TorontoForecast.com - Blog

With many sources of weather forecasts, one could wonder why the predicted forecast high for Toronto varies depending on your source. Of course, one can argue that temperature readings closer to the lake will be much different from one further inland. However, when forecasters predict the maximum temperature, they often refer to what’s expected at the airport. As such, I decided to conduct a study for the next six months as to which weather source has the highest % of having the correct forecast. It is important to remember though that weather forecasting is still a relatively young science and weather prediction is still imperfect. Therefore, this mini-study is not meant to criticize the skill or accuracy of the weather source but hopefully allow us to appreciate how difficult it is to forecast weather. As the adage goes, “Don’t shoot the messenger.”

Methodology: The next day maximum temperature forecast will be obtained at 10:00 pm Eastern Time the night before from the following sources:

• Environment Canada - government run weather agency

• The Weather Network - private Canadian weather company owned by Pelmorex Inc.

• 680 NEWS AM radion station - offers a Weather Guarantee Contest

• CityTV - lead meteorologist Michael Kuss

• GlobalTV - lead meteorologist Anthony Farnell

• Accuweather.com - international private weather company

• Weather.com - private American weather company

The sites were chosen because the weather forecasts are readily available through the Internet and therefore we can compare their temperature forecast high as of 10:00 pm. The official observed maximum temperature will be based from Toronto Pearson International Airport’s (CYYZ) observation as reported from the National Climate Data and Information Archive.

Comparison Calculation: The maximum temperature forecast error will be calculated by subtracting the maximum temperature forecast from the observed maximum temperature, and squaring the result. This method of calculation is called the root mean square error (RMS) where the individiual errors are squared, added, then divided by the number of errors (ie, the number of observations) and the square rooted. In essence, it summarizes the overall error and you maintain the units, degrees Celsius. It will also help us determine which one can provide the most correct forecast rather than the one who can give us the exact temperature forecast.

I will post the results on a weekly basis every Sunday and the final analysis will be revealed six months from now. To start our study, the 10:00 pm forecast for Friday, May 9, 2008 is as follows:

Environment Canada - 16 deg. C

The Weather Network - 14 deg. C

680 NEWS AM radion station - 18 deg. C

CityTV - 14 deg. C

GlobalTV - 15 deg. C

Accuweather.com - 18 deg. C

Weather.com - 16 deg. C

As you can see, the temperature forecast spread is 4 deg. C and it will be interesting indeed to watch their forecasts in the weeks and months ahead.

For Toronto climate information, visit our main site: www.TorontoForecast.com