On July 11th I wrote a blog titled, "Here's Your 28-day Forecast!"
I found a 28-day forecast on-line at pollen.com and had to jot it down to see how awful it was. I thought it would be way off because not only is it difficult to forecast highs/lows/precip 10 days out, 28 days would be quite ridiculous! I, however, was making a wrong assumption apparently.
So here's the forecast from that day:
Tuesday, July 10: Light rain showers, high 81.
Wednesday, July 11: Few clouds. Low 54, high 84.
Thursday, July 12: Few clouds. Low 59, high 87.
Friday, July 13: Few clouds. Low 61, high 88.
Saturday, July 14: Few clouds. Low 62, high 90.
Sunday, July 15: Rain. Low 64, high 87.
Monday, July 16: Rain. Low 66, high 89.
Tuesday, July 17: Rain. Low 67, high 85.
Wednesday, July 18: Few clouds. Low 63, high 84.
Thursday, July 19: Few clouds. Low 61, high 87.
Friday, July 20: Few clouds. Low 61, high 88.
Saturday, July 21: Few clouds. Low 63, high 88.
Sunday, July 22: Few clouds. Low 66, high 88.
Monday, July 23: Light rain showers. Low 65, high 85.
Tuesday, July 24: Scattered T-Storms. Low 64-66, high 80-83.
Wednesday, July 25: Sunny. Low 60-63, high 84-86.
Thursday, July 26: Sunny. Low 60-63, high 80-83.
Friday, July 27: Partly sunny. Low 64-66, high 80-83.
Saturday, July 28: T-Storms. Low 60-63. High 77-79.
Sunday, July 29: Sunny. Low 64-66, high 84-86.
Monday, July 30: Sunny. Low 60-63, high 80-83.
Tuesday, July 31: Sunny. Low 67-69, high 84-86.
Wednesday, August 1: Thundershowers. Low 60-63, high 87-89.
Thursday, August 2: Scattered thunderstorms. Low 64-66, high 87-89.
Friday, August 3: Partly sunny. Low 60-63, high 84-86.
Saturday, August 4: Thundershowers. Low 64-66, high 84-86.
Sunday, August 5: Partly sunny. Low 60-63, high 77-79.
Monday, August 6: Thundershowers. Low 60-63, high 80-83.
So what's up with the colors? Well I highlighted the forecasts which were either right or semi-right. I highlighted the other forecasts that were wrong. I did this through verifying the data at the Burlington International Airport, which I got the forecast for in the first place. Taking a quick glance at the colors, this thing did pretty good! Of those forecasts that were either right or semi-right, 20 of the 28! The forecast was completely wrong 8 of the days. Simple algebra tells us 20/28 equals 5/7 or a 'correct' percentage of about 70%. Now we can argue all day of what is 'right or wrong' with the forecast. Some say if I'm one degree off then it's wrong...which technically is true. However I think it's still close enough to call it right. So there were seven 'iffy' days in this forecast where the 28-day either had the sun condition right but the temperature slightly off, or vice versa. If you take away these 'half-checks' as I call them, then the percentage drops to an accuracy of only 46%. But I gave this thing the benefit of a doubt.
So overall I'm impressed by this 28-day forecast. It won't work every time, but it's not too shabby, at least with this example. Along with the daily data, I also jotted down the forecast for when it would be "coolest" during the day and "warmest" at night.
The forecast for 'warmest lows' was on July 31. Remember this was a July 11 forecast so it was placing it 20 days out. The actual warmest low temperature we recorded in that four week period was Tuesday, JUly 30! That's not bad at all. However the coolest forecast lows were July 28 and August 5, where the actual temperatures reached to 83° and 87°! The actual coolest days were July 18 at 76° and the following week, July 25 at 71°. So in that regards it failed.
What about the next 28 days? Well it features a cooling trend into September. The coldest overnight lows will come on Labor Day with a projected 45° low. It also shows a dry Labor Day weekend and just 11 days with rain out of the 28. Here's the link if you want to see more:
-Meteorologist Steve Glazier