On 23/04/2018 12:43, Keith Harris wrote:
So it is true what Ole said, and the algarithm seems correct based on
the WMO definition of 120W/m2, which is very old...

I'm not sure that the age of the definition has much to do with it (is
it very old anyway?). Isn't the reality that sunshine hours is not
actually a well-defined parameter and can't be because what you're
trying to measure, ie the boundary between bright sun conditions and
not, is subjective. It's a different kind of parameter to eg temperature
or rainfall. So different measurement techniques inevitably give
different answers and probably always will.

The WMO definition is one approach and an appealing one because it
defines a threshold which can be accurately measured with the right kit.
But how well it parallels human feelings about what is bright sun,
especially in hazy or dawn/dusk conditions, is endlessly debatable.

I suppose an alternative definition could be built around just how sharp
a shadow might be, given modern sensors and image analysis software. But
even so, someone will still need to make the call about precisely where
the cutoff is to be made.