Real-time Weather-Climate Discussion and Predictive Insights - 5 October 2005
 

Edward Berry, NWS and Klaus Weickmann, CDC

Since our last discussion (10 August 2005), the signal from the Madden-Julian Oscillation (MJO) has continued to be weak (see the latest Wheeler plot here ).  The MJO signal became incoherent across the Indian Ocean during late May 2005, and much of the tropical convective forcing since then has been from other transient convectively-coupled modes.  Partly as a response to positive equatorial western Pacific sea surface temperature (SST) anomalies ~0.5-1.0C centered around 160E, convective activity has been generally enhanced over the west central and northwest Pacific.  A second weaker area of near equatorial convective activity has occurred episodically across the eastern Indian Ocean, where positive SST anomalies ~0.5-1.0C have also been present. This pattern of anomalous convection affecting both regions has persisted since at least late July.  Despite the weak MJO, there have been two "MJO-like" variations that included features like eastward movement of convection, consolidations of convection over the west Pacific and modulation of western hemisphere convection. The latter played a role in the timing of the tropical cyclogenesis of Hurricanes Katrina and Rita.  SSTs along the equator remain relatively cool across the eastern Pacific and above normal around the date line/central Pacific and over much of the central and eastern Indian Ocean/Indonesia (see link for latest weekly plot).  Part I describes the behavior of the tropical convective forcing and circulation anomalies since July 2005, followed by predictive insights in Part 2.  For information on the status of El Nino and the MJO (including a week 1-2 global hazards outlook) please see the following links: 

Latest CPC ENSO Advisory
 

Latest CPC MJO Discussion and tools

Part 1.   Weather-Climate Overview

Figure 1 (below) shows two panels consisting of Hovmoller plots of  near equatorial (7.5N-7.5S) outgoing longwave radiation anomalies (OLRA) for roughly the period of late June-late September 2005 (bottom panel is more up-to-date). OLRA are used as a proxy for tropical convection anomalies.  As mentioned in previous reports, time filtering is utilized in these Hovmollers to isolate coherent convectively coupled modes such as the MJO, equatorial Rossby modes and Kelvin waves (see 10 August discussion and previous issuances for details).  The two "MJO-like" variations can be seen in the top panel where we have highlighted the date when west Pacific convection was active. The first case shows a consolidation of eastward and westward propagating features over the west Pacific whereas such behavior is less evident in the second case. In both there is the hint of an eastward propagation of negative OLR anomalies into the western hemisphere and this is emphasized by the dashed arrows.   The bottom panel of Fig. 1 shows individual synoptic events tied to the variations discussed above.  More will be said about these shortly.

equatorial olr hov
Figure 1 (recent Hovmoller OLRA plots: equator ; northern tropics. ; southern tropics )


Figure 2 is the same as Fig. 1 except for the northern (2.5-15N) and southern tropics (2.5-15S), respectively.  South of the equator, two flareups of convection can be seen over the Indian Ocean as well as a weak persistent convective signal along 160E.  Extratropical wave activity from the south Indian Ocean contributed to exciting the convective flareups (not shown).  For the northern tropics (Fig. 2, top panel), two features stand out. First, there is some evidence of stationarity to the positive tropical convective forcing along ~130E (bold dark dashed line); and second, there is evidence of coherent propagation into the western hemisphere (lightly dashed arrows). For the 24 July event, the propagating signal moved east at about 7m/s (slightly faster than a composite MJO), and projected onto a MJO between 80-120W during the first week in August.  The 12 September event had a greater contribution from convectively coupled Kelvin waves, two of which can be seen propagating eastward at > 20 m/s.


north and south olr hovs

Figure 2

Returning to Fig. 1, the bottom panel shows some of the synoptic interactions involving the tropics and northern extratropics during the two MJO-like variations discussed above.  The annotated "Hs" and "Ls" indicate the approximate locations of extratropical anticyclonic and cyclonic circulation anomalies, respectively.  The heavy solid lines show the path of three tropical cyclones (Katrina, Rita and Nabi) while the accompanying dotted lines show the path of pre-existing disturbances prior to their reaching tropical depression status.  The two small arrows just west of the dateline show the path of anticyclones associated with two baroclinic waves that propagated and amplified there between about 18-25 August. The wave energy from these developments dispersed eastward and, along with local vorticity sources, helped force the wavetrains highlighted on the figure. The resulting wind anomalies east of the anticyclones around 90W and 70W extended into the subtropics and provided favorable low shear conditions for the intensification of Katrina, as well as some steering of the hurricane. The disturbance that eventually became Katrina had a long history dating back to a flareup of the Atlantic ITCZ on about 12 August.  Note that the wave trains that helped intensify Katrina occur in the wake of the convectively coupled signal that moved into the western hemisphere and may have provided the initial disturbance.  Once the disturbance (which included the remnants of TD#10) came under the low shear conditions associated with the extratropical wave trains, tropical cyclogenesis became more likely.  Archives related to Katrina and Rita can be accessed here.  

The scenario for Hurricane Rita was a bit different.  A baroclinic wave packet that was generated over the west Pacific as part of the second "MJO-like" transient played a key role.  The packet provided both an initial disturbance and favorable shear conditions for Rita.  A detailed examination of the wave packet is beyond this report but we have highlighted three features that were involved in its development; this is based on viewing daily animations of OLR and upper level winds.  The first feature is the flareup of convection over the Indian Ocean.  The  convection anomalies forced twin anticyclones over the west Indian Ocean whose westerly flow anomalies linked with an amplifying anticyclone east of the Caspian Sea on 5 September.  The second feature is Typhoon Nabi, whose track is shown as a solid line ending near Japan on 6 September. On 7 September the upper level anticyclone of Nabi strongly interacts with the upstream amplifying baroclinic wave linked to the Indian Ocean convection.  The interaction appears to initiate the wave packet.  A third feature is the eastward propagation of convection along and north of the equator, highlighted with a double-shafted arrow on Fig. 1.  A further forcing of the packet by this mobile source is also postulated.  A snapshoot of the wave packet is shown on Fig. 1 for 11 September just before the trough along 60W in the Atlantic amplifies.  Figure 3 presents a sequence of plots for 150mb vector wind anomalies from 11 to 13 September and an OLRA plot for 14 September 2005.  The wave packet is centered over the Atlantic Ocean on 13 September and the "C" marks the first sign of a disturbance in the OLR field that can be identified with Rita.  It is located at the southwestern end of a trough whose OLR signature is clearly seen on 14 September.  Weakened shear develops as the anomalous anticylone over eastern North America builds eastward and wind anomalies become more zonal.             

Even though there are differences in the details for the development of Hurricanes Katrina and Rita, both storms formed roughly a week after a coherent convectively coupled signal moved through the western hemisphere.  Favorable shear environments appeared linked with baroclinic waves which amplified near the coast of Asia and propagated phase and energy eastward across North America.      

  wind and olr maps

Figure 3  (most recent 30-day animation of 150mb vector wind anomalies here )

Figure 4 presents another monitoring tool for convectively coupled signals, a near equatorial (5N-5S) Hovmoller plot of 200mb velocity potential.  The blue (red) shading and dashed (solid) black line denote negative (positive) velocity potential meaning upward (downward) vertical motion.  The two "MJO-like" variations are evident on the large zonal scales of 200 mb velocity potential.  The first had a relatively coherent eastward propagating signal that was slightly faster than a MJO.  The second had periods of rapid eastward shifts of negative velocity potential from about 150E-130W during ~ 5-15 September.    These reflect the faster Kelvin activity already mentioned.

200 mb chi hov
Figure 4  (most recent  velocity  potential  Hovmoller here )



Figure 5 reviews the behavior of atmospheric relative angular momentum (AAM) during the past year.  The top panel illustrates the locations of zonal mean anomalies of tropospheric AAM (warm colors positive/cold colors negative), along with the seasonal cycle of the zonal and vertical mean zonal wind (contours). The middle and bottom panels are time series plots of AAM anomalies for the troposphere (to 100 mb) and stratosphere (above 100 mb), respectively. The stratosphere plot shows the current negative phase of the QBO that developed recently.  


last year aam

Figure 5   (most recent plot of  tropospheric AAM here )

Four periods are illustrated numbered 1-4.  Period number 1 (roughly October 2004-mid January 2005) was characterized by 2 areas of tropical convective forcing located across the Indian Ocean and western Pacific.  A Stage 4-1 circulation state of our synoptic dynamic model (SDM) was observed characterized by northern hemisphere midlatitude split flow and three anticyclonic wave breaking events.  The latter were around 10 October, 15 November and 20 December 2004 and "produce" similar zonal mean zonal wind anomalies (see -/+ labels).  A brief event of opposite sign is shown on 5 December 2004.  The split flow is seen in Fig. 5 by the "blues" around 40N and the "yellows" at 25-30N and 55-60N.  During this time tropospheric AAM was increasing while the stratospheric AAM remained constant.  See January 15, 2005 discussion for details.

Period number 2 (mid January-mid March 2005) was characterized by an atmosphere-ocean coupling of tropical convection to anomalously warm SSTs in the central Pacific, which projected onto a central Pacific El-Nino.  Large westerly anomalies developed throughout the tropics and propagated into the subtropics and midlatitudes of both hemispheres.  Meanwhile anomalous easterlies replaced anomalous westerlies at the polar latitudes.  AAM increased very rapidly while the QBO across the stratosphere transitioned to an easterly phase.  The tropospheric response was that of SDM Stage 3 (see February 23, 2005 discussion).   

During period number 3 (mid March-May 2005) the central Pacific coupling ended, and MJO activity resumed.  While there was a weak oscillatory response of the wind fields throughout the tropics and subtropics, there were only weak signals in the extratropics.  In addition, tropical convective forcing shifted into the eastern hemiphere, and contributed to a rapid decline of AAM.  The stratosphere was well into the easterly phase of the QBO.  The circulation state of the atmosphere became dominated by SDM Stage 1 (see June 14, 2005 discussion).

MJO activity weakened by the end of period number 3, and coupled convective variability became "MJO-like" during period #4 and combined with renewed vigor of convection over the west Pacific Ocean.  The details have been presented above.  AAM has continued to decrease, and is ~2.5 standard deviations below the 1979-1998 operational climatology at this time (5 October).  Much of this is the result of deep easterly flow throughout the tropics and subtropics.  The easterly phase of the QBO continues.   

synoptic model

Figure 6 (Description of the SDM is presented here)


2.   Predictive Insights

During the release of the last weather-climate discussion dated 10 August 2005, the atmosphere was in a boreal summer time version of synoptic dynamic model (SDM) Stage 2.  Shortly afterwards, Stage 3 briefly appeared from about 11-17 August, then transitioned to a state consisting of components from SDM Stages 4 and 1.  Our predictions for the period of 10-31 August generally expected the atmosphere to slowly evolve back into Stage 1 from 2, based on stationary tropical forcing across the eastern hemisphere (see August 10 discussion for details).  However, as was discussed in Section 1, the behaviors were much more complicated.   

Nevertheless, the notion of the weather becoming more active across the Rockies and Plains states was reasonable.  Also, unseasonably warm and dry conditions prevailed across much of the central and eastern part of the country.  An important exeption was the very heavy rainfall across portions of the southeast and Ohio Valley due to landfalling Hurricane Katrina.  

Broadly speaking, the predictions for weeks 1-3 did fairly well.  For the specifics on the occurrences of severe storms, please see the SPC storm reports here.  Details on temperatures and precipitation can be found from the appropriate links on this site. 

At this time the signal from the MJO is still very weak.  Satellite imagery (eastern hemisphere, full-disk east Pacific, full-disk west Pacific, full-disk Indian Ocean, full-disk Africa; other images available here) has enhanced tropical convection centered around 80W, across central Africa and over large area of the eastern hemisphere focused around 10N/120E (latest 3-day average of total and anomalies here).  From monitoring it appears this latter region is the result of a consolidation of both eastward moving Indian Ocean and retrograding western Pacific convection.  In fact, refering back to the bottom panel of Fig. 1, the "+" symbol shown at 160E just before 27 September was another convective flare-up over the anomalously warm SSTs (see TAO data here).  A westerly wind event accompanied this development, and the thunderstorm activity has been moving toward the west since then.    

There continues to be a great deal of uncertainty predicting future evolution of tropical convective forcing.  Statistical and numerical models of the MJO are inconclusive (see MJO forecasts, Additional MJO tools and forecasts).  SSTs continue anomalously warm (~+ 5.-1.0C) across parts of the Indian Ocean and central Pacific, as well as the subtropical Atlantic (plot of weekly global SSTAs here).  Monitoring suggests a consolidation of convection may occur near 120E.   This may also signal a strengthening MJO.  The seasonal transition into boreal autumn adds more uncertainty to any predictions for weeks 1-3.  

With the AAM being low and the strongest tropical forcing across the eastern hemisphere, the atmosphere is believed to be in SDM Stage 1.  However, similar to what was observed during northern Autumn 2004, there is a split flow signature across the northern midlatitudes within this base state.  The MJO time scale would suggest evolution into Stage 2 by the end of week 3. Considering the seasonal cycle in the presence of eastern hemisphere tropical convective forcing, the deepest lobe of the polar vortex should intensify across central and east Asia during the next several weeks.  Initially this may place a trough across the central and east Pacific with split flow across North America (ridge across Canada with a closed lows across the southwest and southern part of the country).  However, this flow should consolidate and lead to western USA trough and southeast states ridge by the end of week 2, with the trough shifting toward the Plains by the end of week 3.  The ensemble numerical models give some support to this scenario.  The following prediction is made with extremely low confidence.

Week 1 (6-12 October 2005):  The atmosphere is expected to remain in SDM Stage 1, but with split flow across North America.  Early in this period the most significant rainfall should be across the Gulf of Mexico coast and eastern USA, possibly due to another tropical cyclone after Tropical Storm Tammy.  Please see the latest statements from the tropical prediction center here.  By late in this period the emphasis for precipitation should shift back to the west to locations such as the southern Rockies and central/southern Plains.  Portions of the southern/central Rockies may receive significant snowfall (which could affect the Front Range), depending on the availability of cold air, while the southern Plains has heavy rainfall.  The western states should dry out. While much of the central and eastern part of the country has somewhat cooler then normal temperatures, west and Pacific northwest states should warm to above normal temperatures.  Please see the CPC Drought Monitor for areas of dryness and the latest official outlooks and statements from Storm Prediction Center not only for severe storms, but also fire weather concerns.  

Week 2 (13-19 October 2005):  Toward the end of this period there should be more progressive and baroclinically energetic full-latitude troughs coming onshore into the western USA.  This would favor the western states to cool down while the east and southeast returns to above normal temperatures.  If a storm track develops from the southern Rockies to the Great Lakes, that would suggest more active weather including the possibility of severe local storms and heavy rain across the Plains and winter storm concerns for the Rockies by the end of this week.  

Week 3 (20-26 October 2005):  Should a slowly propagating MJO move into the western Pacific, a transition to SDM Stage 2 would be most probable.  That would favor a ridge from the east Pacific into Alaska with the downstream trough moving into the Plains states.  However, given the large uncertainty, useful temperature and precipitation anomalies for week 3 cannot be specified. 


 Latest CDC Ensemble Forecast

 Latest NCEP Ensemble Forecast

Additional NCEP Ensemble output

Latest Canadian Ensemble Output