The Atmosphere

Good reading: Birney, Gonzalez, & Oesper, "Observational Astronomy", Chapter 7

Transparency/Extinction

The atmosphere absorbs at many wavelengths. There are two major "windows" for ground-based astronomy: optical/near-IR and radio.










Left: transparency (in gray) as a fn of wavelength.




Even in the windows, the transmission can be significantly degraded. This extinction of starlight needs to be corrected, to get the magnitude of a star "at the top of the atmosphere":



where X = airmass and the extinction coefficient k is very wavelength dependant (left, from Stritzinger etal, PASP, 117, 810, 2005).

Extinction coefficients are also, unfortunately, time- and location-dependant! You need to measure them yourself as part of your observing run....



Atmospheric Emission


(image courtesy Kevin Schafer)


Comparison between a night sky spectrum taken at Cerro Paranal-Chile during dark time (lower panel) and one taken in Asiago-Italy (upper panel). Light pollution is clearly visible in the form of Sodium and Mercury emission lines in the blue/visible part of the spectrum. Movie of variable night sky. Media courtesy Nando Patat (ESO); taken from http://www.eso.org/~fpatat/science/skybright/







Refraction and Dispersion


Remember Snell's law?









 Now, the atmosphere is not a sharp transition, but rather a gradual increase in density. We need to integrate over a pathlength to get the total deflection.

It can be shown (Smart, Spherical Astronomy, Chap 3) that, to first order, the correction in position (R) depends simply on the observed zenith distance (z').

zenith dist
 relation
 parameters
z'<45o
 R=C tan(z')
 C ~ 1'
z'<75o
 R=A tan(z') + B tan3(z')
 A=(mu-1)+B
B ~ -0.07
mu=1.0002924

For higher zenith distances, more accurate formulae are needed!


Differential refraction

at large zenith distances (i.e., low altitudes), the refraction angle changes rapidly, distorting angular seperations.

Normally this is a small effect -- <1" for z'<65o -- but is important for precision astrometry.

Differentially refracted sun



Dispersion

The index of refraction is wavelength dependant! Different wavelengths will be refracted different amounts (blue more than red). This is particularly important for spectroscopy, where your slit width can be smaller than the differential refraction.

The solution to this is to rotate your slit to the parallactic angle (ie pointing towards the horizon), which changes over the course of a night as a star tracks across the sky. See Filippenko, PASP, 94, 715 (1982).


Seeing


Watching the image of a star in real time, several changes are obvious:
  • change in brightness (scintillation)
  • changes in sharpness (blurring)
  • changes in position (image motion)
Together, and averaged over an exposure, these effects are collectively known as "seeing."

They produce a blurred profile (point spread function, or PSF) that we often approximate as a gaussian, and the seeing is characterized as a "full width at half max".


Note: sigma and FWHM are not the same!

Also, FWHM is not the same as the size of a star on an image. A bright star will look bigger on an image, because you are seeing further out into the wings of the PSF:




However the real profile is far from gaussian. It has broad extended wings arising from a variety of sources, including diffraction and reflection from the telescope optics, scattering inside the telescope and in the atmosphere, tracking errors, etc...:


Above: gaussian fit to star, plotting log Intensity

PSF of Burrell Schmidt



Gemini North (Mauna Kea) Seeing

Gemini South (Chile) seeing


What's happening? Atmospheric turbulence. Think of the atmosphere as being composed of individual cells; if the cells have a large size, image quality is better. What does "large" mean? Large with respect to the telescope aperture.



Qualitatively, we can consider three sources of seeing:
  • Upper atmosphere (5-10 km)
  • Local atmosphere (atmospheric flow around the observatory)
  • "Dome seeing" (conditions inside the dome and at the dome/outside interface)

Seeing gets worse at high airmass, of course!

How do we minimize problems with seeing?