February 21-23 High-Impact Significant Winter Storm (Updated 2/18)

500 millibar analysis reveals the storm system in question remains a while away from the United States, with enhanced wind readings being found in the Gulf of Alaska. This is the storm system in question, and is the storm system that is at the center of this big mess of confusion amongst the models.

There continues to be a large spread amongst the forecasting models as far as which solution is the most likely to verify. Shown above is the North American Ensemble Forecasting System, or NAEFS forecast for Hour 96, which converts to 4 days away from today. We see our system is in Kentucky at this time. If you've been reading my recent posts, you will understand that there has been a north trend with this storm that some models have had. This trend then resulted in the storm going to Wisconsin and other areas in the Northern Plains/Upper Midwest. It has been within the last 48 hours that we have seen a sudden revolt, so to speak, among the ensemble systems. This includes the NAEFS, the Canadian ensemble system and the American ensemble system. All three are considered big ensemble forecasting systems and have all been in fascinating agreement with the system going south rather than north.

So why is there so much confusion? Well, shown above is the NAEFS 500 millibar height anomaly forecast for 3 days away. Colder colors correspond to low pressure systems, while warm colors are supportive of high pressure systems. The model disagreement appears to stem from how strong high pressure in the Central and East US will be. The high pressure originates from persistent low pressure in the West US. As a result of below normal heights in the West, there must be above normal heights in the East US, as Newton's Third Law of Motion would indicate. These lower heights in the West and higher heights in the East are called the negative Pacific North American index, or negative PNA. In the negative PNA, storm systems that form in the Southwest as a result of persistent low pressure in that area are forced north into the Northern Plains, due to high pressure formation in the Southeast US. The models are catching onto this idea, but why are the ensembles going south? One reason is the presence of high pressure to the east of Greenland, called an east-based negative North Atlantic Oscillation. You can probably guess why the East based negative NAO is called east-based, but the negative NAO implies high pressure is over Greenland. This high pressure then acts to strengthen the subtropical jet stream, which is essentially another jet stream that sits along the Gulf Coast and Southern US rather than the North US (which is the big time jet stream). The subtropical jet stream, which is also called the STJ, is strengthened in a negative NAO. This negative NAO also acts to suppress the Southeast Ridge of high pressure, which would then act to verify the ensembles prediction of the storm sticking to the south.

Now that we've identified the indices at play here, let's look over that 500mb anomaly map and identify why the models are not handling this well. In that image, we can see high pressure centered over Canada and stretching into the US. This stretch into the US is helped by the negative PNA. By Hour 72 (when this image is valid), the high pressure system has started to extend north, leaving a gap of high pressure in the Southeast that is weaker than the high pressure in the Upper Midwest. Some models, like the European model, want to take this system north and have it ride the stronger high pressure wave into the Upper Midwest. The ensembles seem to believe it will go towards the weaker area of high pressure and this take a more southern route. The logical solution would be to take this system to the south, towards the weaker area of high pressure, as the ensembles are showing. I can understand why the models want to take the system more into the Upper Midwest, and it's due to where the system ejects from the Rockies. The farther north it comes out of the mountains, the farther north the track is going to be. Likewise, the further south the system ejects from the Rockies, the further south the storm track will be, like the ensembles are forecasting. In my personal opinion, if the system will be moving out of the Rockies from Southeast Colorado and the jet stream will be flowing towards the northeast (with winds over 100 knots), I could see the Northern track work out.

I combined both possible storm tracks in the image, with the yellow 'L's illustrating the north track and purple 'L's showing the southern storm track. Amounts were put at a wide spread of 2-6 inches for those in the light blue. Most should reach the middle of that spread, while some closer to the dark blue will flirt with the half foot mark. The dark blue contains an even wider spread of 6 to 12 inches. The wide-ness of the forecast spread is due to how low confidence still remains. The system is not yet onshore; we still have another day or two before the system makes landfall on the US. The pink shows amounts surpassing one foot. Southeast NE and north KS will bear the brunt of the system if either track succeeds.

Andrew

February 21-23 High-Impact Significant Winter Storm

Have removed the 'Potential' from the storm title, as the event is within the timeframe of reasonable certainty.

Satellite imagery shows the storm system of interest is in the Gulf of Alaska and nearly onshore in Alaska at this time. Analysis of the infrared field determines that the center of low pressure is on the coast of Alaska, with a very ragged comma shape still present in association with this storm. Surface analysis indicates the storm is in the 990 millibar range, so it is a fairly decent storm at this point in time.

Model uncertainty still remains a priority to deal with at this time, as it does appear there are two possible solutions evolving. Shown above are 500 millibar height anomalies for Hour 120, or 5 days away. Cold colors suggest the presence of low pressure, while warm colors favor areas of high pressure. The bottom image shows 850 millibar temperatures and mean sea level pressure for Hour 132, meaning it is 5 days and 12 hours away. This model is the North American Ensemble Forecasting System (NAEFS), a set of ensembles which I understand is derived from the Canadian and American ensemble forecasting systems. The NAEFS has done very well in the past two years for predicting temperature anomalies during the winter, this will be my first time seeing how it fares with storm systems. The NAEFS is predicting the storm system to ride east-northeast through the Plains and into the Midwest. This is not what recent model trends have been showing- the latest forecast models have taken it further north.

The issue with the models is how strong that high pressure system will be in both Canada and the East US. If we look at the top image of the NAEFS, we can see that the high pressure system is strongest from Canada into Greenland. Models and ensembles differ as to where the center of this high pressure system will be. The NAEFS proposes the high pressure system to be centered in north Canada, which then allows the storm system to progress further east and into the Midwest. The European model, as you will see below, takes a much different track.

This is the most recent forecast from the typically-reliable European model. It is called the ECMWF model in the weather world, but for practical purposes we will refer to it as the European model. This forecast also has two images: the top image once again shows 500 millibar height anomalies, while the bottom image depicts 850 millibar temperatures and mean sea level pressure values. The 500 millibar chart is valid for Day 5 (Hour 120), while the bottom image of mean sea level pressure is for Hour 144, or 6 days away. If we do a comparison between the NAEFS and this European model, it's pretty obvious that the European model is stronger with the high pressure in the East than the NAEFS. We see more oranges stretching down south from Canada and into the Eastern US than the ensemble system. The reason for this is a much weaker connection between the high pressure in Greenland and high pressure in Canada. Almost like a bridge, the European model nearly breaks off the connection between these two masses of high pressure, which would allow the Canadian high pressure system to then push further south and allow the European solution to prevail. If the European model solution did prevail, one would expect lower snow totals with any snow further to the north. Quite an odd solution in my opinion, and we will evaluate it further later on in this post.

Something that needs to be brought up is that strong high pressure system in the Pacific. That big high pressure system, combined with the low pressure in the West US creates what's called a negative Pacific North American index, or a negative PNA. The low height anomalies in the West provoke above normal height anomalies (high pressure) to arise in the East US, which typically can push storms into the North Plains in a similar fashion to the European model. The NAEFS ensemble system also has this high pressure system in the Pacific, but is more widespread than the European model. The more widespread it is, the stronger the high pressure is. As a result of a stronger high pressure, the 500mb heights must become lower and this would push the storm track further to the south. However, we now go back to the previously-mentioned high pressure over Greenland. When high pressure is established to the east of Greenland, it is called and East-based Negative NAO; 'East based' for being east of Greenland, and the negative NAO (North Atlantic Oscillation) implies high pressure over Greenland. The negative NAO provokes the subtropical jet stream to become more active and pushes the overall storm track south. It also suppresses the high pressure in the Southeast, which is why I'm not confident in the north-bound European model. Both images above show the east-based negative NAO, but both have different solutions. It all comes down to which factor has the stronger influence. Because the negative PNA is upstream of the Midwest, it will have a stronger effect than the negative NAO, which is far up in Greenland. However, by the time the storm system is able to reach the Plains, the negative NAO will hold some influence and ought to be able to keep the storm system south, definitely further south than the European model's solution.

I can understand why there have been trends north. Shown above are 500 millibar heights (not anomalies, just the regular values) valid for Hour 96 (4 days away). We can see our low pressure system just beginning to emerge from the Rockies. The issue from here is where it comes out of the Rockies. The NAEFS forecast in this image has the low pressure system in far southeast Colorado. When you get storms coming from that area, there is a decent possibility that high pressure could push it north, and I believe this is where the European model is getting the idea of a northward trend from. Additionally, the storm will be reaching what's called a negative tilt by the time it is exiting Colorado. A negative tilt implies that a storm system's highest vorticity values are pointing towards the southeast, and this is where we find the big difference between the southern solution (NAEFS forecast) and the northern solution (European model).

I put together the NAEFS and ECMWF (European) models together for the same timeframe. They are both forecasting 500 millibar values for 4 days away. The black line illustrates the tilt of our storm system, while the red line illustrates the tilt of the high pressure system in the Central US that is preceding this storm. The issue between the northern and southern solutions is that the European model wants to bring the system to a negative tilt faster than the NAEFS ensembles do. This is shown by the ECMWF image's black line tilted to the southeast, while the NAEFS is more neutral-tilted, meaning the highest vorticity centers are not tilted in any particular direction. In response to the negative tilt by the European model, the storm system digs to the Southeast more. This situation then responds similar to if you were to push some form of a wedge (shovel, for example) into the ground- the shovel (storm system) tries to dig down further south, but the ground (high pressure system) is provoked to move north in response, and this is how the European model gets its northern solution.

Considering there are three different ensemble systems rooting for a southern solution and just one model and one ensemble system rooting for a northern solution (the most recent American model jumped south a bit), I'm going to have to go with the southern solution. I am a bit hesitant, because the European model and its ensemble set are very prestigious and reliable forecasting systems, so it's a very tough forecast to make when it's essentially Europe vs. the world.

Despite this, I am willing to stick with my prediction that was made yesterday for this system. To me, there are three stages of the models. The first is the Long Range track, where the models get a consensus roughly 10 days away from the event. The second stage is what I like to call the '60 Hours of Chaos', where the storm gets within 8 to 5 days away (Hours 180-120 on the models) and the models go in every which way. The third and final stage is the final track, where the models converge on a solution within 5 days of the storm. This final track can sometimes be nearly the same as the first Long Range track. We already observed the first stage, and are now clearly in the second stage. The system will get on land in the next few days, so model solutions should clear up as the time comes closer. I should mention that the Chicagoland area into southern Wisconsin and north Indiana and southwest Michigan may see that 'Accumulation Possible' dropped if dry air manages to start that region with virga (precipitation is being produced, but evaporates in a dry layer of air before it hits the ground) as some sources are suggesting. There will be a lot of dry air in place just a day or two before the event, and this could very well hinder the already-lowered totals. I will most likely have my first accumulation map out either tomorrow or Tuesday; confidence remains low.

Andrew

February 21-23 High-Impact Potential Significant Winter Storm (Posted 2/16)

Storm title is kept the same despite crucial model differences. I do this because, despite said model differences, it is quite obvious that someone will get a thumping of snow out of this storm system. The question right now is: Who?

Satellite imagery indicates that our storm system is moving into the Gulf of Alaska at this hour, with elevated levels of infrared imagery suggesting the center of this storm system is located around the Aleutian Islands. Pressure observation charts suggest the storm system has a minimum pressure of 988 millibars, a pretty solid storm system. The comma shape of the infrared imagery depicts the storm as a well-functioning system that will certainly be able to make some noise on land, both on the wintry side and the severe weather side.

This continues to be an unresolved issue between the European model and several other reputed models. Shown above is the Day 5 forecast for 500 millibar height anomalies. Cold colors indicate low pressure systems, while warm colors display high pressure areas. We see our storm system in the Southern Plains at this time, with high pressure in the East US and additional high pressure near Greenland. This high pressure system in the East US is what is causing such intense model differences. Both the European suite and American model suite agree that the storm will be ejecting from the Rockies in the position shown above in the apricot color. That's where the similarities stop. The European model believes high pressure in the East will force the storm system to move north-northeast, cutting out many regions of the Midwest from snow that was previously shown by this model.

These are the American ensembles for the same timeframe and same parameters as the European image above. We still see our storm system in the Southern Plains, although there is no longer an apricot shade. This is to be expected, as the ensemble members begin to diverge at this stage and therefore the mean solution is not as strong as the single European model. In comparison, the high pressure in the East US is weaker and centered further north in Canada than the European model has it. This is a crucial difference, as the weakness of the high pressure does not force such a northward movement anymore. It does move northeast, but nearly as drastically as the European model. Also of interest is stronger high pressure in the Pacific than in the European model. This stronger high pressure would force the storm further south.

I'm having trouble understanding why the European model takes the storm directly into the high pressure system. It is apparent that we will see some northward movement in response to persistent low pressure anomalies in the West and high pressure anomalies in the Northeast Pacific. These two factors make up the negative Pacific North American index (PNA). The negative stage of the PNA responds to the low pressure anomalies in the West by provoking high pressure in the Southeast US, and this does correlate with the issue the models are having with the high pressure in the East US. If it was just the negative PNA, we would most definitely see the solution being shown by the European model. However, we now go back to the previously-mentioned high pressure over Greenland. When high pressure is established to the east of Greenland, it is called and East-based Negative NAO; 'East based' for being east of Greenland, and the negative NAO (North Atlantic Oscillation) implies high pressure over Greenland. The negative NAO provokes the subtropical jet stream to become more active and pushes the overall storm track south. It also suppresses the high pressure in the Southeast, which is why I'm not confident in the north-bound European model. Both images above show the east-based negative NAO, but both have different solutions. It all comes down to which factor has the stronger influence. Because the negative PNA is upstream of the Midwest, it will have a stronger effect than the negative NAO, which is far up in Greenland. However, by the time the storm system is able to reach the Plains, the negative NAO will hold some influence and ought to be able to keep the storm system south, definitely further south than the European model's solution.

It's times like these when you have to stop looking at the models and think about how the atmosphere works. In times like these, I like to use my aforementioned river analogy. If this storm was a leaf that had fallen off a tree and into a river, would it want to move towards the rock in the river or towards the open river channel? If you think it wants to move towards the open channel, you've just explained the big paragraph I typed above. And in all honesty, the models are worthless. Until the storm gets on land, we might as well be trying to throw a dart, in the dark, blindfolded, spun around several times, with the correct solution being the dartboard bullseye three millimeters wide and a mile away from you. But if you're going to look for hints, always look at the ensembles. They stand the best shot of being right from a time this far away, which is another part of the reason why I support the cross-country solution.

The 12z run of the American model shows very heavy snow in the Plains. I am very confident that Nebraska and Iowa will certainly cash in on heavy snowfall. Even South Dakota could receive some extreme snowfall. Amounts could very well exceed 18 inches based on snow-to-water ratios above the usual 10:1 (10 inches of snow equals 1 inch of water). We could see ratios flirting with 20:1, meaning there would be lighter, fluffier snow, but a lot more of it. Towards the Midwest is where confidence drops. The American model believes in a sharp cutoff near Chicago. A sharp cutoff will happen, but this is the first time the American model proposes a cutoff zone so far north. The American ensemble system has been more persistent in the rain/snow line remaining in central Illinois. This is the same solution previously shown by the American model, up until now.

As of now, the European model is a far north outlier, and the American ensembles have been very consistent. I preferred the American ensemble solution, because I don't buy into an extreme far north solution. To me, there are three stages of the models. The first is the Long Range track, where the models get a consensus roughly 10 days away from the event. The second stage is what I like to call the '60 Hours of Chaos', where the storm gets within 8 to 5 days away (Hours 180-120 on the models) and the models go in every which way. The third and final stage is the final track, where the models converge on a solution within 5 days of the storm. This final track can sometimes be nearly the same as the first Long Range track. We already observed the first stage, and are now clearly in the second stage. The system will get on land in the next few days, so model solutions should clear up as the time comes closer.

I will update daily here on the blog, and much more often on our facebook page at www.facebook.com/TheWeatherCentre .

Andrew

February 21 Potential Severe Weather Event

The Storm Prediction Center has outlined an area of severe weather for Day 6, or February 21st.

Prognosis... Strong storm system will be ejecting from the Rockies on the 20th-21st. As the storm system moves east, the subtropical jet stream and Pacific jet stream will merge into a single, powerful jet stream that will be centered over the outlined region. Current thinking is that the abundance of moisture and strength (albeit weakening) of the storm system will provide a base for severe weather on the 21st. Tornadoes and damaging winds would certainly be possible, but model uncertainty forbids me from making any definitive conclusions on threats for this day.

I will have more updates in coming days.

Andrew

February 21-23 High Impact Potential Significant Winter Storm

Have downgraded system from Blizzard to Significant winter storm on account of recent model trends. Will maintain rest of storm title.

Infrared satellite imagery of the Pacific Ocean shows the storm of interest near the Aleutian Islands. We can see some convective activity associated with the front end of the storm system, characterized by a nearly vertical line of elevated infrared values. The storm itself remains far out to sea, away from any nearby weather stations in the Mainland that would otherwise help figure out model forecasts. We will need to wait until Tuesday or Wednesday for this system to crash ashore and the National Weather Service's weather balloons to get inside the storm and help out model forecasts.

Previous model agreement has been wiped away as we enter the dreaded timeframe where the models go every possible solution before they convene on one solution. This process typically takes a few days, but can last more than 5 days. I expect we will see model agreement come Wednesday evening when the system gets onshore. Model disagreement appears to be stemming from the forecasted strength in high pressure that will be over Canada and into the East US when the storm digs into the Southwest. When this storm system hits the West Coast and shifts south, it is typical to see high pressure arise in the East. The big question is, how strong will this high pressure be? A stronger high pressure would result in a more northerly track, while a weaker ridge would indicate a more cross-country, west-to-east storm track. This cross country idea is what earlier model solutions were showing. This dilemma is shown on the 500 millibar forecast from the American ensemble prediction system above. Our storm system is shown in the West US, with high pressure in the East. In this situation, the base of the high pressure system is in Canada, so the storm takes a more cross-country track. The actual American model has the base of the high pressure system in the Great Lakes, which is how we get that north track, as shown below:

My thinking is that this storm is going to hit areas that the models were in good agreement yesterday. The reason? Well, let's go in order here. First, we'll start in the West US. We have low pressure in the area, and that can instigate high pressure response in the Southeast US, which would go hand in hand with the American model solution above. However, looking at Greenland in the upper part of both images, we see high pressure to the east of Greenland. This is what's known as an east-based negative NAO. You can probably figure out why it's named east-based, but the negative NAO infers high pressure is in place in Greenland. This high pressure can then force the subtropical jet stream (located along the South US) to act up, and this brings the storm track further south. Also, it can oppress formation of the Southeast Ridge that the West Coast low pressure is trying to do. These two indices that are battling for the Southeast may just cancel out and we could see that cross-country solution as I suggested. The third (and probably most important) reason is that storms don't head right into high pressure systems. Don't get me wrong here, they can, but in a river, does water try to go to the place where it will get stuck? No, it goes towards the most open part of the river to keep on moving. This means that the low pressure system would theoretically not want to go north, right into the Canadian high pressure system. Rather, it would want to travel west-to-east (cross-country track), where it would encounter less high pressure anomalies. Adding onto that, if we see the base of this high pressure will be in Canada, the weaker high pressure anomalies in the Southeast would not be able to properly force the storm system north, like recent model solutions have been showing.

It's times like these when you have to stop looking at the models and think about how the atmosphere works. In times like these, I like to use my aforementioned river analogy. If this storm was a leaf that had fallen off a tree and into a river, would it want to move towards the rock in the river or towards the open river channel? If you think it wants to move towards the open channel, you've just explained the big paragraph I typed above. And in all honesty, the models are worthless. Until the storm gets on land, we might as well be trying to throw a dart, in the dark, blindfolded, spun around several times, with the correct solution being the dartboard bullseye three millimeters wide and a mile away from you. But if you're going to look for hints, always look at the ensembles. They stand the best shot of being right from a time this far away, which is another part of the reason why I support the cross-country solution.

I maintain the high impact part of the storm title because I do believe that there will be an intense snowstorm somewhere with the storm. The system in question will tap into pure Gulf of Mexico moisture that will allow heavy snow to fall across a potentially wide swath of land. Right now, a North Platte, NE to Cedar Rapids, IA swath is looking good for the most snow, with similar amounts east until Gary, IN.

There will be more developments as we progress through the weekend and into next week. My thoughts will undoubtedly change, but that's what weather does: change. Please don't ask for your location's snow, I will not answer if you ask.

Andrew