by Doug Schneider (@Race4caster)
Any meteorologist who takes his job seriously will check his forecast against what actually happened. Experience is the best teacher. Verification is how we learn what went wrong, what went well, and how we can do better. Scott and I want to be open about our forecasts, and show you how we're doing. At regular intervals through the race season, we'll post a summary of how our forecasts have verified.
A few points of note: PoP is probability of precipitation, or the percent chance of rain that is forecast. For more on what that means, see this post. Since most race tracks don't have any weather instrumentation that is publicly available, we have to use the nearest available weather station, usually a nearby airport. This is the "verification point", named by its three-letter identifier. In this case, SAC NWR is a weather station located about 15 miles SE of Thunderhill Raceway. There may be times when the observations at the verification point differ from the track, but the difference should not be significant most of the time. We verify precipitation based only on if measurable rain fell at the observation site, and not based on how much rain fell. Thus, the "observed precip" row will either be Yes or No.
The rainfall forecast was pretty much right on the money for the 25 Hours of Thunderhill. I had the chance of rain in the right time periods with my initial forecast, and I correctly raised the PoP forecast for Friday night as it approached. In a tweet, I said rain would begin to move in between 4 and 6 pm, and it began around 5:30 pm. 0.17 inches of rain was recorded, which was a little below the range that I had forecast in a tweet (0.25-0.5). The entire race was dry, as I had forecast from the beginning.
The temperature forecast was good each day, but off at night. Temperatures did not drop nearly as much as I expected. Probably my biggest forecast error, and one that does not show up in this spreadsheet, was wind. Observed wind speeds on Friday were 10-15 mph, and 15-20 mph with gusts to 25 mph Friday night. I had forecast 5-10 mph winds during these periods. The higher-than-expected winds may explain why the low temperature forecasts were off. Higher winds at night create more mixing of the air near the surface, which prevents cold air from settling at the ground.
Since rain is probably the most important part of the forecast to most people at a race, I'm pleased with how my forecast verified.
A few points of note: PoP is probability of precipitation, or the percent chance of rain that is forecast. For more on what that means, see this post. Since most race tracks don't have any weather instrumentation that is publicly available, we have to use the nearest available weather station, usually a nearby airport. This is the "verification point", named by its three-letter identifier. In this case, SAC NWR is a weather station located about 15 miles SE of Thunderhill Raceway. There may be times when the observations at the verification point differ from the track, but the difference should not be significant most of the time. We verify precipitation based only on if measurable rain fell at the observation site, and not based on how much rain fell. Thus, the "observed precip" row will either be Yes or No.
The rainfall forecast was pretty much right on the money for the 25 Hours of Thunderhill. I had the chance of rain in the right time periods with my initial forecast, and I correctly raised the PoP forecast for Friday night as it approached. In a tweet, I said rain would begin to move in between 4 and 6 pm, and it began around 5:30 pm. 0.17 inches of rain was recorded, which was a little below the range that I had forecast in a tweet (0.25-0.5). The entire race was dry, as I had forecast from the beginning.
The temperature forecast was good each day, but off at night. Temperatures did not drop nearly as much as I expected. Probably my biggest forecast error, and one that does not show up in this spreadsheet, was wind. Observed wind speeds on Friday were 10-15 mph, and 15-20 mph with gusts to 25 mph Friday night. I had forecast 5-10 mph winds during these periods. The higher-than-expected winds may explain why the low temperature forecasts were off. Higher winds at night create more mixing of the air near the surface, which prevents cold air from settling at the ground.
Since rain is probably the most important part of the forecast to most people at a race, I'm pleased with how my forecast verified.
