Re: [Flightgear-devel] Flying over an island

[Melchior F. <mf...@ao...>]

* Torsten Dreyer -- Monday 11 February 2008:
> You might also take weather phenomena into account. When there is
> something like FG (fog) [...] There might be something like blown
> sand [...] 

True, but I'm mostly interested in reconstructing visibilities >10km,
so I want to try to estimate values based on known visibilities in
that range. Values below are only interesting for curve fitting,
and there the few sand storms will only be noise.



* LeeE -- Monday 11 February 2008:
> I think I'd suspect the 110 miles figure (if that's a ground level 
> value) as well, not only because that's a lot of atmosphere to see 
> through but also because of curvature.

According to google searches the 110 are realistic for the East
coast (as a maximum), while visibilities in the West coast of the
USA are much lower. But that should already be contained in the
humidity. (Kind of, as we should really look at size and concentration
of any particles. But humidity is all we have.)



> I tried a quick Google to see if I could find any rules/formulae for 
> visibility due to atmospheric conditions but didn't hit anything.

There are several hits that confirm the correlation of rel. humidity
and visibility. There's even a paper for the airport in Tokyo that
explores the relation and makes predictions based on that. But only
for up to 4000 m, which can probably not be extrapolated.



> It'll be interesting if you can come up with rules or a formulae 
> from your analysis of a large set of METAR data.

The data set is ~100.000 entries, but while the method should be
appropriate, my first graph didn't look promising. The rel. humidity
calculation is relative simple in fgfs (and my test), and I might
have to re-think that. That's why it took longer than my promised
"later today".  :-)  (Also I've had other stuff on my list. But
I'll revisit that topic.)

m.