Weather models are a confusing if not completely complex topic!
In this nearly 20 minute video, we break down what models are, their strengths and weaknesses, and how you can use them to develop better storm forecasts. This guide is meant to be an introduction to weather models with a more theoretical take on their uses with subsequent, more complex lessons beyond the scope of this one.
What is a weather model?
Weather models are tools and are not perfect in their predictions of the atmosphere. Reading models literally is a one way ticket to failure. Your approach should instead approach models as the imperfect tools they are. These complex weather programs run on pre-determined variables like current observations, radar, etc. — and the biases that went into their design will result in biased forecasts that you will eventually pick up on as you go along in your journey.
Models are only as good as the data they are fed, and sometimes there are gaps in available data which could lead to imperfect runs of models.
The general rule is don’t trust weather models, but use them.
Some terminology to know for weather models
There are important terms you’ll know to become familiar with when it comes to models:
- Runs: These are when weather models typically execute as a program. Just like you run an app, weather models have to be initialized.
- Ensembles: When you hear ensembles discussed, these are multiple runs of individual models that are then averaged into a single run. This can be helpful for viewing long-term forecasts.
Types of weather models
Global models like the GFS and ECMWF run predictions over the entire globe. Regional models like the NAM and HRRR will only predict weather within a smaller region of the globe. The latter oftentimes rely on global models as base-line inputs to model their regions.
Weather models will simulate the atmosphere at different resolutions. Generally these are larger like 12km for some global models and down to 3km or even 1km for regional models. The 3km and smaller models are typically considered high resolution. As a reference the infamous HRRR is a 3km resolution model.
For now, models with lower resolutions will run further into the future. This is helpful so you can begin to anticipate global weather patterns a week or more away. However, higher resolution models are helpful for forecasts the same day — which has obvious implications for the storm enthusiast.
So how accurate are weather models?
Weather models are accurate to a point — again take them literally but not seriously.
The accuracy of models depends on how you use them. If you try to take a single GFS run verbatim for the position of the dry line a week away — you will always run into issues. However, taking a look at trends on that model over multiple runs will allow you to get a general idea of what might just happen.
Three rules of weather model usage
Rule 1: Models are a single tool to use amongst many in creating forecasts. Observational data is as valuable if not more so, especially when considering shorter-term forecasts over the next few hours.
Rule 2: Always start looking at models after you’ve looked at what’s actually going on. This is especially true with high resolution models. These models sometimes miss huge factors in the day like an ongoing storm complex.
Rule 3: Never trust the model (fully at least). Never blindly believe models as they are liars, for real. But if multiple weather models are saying the same thing, you can trust things a bit more.
You must log in to post a comment.