Why are there Two Hurricane Models, the European Model and the American Model when physics and statistics are the same everywhere?
In: 293
And here I thought this was going to be a debate about the early history of meteorological forecasting, and the issues between the U.S. and Cuba in the late 1800s/early 1900s carrying on through today… silly me.
But here’s a corollary question – the U.S. traditionally flies into hurricanes to gather data, which is then applied to the spaghetti models. Do European models gather their own data in similar fashion or take that data into account (i.e., is it shared with European weather agencies? I’d like to think so, but you know what they say about assuming…). Wondering if this is a contributing factor to the differences as well.
The reason for this is likely due to differences in methodology and/or data used by the two groups of meteorologists. Additionally, weather patterns can vary significantly from one region to another, so it makes sense that there would be some discrepancies between the European and American models. Ultimately, though, both models are based on the same underlying principles and should produce similar results.
Basically because physics are far from settled. There’s an entire half of physics in the universe we don’t even recognize [because we don’t accept nature for what it is, or us for what we are.](http://www.urbanagandenergy.org/soe/) We don’t really understand the whole of how or why weather works the way it does, so all predictions are made by someone’s assumptions being modeled by high power computers. Inside 1.5 days we do quite well and models match and make 75% accuracy. Outside of that the matches and accuracy decline sharply because nature is far more complex than we recognize. There are factors that effect weather development that aren’t included in anyones models.
Computing power per capita. The United States and surrounding countries get gangbanged by 5-12 hurricanes a year, so we care more about figuring out a general course so we can just be ready for the gaping afterwards. So, we have to buy a lot of computers to figure out which area is gonna get abused the most, so they can start moving hundreds of thousands of palettes of medical, food, water, etc supplies to the GENERAL area of a 50-100 mile wide gaping hole.
Europe gets maybe two hurricanes a decade, so they only need a MacBook Pro and some fancy drawings in photoshop.
While Europe is trying to be accurate, the Americas are just trying to figure out who drew the short straw with this hurricane and is getting pounded for 3 days.
Because 50-100 miles wide could mean that as Florida’s tip is getting wet, Cuba is taking a beating. We have a lot more people affected by severe weather in multiple states, countries etc over a 1000 mile track.
Europe has to figure out “Are we getting some bad rain.”
America has to figure out who needs triage
Because the word “physics” is a bit ambiguous. Sometimes it refers to the activity of the physical universe (which is always the same everywhere) and sometimes it refers to the *scientific study* of the physical universe. Theoretically, an infinite number of different “models” (mathematical methods and sets of statistics) can be used to describe/predict what happens in the same physical universe. When dealing with very complex (chaotic and only partially understood) systems like weather events, the more models, the better, and we can use whether (pun not intended but implicit) multiple models predict the same results as an indication of the reliability of the prediction.
There are so many things to account for that it’s impossible to do so and still be able to process it.
Not all of the math in these things are closed form equations, so you have to numerically solve using arrays or loops for an infinite number of terms.
Since this is computationally expensive, people have come up with simplifications, or “models” that will simplify the math extensively with minor trade offs. Sometimes the particular simplification does great at capturing certain details, but other details are less accurate.
Each method for simplifying has tradeoffs. But none of these models are based on a single thing. They are incredibly complex systems and each model will have different inputs than another model based on what their simplified equations require.
So with different inputs, different simplifications that allow us to process something in a realistic time frame, you get different results.
Usually they agree roughly. But the further you try to look ahead, they will all become inaccurate quite fast.
So meteorologists will run some/all models and the red cone you see in hurricane trajectories is actually all the models, with various inputs and it’s results all superimposed on top of each other, then filled in so you get an area of effect.
Notice, right next to the hurricane, the cone is narrow. The models agree pretty closely. The further out, the more spread the red cone is as assumptions begin to break down across all models. The further out they try to model, the less inside that “sweet spot” the model’s creators originally focused on.
Ex-meterology student here, the different models “weight” atmospheric data in different ways, which causes different outcomes.
For example, if the Euro model weighs a low pressure developing deeper, that would impact the direction of a hurricane may travel. If another model says the low pressure won’t develop that much, then the hurricane goes somewhere else. Usually, the models will get in agreement the closer to the forecast time happens, which is why an official Hurricane warning only happens like, 36 hours out.
Because they can neither perfectly observe nor perfectly model the physics, so both are based on the statistics with broad, approximate strokes of physics mostly informing which batches of statistics to look at.
Because there are differences in the physics of different areas, the models derived largely for those areas will differ in order to be most predictive.
There are different ways of estimating and calculating, and different data inputs to pay attention to. Hurricanes are chaotic systems, so small differences in each step can add up to different and even contradictory results.
Hopefully they start throwing out the underperforming models.
Read Chaos by James Gleick it’ll provide the answer in easy terms. Especially the segment on Edward Lorenz. Physics is the same but to model atmospheric phenomena is complex requiring statistical data and huge numbers of variables. Weather phenomena are driven by nonlinear dynamics in which there is sensitive dependence on initial conditions. There are more than two models. And meteorologists run many models and then look at the probable outcomes from which a most likely outcome is selected. It’s just the European model seems to predict weather characteristics more accurately. Both models use the same physics and thermodynamics. Outcomes depend on the number of iterations the model is run and given the complexity of variables involved predicting weather is not going to be precise all the time. Also there are differences between what the model is designed to predict, how far into the future it predicts how frequently it is recalibrated etc. When one talks about a model they need to realize it’s not like using the ideal gas law and plugging in the knowns to get the unknown result. They use numerical approximations for nonlinear formulas and then run the model iteratively over and over. So many times such that it requires super computers to run all the calculations, and then look at the most probable results.
Latest Answers