Implement flexible extinction model

[dkirkby]

The idea here is to parameterize the extinction function in terms of, e.g.

• air mass (Rayleigh scattering & O2 absorption)
• water vapor
• ozone
• aerosols
• clouds

This could probably be as simple as running MODTRAN4 to tabulate some representative configurations and then linearly interpolating to the desired input parameters.

The purpose of this enhancement is to identify which of these parameters is likely to have the biggest effects on SNR at KPNO. We might then decide to model these parameters in survey simulations and review if / how they are being recorded on the mountain.

[dkirkby]

The following plot from David Burke's talk at the recent DESC workshop shows the various model components calculated by MODTRAN:

[dkirkby]

The open-source LIBRADTRAN library might also be useful: http://www.libradtran.org/doku.php