, leading to a recent increase of global atmospheric angular momentum (latest plot).  Split flow still persists across the northern (and southern) hemisphere.  Since the MJO signal is weak, numerical and statistical predictions of a MJO cannot be considered useful at this time (see latest MJO forecasts ).

Confidence in the upcoming week 1 outlook is good, due to generally good numerical model agreement and the assessment of the current weather-climate situation. For weeks 2 and 3 confidence is low due to the lack of a strong MJO signal at the present time and to the events discussed in the previous section.  Our predictive insights are given below.

Week 1 (15-21 January 2005):  The general increase of westerly flow throughout the northern subtropics is expected to lead to a progression of the recent anomalies across the Pacific-North American sector.  By the end of this period the ridge should progress to around 115-120W, as part of a split flow pattern across western North America.  Troughs are expected to be located around 150W and 70W.  This state has a weak to moderate projection onto the positive phase of the PNA.  Over the lower 48 states, a subtropical jet should continue across the southern states with the polar jet extending from the Pacific Northwest to the Mid-Atlantic states.  This situation would be expected during Stage 3 of the SDM (see Fig.3 for the locations of the tropical convective forcing).  

The weather across the continental US should become generally mild and tranquil during this week 1 period.  At least the western two-thirds of the country should experience above normal temperatures by the end of next week, while the east coast may be a bit colder than normal.  There should not be much Arctic air in the country by the end of this period.  Most of the country should have little precipitation.  The important exception may be the Pacific Northwest states, which could have significant precipitation toward the end of this period.  Additionally, this precipitation may start to move south.  Also, there may continue to be light-moderate snow events across the Great Lakes (in addition to episodic heavy lake effect snows) and Northeast states with "clipper" systems (cyclogenesis off the New England coast will have to be monitored). 

Week 2 (22-28 January 2005):  This is the period when the general zonal mean westerly flow across the subtropics may weaken, as tropical convective forcing could become quite intense across the Indian Ocean and perhaps even Indonesia, in addition to remaining strong across much of Brasil and southern Africa.  A circulation structure similar to Stage 4 of the SDM would be expected.  For the USA that would mean the heaviest precipitation would shift southward along the west coast, including California, by the end of this outlook period.  Precipitation may also start to increase for the Rocky Mountain states, especially west of the Continental Divide.  The rest of the country may have only light precipitation.  While the western states start to cool, the eastern and particularly the southeast part of the country may shift to warmer than normal temperatures.   

Latest CDC Ensemble Forecasts

Latest NCEP Ensemble Forecasts

Additional NCEP Ensemble Output

Latest Canadian Ensemble Output