Lately WXSIM seems to not be doing well with forecasting Day time highs. I take an automated forecast at 07:00 and at 19:00. The morning forecast has a fairly low accuracy level while the evening forecast does quite well in comparison with the NWS for my area.
You can see the comparisons at http://www.barrowcountyweather.com/wxforecast_compare.php
Any suggestions?
Many thanks!
Al…
Hi Al,
generally afternoon forecasts work best, and morning ones are iffy at times, due to the difficulty of properly characterizing the air mass when local inversions may exist. Another issue, though, is cloud cover. Have you checked to see whether that’s been forecast accurately? Often temperature errors are secondary to cloud cover errors.
Tom
Hi Tom,
Thanks for your quick response. I am not sure what you mean by or how I check “…checked to see whether that’s (cloud cover) been forecast accurately?”
Al…
One way to check is simple observation of the sky and looking back at the forecast wording, but that isn’t too precise most of the time. If you have a solar radiation sensor, you can do a direct comparison using wret.exe. I might post something like that here as an illustration soon.
Tom
I just had a 0600 run pull a 91F hi. I just don’t think so. So, I checked the cloud cover and it’s showing pcldy when it’s dense ovrcst. So, I’m going for a manual run.
update
Ticked import cloud cover - Still 91F for auto run.
manual 77F
Are you saying that the forecast is showing partly cloudy when it’s dense overcast, or that the initial data is? As far as importing cloud cover as local data, first note that’s only for users with a solar radiation sensor in the daytime, and it’s for the calibration run and initial cloud cover. It has almost no effect on forecast cloud cover.
Feel free to send me custinit.txt, latest.wxf, and wdata.txt and I can take a look.
Tom
Just the initial import cloud cover. I do have a solar sensor, so now I’m utilizing cloud cover under Current Data.
Lee,
Since you have a solar sensor, you’ll be able to do after-the-fact comparisons, to see how cloud cover turned out. It will be interesting to see how the forecast high of 91 happened.
Tom
Lee,
Can you show the temperature comparison as well?
Tom
