Weather Forecast

You might think that weather forecasting requires a degree in meteorology. Well, perhaps a degree in statistical analysis would be more useful. Here is a new way to forecast the weather with less knowledge, but greater accuracy.
Here in Canon City, Colorado, on Friday, February 2, 2007, I brought in my Newspaper from the porch when it arrived, at about 3 in the afternoon. I opened the newspaper to the page with the weather forecast. I was wondering how cold it would be on Saturday.
13 degrees Fahrenheit was the high temperature forecast for Saturday. This was way too low, I figured. I checked the forecasts on television and on the internet. They said we would reach 23 or 27 degrees the following day. These forecasts were also too low, I knew. I told my wife it would be in the 30s or higher. What was the actual high temperature the next day? 53 degrees Fahrenheit.
That's not a typo, by the way (how do they keep their jobs?). Weather forecasting "experts" were off by as much as 40 degrees - for a simple 24-hour forecast of the high temperature. Why were they so far off, and how could I be better than them at forecasting the weather for Saturday?
Well, I don't have an answer for the first part of the question. Weather here is certainly more unpredictable than in many places. Perhaps the meteorologists follow there computer models too slavishly, even when experience and intuition tell them to adjust a forecast up or down.
I can answer the second part of the question. My forecast was closer because the "experts" were so consistent in the errors they made. I had counted something like 15 out of 20 days when all the various weather forecasts predicted a high temperature that was 5 degrees or more lower than the actual temperature. Seeing that, all I had to do was take the forecast (the one predicting the highest temperature) and add five or six degrees.
A New Weather Forecasting Idea
The consistency in the direction of their errors was the key to my better forecast. In other words, they weren't forecasting too high one day and too low the next. They were making their errors in the same ways repeatedly.
The next logical question is whether errors are as consistent in other parts of the country. Looking at the statistics could answer this. One could check the forecast highs and lows for the last 365 days, and check the actual temperatures for those days. One could also see what the predicted probabilities of rain or snow were, and then note what actually occurred.
For example, suppose a forecaster predicted a 50% chance of rain 24 times, but it actually rained 18 times. Perhaps he had the best data, but he was too conservative in its application. This may not be a one-time problem. This can be determined by doing more statistical analysis. If his error was consistent, you could know nothing about weather forecasting and provide a more accurate forecast simply by saying "A 75% chance of rain tomorrow" every time he said there was a 50% chance.
That's the essence of how this new forecasting idea works. You first gather statistical information on the forecasts of several meteorologists or weather forecasting services. Then you compare these forecasts to the actual weather that happened, and look for any consistent errors. Ideally you would want to create a computer program, the idea being that as you enter each of these forecasts into it, they are adjusted for any known tendencies. The result should be more accurate weather predictions.
An example might make this clearer. Suppose that over the last year Forecaster A has been forecasting a high temperature that averages 4 degrees over the actual high. The computer would adjusts his forecast down four degrees. Perhaps a more sophisticated analysis shows that Forecaster B is consistently predicting too high of a probability of rain in the fall, but too low of a probability of rain in the summer. Once this is discovered and programmed in, the computer can adjust the forecast for these factors. For greater accuracy, the adjusted forecasts of three or more sources could be averaged.
You wouldn't need to know anything about weather forecasting. The underlying idea is that even when experts have the best knowledge and data, they can apply it incorrectly, and do so consistently. Perhaps some television stations will soon get rid of their meteorologists and take advantage of this new weather forecasting idea: "And now it's time for your electronic weather forecast, from our Statistical Analysis Weather Machine..."