Posted by Wick Smith on August 24, 2009 at 2:45 PM
A local astronomy buff told me about a great site for determining specific hourly sky conditions in areas in North America. It will provide you hourly predictions on the expected cloud cover and visibility forecast. I've found it to be extremely accurate in the locations I monitor. It is great for determining whether or not to get up early for a sunrise shot, or for other shots that are dependent upon sky conditions. It only forecasts out about two days for accuracy reasons.
This tool is not a bad thing, any information is good so long as the limitations are understood. It is unfortunate that no one (to my knowlege) has put in any serious effort in the US to develop a good sky model, if my programming skills werent abysmal id probably do it myself (or if someone provided me with a huge grant )
The tool was developed for astronomers, so IMO the value to photographers is somewhat limited. The only graphic i can see that may be of use is sky coverage. Unfortunately, theres no way to tell what sort of clouds we are dealing with here. I dont know what parameterizations are used here (i may email the CMC for info) but the disclaimers make one take pause: it specifically mentions that low clouds and thunderstorms may not be accurately accounted for, which means the convective and boundary layer schemes probably arent there. This limits the value to a degree. In my opinion, one is better off looking at charts of relative humidity at various levels: this will give one a better understanding of the cloud layers involved, and really it is cloud type and levels that are more significant to the photographer if you ask me. It takes a bit more studying and willingness to learn, but the info is there. For one stop shopping, this astronomical tool is worth a look.
Clouds, sky coverage etc are notortiously the most difficult things to predict in the weather. Models such as the one this tool is based on (the Canadian Regional Gem--too bad a local U.S. model is not available!) have been known to do a pretty lousy job forecasting clouds. Which is a good thing, cause it keeps me employed! A forecaster typically looks at numerous models and parameters in making a forecast, not just one canned result. This particular model run could be completely out to lunch, and you'd have no way of knowing that unless you compared it to other models--which you can't do with this tool! I just looked at the most recent cloud forecast for my house, and it said 70% sky coverage. Actually, it is 100% and raining. FWIW, the NWS graphical forecast had 90% coverage for the same time and place. Just be cautious in using tools such as this, there is no magic formula in them.
Posted by Stan Rose on August 24, 2009 at 10:19 PM
Allow me to expand on this a bit.
Model forecasts of clouds are not limited to this one tool. What makes this particular tool appealing is that somebody pakaged it up to make it a bit more user friendly for astronomers (unfortunately photographers were not included ) But there are many many sources of information if you do a bit of searching and can stomach a bit of technical whiplash. SInce i went to grad school at UW, Im familiar with their regional MM5 model (I ran a few myself and worked on some of it's physics). Here is an ensemble forecast (meaning, multiple models, rather than just one) of "integrated column sloud water" which can be interpreted as clouds for the sake of this example (info from http://www.atmos.washington.edu/~ens/ewme.cgi
It is a 12 hour forecast from this morning for 5 pm this afternoon. The colors can be taken to represent clouds.
Posted by Stan Rose on August 24, 2009 at 10:24 PM
Now,, here is the visible satellite, verification 12 hours later, at 5 pm this afternoon. Note that this for the Pacific Northwest only, doesnt help me in Colorado much, but i'm just pointing out that there is more information out there. There are also regional models run in Colorado or your local area that also forecast similar parameters. Personally, I think information like this is more valuable than just a little gray box, but unfortunatley it does take a willingness to dive in and get dirty. If I get some free time (fat chance) maybe I will come up with a list of links that lead to model information like this, since photographers seem to be hungry for this information.
Posted by Wick Smith on August 25, 2009 at 8:53 AM
Stan,
Click on the "little grey box" for any future hour on the graph. It will show you a projected cloud cover for that hour for that specific spot on the map. It's not perfect, but I find it highly accurate. If it doesn't give you all you want, that's OK, too. I thought it would be as helpful to some others as it is to me.
Discussion
7 replieshttp://cleardarksky.com/csk/
Nancy
www.NancyWagnerPhotography.com
This tool is not a bad thing, any information is good so long as the limitations are understood. It is unfortunate that no one (to my knowlege) has put in any serious effort in the US to develop a good sky model, if my programming skills werent abysmal id probably do it myself (or if someone provided me with a huge grant
The tool was developed for astronomers, so IMO the value to photographers is somewhat limited. The only graphic i can see that may be of use is sky coverage. Unfortunately, theres no way to tell what sort of clouds we are dealing with here. I dont know what parameterizations are used here (i may email the CMC for info) but the disclaimers make one take pause: it specifically mentions that low clouds and thunderstorms may not be accurately accounted for, which means the convective and boundary layer schemes probably arent there. This limits the value to a degree. In my opinion, one is better off looking at charts of relative humidity at various levels: this will give one a better understanding of the cloud layers involved, and really it is cloud type and levels that are more significant to the photographer if you ask me. It takes a bit more studying and willingness to learn, but the info is there. For one stop shopping, this astronomical tool is worth a look.
Clouds, sky coverage etc are notortiously the most difficult things to predict in the weather. Models such as the one this tool is based on (the Canadian Regional Gem--too bad a local U.S. model is not available!) have been known to do a pretty lousy job forecasting clouds. Which is a good thing, cause it keeps me employed!
Model forecasts of clouds are not limited to this one tool. What makes this particular tool appealing is that somebody pakaged it up to make it a bit more user friendly for astronomers (unfortunately photographers were not included
http://www.atmos.washington.edu/~ens/ewme.cgi
It is a 12 hour forecast from this morning for 5 pm this afternoon. The colors can be taken to represent clouds.
Click on the "little grey box" for any future hour on the graph. It will show you a projected cloud cover for that hour for that specific spot on the map. It's not perfect, but I find it highly accurate. If it doesn't give you all you want, that's OK, too. I thought it would be as helpful to some others as it is to me.