[GRASS-user] Possible to model exatmospheric radiation with r.sun?

[Hamish]

Dylan wrote:
> I would be interested in the result, and happy to test.

maybe testing is as easy as running the test for elevation.10m
in the spearfish dataset, or the r.volcano example from the r.sun
wiki page  http://grass.osgeo.org/wiki/R.sun
with lin=1.0, and checking the value you pointed out in the
r.info metadata?

> Any tips on getting started with the coef maps and GRIB
> (not sure what that is)?

GRIB is "GRIdded Binary", and is the standard output raster
format for atmospheric/wave model forecast data. GDAL can read it
(although AFAIK /still/ suffers from a 1/2 cell grid registration
bug so needs a r.region correction after import)
Interfacing with that data is parthly what I wrote d.barb to work
with, although d.rast.arrow will partly do the job as well.
Other free software like zyGrib.org and OpenCPN.org can be used
to visualize it.
I've just now written up a 5 minute quickstart for using zygrib,
see the link at the end of this page:
http://adhoc.osgeo.osuosl.org/livedvd/docs/en/overview/zygrib_overview.html

zygrib does a really nice job of animating cloud cover, give it
a try!


how how to translate that to the beam and diffuse coefficients
for r.sun ... hmmm ... well I guess the beam value would be the
opacity of the cloud ceiling, and diffuse would be the opacity
of the fog.
The man page says they want to be scaled from 0.0-1.0, and
r.rescale can do that from the r.in.gdal'd GRIB data. I assume
0.0 is clear sky but I'm not sure what 1.0 would be equivalent
to.. a grim black sky as you might get under a thundercloud?
Probably just the GRIB file's cloud-cover max is highly variable
and not appropriate.


Hamish