Heading into winter, 5 myths about forecast models

As part of our winter forecasting feature this autumn, we’re releasing a few posts with information and tounge-in-cheek opinions about winter and meteorology. The first in the set comes this afternoon, where our forecasters sat down to compile a list of the 5 biggest forecast model myths as we head into winter. You can check them out below, to get a quick primer on what to expect over the next several months.

  • #5: Forecast models always have the right idea

    Lets start with the initial misconception of forecast models. These models are made to be used as guidance (you’ll see that word several times in this post), and that seems to get lost in the shuffle especially during the winter months. Listen — these are computers that are trying to simulate the atmospheric processes which are fluid. Using them as guidance, not taking them as fact, is essential. It is important, and our job as meteorologists, to discern between the models that have the right idea — and the ones that don’t.

  • #4: Convective feedback always messes with models

    Here’s a pretty classic one that we get during the winter months. More often than not, when we are tracking a potential coastal storm system, the baroclinic zone off the Southeast/Mid Atlantic coast will play games with the forecast models. You’ll sometimes be able to see convective (or grid-scale) feedback off the Southeast Coast, or sometimes even in the Gulf of Mexico. But this cannot always be blamed for adjusting forecast model solutions. Models are designed to be able to handle some sort of convective feedback and apply it properly.

  • #3: The NAM is best within 48 hours

    Oh man. We’re not even sure where to start with this one. While the NAM is a very useful model in certain scenarios, there seems to be a general opinion developing that it has the best handle on the situation inside 48 hours. The fact is, each system and atmospheric set up is different. And the NAM may or may not have the best handle on it. We’ve seen plenty of colossal failures by the NAM inside 48 hours (take February 2013’s winter storm as an example), but plenty of good forecasts as well.

  • #2: The GFS is always suppressed/out to sea

    This one is a little more tricky, for sure. First off, the GFS had a known and well-documented bias to suppress east coast cyclogenesis during the early to mid 2000’s. This was noted by several reputable sources. Since then, however, the model has undergone several upgrades and this bias is much less prevalent. It is important to remember, again, that each set up is different. So the GFS may or may not have the right idea when it is showing your beloved blizzard tracking over Bermuda at 96 hours.

  • #1: The Euro is always right

    Ah, the unbeatable behemoth of forecast models: The ECMWF. While more often that not, it will verify better than any other global model (it does, after all, have an unfair advantage with 4VAR initialization scheme) — it will not always be correct. One of the biggest mistakes forecasters or hobbyists will make during the winter months is to go straight to the wall with the ECMWF’s forecast. The pattern, teleconnections, and meteorology behind the setup will always be more important than any forecast model. Sure, the Euro has had its great performances (Sandy, as example)…but it has had its bad ones as well (last week’s cutoff nor’easter). As always, meteorology — not modelology.