Welcome to my blog – it will be updated as any threat for an east coast snowstorm, particularly in the I-95 corridor from DC to NYC, including the Delaware and Lehigh Valley’s, arises.
It’s a work in progress, but I felt this was a better way to express my thoughts without dealing with all the BS on other sites. Feel free to comment, positive or negative, but if you disagree, prepare to back it up – I welcome criticism and disagreements as long as they are presented in a logical fashion (those of you who have followed me for the past few years know what I’m talking about).
I reposted the blog from a few days ago below more for its archival value. Things have obviously changed some since that time (see the most recent blog today):
When considering the possibility of an east coast snowstorm, there is a method to my madness:
First, I look at the indices and teleconnections, and the pattern they’re likely to yield;
Second, I look at the models, giving me clues as to the various possible scenarios the pattern may dictate; and,
Third, I try to find analogs from the Kocin and Ucinelli snowstorm bible, Northeast Snowstorms.
I. So, to start, let’s look at some of the major indices and how they’re shaping the pattern:
(1) La Nina.
So far this season La Nina hasn’t had the traditional effects one would expect – generally warmer than normal sea surface temperatures (SST’s) in the eastern Pacific leads to a milder winter with a lack of major events. Icebox’s famous claim of no major Miller A’s this year has already fallen by the wayside more than once this year, however, playing the odds, he was absolutely right. Before this year, they just didn’t happen. There was only one in the last 100 years or so of records I believe. But this year is far from normal. One thing that I had pointed out in the beginning of the year was the type of La Nina we were/are experiencing. Just as important (if not more important) is the type of La Nina we’re in, i.e. an east versus west based event. The importance is best illustrated by the composite temperature anomalies below:
Clearly, an east based La Nina is more favorable for a colder, snowier winter than a west based (and a west based El Nino is much more favorable for us as well – which, by the way, was what we had during 2009-2010). This year’s La Nina has definitely been an east based type:
Looking at the above temp anomalies, one can clearly see 2009-2010’s pattern yielded more widepread cold throughout the region, while the east based La Nina of this year shows more to perhaps slightly below temps for the region. Clearly, the cold has been more severe for the majority of the winter thus far, and there are obviously more factors involved than just La Nina. Another important point to look at is also illustrated in the second map above – one can easily see the disappearance of the warm SST’s west of central America as we headed through January. La Nina is weakening, fairly rapidly, especially in the last few frames. The affect of this weakening will take time to be seen across the globe. But again, there’s obviously other forces at work here.
(2) North American Oscillation (NAO).
The NAO is an index which measures the degree of blocking out in the Atlantic. Points on the NAO:
1) The more negative the index, the larger and stronger the blocking high over the northern Atlantic, and more blocking, the slower storms, and everything on the weather map for that matter, move over time. The NAO was negative for a large part of last winter, and most of this winter so far. In addition to slowing things down, a negative NAO also provides an effective railroad for Arctic air into North America.
2) There are major differences between a west based and east based NAO. If the -negative NAO is more west based (large high pressure centered around Greenland), such blocking would prevent an inland storm or Lakes cutter. If east based (around the UK), an inland storm is definitely on the table.
3) For clues to where the NAO may be headed in the coming weeks, look to the stratosphere. When the straosphere warms, it is a strong signal that blocking (-NAO) will follow within a few weeks. It also argues for a more west based NAO when the blocking does occur. Here’s the current state of the stratosphere below:
The graph above is a little busy, but basically the warm colors above indicate warming of the stratosphere. You can see that late in January the colors were more green/blue, indicating cooling. Warmer colors were seen earlier in January – and no surprise, the NAO was negative during that time up to just a short time ago (see the graph below). The NAO is now postive, corresponding to the cooling in late January seen above (remember there is a lag of a few weeks). One can also see that stratospheric warming is again taking place, suggesting a return of blocking in the coming weeks. The NAO forecast below shows two important dips. The first is around the time of our storm, which will be discussed shortly The second is toward the end of the forecast period. This second dip may be the return of blocking as a result of the stratospheric warming we’re seeing – unfortunately the forecast only goes out 2 weeks, but you get the idea.
4) The strength of a negative NAO and direction its headed are both equally important. If the NAO is too strong, say -2 or lower, cold air floods the area, and this dense air pushes storms way to our south, and suppresses them, the so-called southern slider. Ideally, we want one of two things for a big storm – the NAO somewhere between 0 and -1, and/or a trationing period, in other words a forecast of the NAO going from negative to neutral, or postive to neutral. Big storms often occur in these states of transition, which is why, if you look at the NAO graph below, that first dip around Feb 10 is important – right around the time of our storm.
So taking all of that into account, it’s no surprise that storms have come closer to the coast in the last few weeks, and we’ve even had a storm cut west of us last week, as the NAO has been skirting the neutral line and was positive for the Chicago blizzard.
(3) Arctic Oscillation (AO).
The AO is another index which more directly measures the degree of Arctic cold in North America. The more negative the index, the colder it will be over the United States, as arctic air discharges over the continent from the north pole. Last year, we had a near record to record negative AO, and subsequently the winter of 2009-2010 was largely very cold. If one just looks at the graph below for this year, the AO has been negative for the majority of this season as well, until only recently. The recent spike to positive has coincided with our relatively warmer temps, and increasing mixing with storms over the past few weeks. One can also note the forecasted sharp slope downward corrsponding to the Arctic outbreak for the next few days, and a brief return to a mildly negative state:
(4) Madden Julian Oscillation (MJO).
This index is complicated, and highly erratic at times, but it’s essentially an indirect measure of the strength of the subtropical jet. Now, in La Nina years the Pacific/Polar jet stream is by far the dominant player. However, the short term forecasted MJO can be helpful, as there are ceratin phases which lend support to a major east coast storm. When in the “circle of death” in the center of the graph below, the MJO doesn’t have as much influence on our weather. When in phases 7, 8, or 1 however, it’s another indirect yet helpful signal for an east coast storm. Right now, the MJO is forecasted (the forecast is in green below, with the individual ensemble spread in the grey shaded area) to remain in the circle of death, so it’s influence is basically a push for our storm next week:
5. Pacific North American Pattern (PNA).
This index measures the degree of ridging over the western US. One of the key factors in a major east coast storm is a western ridge centered around the Idaho/Montana area of North America, which leads to surface low pressure tracking somewhere along the benchmark (40 N and 70 W) off the east coast, and allowing for an I-95 snowstorm. The stronger the ridge, the more positive the PNA. The foreast for the next few days is somewhere around the neutral area:
Below is the upper air pattern for late next week from European ensemble mean. You can see there is a western ridge centered just west of the west coast. Notice I said west coast, not Idaho/Montanna. This is an important distinction. Storms that have occurred recently have had a ridge position around the coast rather than farther east, and surface low pressures have been tracking closer to, or on the east coast, and WEST of the benchmark, as a result. This leads to more mixing/slop events, which we’ve seen more of recently.
II. After looking at the major indices and teleconnections, I’ll look at the models, and have included my 7 rules regarding models below:
(1) The overall pattern hold more weight than individual model runs (ie, please see above).
(2) The foreign models (Euro, CMC, UKMET), particularly the European, are vastly superior to the American models (GFS, DGEX, NAM) in the medium range (from about 6 days out to 48 to 72 hours before the storm). If you’re going to look at the models at all, these are the ones that carry the most weight. The GFS has far and away been blown out of the water by the European in the medium range over years, and is often called Dr. No, because when the GFS says yes, the Euro will often come back and say “no!”
(3) The upper air maps (500 mb) are as if not more important than the surface maps in the medium range (ie the western ridge position, above). They’re the “behind the scenes” of a movie production, with the surface maps the actual film.
(4) The GFS and NAM become more important as the storm gets closer, within 48 to 72 hours away. The Euro actually takes a backseat to these two within 2 days if you can believe that.
(5) We all talk about trends, and yes they are important. But just as important is to recognize trends within the same model, not between models. It is far more important for one model to move towards a solution over several runs than a group of models to move to a solution over one run. This is because of the physics behind the models, and their inherent biases. Better to compare apples to apples than apples to oranges, to put it simply.
(6) The flip flop rule. I know Rob has harped on this, but being most of us salivate like wolves around a dying lamb when each run comes out, I thought it prudent to bring it up again. The models tend to go toward a solution during the 5-10 day period, lose it, and then pick it up again around 48-72 hours out. Do not be surprised if this happens again. Only the storm of ’93 had very good model agreement from a week out right up to the storm, and even that cut a little farther west than originally forecasted at the last minute.
(7) The ensembles. These are a group of individual model runs on a given model suite, which are used to generate the operational run, which is what you see when clicking at say hour 60 on the GFS. Now, the op run is not an average of all the ensembles, but all are used to create it. It can be helpful to look at the ensemble mean or even each individual member to get an idea of spread of possible solutions for a storm for a particular model run. They can sometimes give you clues as to where the op run may go in the future, as it sometimes can lag behind the ensembles/ensemble mean.
III. Regarding my third point in the beginning, analogs from the Kocin book, I haven’t had a whole lot of time to go through these. Two days ago Joe Bastardi went through around some pretty impressive analogs for next week, mostly relating to storms in the past that have seen rapid deepening of surface low pressure leading to severe blizzards, ie, Midwest blizzard of 78, March 93, and Jan 96, when the models were consistently showing a rapidly deeping storm for next week. They aren’t showing that now – although anything’s possible, they may come back to it (see the flip-flop rule above, although this usually corresponds more to the American than the foreign models).
If I had to go with my gut, given the transitioning NAO from positive to NEUTRAL (not negative), a transitioning AO to mildly negative numbers, and a neutral PNA with a ridge axis on the west coast, as well as the trend of recent east coast storms to be closer to the coast, I think the storm trends west over the next few days and mixing/slop becomes an issue at least somewhere in our forecast area, including the Lehigh Valley. But that’s just my gut when looking over everything. We’re 5-6 days away, so we really won’t know much more until at least Monday or Tuesday, when the early week storm moves out and the disturbances responsible for the storm late next week make it onto land for better data sampling.
That’s it for now. Sorry for the long read.