Wednesday, April 24, 2013

How Do We Predict the Weather?


By Olivia Humes, Grade 11, Imagineers

Imagine it’s the late 18th century. You’re in the backwoods of Pennsylvania, which has just recently become a part of the world’s newest country--the United States. Since you live in a settlement many miles away from Philadelphia, your income and success this year completely depends upon your harvest, which in turn, depends upon the weather. But how do you know if you’ll get enough sunshine or rain to produce a good harvest? How do you know when the next storm will blow in?

The short answer is, you don’t.

Professor Cliff Mass is somewhat of a weather celebrity in the Pacific Northwest. He runs a weather blog, a radio show, and creates cutting-edge weather forecasts at the University of Washington. I got the opportunity to listen to him lecture on the history and future of mathematical weather forecasting last Sunday, April 14th.

He explained that in the early days, weather predictions were qualitative rather than quantitative. Sayings like “red skies at morning, sailors take warning” or “clear moon, frost soon” were often used to predict the weather of the next few days. Though many of these sayings reflect the cause and effect relationships of the underlying processes that change our weather, they offer little specific predictions, and only allow general actions to be taken. The first hint that quantitative weather prediction could be used to predict future conditions came when Benjamin Franklin, our most meteorologically inclined Founding Father, discovered that storm systems moved in a roughly west-to-east pattern across the country. But for a long time, the precision of our early instruments and the availability of data limited the effectiveness of weather prediction.

To a large extent, these problems still exist today. Even with computational models developed in the 1920’s using physical equations, such as laws governing the behavior of gases, and the almost unimaginable computing power available literally at our fingertips today, the accuracy of our atmospheric predictions aren’t quite on point. We’ve accepted this as part of everyday life, despite the confident forecasts delivered to us via our local news stations and smartphones. A lack of weather reporting stations in and over the Pacific Ocean can have a big effect on the accuracy of our reports, particularly on the West Coast, where storm systems arrive directly from over the water.

But even with perfect global weather coverage and infinite computing power, weather prediction is inherently difficult. Feedback loops in the atmosphere, where individual processes either enhance or reduce each other’s effects, create a chaotic, nonlinear system where the smallest inaccuracies in measurement can dramatically affect the outcome of a whole day’s weather.

These inaccuracies can be as small as observational errors intrinsic to the equipment we use. This principle is known as the butterfly effect, but Professor Mass likens it to a pinball game. Imagine a pinball game, but one in which you have no control over the paddles, and can only pull back the spring at the beginning of the game to start the ball’s motion. No matter how accurately you try to recreate a shot that resulted in a huge score, you’re extremely unlikely to recreate that lucky shot if your friends ask you to prove it to them later. The weather, like the pinball game, is a similarly chaotic system.

The approach that Mass takes with his predictions, he explained, is to run his weather models over and over, varying the initial conditions slightly, within observational error. Then, each result is averaged, and a general forecast of probabilities is created. These forecasts prove to be both more accurate and more useful to us, and allow us more realistic expectations of the next day’s weather. After all, how do weather forecasters know when the temperature will be exactly 70°?

Now imagine it’s 2013 again. Imagine you’ve just taken out your phone to check the weather reports. Looks like there’s a 75% chance temperatures will be between 80 and 85 degrees. Put on some sunscreen. It’s time to hit the beach!

Read Professor Cliff Mass’s blog here:


Happy Earth Month!

No comments:

Post a Comment