Embedded Files
One day, Frank was given a weather computer. The packaging claimed that the computer could forecast the weather anywhere and everywhere in the world. Furthermore, the box claimed that it did its forecasting based solely on the laws of physics.
Frank was skeptical. He knew that it was very easy to get a computer to forecast the weather. For example, one could program a computer to always forecast a high of 70 and a low of 50. It's a forecast, but not very useful. A more accurate forecast would have the computer look up the climatological averages for whatever site it was forecasting for. More accurate than random guessing, but still not very impressive.
Frank didn't believe that this small computer (with a Pentium 2 processor) was actually going to make forecasts based on the laws of physics, which is how the National Weather Service supercomputers make their forecasts. So he decided to try it out on a simple problem to see if he could figure out how it worked. Choosing something practical, he asked it to forecast the temperature on the 16th green at the TAMU Golf Course. (Frank was a very good golfer.)
The day Frank chose had calm winds, so Frank figured the computer should have a very easy time with the forecast. But it was still much more complicated than he expected.
First, the computer wanted to know what height to forecast the temperature for. Would it be right at the ground, a foot above the ground, four feet above the ground, or twenty feet above the ground? Frank picked four, figuring that people felt temperatures mostly at chest level.
Then the computer wanted to know what the current temperature at four feet was. Frank thought this was a reasonable request, it being hard to forecast the future if you don't know what the present is. So he took a thermometer, measured the air temperature, and typed the value in to the computer.
Then the weather computer wanted to know the temperature at one foot and twenty feet. Again, Frank thought this was reasonable, so he measured the temperature at ankle level, then tied the thermometer to a helium balloon and raised the balloon to a height of 20 feet. He let it sit up there for a while, then lowered it quickly and read off the temperature.
Frank had just taken his first upper air observation. Admittedly, 20 feet is not exactly upper air, and most weather balloons are released into the air and radio their data back to the ground, but the principle is similar.
This satisfied the computer's desire for initial conditions. The initial conditions are possibly the most important part of the forecast. Consider this analogy: if you want to forecast the position of a baseball, knowledge of the laws of physics (and/or experience) is helpful, but the most important thing is knowing how hard and in which direction the ball was thrown.
Report abuse