Real-time Weather-Climate
Discussion and Predictive Insights
Edward Berry, NWS and Klaus Weickmann, ESRL/PSD
Since our last discussion (
The signal from the Madden-Julian Oscillation (MJO) has
continued to be weak, as indicated by monitoring tools such as the Wheeler plot
(shown here)
and the coherent modes Hovmollers (here). Tropical
convective forcing became strong across the region of
Part I presents an overview of the SST and tropical convective evolutions since
late 2005. One purpose of this report is to extend the February 15th
2006 discussion, which documented a transition from El-Nino to La-Nina
over a period of about a year. Here we will give particular emphasis to
the global circulation transitioning to a La-Nina base state and the
utility of the Global Synoptic-Dynamic Model (GSDM) to diagnose and predict
that evolution. A synopsis of the current weather-climate situation with
predictive insights follows 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 MJO Discussion and tools
Part 1. Weather-Climate Overview
Figure 1 (below) shows two time-longitude sections of near equatorial five-day averaged SST (left) and SSTA (right) in deg. C. The life cycle of the central Pacific warm event (denoted by "EN" with SSTAs in excess of 1C) is seen from boreal summer 2004 extending into spring 2005. SSTs of 29C and greater, a threshold for supporting persistent tropical convection, extended east of the date line (vertical dash line). The evolution toward La Nina (LN) is seen from fall 2005 up through February 2006, with the 29C and warmer SSTs well to the west of the date line. During March, surface westerly wind anomalies appeared across the equatorial central and east Pacific linked to an eastward shift of the tropical convective forcing (discussed below). These westerly wind anomalies initiated a weak Oceanic Kelvin wave-like response, leading to a warming of the basin. Warm SSTs shifted east from the western Pacific and it warmed to at least 27C across most of the equatorial Pacific Ocean (indicated by orange dashed line on Fig. 1). However, the 29C isotherm remained well west of the dateline for the 7th consecutive month.
Figure 1 (Hovmoller plots of SSTs and SSTAs; latest images of various TAO/TRITON SST monitoring tools are here )
Figure 2 (below) is a time-longitude plot of the near equatorial (7.5 N-S) anomalous outgoing longwave radiation (OLRA) field since late fall 2005. Recall we use OLRA as a proxy for deep tropical convection. Multiple evolutionary behaviors of the tropical convective forcing have been annotated, particularly to emphasize important global circulation responses.
Recall from the February 15, 2006 discussion the tropical convective forcing
consolidated around 120-140E during December 2005, attributable to the emerging
cold La-Nina event (indicated by the black horizonal dashed line ~ December 10
on Fig. 2). Since then, at roughly 30-day intervals, eastward movements of the
tropical forcing have occurred, including contributions from the MJO
(phase speeds ~5 m/s) and other convectively coupled modes (phase speeds ~10-15
m/s). Large scale tropical thunderstorm clusters organized across the
eastern IO, propagated east to roughly 140-160E, then stalled and even
shifted back to the west. Convectively coupled Kelvin waves (Kw) tied to these
eastward movements subsequently propagated into the WH. The SST
distribution due to La-Nina was believed to be responsible for the convection
to stall, which was at about the eastern periphery of the 29C and warmer SSTs.
The "A-E" annotations with the blue brackets on the left side of Fig.
2 subjectively define periods when persistent global circulation anomalies
occurred. These periods can also be placed in a particular stage of the
GSDM. The slanted heavy dashed lines from left to right illustrate the eastward
movement of the tropical convection, numbered 1-5 (time increases toward the
bottom), with the vertical lines denoting the stalling. The right
purple brackets show the interval between each event in days, while the green
squares isolate a few of the WH tropical convective signals.
Period "A" started about the time of the Eastern Hemisphere
interannual consolidation of tropical convection in early December 2005 that
continued into the first few days of January. During this period event #1
moved east into the west Pacific by mid-December, following by a westward drift
of less enhanced tropical convection denoted by the dotted line. A fast
signal propagated into the Western Hemisphere, partially through wave breaking
processes from Pacific Ocean synoptic developments. Period A had an
inconsistent circulation-tropical convective signal and thus is classified in
the GSDM as Stage 1 for purposes of convection and Stage 3 for circulation,
more on this later.
Time period "B" is when the seasonal atmospheric circulation
transitioned to La-Nina conditions. The event is similar to other boreal
winter, seasonal "wakeups" to
the presence of SST anomalies. The most famous is of these is the
"El-NoShow"of January 1998. The GSDM framework was generally
followed, but in this case with a La Nina "flavor". As tropical convection emerged into
the Eastern Hemisphere (EH) during early January, the EAJ weakened, and GSDM
Stage 1 finally materialized in the circulation by about the middle of January.
Eastward shift # 2 came into the western Pacific by about the 20th,
consisting of a KW and a moderate signal of the MJO. Toward the end of
January tropical convection became very intense along the SPCZ, with OLRA
~minus 70 W/m**2 and less (not shown). The vertical dashed line shows the
stalling of the main convective signal although some wave forcing does spread
east and excite convection over the Western Hemisphere (WH). A rapid
increase of zonal mean westerly flow occurred throughout the tropical and
subtropical atmosphere leading to GSDM Stage 3.
The WH convection moves east and consolidates with a westward-propagating convectively coupled Rossby mode across the IO by mid-late February 2006 to initiate Period "C".. GSDM Stage 1 develops, which is a return to stronger La Nina conditions. As observed, for instance, from time-longitude sections of 250mb meridional wind anomalies from 30-60N, Event #3 primarily produced amplitude variations of synoptic events.
The GSDM Stage 1 pattern began to break down toward the end of March tied to
convective Event #4. A complex chain of events transpired that
likely included impacts from the seasonal cycle. Possibly forced by
extratropical wave trains, a westerly wind event (WWE) across the central and
eastern equatorial Pacific initiated a weak oceanic Kelvin wave.
That lead to a rapid warming of the cool SSTAs across the equatorial
cold tongue and an overall weakening of the La-Nina SST pattern.
After about March 21st, the persistent trough along the USA west coast
shifted back into the East Pacific.
A well-defined KW initiated near the end of Event #4 contributed to a WH signal
and yet another IO/Indonesia consolidation, convective Event #5. GSDM
Stage 4 would define the time denoted by Period "E". In Section 2 we
discus what has occurred since 23 April 2006.
Figure 2 (Hovmoller plot of near equtorial OLRA; latest image here; additional plots here )
As illustrated in Fig. 3 (below), a time-averaged position for the center of the enhanced convection was ~120-140E, which would be roughly GSDM Stage 1. However, the East Asian Jet (EAJ) was expanded across the North Pacific, which would be the expected circulation from GSDM Stage 3. An expanded EAJ would is not statistically consistent with EH tropical forcing, suggesting nonlinear eddy feedback processes from the midlatitudes contributed to the maintenance of that jet. This period can be thought of as a GSDM Stage 3 circulation with GSDM Stage 1 convection.
Figure 3 (below) presents two averages of 250mb vector wind
anomalies and OLRA to further describe important aspects of the 2005-06 cold
season. The top 2 panels are for ~December 2005, roughly period
"A". Here we see the extended EAJ/North Pacific Jet with
anomalies in excess of 24 m/s with the tropical convective forcing ~120-140E.
A low amplitude wave in the westerly wind anomalies may represent a
Rossby wave excited by the Indonesia convection (see this link for a
nice write-up). There is a secondary wavetrain that emanates from the
central North Pacific through North America and into subtropical western
Africa. The latter would project onto a positive phase of the PNA and
negative phase of the NAO. We speculate that synoptic developments within
the base state of the low amplitude Rossby wave contributed to secondary
forcing of the wavetrain over North America and the Atlantic Ocean. Notice the
large anticyclonic circulation anomaly across northern Asia (which would lead
to anomalously high MSLP), as well as the positive OLRA in region around
Hawaii. The east-west negative OLRA along and off the USA west coast
reflects the precipitation enhancement as the anomalous North Pacific Jet
impacted that region. GSDM Stage 3 would characterize the circulation for this
period.
The bottom panels are ~March 2006, periods "C" and "D".
The OLR shows a similar enhancement as during December 2005 over
Indonesia and suppression across the central and east Pacific. The zonal
scale of the active convection is slightly larger. However, the
circulation anomalies are completely different. The EAJ is retracted with split
flow across the North Pacific. There are twin subtropical cyclones near
the date line, with the northern member contributing to flooding rainfall
across Hawaii. The negative OLRA signal of ~ minus 40 W/m**2 is a
footprint of that situation, as is the weaker (north-south) signal along the
USA west coast. In both of these situations much of the west coast
received substantial precipitation. The coherent Rossby wave train
residual for March 2006 would project onto the negative phase of the PNA. Note
that it appears to emanate from the active convective region near
Indonesia. The SST, tropical convection and circulation response projects
onto a La-Nina and GSDM Stage 1.
Figure 4 (below) shows plots of
vertical and zonal mean anomalies of tropospheric relative angular momentum
(AAM), top panel, a time series for the global AAM anomaly, middle, and finally
the flux convergence of the AAM transports, bottom. These are generated
from the reanalysis data using a 1968-1997 climatology. The second level
of shading in the top panel represents a vertical and zonal mean zonal momentum
at 30N of ~3 m/s, while the second level of shading in the bottom panel
represents a vertical and zonal mean zonal momentum tendency at 30N of
~0.4m/s-day. The periods "A", "B" and "C" are the
same as above. In the context of the earth-atmosphere angular momentum
cycle (budget), the greater the globally averaged westerly (easterly) flow, the
higher (lower) the AAM. Loosely speaking, for Fig. 4, the warm (cool)
colors indicate zonal mean westerly (easterly) anomalies and flux convergence
(divergence) of AAM transports. Convergence (divergence) of AAM
transports would lead to more westerly (easterly) flow. Recall that, in
the northern hemisphere, depending on the details of the wind fields,
northeast-southwest (northwest-southeast) tilted troughs and
ridges transport AAM poleward (equatorward).
The top panel shows a persistent pattern of zonal mean easterly flow anomalies
that start in early December 2005 near the equator and shift slowly poleward
into the subtropics over the following month. A more rapid poleward shift
during January 2006 represents the moderate MJO associated with Event #2.
A persistent pattern of zonal mean zonal wind anomalies with easterlies in the
subtropics then follows. Zonal mean westerly anomalies with the extended
EAJ during December 2005 also move poleward to 45N.
During period "A", the global AAM was ~1 standard deviation
below the 1968-1997 climatology (middle panel) and strong flux convergence of
momentum was occurring into ~30-35N (bottom panel, dashed horizontal purple
line). Eddy feedbacks and east Asian topography likely played a role in
maintaining the extended jet despite the Indonesian convective forcing, as
discussed for Fig. 3.
During "B", as convective envelope #2 moved east, westerly flow was
added to the tropical and subtropical atmospheres (working with other processes
described by the GSDM). That caused the zonal mean anomalous easterlies
to propagate poleward with westerlies replacing them. As indicated by the
dashed red line, by about January 26 the global AAM anomaly quickly become
positive, meaning there was a maximum positive tendency of AAM. GSDM
Stage 2 would best describe the global circulation at that time. The
zonal mean anomalous easterlies being forced into the midlatitudes by event #2
and the positive AAM maximum are believed to have been critical phyical
processes for the global circulation to finally behave as would be expected
during a La-Nina, which was period "C". In contrast to
"A", there was flux divergence of AAM from the midlatitudes during
"C"; again, in addition to other behaviors. GSDM Stage 1
described the situation during "C".
Figure 4 (Plots of, top panel: vertically and zonally averaged tropospheric AAM
anomalies; middle: globally averaged AAM anomaly; bottom: flux convergence of
AAM transport. Latest reanalysis plots here
and here;
additional plots here.
See text for details.)
We describe nearly one cycle of the SDM going from Stage 3-4
to 4-1 to 1-2 to 2. On February 12, the tropical convective forcing was
starting to come back into the Eastern Hemisphere (see Fig. 2). GSDM
Stage 3-4 would still best describe this situation. However, the tendency
of AAM was approaching a negative minimum and a transition to GSDM Stage 4 was
imminent (links to yearly AAM plots here).
The anticyclonic gyre over northwest Canada was continuing to retrograde
as the EAJ retracted. A wave energy dispersion interacting with this
evolution lead to an anomalously deep eastern USA trough and subsequent
significant winter storm for many of the major mid-Atlantic and northeast
cities.
On February 19, the circulation was transitioning from GSDM Stage 4-1 to Stage
2. A severe Arctic outbreak affected most of the western and central USA,
which was in sharp contrast to the mild 2005-06 winter. March 12 is an
example of a classic rendition of GSDM Stage 1, typical of period "C"
discussed above. AAM was near a negative minimum (~2 standard deviations
below the 1968-1997 reanalysis data climatology) and there were twin anomalous
subtropical anticyclones near 120E and twin cyclones over the east Pacific.
A RWD linked to the subtropical anticyclones led to the development of an
anomalously deep and intense western USA trough. The latter was
responsible for at least 1 widespread outbreak of severe local storms across
the Plains (see SPC storm reports here).
March 12 also had large anomalous anticyclonic circulation anomalies
across northeast Asia and west of Alaska.
By April 2 the GSDM Stage 1 response was weakening, tied to the fading La-Nina.
As shown by Hovmoller plots (do plots here) of fields such as 250mb
meridional wind anomalies, the trough position which was along the USA west
coast shifted back into the eastern Pacific after March 21 (leading to period
"D"). The shorter wavelengths are likely the result of a
seasonally weakened meridional temperature gradient, which may have combined with
the emerging western Pacific (event #4) convection to lead to the retrogression
of circulation anomalies (and a weakening of the northern member of the twin
subtropical cyclones responsible for the flooding Hawaii rains).
Toward the end of April important changes to the circulation occurred.
Tropical convection redevelops over the Indian Ocean and Indonesia
leading to event #5 shown on Fig.2. On April 23 the upper level anticyclones
are near Indonesia and westerly flow anomalies are evident over the northern tropics
(~GSDM Stage 1). By April 28, the westerlies have moved into the Indian Ocean
and the anticyclones have moved east, suggesting GSDM Stage.
Figure 5 (Sequence of selected daily mean 250mb vector wind anomalies. Latest animations of these and other fields here)
2. Predictive Insights
SSTAs across most of the central and east equatorial Pacific are within .5 deg C of normal, with SSTs ranging from ~26C near 100W to 29C and warmer just west and southwest of the date line. SSTs of 29C and higher also extend from the southwest Pacific into the IO. At depth, anomalies around plus 1-2C extend from 50-250m east-west along the equatorial cold tongue, meaning a slightly deeper than normal oceanic thermocline. SSTAs from the IO into the west Pacific are at least plus .5-1C, with the Pacific horseshoe experiencing ~ plus 1-2C. The Caribbean into much of the NTA also has SSTAs ~ plus 1-2C. Latest prediction from CPC expects ENSO-neutral conditions to prevail during the next 3-6 months (see latest TAO data here, ESRL/PSD data here ).
At this time the signal from the MJO is weak with the WH upper tropospheric
divergence signal beginning to re-emerge into the IO. Satellite imagery (eastern
hemisphere, full-disk west Pacific, mtsat, IO, Africa; other imagery here ) still has enhanced convection
extending from the east Pacific ITCZ across tropical South America and into
especially central Africa. However, during the last few days there has been a
rapid increase in the tropical convection across the very warm SSTs of the
central IO particularly near the Bay of Bengal. Tropical convection is
less organized from the west Pacific toward Southeast Asia (with the important
exception of Typhoon Chanchu; see here).
Convection has also become a little better organized along the SPCZ over
the warm South Pacific SSTs east of Australia. Suppression, which has been over
the IO for a few weeks, has abruptly shifted into the northwest Pacific Ocean
(latest 3-day averages of OLR total and anomalies, and other data here ).
Statistical and numerical models of the MJO (see ESRL/PSD MJO tools , BMRC MJO tools, CPC MJO tools) generally support the notion of enhanced convection continuing to organize across the IO from tropical Africa by about early week 2, with suppression generally north of Indonesia. SSTs have also remained slightly above average across portions of the central and southern Indian Ocean (~ plus 1-2C), which can be a precursor to a re-emerging MJO into the EH. Finally, SSTs are ~ plus 1-2C above normal across the tropical southwest Pacific, as well as the subtropical North and South Pacific and Atlantic Oceans. These regions will also contribute tropical convective forcing at times. The seasonal cycle will play a role to whatever evolutionary route the tropical convective forcing takes during the next few weeks.
Since April 28th the interplay between SSTs, tropical forcing and circulation response-feedbacks have produced a situation best described by GSDM Stage 3. Recently, with the tropical convection returning to the IO, the global circulation is believed to be transitioning to GSDM Stage 4. Plots of recent daily mean 150mb and 250mb daily mean vector wind anomalies clearly present a signal of twin subtropical centered anticyclones ~ 20 W as well as the lingering pair just east of the date line. Subtropical cyclones are present across the east IO with upper tropospheric wind anomalies ~20 m/s. The North Pacific Jet has retracted (with amplification) during the last several days.
AAM anomalies are ~.5 standard deviations above the 1968-1997 climatology, with anomalous zonal mean westerly flow across particularly the SH tropical and subtropical atmospheres, and the midlatitudes of both hemispheres. Anomalous zonal mean easterly flow exists along ~30N and 30S, as well as the NH polar latitudes. The latter are actually a residual of the "bursting anticyclones" that started in March 2006(discussed above).
AAM tendency has again become positive (~ 2 standard deviations; see plot ), with much of that due to positive torques from north-south mountain ranges including those in East Asia and the Andes (see plot). The frictional torque remains negative, and will remove the added zonal mean westerly due to the mountains (see plot ). In the context of the GSDM, right now we are seeing a submonthly component helping to maintain existing anomalies. However, with the IO convection intensifying, we believe Stage 4 is most probable by the end of week 1, with a transition to GSDM Stage 1 during the week 2-3 time scale. Of course, the details (timing, location and amplitude) are unclear after roughly day 3.
For the PNA sector, and particularly the CONUS, a ~140W trough, ~110W ridge and eastern USA trough still looks like a good bet for much of week 1. However, these synoptic features are expected to progress by the end of this period as the circulation evolves into GSDM Stage 4. Nearly all models and their ensembles have locked onto this prediction, especially the CDC and NCEP ensembles. The current high latitude retrogression of anticyclonic circulation wind anomalies extending from the North Atlantic Ocean (leading to a negative NAO) into eastern Canada will be a contributor to the weather across the CONUS not only for week 1, but afterwards as well.
As the North Pacific Jet retracts and interacts with the retrograding anticyclone, a ridge across the central Pacific possibly extending well into the Arctic is probable as week 2 approaches. A downstream stronger trough should begin evolving just off the USA west coast, linking with a STJ extending into the southwest and south central states. A ridge is probable across the central USA as a trough lingers across the east and northeast states. During the week 2-3 time-scale, an active GSDM Stage 1 pattern may return although farther north than March due to the seasonal cycle. Experience tells us that the numerical models will likely struggle again as the tropical convective forcing comes back into the EH, particularly should a GSDM Stage 4-1 transition occur. In fact, inconsistencies between model forecasts have already been increasing during the last 3-5 days. The following are images of model output from the CDC ensemble.
Figure 9. (Week 2 ensemble mean 500mb height anomaly from the CDC ensemble; latest forecast)
Figure 10. (Week 2 calibrated tercile probabilities for temperature from the CDC ensemble; forecast )
Figure 11. (Same as 10 but for precipitation; forecast)
Week 1 (17-23 May 2006): Overall, still not a very active pattern for much of the USA. Roughly the eastern third of the country looks to have below normal temperatures while record maximum temperatures are probable for locations such as the Intermountain west and perhaps the Front Range into the Southern Plains. Most shower and thunderstorm activity will be east of the Mississippi River, with little precipitation to the west. While the Northeast states gradually dry out, rainfall should increase across the Pacific Northwest. Even though moisture transport from the Gulf of Mexico will be minimal, the possibility of one or two episodes of strong northwest flow thunderstorms for the Northern Plains/Upper Mississippi Valley and Great Lakes cannot be dismissed. Finally, record low minimums and late season frost may be a concern for at least the upper Great Lakes States.
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. Finally, the CPC
CPC
USA Hazards Assessment for offers additional insights not only for possible
week 1 high impact weather, but week 2 as well.
Week 2 (24-30 May 2006): Please see Figs. 9-11. We think these figures are reasonable; however, precipitation probabilities should be increased from the Northern Plains into the Great Lakes as well as the Pacific Northwest. GSDM Stage 4 would be probable by the start of this period, possibly transitioning to GSDM Stage 1 by the end. As Gulf of Mexico moisture transport improves (SSTs are above average there, ~ plus .5-1C), severe local storms may become an issue for much of the central USA, particularly centered on locations such as Nebraska-Iowa. Warmth should also spread eastward from the Central and Southern Plains into the Deep South. In fact, portions of the Southern Plains may see record highs with temperatures in excess of 100F. Near normal to slightly cool conditions may persist across the east and particularly for the Northeast while below normal temperatures spread across the Pacific Northwest.
Week 3 (31 May- 6 June 2006): A transition from GSDM Stage 4-1 and possibly to Stage 2 is a concern for this period and may continue into week 4. That would mean an active southwest flow storm track pattern for the Plains as troughs come into the western states and then move northeast into the Upper Mississippi Valley. Tropical moisture transport from the Gulf of Mexico should be plentiful by this time. The weather would be similar to what was observed during March, only farther north and west. A focus for severe local storms may occur from the Northern/Central Plains into the western Great Lakes States.
Additional NCEP
Ensemble Output
Latest
Canadian Ensemble Output
Latest
Deterministic ECMWF Forecasts
Appendix (Interesting synoptic details)
Figure A1 (below) complements Fig. 3. Panels 1 and 3
are anomalies of mean sea level pressure (MSLPA). Notice that during
December 2005 there were large positive anomalies of MSLP across most of
northern and eastern Asia with negative anomalies across the central Pacific
Ocean (absolute magnitudes ~ 15mb and larger). That situation was
essentially reversed during March 2006; however, the anomalies were not as
large both in terms of magnitude and spatial coverage. The
point, for December 2005, is that the additional westerly torque (source
of westerly angular momentum) imposed onto the atmosphere
by anomalously high MSLP on the eastern slopes of north-south mountain
ranges of the Asian topography (positive mountain torque) likely contributed to
the extended EAJ (and anomalous flux convergence ~30-35N). The opposite
would be true for March 2006.
In reference to the bottom panel of Fig. 3, recall the above discussion contrasting the divergence of the AAM transports between roughly periods "A" and "C". The bottom panel of Fig. A2 (below) is an extension of that record. During GSDM Stage 1 situations it is generally observed that AAM is fluxed from a subtropical sink into a midlatitude source. However, we have observed ~10 day meridional variations of these transports during GSDM Stage 1 which are linked to midlatitude synoptic eddy exhange processes. Figure 6 presents a fairly well defined situation of such an occurrence during the first part of April (blocked off by the black vertical lines).
The top 2 panels are plots of daily mean 250mb vector wind anomalies for the dates shown. Anticyclonic and cyclonic circulation anomalies are annoted as red "Hs" and "Ls", respectively. The magnitude of the circulation anomalies is indicated by the scale on the bottom of the middle panel. On April 6, as part of wave trains trains dispersing across central and south Asia into the central subtropical Pacific Ocean, a northwest-southeast orientation of at least 1 pair of anomalous trough and ridge axes was present (red dashed lines). That tilt would favor an equatorward flux of AAM, as seen by the downward pointing arrow on the bottom panel of Fig. A2. On April 12 not only was the orientation of the anomalous troughs and ridges southwest-northeast, but more pronounced than the tilts seen on April 5 (non-linear behaviors). AAM was strongly fluxing poleward at that time, with the maximum convergence around 50N.
Figure A2 (Sequence of
two selected 250mb daily mean vector wind anomalies and AAM flux divergence)