How the Tropics Induce Major Pattern Changes in the Midlatitudes

[phil882]

I originally was just going to include this in the Medium Range Forecast discussion, but after starting to write out everything, I realized this actually might go a long way towards explain just why the MJO has such a significant influence on our shifting atmospheric patterns over the United States and elsewhere. Hopefully, it represents a small glimpse on just why the tropics are so important when forecasting in the medium range, looking for pattern teleconnections.

I hope I'm not stepping on anyones toes here. HM has already brilliantly described the global regime shift that will likely allow this MJO episode to propagate. I'm merely trying to explain how the MJO induced convection will affect the mid-latitude pattern.

So here I hope to explain why I think we will see a major shift in the the weather pattern across the United States for at least the first 10 or so days of February.

As most know, the shifting of the MJO into phases 8, 1, and 2 are typically favorable for cooler temperatures over the Eastern United States during winter. Indeed, it looks like we may be looking towards such a shift in the upcoming week. However, its also important to understand why this allows the pattern to be favorable for increased troughing over the eastern United States. I'll try to break it down in an easy to understand way that should explain why we be about to see a significant shift in the overall pattern, one that we have yet to witness thus far since the start of winter.

First, lets look at look at what the 250 hPa height anomalies looked like back on the 16th of January. All of the following maps are from the GFS Analysis from Kyle Griffin's model page at the particular time interval noted on the image.

Note in this first image all the negative height anomalies over the tropics. This depression of the height field is related to the lack of convective activity over the region. This is actually well represented by the negative 250 hPa temperature anomalies over the tropics, since the tropopause is typically higher than 250 hPa in this region and lower temperatures at this pressure level can be attributed to a lower tropopause and thus lower heights. Note also the displacement of the mean 250 hPa ridge to the west towards the Western Pacific with the apex near the longitude of Japan. This fact is important, as the ridge center in this region also allows the focus on the Pacific Jet to be just off the Asian coastline, with a jet streak of 100+ m/s (200+ knots!) (as seen below).

This jet streak allowed for repeated cases of surface cyclogenesis across the midlatitudes over the Western Pacific thanks to a favorable jet streak configuration. However, such a pattern is not favorable for an east coast trough because this configuration leads to an amplified ridge aloft located over the the Aleutians. Following the teleconnections downstream, this typically leads to a west coast trough and then an east coast ridge, which was generally the case in the analysis. This has pretty much been the mean state of the atmosphere for the majority of January, mainly thanks to the combination of cool La Nina SSTs suppressing heights over the Central and Eastern Pacific combined with a MJO that seemed eternally stuck in phases 5-6 which only further helped to amplify this signal.

Ok now lets flash forward to the 120 hour GFS forecast valid for February 4th at 00z. Note that we have pretty much seen a pattern reversal regarding the 250 hPa heights.

Wow, there is now a broad trough centered just west the Aleutians with negative height anomalies, while there is now a broad and pretty anomalous ridge over the tropical Central Pacific. This is in response to a progression of the MJO as seen by the now very anomalous positive 250 hPa temperatures over the Central Pacific. If the GFS is to be believed (still a major caveat at this time frame although it does have the ECMWF's support) this would be the first major forward progression of the MJO since meteorological winter started. Another important feature is that this MJO completely overwhelms the La Nina signal over the East Pacific, with little evidence of negative 250 hPa temperature anomalies across the tropics in the Central and Eastern Pacific. Previous MJO episodes this winter have failed to create a dent in this dome of negative 250 hPa temperaturtes in the deep tropics across the Central and East Pacific, but this future MJO progression has the potential to break the chain.

So how does this affect the response downstream across the United States. Take a look at how our jet stream has changed with this renowned progression of the MJO.

The 100+ m/s jet streak is now located over the Central Pacific (where the subtropical ridge apex is located at 180 degrees) and is no longer limited to the Asian coastline. In addition, the jet streak is much larger in size, owing to the increased gradient between the strengthened subtropical ridge responding to the MJO and the stronger polar trough located just west of the Aleutians. This new configuration of the jet streak tends to favor surface cyclognesis over the Aleutians. Once again following typical downstream teleconnections, this forces a ridge over the western North America coastline and then an East Coast Trough. The signal doesn't develop immediately, but gradually by 180 hours as the MJO progression continues.

So thus, we have reasoned why, based on a tropical perspective (mainly following the MJO), we see certain temperature and trough configurations in the Pacific which translates down to the United States.

The main idea that I want to drive home (and many other mets here have alluded to this repeatedly) is that convection (or lack of convection) in the tropics invokes an important mid-latitude response.

Where deep convection in the tropics occurs, it will force heights in the mean subtropical ridge to rise. This height rise will likely result in a strengthening of the jet stream producing an increased wind maximum aloft at that same longitude of the convection. Using a linear jet streak configuration, enhanced divergence aloft in the left exit region of a jet streak tends to aid cyclogenesis. Where cyclogenesis occurs, there is a downstream response of ridging. These same ideas don't just apply to the MJO, but also to La Nina and El Nino episodes as well.

[LongRanger]

+1 nice writeup

[phil882]

To add fuel to the proverbial fire, it seems that there is potential this MJO episode might be one of the strongest since records have been kept in phase 7-8. The plume diagram is literally off the charts (+4 Sigma!). This is a mightily impressive MJO episode. The ECMWF is much more tempered, but still is a rather impressive 2.5 sigma in phase 7.

GFS MJO Forecast:

ECMWF MJO Forecast:

[CoastalWx]

Good write up Phil. Nice way of looking at how the upper level features adjust to this burst of tropical convection and associated latent heat release.

[BethesdaWX]

Thanks a ton for making this thread, I really enjoyed reading it. Do you suspect the -AO as the culprit for allowing the progression of sustained convection?

[Bryan Wood]

I enjoyed reading this, thanks for posting! Is there a certain quantity of storms (and/or location of them) that you look for to push the mid-latitude response and the heights in the ridge to rise?

[phil882]

Thanks a ton for making this thread, I really enjoyed reading it. Do you suspect the -AO as the culprit for allowing the progression of sustained convection?

No problem, thanks for reading

Again I'll defer to HM's excellent post explaining the AAM's role in causing a global regime shift which is allowing this most recent MJO episode to propagate into phase 7-8. He pretty much called for this a week ago!

http://www.americanw...ost__p__1315283

[phil882]

+1 nice writeup

Good write up Phil. Nice way of looking at how the upper level features adjust to this burst of tropical convection and associated latent heat release.

I enjoyed reading this, thanks for posting! Is there a certain quantity of storms (and/or location of them) that you look for to push the mid-latitude response and the heights in the ridge to rise?

Thanks guys!

As for the quantity/location of storms, its key to know where you see the mid-latitude storms start to intensify, because this leads to the downstream ridge/trough configuration given by baroclinic growth. In the original post, I highlighted that depending on where your jet streak is located, you will find different areas where mid-latitude surface cyclones start to intensify. For most of January, this took place over the west Pacific in the left exit region of a jet streak that hugged the Asian coast. This lead to ridge amplification over the Aleutians. The GFS is showing that in February as the MJO moves across the Pacific, we should see the jet streak shift eastward into the central Pacific. This will allow surface cyclones to intensify over the central and east Pacific and amplify the 500 hPa flow over the west coast of North America.

[Geos]

Great write-up! Good to hear the MJO is finally progressing along and breaking out of it's stagnant regime.

[OKpowdah]

I want to bump this, and say excellent discussion Phil!

Also, rather than start a new thread, I wanted to pose a question here ...

What are some questions remained to be answered about the MJO in particular?

I know the basics and I also understand there is uncertainty with some of the driving mechanisms behind the MJO

[am19psu]

What are some questions remained to be answered about the MJO in particular?

What exactly is the MJO? It's not predicted by any of the equations of motion (unlike, e.g., Kelvin waves)

What causes it to form and decay?

What influences its phase speed?

What influences its magnitude (though I think this one is a bit easier since it's convectively coupled)?

[phil882]

What exactly is the MJO? It's not predicted by any of the equations of motion (unlike, e.g., Kelvin waves)

What causes it to form and decay?

What influences its phase speed?

What influences its magnitude (though I think this one is a bit easier since it's convectively coupled)?

I should have time a little bit later today to try to add a little bit to these questions. We had a nice tropical discussion this morning with Dr. Roundy and there are a lot of interesting things going on with this upcoming episode. He ultimately agrees that this MJO episode should help to enhance Western North American ridging... and the latest ECWMF medium range products agree that a strong PNA ridge is likely to set up beyond the day 5 range.

[QVectorman]

What are some of the main parameters used by models to forecast the phase?

Could you briefly explain Roundy's approach to diagnosing the MJO for his OLR anomoly maps? I look at them all the time along with the NCEP data and by eye balling a map of OLR my idea of where the MJO should be doesn't match the actual, so obviously I don't understand the classification process.

Could you share a few of the interesting tid bits mention by Roundy for this up coming wave?

Thanks!

[dtk]

What are some of the main parameters used by models to forecast the phase?

Could you briefly explain Roundy's approach to diagnosing the MJO for his OLR anomoly maps? I look at them all the time along with the NCEP data and by eye balling a map of OLR my idea of where the MJO should be doesn't match the actual, so obviously I don't understand the classification process.

Could you share a few of the interesting tid bits mention by Roundy for this up coming wave?

Thanks!

I'm not sure I understand this question, unless you are talking strictly about the statistical models.

The dynamical models (GEFS, ECM, etc), are discretized versions of the governing equations. Some of the dynamical models will have more success than others based on their characteristics (active coupling to an ocean model, convective parameterization schemes, resolution, etc. ?). How the phases are actually predicted depends on how the dynamical models forecast the components that go into those diagrams (OLR, 200/850 zonal winds, etc.).

[QVectorman]

I'm not sure I understand this question, unless you are talking strictly about the statistical models.

The dynamical models (GEFS, ECM, etc), are discretized versions of the governing equations. Some of the dynamical models will have more success than others based on their characteristics (active coupling to an ocean model, convective parameterization schemes, resolution, etc. ?). How the phases are actually predicted depends on how the dynamical models forecast the components that go into those diagrams (OLR, 200/850 zonal winds, etc.).

Ah...ok, I didn't stop to think that the phase was just forecasted from the output of OLR, 200/850 zonal winds etc. from each model. I thought maybe things such as AAM, mountain torques, frictional torque, GWO, rossby waves, kelvin waves may have been factored in and given more weight. I wasn't sure, I am completely oblivious to how the MJO forecast is generated.

[phil882]

What are some of the main parameters used by models to forecast the phase?

Could you briefly explain Roundy's approach to diagnosing the MJO for his OLR anomoly maps? I look at them all the time along with the NCEP data and by eye balling a map of OLR my idea of where the MJO should be doesn't match the actual, so obviously I don't understand the classification process.

Could you share a few of the interesting tid bits mention by Roundy for this up coming wave?

Thanks!

I don't know if you were already previously aware of this, but In order to be able to better identify the MJO, the OLR data is actually filtered. In order to do this, the total spectrum of OLR data is organized in a log time vs. wavenumber plot. Next, the data is divided by the background spectrum in order to normalize the data (figure below). From this point, you can see how the OLR data starts to cluster into different categories (Kelvin waves, Mixed Rossby Gravity waves, Equatorial Rossby waves, TD like waves ect.) However, also notice towards the bottom of the graph between wave-numbers 0 and 5 and a period of more than 20 days lies the MJO. It is this OLR data that is filtered to discover the MJO anomaly that you see in a lot of the plots on roundy's web page.

Source: Schreck et al. (2011).pdf >>> A good read for a lot more information!

Now great, we can now see the OLR signal that is exclusively the MJO. But how does this actually help us? Well now that we have divided up the OLR signal into different wave phenomena, we can actually see them all on different plots and see how each different type of equatorial wave influences the total OLR.

http://www.atmos.alb...smovrec523.html >>> Click to view animation of figure below.

The link and image above are a really useful tool to show the past, present, and future evolution of the MJO in relation to its OLR anomalies. The useful thing is that in addition to the MJO, you can also see the OLR contribution of other waves which may suppress or enhance the negative OLR signal of the MJO. Currently, you can see that while the -OLR associated with the MJO is strong, it is currently being countered somewhat by the +ORL anomalies of the inter-seasonal cycle (La Nina) as well a suppressed phase of a kelvin wave which is also producing +OLR anomalies over the region of the MJO. While the inter-seasonal cycle isn't going away anytime soon... the kelvin wave will quickly move out of the way and allow the total -OLR anomalies to expand eastward which will allow the MJO to really show itself in the next few days. For more information on exactly how Roundy calculates these figures and also makes forecasts a month into the future (including their skill... which is quite good!) check this link out.

There is another thing that is worth noticing if you view the animation above. Notice how the MJO is not symmetric about the equator. Right now the strongest -OLR anomalies are focused in the southern hemisphere. This was actually one of the reasons why the MJO using the RMM1/RMM2 phase space was not propagating out of phases 5/6. The MJO was propagating eastward, but it was not centered at the equator, and the -OLR anomalies actually shifted southward over the second half of January allowing +ORL anomalies to form at and north of the equator, which effectively canceled out the signal when you take a longitudinal average. The good news with this MJO episode though is that the MJO is propagating eastward and equatorward with time (at least according to the statistical forecast) which should allow this episode to thrive past phase 6.

As for what Roundy had to say about this current episode... hes certainly interested in the current asymmetrical appearance of the MJO currently and whether or not it will actually become more symmetric at the equator like the forecast projects. For those that like cold in the east, a more symmetrical MJO or even a MJO asymmetric in the northern hemisphere would be preferred for a more significant impact on the mid-latitude pattern in the United States. One other nugget worth sharing is that the reason why the GFS and ECMWF have different magnitudes with respect to the MJO has a lot to do with how much they are developing the TCs over the South Pacific. The GFS is much more bullish on tropical cyclogenesis which would tend to enhance the MJO signal. The ECMWF has been much more bearish with development, and thus the -OLR anomalies are not as impressive and thus a weaker amplitude MJO.

What exactly is the MJO? It's not predicted by any of the equations of motion (unlike, e.g., Kelvin waves)

What causes it to form and decay?

What influences its phase speed?

What influences its magnitude (though I think this one is a bit easier since it's convectively coupled)?

As alluded to, these questions for the most part still remain a mystery, although there are some rumblings on how closely tied Kelvin waves and the MJO are in actuality. There honestly isn't that much separation in the time vs. wavenumber charts and they both share the same dynamical structure when convectively coupled. Some have made an argument that the MJO is just a large Kelvin wave or make up of a multitude of different Kelvin waves. However, most of this is rather speculative at this point, and there hasn't been any published literature on this idea... but this might be the direction some of the research is pointing in the near future, perhaps re-writing the rule book on the MJO.

[QVectorman]

Wow! A ton of great info. Thank you so much for taking the time to write this up! I know it had to take a while, much appreciated!

[Big O]

I don't know if you were already previously aware of this, but In order to be able to better identify the MJO, the OLR data is actually filtered. In order to do this, the total spectrum of OLR data is organized in a log time vs. wavenumber plot. Next, the data is divided by the background spectrum in order to normalize the data (figure below). From this point, you can see how the OLR data starts to cluster into different categories (Kelvin waves, Mixed Rossby Gravity waves, Equatorial Rossby waves, TD like waves ect.) However, also notice towards the bottom of the graph between wave-numbers 0 and 5 and a period of more than 20 days lies the MJO. It is this OLR data that is filtered to discover the MJO anomaly that you see in a lot of the plots on roundy's web page.

Source: Schreck et al. (2011).pdf >>> A good read for a lot more information!

Now great, we can now see the OLR signal that is exclusively the MJO. But how does this actually help us? Well now that we have divided up the OLR signal into different wave phenomena, we can actually see them all on different plots and see how each different type of equatorial wave influences the total OLR.

http://www.atmos.alb...smovrec523.html >>> Click to view animation of figure below.

The link and image above are a really useful tool to show the past, present, and future evolution of the MJO in relation to its OLR anomalies. The useful thing is that in addition to the MJO, you can also see the OLR contribution of other waves which may suppress or enhance the negative OLR signal of the MJO. Currently, you can see that while the -OLR associated with the MJO is strong, it is currently being countered somewhat by the +ORL anomalies of the inter-seasonal cycle (La Nina) as well a suppressed phase of a kelvin wave which is also producing +OLR anomalies over the region of the MJO. While the inter-seasonal cycle isn't going away anytime soon... the kelvin wave will quickly move out of the way and allow the total -OLR anomalies to expand eastward which will allow the MJO to really show itself in the next few days. For more information on exactly how Roundy calculates these figures and also makes forecasts a month into the future (including their skill... which is quite good!) check this link out.

There is another thing that is worth noticing if you view the animation above. Notice how the MJO is not symmetric about the equator. Right now the strongest -OLR anomalies are focused in the southern hemisphere. This was actually one of the reasons why the MJO using the RMM1/RMM2 phase space was not propagating out of phases 5/6. The MJO was propagating eastward, but it was not centered at the equator, and the -OLR anomalies actually shifted southward over the second half of January allowing +ORL anomalies to form at and north of the equator, which effectively canceled out the signal when you take a longitudinal average. The good news with this MJO episode though is that the MJO is propagating eastward and equatorward with time (at least according to the statistical forecast) which should allow this episode to thrive past phase 6.

As for what Roundy had to say about this current episode... hes certainly interested in the current asymmetrical appearance of the MJO currently and whether or not it will actually become more symmetric at the equator like the forecast projects. For those that like cold in the east, a more symmetrical MJO or even a MJO asymmetric in the northern hemisphere would be preferred for a more significant impact on the mid-latitude pattern in the United States. One other nugget worth sharing is that the reason why the GFS and ECMWF have different magnitudes with respect to the MJO has a lot to do with how much they are developing the TCs over the South Pacific. The GFS is much more bullish on tropical cyclogenesis which would tend to enhance the MJO signal. The ECMWF has been much more bearish with development, and thus the -OLR anomalies are not as impressive and thus a weaker amplitude MJO.

As alluded to, these questions for the most part still remain a mystery, although there are some rumblings on how closely tied Kelvin waves and the MJO are in actuality. There honestly isn't that much separation in the time vs. wavenumber charts and they both share the same dynamical structure when convectively coupled. Some have made an argument that the MJO is just a large Kelvin wave or make up of a multitude of different Kelvin waves. However, most of this is rather speculative at this point, and there hasn't been any published literature on this idea... but this might be the direction some of the research is pointing in the near future, perhaps re-writing the rule book on the MJO.

Great information.

In my opinion, this topic should be pinned.

[CoastalWx]

That's very informative Phil. Awesome.

[NavarreDon]

Goodness....great job Phil! Thanks for taking the time to post it all, but can't you just come back to Asheville & forecast our weather?

[phil882]

Verification Time... lets see how well the 120 hour GFS fared with the most recent analysis.

It seems that the GFS was a bit too bullish with the subtropical ridging over the Central Pacific. Looking at the 250 hPa anomalies, you can see this was likely the result of weaker than expected positive temperature anomalies (+ 1 to 2 sigma instead of +3 to 4 sigma). However, you can also see that the GFS was also too warm further west over Indonesia, which has resulted in a weaker subtropical ridge near Asia. Thus the upper level temperature anomaly does seem to be having a significant impact on the upper level ridging (which is not surprising).

Maybe the more important take home message is that the subtropical ridge forcing isn't as strong as forecasted, and that brings into question if the GFS is over-forecasting the amplification of the ridging. A good way to diagnose this is to look at a dprog/dt (change in model prognosis over change in time) over the Central Pacific.

http://www.atmos.albany.edu/student/kgriffin/maps/dprog/F006/500vort/npac/500vort_npac_dprog.html

Note that as the animation moves forward we have a robust 588dm ridging that is rather persistant in the model analysis up to the 72 hours forecast for 00z on the 4th of February. However the final three days of forecasting has revealed a pretty noticeable weakening of this ridge which seems much in line with the weaker than expected +temperature anomalies aloft.

So what does this mean in lamen? It suggests that the convective signal with this upcoming MJO isn't as intense as forecasted. That doesn't necessarily mean the MJO won't be of significant amplitude (in fact it could have little baring on the intensity of this upcoming MJO phase shift) but I'm suggesting it could have an impact on how much the MJO will influence our weather in the mid-latitudes.

I am very curious to see how this plays out... even with the weaker latent heat forcing, if its still a positive anomaly over the central and east pacific... it should still support a +PNA.

[QVectorman]

Phil, what is the lag time for each phase of the MJO specifically, 7,8,1 to US weather. I thought I read 7-14 days, but here they state they found it to be 5-7 days for the 2009/2010 winter from max convection in each phase to max development of a trough over the eastern US. And do you know if they use a lag time in the CPC temp/precip composites? Thanks!

[GaWx]

Phil, what is the lag time for each phase of the MJO specifically, 7,8,1 to US weather. I thought I read 7-14 days, but here they state they found it to be 5-7 days for the 2009/2010 winter from max convection in each phase to max development of a trough over the eastern US. And do you know if they use a lag time in the CPC temp/precip composites? Thanks!

I did get confirmation that there is no lag time assumed in the various MJO phase U.S. temp/precip anomaly maps produced by Earthsat. So, when Earthsat shows a map for average U.S. temperature anomalies associated with, say, phase 8, that map is based on the actual average U.S. temperature anomalies of all days in the database when the MJO was actually in phase 8 (i.e., no lag). Also, these Earthsat maps look similar to the CPC maps. In other words, phases 8, 1, and 2 maps look below normal in most of the E US temperaturewise on both. Furthermore, I've never seen a mention of any lag when any of these maps are shown. So, unless told otherwise about some specific lag period, I always assume no lag is assumed in any U.S. temperature or precip. maps associated with any particular MJO phase. Now that being said, I also know these are just longterm averages and that the actual results for the U.S. associated with any particular future MJO phase will not match exactly and could even be totally different (example warm E US instead of cold for any one composite for the entire period covered by a particular phase 8). That is largely because there are so many other factors like AO, NAO, ENSO, etc., that could easily skew what actually happens, and I realize that these maps were produced by just looking at an average of the actual anomalies for all, say, phase 8 days in the database (again with no lag). In reality, more days were likely cold as opposed to warm thus producing an average anomaly of cold. But I always keep in mind that it could still be warm based on there having been some actual warm periods.

I always like to think of the various indices as tools rather than crystal balls.

[LocoAko]

Great OP and subsequent discussion! I don't know very much about the MJO but this was nicely done. As I learn more and more I'm always sort of impressed at just how much of a role latent heat plays in affecting, well, almost everything. Good stuff. Sucks that the verified temperature/ridge anomalies aren't what was anticipated, though.

[phil882]

One other nugget worth sharing is that the reason why the GFS and ECMWF have different magnitudes with respect to the MJO has a lot to do with how much they are developing the TCs over the South Pacific. The GFS is much more bullish on tropical cyclogenesis which would tend to enhance the MJO signal. The ECMWF has been much more bearish with development, and thus the -OLR anomalies are not as impressive and thus a weaker amplitude MJO.

Time for some verification from this earlier expectation... We now have two TCs currently in the South Pacific basin... Jasmine is a 100 knot cyclone, while Cyril is a 45 knot system. This verification supports the GFS's interpretation of the MJO about 5 days ago, and indeed, the GFS has done much better than the ECMWF with regards to the MJO propagation so far. This is likely why the models are coming around to the idea that the MJO signal will continue to propagate into phases 8 and beyond over the next week.

[HM]

Great OP and subsequent discussion! I don't know very much about the MJO but this was nicely done. As I learn more and more I'm always sort of impressed at just how much of a role latent heat plays in affecting, well, almost everything. Good stuff. Sucks that the verified temperature/ridge anomalies aren't what was anticipated, though.

I agree, great post/thread. This MJO wave is perfect for seeing the connection to the Mid Latitudes.

Thanks for the kudos Phil!

[DaculaWeather]

Awesome Phil! Really good info.

[QVectorman]

I did get confirmation that there is no lag time assumed in the various MJO phase U.S. temp/precip anomaly maps produced by Earthsat. So, when Earthsat shows a map for average U.S. temperature anomalies associated with, say, phase 8, that map is based on the actual average U.S. temperature anomalies of all days in the database when the MJO was actually in phase 8 (i.e., no lag). Also, these Earthsat maps look similar to the CPC maps. In other words, phases 8, 1, and 2 maps look below normal in most of the E US temperaturewise on both. Furthermore, I've never seen a mention of any lag when any of these maps are shown. So, unless told otherwise about some specific lag period, I always assume no lag is assumed in any U.S. temperature or precip. maps associated with any particular MJO phase. Now that being said, I also know these are just longterm averages and that the actual results for the U.S. associated with any particular future MJO phase will not match exactly and could even be totally different (example warm E US instead of cold for any one composite for the entire period covered by a particular phase 8). That is largely because there are so many other factors like AO, NAO, ENSO, etc., that could easily skew what actually happens, and I realize that these maps were produced by just looking at an average of the actual anomalies for all, say, phase 8 days in the database (again with no lag). In reality, more days were likely cold as opposed to warm thus producing an average anomaly of cold. But I always keep in mind that it could still be warm based on there having been some actual warm periods.

I always like to think of the various indices as tools rather than crystal balls.

Thanks for finding that info out about the composites. There is a lag time from the phase until a ridge or trough develops in the continental US. It's found to be 5-7 days according to the latest research. But years ago they use to say it was 7-14 days. Which makes sense because you need time for the atmosphere to respond and something to develop and then it has to travel downstream. It takes an entity 3 to 4 days to cross the US and the physical domain of phase 7/8/1 is at least two and a half times that distance away from the center of the US. Nothing is instantaneous. So it would make more sense to me to use a mean lag time in composites.

[GaWx]

Thanks for finding that info out about the composites. There is a lag time from the phase until a ridge or trough develops in the continental US. It's found to be 5-7 days according to the latest research. But years ago they use to say it was 7-14 days. Which makes sense because you need time for the atmosphere to respond and something to develop and then it has to travel downstream. It takes an entity 3 to 4 days to cross the US and the physical domain of phase 7/8/1 is at least two and a half times that distance away from the center of the US. Nothing is instantaneous. So it would make more sense to me to use a mean lag time in composites.

You're welcome. It sounds like you're thinking that 7-8-1 is the trio of phases associated with later generating eastern U.S. troughing. However, 8-1-2 is colder on average than 7-8-1 (when no lag assumed). Phase 7 is slightly warm on average while it is occurring due to a tendency for E US ridging then while phase 8 averages cold (again with no lag assumed) as the troughing then tends to make its way into the E US. So, since there often are ~5-7 days to go from the start of phase 7 to the start of phase 8, maybe phase 8 tends to be cold due mainly to the often preceding phase 7 first generating the pattern leading to troughing reaching the E US at around the time phase 8 starts. On the same token, maybe phase 1 tends to be cold due to the lagged effects of phase 8, phase 2 tends to be cold due to the lagged effects of phase 1, etc.

Opinions?

[QVectorman]

You're welcome. It sounds like you're thinking that 7-8-1 is the trio of phases associated with later generating eastern U.S. troughing. However, 8-1-2 is colder on average than 7-8-1 (when no lag assumed). Phase 7 is slightly warm on average while it is occurring due to a tendency for E US ridging then while phase 8 averages cold (again with no lag assumed) as the troughing then tends to make its way into the E US. So, since there often are ~5-7 days to go from the start of phase 7 to the start of phase 8, maybe phase 8 tends to be cold due mainly to the often preceding phase 7 first generating the pattern leading to troughing reaching the E US at around the time phase 8 starts. On the same token, maybe phase 1 tends to be cold due to the lagged effects of phase 8, phase 2 tends to be cold due to the lagged effects of phase 1, etc.

Opinions?

Yes, I understand what the pictures show on the CPC. But if you research it a little more, here in the forums the past few days you can see the data that there are more storms in 7/8/1 than from phase 2. Also this paper confirms the connection to 7/8/1. And a variation of Wes image from the -NAO thread below...Shows 7/8/1 more times than not leads to a -NAO thus -AO with a lag of 8-14 days. Which would explain why 2 shows a colder pattern in the composites.

Stole this from ChicagoWx in the subforum but good info showing 7/8/1 being more favorable for snow for Chicago.

I went back and looked at all 6"+ snowstorms at O'Hare Airport since 1974-75. And with that, I tried to tie in what phase the MJO was in for all of those snowstorms. For the most part, the phase is the same for each day of the storm. In the event that the phase changed, I used the first day's number.

ORD data from the Utah Climate Center and the Chicago National Weather Service.

MJO text data used: http://www.bom.gov.a...4toRealtime.txt

MJO phase diagrams: http://www.cawcr.gov...sediag.list.htm

Date, snowfall amount, and MJO phase of each day of the storm in order.

Conclusion, it seems phases 6, 7, 8, and 1 are the most favored overall (chart below). If we take out the Oct, Nov, and April storms/starts, it goes like this: phase 7 (8), phase 6 and 8 (7), phase 1 (6), phase 2 and 3 (5), phase 4 (4), and phase 5 (3). Of course there are many variables to look at when it comes to individual storms, but I just thought it'd be interesting to see this data.