Cumulus Support

I hope someone can clear this up for me as I am a little confused.

I used to record bright sunshine hours using a traditional sunshine recorder where it burnt a trace onto a piece of card and I counted the burn marks to make the sunshine hours for that day.

Since I installed my Solar unit on my Vp2 I stopped using it after a period of time as the VP2 figure computed matched the old fashioned way. I was using something called weatherlink reporter to compute this info. Within this programme you could set the "bright sunshine " figure which I had at 120 w/sqm. In other words it only recorded sunshine above that amount and worked out a sunshine total for that day.

With cumulus we have two settings on the screen available to change
Threshold and Transmission factor

What are folk using as their settings ?
I am using the default settings of 75% and 0.80 but the results I am getting back on the daily sunshine total are very different to the weatherlink reporter result

eg Yesterday
WLR- 0.7hrs ( mainly cloudy all day)
Cumulus 2.3hrs

Could the sunshine total be shown in the monthly weather report ?

Many thanks

A fixed 120 W/m2 threshold is completely useless in this case, so Cumulus will hopefully never agree with those figures. A fixed threshold only works when your sensor measures direct irradiance. The Davis solar sensor measure global irradiance, which varies throughout the day (it depends on the position of the sun in the sky). So when the sun is low in the sky, the value could be below 120 but still the sun is shining. Cumulus works out the theoretical maximum value at any given time of day (taking into account your location and the date) and compares the current reading with that.

So, basically, you will have to tweak the figures to get good results, but comparing the results with something which uses a fixed threshold isn't a good way of checking your results.

Could the sunshine total be shown in the monthly weather report ?

Yes, I have plans to add more data everywhere; I'm getting there as fast as I can... Actually, I thought I had already done this one, but clearly not.

Thanks Steve

So the previously inputted longtitude and latitude etc have been taken into account. This is similar to the WLreporter which asks for the same.

For my C.O.L report I would like the following
Daily Mean MJ/m2
Max Daily MJ/m2 and date
Min Daily MJ/m2 and date
1hr Maximum w/m2 and date/hour starts eg 815w/m2 at 12noon on the 24th

For my daily report and COL summary, number of bright sunshine hours, days with no sun, sunniest day etc in hrs.

Interested to know what others have as their Cumulus solar settings - tks as ever

Paul

Paul C wrote:So the previously inputted longtitude and latitude etc have been taken into account. This is similar to the WLreporter which asks for the same.

It seems odd that it asks for your location, but then uses a fixed threshold.

For my C.O.L report I would like the following
Daily Mean MJ/m2
Max Daily MJ/m2 and date
Min Daily MJ/m2 and date
1hr Maximum w/m2 and date/hour starts eg 815w/m2 at 12noon on the 24th

For my daily report and COL summary, number of bright sunshine hours, days with no sun, sunniest day etc in hrs.

So not much at all, then

I'll see what I can do, but I don't know when I'll be able to do that little lot.

I don't understand this one:

1hr Maximum w/m2 and date/hour starts eg 815w/m2 at 12noon on the 24th

Is it the maximum of the set of solar readings taken on the hour?

Interested to know what others have as their Cumulus solar settings

I'm using

SunThreshold=75
RStransfactor=0.87

but I haven't had chance to do any validation.

I have my own script (VBScript) that calculates the theoretical radiation and compares it with the actual value registered by the Davis sensor. I use it to create a 48h graph showing both values and also to calculate a "current weather" icon, that appears on my main webpage and in the "sticker" you can see in my signature, based on the ratio between both values among other variables. If you take a look to the graph in a sunny day you'll see that both values match pretty well.

However, I've been taking a look to the same ratio calculated by Cumulus and it usually gives higher ratios, normally well above 100%. In other words, it looks to me like it calculates a too low theoretical value, at least in my case.

I'm telling you this because it might explain what Paul is also observing in terms of sunshine hours on cloudy days (the effect on very clear days might not be noticeable).

Steve, may I ask you what model you are using for the theoretical calculation? In my case I use Bird & Hulstrom's model, although a simplified model that I was using earlier based on approximate formulas I got from a book still delivers pretty good results using the appropriate local adjustment factors.

Needless to say that this is an interesting subject for me too

Meteoalmoster wrote:Steve, may I ask you what model you are using for the theoretical calculation?

I'm using Ryan-Stolzenbach. I got the information from here: http://www.ecy.wa.gov/programs/eap/models.html - see the SolRad link. I'm confident that my code is correct compared to the spreadsheet, but I too thought the figures too low, which is why I am using a transmission factor of 0.87 - even with that setting, my peak values are higher than the theoretical max.

Perhaps I am not doing the correct thing with the result of the formula? If you'd like to take a look and offer any advice, I'd be grateful.

1hr Max

" Observers are asked to quote 1hour maximum values for a 60 minute period commencing on the hour if possible "

I suppose the strongest solar hour in the month.

Its a strange request and only 18 weather stations report the total, half of which are VP2 stations.

I can perhaps send you a copy of the data table if you dont subscribe to the COL mag

Sorry for the requests,

steve wrote:

Meteoalmoster wrote:Steve, may I ask you what model you are using for the theoretical calculation?

I'm using Ryan-Stolzenbach. I got the information from here: http://www.ecy.wa.gov/programs/eap/models.html - see the SolRad link.

From a different source in my case, but it looks like we ended up finding the same information .

I've played a little bit with the Excel file pasting real (1') data from my station into it. I picked a recent day in which we had an especially clear sky and it looks like Ryan-Stolzenbach model gives very similar, almost identical results to Bird-Hulstrom (with my adjustments, see later) if I choose an adjustment factor around 0.835. In this case both models match pretty well between each other and with real data.

In fact, I also had to tweak a little bit the "default" values for the B-H model, and this one has more switches you can play with (up to 6 parameters). I had not tried the R-S model but it looks like all of them can provide good results with the appropriate tuning. A different question will be if such tuning will be valid throughout the year and under different conditions. I'll keep an eye on it.

I still cant get to grip with this Bright Sunshine Total - large discrepancies between Cumulus and Weatherlink Reporter (WLR)

January 2011 so far
Cumulus 91hrs
WLR 24.9hrs

Cumulus is set at Sunshine Threshold of 80% and Transmission Factor at 0.83 ( to be honest I dont fully understand what they both mean or would do if I altered them)
WLR is set at recording a Bright Sunshine Total when solar reaches above 120 w/sqm.

As a result of have got my old Sunshine Recorder out and got ahold of some new cards and I intend to carry out some form of comparison in February and March to see how they compare.

A question or two has arisen
1. What is the best way to check Cumulus each day to obtain a daily sunshine total.
2. Should I reduce the settings during the winter for WLR to say 100 w/sqm

cheers
Paul

Paul C wrote:WLR is set at recording a Bright Sunshine Total when solar reaches above 120 w/sqm.

As I said above, you can't take the solar figures produced by the Davis solar sensor and apply a fixed threshold and hope to get any meaningful results. The threshold varies depending on your location, the time of day, and the date. The 120 W/m2 figure which Weatherlink uses is meaningless.

1. What is the best way to check Cumulus each day to obtain a daily sunshine total.

The daily total is in dayfile.txt

Thanks Steve .

The weatherlink "reporter" is seperate to weatherlink - I could send you the programme if you wish, or the link to it and see what you make of it ? It calculates data rather than a weather programme like Cumulus etc.

edit - looks like its still available - http://www.prodata.nildram.co.uk/

I've looked at the download page. It doesn't seem to use the Weatherlink 120 W/m2 threshold, it says:

"Sophisticated algorithm for calculating Bright Sunshine duration from solar radiation data"

I know that John Dann understands the problems with getting sunshine hours out of the Davis sensor (much more than I do, in fact), so I wouldn't expect it to use the fixed threshold, and so it seems, it doesn't.

But you won't get sensible results from the Cumulus calculation without finding the right settings. And I know that there are issues with early and late in the day when the theoretical max is very low. I believe John incorporates a period at the start and end end of the day which effectively gets ignored. I need to do something similar in Cumulus when I get the time.

Part of the settings asks for
Latitude/Longitude and time zone etc

Also asks for "minimum Intensity for BS evaluation (w/sqm) - which I have at 120sqm.

May be worth running the programme at your end and see how your data compares with Cumulus

(Clearly Cumulus is superior ) - its just the solar aspect that interests me at the moment.

I have now re-installed my Casella Sunshine Recorder Mk3 and awaiting the sun to shine, which fortunatley it is now doing

thanks

Paul C wrote:Also asks for "minimum Intensity for BS evaluation (w/sqm) - which I have at 120sqm.

This is probably how John is handling the start/end of day problem; that seems like a better method than ignoring a period of time. 120 seems a bit high to me.

May be worth running the programme at your end and see how your data compares with Cumulus

I have no Weatherlink data to feed it. And no sun

I have now re-installed my Casella Sunshine Recorder Mk3 and awaiting the sun to shine, which fortunatley it is now doing

Good - it'll be interesting to see a comparison of all three methods; then you should be able to tweak the Cumulus settings to get close to the CS.

I've added a 'minimum' setting to the latest build to cater for the tricky start and end of day periods. Default zero, because I haven't a clue at the moment what a sensible value would be.