Calculating sunshine hours from solar radiation

And you're sure that all the stations you're getting radiation data from are using such pyranometers? AFAIK most non-official stations at least, just use the solar panels used to power the station itself to "measure" sun intensity.

Would think less for Littlefork. Likely around 2200.

I notice this as well ("Foehn effect").

But does cloud cover "spill" over the GWS metropolitan area? I thought the clouds don't make it down the mountains. 

Couple more places:

Carlos Avery, MN (near Minneapolis): 2279 hours
Bonners Ferry, ID: 2216 hours
King Coulee, MT (near Glasgow in the NE corner): 2794 hours
Las Cruces, NM: 3434 hours

The Minneapolis value seems too low but the others seem about right.

Couple more:

Redding, CA:                       3266 hours
Scipio, UT (central Utah):     3220 hours
Helena, MT:                         2579 hours 
Coville, AK (in the SW part): 1622 hours

Ok I think I finally found a representative station with more than a few year old POR.

I looked at all the nearby CO Ag Mesonet stations and the two closest RAWS and the Ault Mesonet station seems as good as it’ll get. It’s 17 miles due east of Fort Collins, it’ll be a bit more continental and less snowy but sunshine hours shouldn’t change too much. The nearer stations all had either inaccurate or missing data or few years.

Ault has data since 1993 with only the scattered missing month. 

Here’s the averages I calculated:
Month/Hours/%

Jan    170   56.9%
Feb    199   66.0%
Mar    260   70.1%
Apr    252   63.0%
May   273   60.8%
Jun    318   70.4%
July   323   70.6%
Aug   297   69.8%
Sep   267   71.6%
Oct   227   66.1%
Nov  181   60.9%
Dec  146   50.8%

Annual  2913 hours

March is surprisingly high, but the pattern correlated with other stations. And it’s not going to be randomly high over 25 years. December might have some issues with shading as it’s suspiciously low but otherwise accurate. The 2900 hours annually is right in line with what it should be.

The sunniest complete year was 3047 hours in 2000 and the cloudiest was 2741 hours in 2009.

Tommyfl where do you obtain solar radiation data from? Is there a database which has global records of this?

Ault clearly has shading issues in winter. I just looked at Briggsdale about 20 miles east of Ault and 37 miles from here.

The two stations are within 10 hours from February to October and Ault is sunnier than Briggsdale a few of those months. However, Briggsdale is 17, 35, and 28 hours sunnier in the other three months, which is significant by percentage. The difference is inexplicable except that Ault isn’t accurate those three months, and we don’t get 50% sun in December here so Ault is clearly the inaccurate one.

Briggsdale goes back to 2002 but has lots of missing data, mostly in the summer months. Since they’re so close in the warm season, I think using Ault March-September and Briggsdale October-February would be pretty accurate.

This method gives 3002 hours which I think is about 40-80 hours too high but none of the Fort Collins, Greeley, or RAWS stations are anywhere close to accurate and there’s nothing else closer.

I don't think it's possible to download a Highcharts graph as text unless the developer has enabled that option. See how the top-right menu allows download as CSV and viewing as a data table on my site, but not on the CSU page. 142.11.236.169/php/weatherstation.php

However, since you can download solar radiation data at 5-min intervals, you can use that data and equation 28 (www.fao.org/3/x0490e/x0490e07.htm#radiation) to calculate average Ra for each 5 min period. It's quite a involved calculation but it can be done. Once you have Rs/Ra, percent possible sunshine can be calculated for each observation easily. I could make another calculator to use 5-minute data, but it would take some time.