[GRASS-user] r.sun annual radiation

[Salvatore Mellino]

Hi Hamish,

now I can't verify in which way the sun map made with arcGIS was done (albedo, atmospheric turbidity…) maybe the differences depend on terrain shadow effect that I have not considered in grass map.

If I want to add this effect to your script for global radiation, how can I do?
And moreover,  will your script calculate automatically  slope and aspect from DTM or it use the default values?

Thanks

Salvatore


Il giorno 06/set/2011, alle ore 05:13, Hamish ha scritto:

> Salvatore wrote:
>> thanks for the script, it works. I want to be sure that the
>> result is the global radiation map expressed in W/m2/yr (or
>> it is W/pixel/yr?). I compare it with ones made with ArcGis
>> and the values are differents. 
> 
> how much different? just by a few percent or wildly different?
> 
> 
> I have just done some back-of-the-envelope calculations based
> on Spearfish's latitude on Sept 21st and the solar constant,
> both instantaneous at 12 noon, and a rough integration over the
> day, and both answers of my very crude calcs come within 3% and
> 8% of the presumably much more correct r.sun results. (and I'd
> expect the atm scattering to take more out in the dawn/dusk
> hours so that increase to 8% diff for the daily sum doesn't
> worry me) So they would seem to me to be in the correct order of
> magnitude.
> 
> 
> It is not surprising that ESRI is using a different method/
> algorithm than r.sun uses, and if so not surprising that the
> answer would be slightly different. Can you hold the albedo and
> atmospheric turbidity identical in both methods?
> 
> the basic integration of flux from a point source following a
> curve for a known geometry should be rather similar, but the
> fancy atmospheric and ground albedo adjustments is where I'd
> expect to see the differences and interesting science.
> 
> 
>> Another question…the albedo effect is excluded?
> 
> the albedo in r.sun is either given by a raster map containing
> ground coefficients, or the "alb" parameter, which defaults
> to 0.2.
> 
> 
> Hamish
>