A review and documentation of the snowstorm for March 3, 2014. (Weatther Ranger was saying no storm was coming, all others were saying a big storm was coming)
Upon arriving at my son’s day care on the morning of Thursday, February 27, 2014, I was quickly asked about the giant snowstorm heading our way. I was puzzled. Giant snowstorm? What are they talking about? I asked for details. Apparently it was pointed out by broadcasters on local radio in State College that there would be a snow storm on March 2nd through the 4th that would affect a large portion of Pennsylvania dramatically. I knew there was a potential for a storm coming in that time frame, but had not yet reviewed all the models for that period, so I was a bit surprised and wanted to know what they were talking about. I quickly went home and set down at my PC to check. At that moment I also got a ping on Facebook alerting me to a message. It was from a neighbor with a link to a picture: the ECMWF(European) model output from the previous night’s solution (the 00Z, or 7 PM EST on the 26th). The image showed a large swath of heavy snow–greater than 20 inches–for a good chunk of the state by the time the storm would pass on Monday night the 3rd. Several radio station forecasts had mentioned that anywhere from 6 to 22 inches of snow would fall across the state by the end of Monday the 3rd, about 5 days away. THIS WAS APPARENTLY THEIR CHOSEN SOURCE! A review of all the models did show a fairly significant storm for the eastern portion of the U.S., but exactly where the snow would fall and how heavy it would be was varying between the models and in the ensemble solutions as well (ensembles are the same models run with different conditions to see what happens). So that far out, there was already a lot of uncertainty about who would get what. It was more certain there would be a snow storm somewhere in the East, but where and how much couldn’t be predicted reliably.
Click to enlarge
This figure is the heavy snow as depicted by the ECMWF model, 007 on the 27th of February 2014.Click to enlarge.
On Thursday night the model solutions had already begun to show a trend in where the storm would go and who would get the worst of it. That trend was led by the Canadian (CMC) model. It showed that a chunk of Arctic air would rotate southward and around the strong Polar Vortex over eastern Canada and push the energy arriving from the West Coast southward into Virginia. And finally by Saturday, March 1st the consensus of most of the models continued this trend, however they also indicated that a fairly large portion or Pennsylvania would receive substantial snow, with up to 4 to 5 inches indicated by the ECMWF for State College as late as the 00Z run on March 2nd (7PM March 1 EST).
Click to enlarge
This figure is the results from the 00Z ECMWF on March 2nd(7 PM March 1 EST). Total snowfall up through 7 PM on March 3rd. Click to enlarge.
The final result was a far cry from this solution (see last figure below), and it appears that the GFS (U.S. Global Forecast System) and the CMC combined solutions from as early as 1200 UTC on the 1st (7AM EST) had it pretty well focused on Virginia and areas well south of Pennsylvania. All through the weekend it was noted that many media sources were still calling for “ground zero” to be in Pennsylvania and the maps on online with NBC News, as depicted from the Weather Channel, showed as late as 9 PM EST on Sunday March 2nd that much of Pennsylvania would still receive half a foot of snow (see figure below).
Image from NBC News online, copied at 9 PM EST on March 2. Still showing a significant amount of snow for PA and much of the northeast that received next to nothing. DC did not get 8 to 12 inches, but in fact 3 to 5 inches.
This meteorologist as early as Saturday morning on March 1st had reduced his prediction for State College to 3 to 6 inches, then reduced it the next morning to 2 to 4 inches, and finally on Sunday afternoon informed his subscribers that no significant snow would fall in State College, and the forecast was indeed a “bust”! Further, it was stated as early as Friday night the 28th that confidence in the forecast was very low and that “This storm could easily move south and bypass Pennsylvania altogether.” Hmmmm, hindsight says he should have jumped on that intuition–note to self.
Total Snow for event 3-4-14, click to enlarge
Snowfall amounts as of the Morning on March 4th received on March 3rd. You can see the difference from the prediction.
So what has led to this seemingly endless supply of inaccurate predictions so far in advance? And what would be a responsible way for getting the word out that a storm is coming, but not trigger panic in some, and also not trigger the “stone throwing” at us meteorologists for getting it “wrong”? Here are some thoughts.
The ratings for commercialization of weather information. What? How do commercials cause this to happen? The process begins with ratings, which drive the advertising dollars. If a media source can crank up there numbers, they can charge more for advertising, thus make more money. What is the best way to crank up ratings? Make whatever is being talked about exciting–and nothing excites, or panics, a large number of people more than bad weather, especially a storm that could dump 20 inches of snow. The media outlets pounced on it so big that they could not let it go, even after it became obvious that PA would not get hit that badly and the storm would, indeed, shift south. Well, why didn’t they just shift their maps? I believe the meteorologists did, but the adjustments did not make it up the chain of communication in time for the news and their recording deadlines, plus it would wreck the hype they had already built. So you the viewers and users are left with inaccurate information for a variety of reasons. One big one is too much hype to drive the ratings; another is too many pipes and valves between you and the source of information: the meteorologists.
What about the National Weather Service? That system is a bureaucracy caught up in rules and procedures (see the various manuals and thousands of pages on how to do things, including forecasting the weather). The highly trained meteorologists on the front line are also caught up in rotating shift work–every 7 days or so your body is on a different schedule, including some overnights (12 PM to 8 AM). Some people claim they can handle it. I experienced it when I, too, was part of this crowd. No one handles rotating shift work, it handles you! The quality of product suffers dramatically, though often a step better than the media sources referred to earlier. In this event my subscribers knew they were getting no snow, when the NWS still had advisories and warnings out for a good chunk of the state for an additional 6 hours plus. In my point of view, most of the time the NWS does a great job of issuing severe weather warnings and advisories, and that should be their only job, keeping the public safe from harm’s way. For quality and timely forecasts, please leave that to an industry niche truly driven by the users, where forecasters sleep at night. No forecasting needs to be done at 1 AM unless it is for extreme weather in the short term.
The rip and read fiasco: The local radio and to some degree television stations as well do what the business calls “rip and read” for weather forecasting. They basically take the NWS forecast and read it to their listeners. There is absolutely little or no connection from an information level to the accuracy of the product. Day after day it is very common for the forecast to be -“out of date”- sometimes by several days, or worse and absolutely does not jive with what is going on in the real world. That can happen for many reasons, too many to get into here. But it is a major source for the “stone throwing” at us meteorologists. It happens at least once a week and often much more at several radio stations around the region. There are some stations that also record or play recordings made by professional weather forecasting organizations. For many reasons, those recordings do not jive with what is going on with the weather, sometimes a by a lot! The most often issues is technical, and sometimes that is at the recording end. But errors are fairly common and an inherent process of the system.
LASTLY: Do you think those inexpensive or free smartphone apps give you the best forecast? Those predictions come directly from what is called MOS, Model Output Statistics (direct from the computers), or some geographical derivative of that output. Rarely does any human being ever interact with that data, except sometimes for the largest cities. When you download those apps, they are following you everywhere and sharing all your information to the commercialized world. If you are tired of “Big Brother’s” compiling of all your personal data into a marketing information mosaic, consider getting your weather information from a source who could care less about the shoes you just bought or your last night out on the town. These apps help the “system” do just that–while making a ton of $$$ selling your information.
My point here is that if you want quality weather forecasts that are as accurate and timely as possible without the hype, it might be worth you spending a little money, just a little, to get those services. How much would it cost? For a whole year’s worth of service including e-mail and text message notifications for alerts to changes and potential arrival of severe weather, it would cost less than a dinner out for two at a medium-priced restaurant, not including tip. The service also provides video explanation (on a platform of your choice, including smart phones!) of what is happening, access to an archive of significant events and a solid resource committed to helping educate you and our community about how the atmosphere functions and interacts with all life on our home. The value for your dollar will also keep growing with time as the service that can be provided expands.
These local businesses make Weather Ranger possible, please consider their products and services when the desire or need arises, and mention Weather Ranger when you do: