Raster Solar Radiation - Shadow Values?

[JohannesBierer]

Hello,

I want to use the raster solar radiation tool to compare the impact of tree removal measurements on solar intensity. To show that the solar radiation is bigger in tree removal areas I use a digital surface model before and after the measurements. With the solar radiation then I use focal statistics to find the minimal values in the given area where vegetation grows because I want to find the shadow values? In areas with no vegetation i keep the calculated solar radiation values.

What do you think, is that a possible solution?

Any ideas are welcome?

[RyanDeBruyn]

What does the DIFF provide your where you subtract the two results to show where the significant areas of change are?

Using the focal statistics with the specified neighborhood seems like it would highlight those areas (cells) with lowest radiation (ideally because those were more obstructed) thus showing a "shadow".   For example if you had one large tree in the middle of a flat surface and ran the analysis for 1 hour, you would expect to see the effect of the shadow.  Longer overall time periods may be more generalized since the sun position is changing.

Note the output of Raster Solar radiation is the total or global radiation and is a combination of direct and diffuse radiation output.   If you haven't already you may want to try with the optional output "direct radiation raster" since this will highlight the effect of shadows cast by surrounding topography more than the diffuse (which can be from many directions). 

Also the direct duration raster will show the change in the amount of direct sunlight (hrs) for your analysis time configuration (start and end).  To show those cells which recieve "less or more" direct sunlight.

What is your analysis cellsize?
What is the time configuration (start and end)?

As an example, this was my result of the focal statistics (3x3 rectangle minimum) on the direct radiation output. I ran an analysis on a 1meter DSM  for 1 hour in January 1 in Los Angeles.  Areas in black show values = zero (no light). 

[JohannesBierer]

Thank you for your answer.

I used also a cellsize of 1m but in a longer time period. I think I used a hole vegetation period. My area include big trees, smaller trees and no trees, so I tried to represent this with different classes of shadows. But I stopped working on this because it were only assumed values.