Measuring Red Shift with Accuracy

[wolram]

We have measures that define the volt, amp, kilo, pressure, etc, etc, compared to these
measures how accurate are red shift figures? i know that tolerances are only guaranteed
in a machine shop if the temperature is within limits, and all the above can be tested time and time again, but how do we measure red shift with such accuracy?

 

[marcus]

I googled "spectroscopy" and got
http://en.wikipedia.org/wiki/Spectroscopy.

It seems like a pretty good article, although it lumps together the cosmological redshift and the Doppler shift

(one is due to stretching during travel, the other is due to source motion, but the shift in wavelength is measured the same way)

basically the tool is a PRISM or something analogous to a prism called a diffraction grating which you could think of as a lot of parallel scratches on a thin flat glass plate that act like a lot of thin parallel prisms.

the diffraction grating (like a prism) spreads the light out into its different wavelengths
so you get a band that looks like a cross-section of a RAINBOW

a rectangle with lots of different color-stripes-----it is called a "spectrum"

the color stripes are called "spectral lines"

this band of color-stripes is made to fall on photographic film, or on a modern solidstate CCD electronic substitute for film----so they take a picture of the spectrum
and they LOOK WHERE THE LINES ARE

it is like a fingerprint of the star

what lines, tells what atoms are in the star----like if it is an recycle-material star with some sodium or iron in
or if it is a fresh-material star with mostly pure hydrogen and helium

all these things have distinctive lines

and they measure carefully if the lines have been SHIFTED to a slightly different wavelength.

spectroscopy is done very accurately in Earth laboratories and they know these wavelengths very accurately, using light from flames and electric discharge etc.

so all the astronomers need to do is stick a telescope into the picture (between the "flame" and the "prism")

it is one of the cooler things humanity is done----a way of smelling what is cooking on the surface of stars

 

[f95toli]

I am not familiar with the exact techniques used to measure red shift.
However, I imagine that the precision is so good simply because what you are measuring is essentially frequency; and frequency/time can be measured with EXTREMELY high precision. How high depends a bit on the time spans etc involved (for long times drift becomes a serious issue) but something like 1 part in 10^15 should be possible for all frequencies/wavelenghts involved. Hence, I don't think precision is an issue.

Modern clocks can measure time with an extremely high precision and in just a few years will have reached a point where the precision will be limited by how accurately we can determine the position of each clock on earth; this is needed to account for relativistic effects so that the clocks that are part of standard time can be compared to each other.