Positive AMO Losing Grip

The positive Atlantic Multidecadal Oscillation (AMO) appears to be losing its grip on the waters near Greenland in the past month.

Above shows a side-by-side comparison of water temperatures on October 17th and November 14th. If we look on the left side (October 17), we see a wide swath of above normal water temperatures stretching from Greenland, through northern Canada and to the waters south of Greenland. The waters off the East Coast also showed heavy warming during that time period. If we look to the right image, taken on November 14th, we now see neutral water temperatures in place around much of Greenland and in Northern Canada. However, waters remain warm to the east of eastern Canada, and have actually intensified off the coast of the Northeast US.

What is setting up here is something called the Atlantic Tripole. In a nutshell, the Atlantic Tripole is shown by cool waters off the coast of the Northeast and East Coast, warm waters in the Gulf of Mexico, Caribbean and southeast Atlantic, and warmer than normal water temperatures in the Greenland waters. Below are temperature and precipitation anomalies during the Atlantic Tripole.

Temperature Anomaly

Precipitation Anomaly

Andrew

November 27-29 Storm: First Call

This is my first call on the November 27-29 potential winter storm system. Let's start with the factors that are going into my first call.

This is the latest ECMWF model. This is valid for the morning of November 26th. It shows our storm system currently in southwestern Oklahoma, with a minimum pressure of 1006 millibars- a pretty weak system overall. Notice the cold air far up in the North Plains. If that were to be pulled south by this storm system, many locations could see absolutely frigid temperatures.

2 days later, on the morning of November 28th, the storm system has moved due east and is now out to sea, hugging the East Coast. Considering we will be in a negative NAO, this is not surprising, but the due east movement of the storm is pretty sketchy, especially considering the ECMWF has been supporting a North Camp (system affects Midwest, Great Lakes) solution for so long.

By the way, so you know, the North Camp was described above. The South Camp is for models that take the system through the South US and up the East Coast. They are called 'camps', because the models will group together for some time (like a camp of people) under one forecast solution (like a group of people with a common cause).

Here's the run-down of the models for camps.

North Camp: GGEM, GGEM ENS, NAEFS ENS
South Camp: ECMWF, GFS, GFS ENS

Based on how the models are treating the system heading out of the Rockies, and the 4 day lag between observed NAO conditions and effects on North America, the system should be forced south, providing a rainy scenario for the region. However, when the storm shoots north and strengthens a bit near New England, this precipitation will turn to snow. Best accumulation will occur nearest the Great Lakes as cold air behind the system brings about lake effect snow potential.

Andrew

ENSO Update: New Forecast Models Show Neutral ENSO Winter

The latest update from the IRI/CPC (International Research Institute/Climate Prediction Center) shows that the wide array of models used in forecasting ENSO phases is looking neutral for this winter.

The individual models are combined into the ensemble mean, shown as a yellow line on the chart. This yellow line keeps the ENSO state where it is at the moment with fluctuations here and there. This prospect isn't all that surprising, seeing as the Madden-Julian Oscillation (MJO) has not been showing any love to the ENSO regions recently that would otherwise bump up the prospects for an El Nino.

You're probably looking at all those models and wondering 'Which one is the best?' Well, the IRI has an archive of compiled forecasts of each model and observed ENSO conditions superimposed. I went through each model and found the most accurate one of the bunch, shown below:

This model is called the ESSIC model. Statistical models are shown in yellow and dynamical models in green. Don't worry, those fancy terms don't need to be defined at the moment. The ESSIC's previous forecasts are in red, and observed conditions are superimposed in black. It appears that the ESSIC model did very well in forecasting the predicted ENSO states, whereas other models had ridiculous warm or cold bias' that rendered their forecasts completely useless.

In summary, the ENSO conditions should stay in neutral for this winter. This means an increased chance for a cool Plains, a stormy south Plains and wet Southeast-East Coast.

Andrew

November 27-29 Potential Winter Storm (Updated 11/20)

This is a post regarding the November 27-29 storm potential. Let's not waste any time and get right into it.

This is the 12z GFS model forecast for the early morning hours of November 27th. This chart combines isobars, areas of high and low pressure, the blue rain/snow line, and precipitation all in one. The GFS is forecasting our system to be resting in the Gulf Coast in the early morning of the 27th. Precipitation looks to be ongoing across a wide swath of that area, with Illinois, Indiana and Missouri possibly getting snow. Mississippi, Louisiana and other states in the same area would likely get rain. This low pressure system is forecasted to be as low as 1015 millibars- pretty high, indicating a weak system.

12 hours later, we are in the afternoon of the 27th. At this point, the storm system has moved northeast and strengthened ever so slightly from 1015mb to 1014mb. Not a very significant change. Precipitation is now very much spread out across the East US. There could be snow ongoing in Pennsylvania, Indiana, Ohio and western New York if this verified. The mountain influence would also favor snow in Virginia. Take note of how the blue rain/snow line is not pushing south in the wake of this storm system. If we take a look farther east, we see a bowing formation of observed precipitation, telling me that this is likely a cold front. If it is, why is the cold air not pushing in behind? A very mysterious case.

Now, we have moved forward to the morning of November 28th, when many are just getting up. Our storm system has now jumped ship and is offshore, with a low pressure of 1000mb. A fair drop from earlier, but still a weak system. Precipitation now extends well offshore, but snow is abundant across the Northeast, with New York and Pennsylvania getting the best inland snow. Vermont and New Hampshire also get in on the action. However, as I showed above, the blue rain/snow line is still not filling south behind the system.

I can't say I'm liking the GFS' solution at the moment, because the GFS has not been as good as the ECMWF model, which we will take a look at next.

This is hour 144 of the ECMWF, also known as the morning of November 26th. We are already seeing a large time difference between the ECMWF and GFS, something that tells me these two models are done agreeing and are trying to get their way. The ECMWF has our storm system on the Oklahoma-Kansas border at a minimum pressure of roughly 1001 millibars. The ECMWF has the system further north and stronger than the GFS at this point. Notice the cold air remains to the north. This is not surprising, seeing as the system has just exited the mountains, where it lost strength and thus its ability to pull in air behind it.

Now, we have gone forward to the morning of November 27th, just 24 hours later. The ECMWF is predicting this system to be in Indiana at this point, with a minimum pressure reading of 1006 millibars. This means that the system has actually weakened as it has gone through the Plains. Note the cold air now flowing in behind the system in the Plains. This is what the GFS is not showing, and I believe that it should be. What is also interesting is that this system and another down on the Arklatex borders are sharing isobars, indicating that they may be trying to come together.

So we now have two 'camps' (different solutions by different models are frequently called camps in the weather world, with one camp taking one scenario and another camp taking a different scenario). The ECMWF is in the north camp (takes the storm north), while the GFS is in the south camp (takes the storm south and up the coast).

Sure, models are good, but we can never forget the ensembles.

This is the GFS Ensemble mean for the morning of November 27th. Unfortunately, the blue rain/snow line did not print for this image. The ensembles are showing abundant precipitation through the Ohio Valley and down into the Gulf Coast. A wider coverage of precipitation near the coast tells me that the ensembles may be favoring that area for the system's placement, just like the GFS model itself.

We now fast forward 12 hours to the evening of November 27th, where we find the blue rain/snow line has returned. Precipitation is most abundant from the Ohio Valley into the Mid Atlantic and Southeast. The placement of that abundance tells me that the low pressure system is taking the route that the 12z GFS took. That is a big victory for those in the Northeast wanting snow to fall from this system.

Another 12 hours later, on the morning of the 28th, the low pressure system has indeed moved up the coast, and it looks like heavy precipitation has hit the Northeast (possibly as snow) as the system moved up and out of the region. Because the ensembles were so indecisive today compared to yesterday as far as low pressure placement goes, I'm not prepared to make a comment on these ensembles until I see the ECMWF Ensembles.

North Camp: ECMWF
South Camp: GFS, GFS ENS

The ECMWF ensembles are currently having this storm system in southwest Oklahoma, a bit further south than the ECMWF model itself. The system has a minimum pressure of 1009 millibars, which is about as strong as the GFS model. Keep a very close eye on that ridge in the Southeast. That ridge is what will make or break this storm for one area or another. Right now, the ECMWF Ensembles appear very concentrated on the idea of the ridge being present during this timeframe- a good sign for Midwesterners.

24 hours later, on the morning of November 27th, the ECMWF Ensembles have this system in the Midwest/Ohio Valley with a minimum pressure of 1015 millibars, meaning that it has weakened as it moved from the Rockies to the Midwest. The Southeast ridge has now moved offshore, but its effects are still present so that the storm was moved north. This does make the ECMWF ensembles a member of the North Camp.

Latest Camp Stats:
North Camp: ECMWF, ECMWF ENS
South Camp: GFS, GFS ENS

I took the liberty of looking at a few other models for their take on the situation, and I have sorted them into camps for ease of access. Here is the latest model camp stats.

North Camp: ECMWF, ECMWF ENS, GGEM (CMC), 
South Camp: GFS, GFS ENS, DGEX, 

It would appear that the North Camp has a slight lead. But let's now look at how this could all play out and not just stare at the models.

This is the 500mb height anomaly forecast from the GFS model (the same one we saw above), valid on the morning of November 27th. You might look at it and see lots of colors, but check out the upper right hand corner. See that big red splat to the east of Greenland? Not only is that a negative NAO, it is an east-based negative NAO. What's the difference, you might ask?

Image from NCSU

This image above shows height anomalies of an east-based negative NAO and a west-based negative NAO on the left hand side. The right hand side shows temperature anomalies for their respective bases. When I say 'east-based', it is in relation to Greenland. In this case, an east-based negative NAO would have the ridge of high pressure (the negative NAO) to the east of Greenland. In a similar fashion, a west-based negative NAO allows the ridge to be west of Greenland. Both types of negative NAO are different from each other. In west-based negative NAO's, the Northeast can expect to see the storms the most often. In an east-based negative NAO, the Midwest tends to get the best storms. Do you see why I'm skeptical, with the GFS showing an east-based negative NAO but bringing the storm up the East Coast?

Also, if you look back to the 500mb height anomaly image, you'll see that a positive PNA is trying to form, shown by a ridge in the northeast Pacific. A positive PNA is best for the Midwest for storms, enhancing my suspicion of the 12z GFS.

In conclusion, this system remains quite a ways away. However, I feel that the ECMWF and the ECMWF Ensembles have the best handle on this system at the moment. I pointed out flaws in the GFS, and I will discard its scenario for that reason.

Andrew

Model Ensembles Supporting Plains, Midwest Snowstorm

It appears that the GFS Ensembles and ECMWF ensembles are in very good agreement on the potential November 27-29 winter storm (see full post here).

This is the forecasted situation for the morning of November 28th, showing 12 hour precipitation, isobars, high and low pressure areas, and the blue rain/snow line superimposed on all of it. The storm system we are watching is forecasted to be in the eastern Great Lakes at this point, and if we follow the 12 hour observed precipitation, it seems possible that snow could actually fall in the Plains, Midwest and Ohio Valley into New England. The heaviest snow would likely be in the Midwest and south central Plains (KS/NE).

The ECMWF ensembles are showing the exact same location and exact same strength as the GFS Ensembles- remarkable for a situation more than 7 days away. This image is valid for the same time as the GFS ensembles, which is November 28th in the morning. However, this time the ECMWF Ensembles show the 850mb temperatures, which are still following the GFS Ensembles. Imagine the line between the light blue and light green is the equivalent of the blue rain/snow line on the GFS Ensembles, and you can see how remarkably similar these two ensemble systems are.

I will connect all of this in a super-post on this potential tomorrow.

Andrew