Astronomy · Observations that support cosmology

Measuring redshift before interpreting it

Identify a common spectral shift while separating wavelength calibration from the physical interpretation of a measured redshift.

Continue in the approved chapter order. Necessary vocabulary, quantities and assumptions are explained in the lesson.

Back to the chapter

After reading, you can

Define a dimensionless redshift with identified features and consistent wavelength conventions.
Separate additive wavelength offsets from a multiplicative scale ambiguity.
Report redshift evidence before assigning velocity, distance or cosmic interpretation.

Redshift begins with a comparison of wavelengths

For an identified spectral feature, define z = λobserved/λemitted − 1. The two wavelengths must use the same unit and compatible conventions, such as vacuum wavelengths for both. A positive z means the observed wavelength is longer; the ratio and z are dimensionless.

The emitted or rest wavelength comes from a feature identification and reference information. A single peak in a noisy spectrum can have several plausible identities. Blends, sky residuals, instrumental defects or source-dependent line profiles can complicate a centre measurement. Choosing a rest line because it gives an expected redshift is not an independent confirmation of that choice.

Several features and the surrounding spectral pattern can test whether one shift describes the candidate identification. Relative strengths and continuum information also require a source and instrument model. They can help distinguish alternatives without making every template assumption certain.

Separate the measured spectral evidence from the interpretation of its shift. Local motion, gravitational effects, cosmic expansion, observing frame and wavelength calibration enter different parts of that interpretation. A numerical wavelength ratio is not automatically a distance or speed.

An additive offset can produce inconsistent naive line shifts

Consider three original hypothetical correctly identified vacuum-rest features at 400, 500 and 700 nm. Assign observed detector wavelength indications 480.6, 600.6 and 840.6 nm. These invented lines are not assigned to actual atomic species or a measured galaxy.

Use the exact toy mapping y = g(1 + z)λ + b. Here λ is rest wavelength in nm, y is the indicated observed wavelength in nm, g is a dimensionless multiplicative wavelength scale and b is an additive offset in nm. The model stipulates one common shift and affine calibration across this range.

Dividing each uncorrected y by its rest wavelength gives naive shifts 0.2015, 0.2012 and approximately 0.200857143. The lack of a single value is compatible with the common additive offset in this toy. It does not by itself establish three different physical motions.

If g = 1 and b = 0.6 nm are independently known, subtract b before taking the wavelength ratio. The corrected observed wavelengths are 480, 600 and 840 nm, each giving 1 + z = 1.2 and z = 0.2. The correction is meaningful because its unit and role match the stipulated model.

Original hypothetical vacuum-wavelength pattern. The last column assumes independently calibrated g = 1 and b = 0.6 nm; no real species or source is claimed.
Rest λ (nm)	Indicated y (nm)	Naive y/λ − 1	Corrected z
400	480.6	0.2015	0.2
500	600.6	0.2012	0.2
700	840.6	0.200857143	0.2

A common fitted slope can still combine two unknowns

Without independently knowing b, compare two distinct rest wavelengths. Their difference removes the additive term: (600.6 − 480.6)/(500 − 400) = 1.2. Thus the fitted slope is m = g(1 + z) = 1.2, and the intercept is b = 480.6 − 1.2 × 400 = 0.6 nm.

The third feature agrees with that affine relation: 1.2 × 700 + 0.6 = 840.6 nm. It checks consistency of the supplied pattern with the declared mapping. It does not separately determine g and z, because they enter the slope as a product.

For example, the different scale g = 1.001 and shift z = 1.2/1.001 − 1 ≈ 0.198801199 predict exactly the same three indicated wavelengths with the same intercept. The calibration ambiguity remains even with exact arithmetic and more precision at these features.

An independent wavelength reference can constrain the instrumental scale and offset, subject to its own uncertainty and observing conditions. The VIM calibration definition emphasizes relationships with reference quantity values and uncertainty. This hypothetical affine model illustrates the need for that information; it is not a universal description of every spectrograph.

A reliable redshift is more than a best-fit label

The SDSS DR17 redshift documentation gives a concrete dated approach: compare spectra with shifted source templates, using measurement uncertainties and considering alternative template classes and redshifts. The reported result depends on that comparison set and model.

The same documentation records warning information and alternative fits. Weak features, unreliable spectral pixels or nearly competing solutions can make a result ambiguous. A warning-free result is useful quality information, not logical proof that the identification and every calibration assumption are correct.

For actual use, retain the spectrum, uncertainty and mask information, instrument resolution, wavelength solution, template identification, fitted shift, alternatives and observer reference frame. Inspect whether different features originate in the same relevant source component and whether blends or local motions affect them.

Uncertainty in line centres, rest references and calibration can be correlated across features. Repeating a ratio with several lines does not automatically turn a common wavelength-scale uncertainty into independent noise that averages away. Its effect belongs in the measurement model and uncertainty analysis.

Redshift evidence can support several subsequent questions

In a specified expanding background, the cosmological component can be related to the ratio of scale factors for comoving emission and observation. That is a physical-model interpretation, with local contributions and observing-frame corrections treated consistently.

A distance inferred from redshift additionally requires an expansion and geometry model and a definition of distance. Luminosity, angular-diameter, comoving and light-travel distances need not coincide. A small-redshift velocity or distance approximation should not be extended to arbitrary z merely because a catalogue displays a velocity column.

Report what the spectrum supports first: identified features or templates, calibrated wavelength convention, shift and uncertainty, frame and remaining alternatives. Then state the conditions under which that result enters a motion, source or cosmological inference.

This chapter develops those observational links while the separate Cosmology course develops cosmic models in its own sequence. The next lesson considers calibrated distance indicators, including why a fixed observed photometric band needs a stated spectral correction before comparison with a rest-band luminosity.

Worked examples

Recover a common shift from one calibrated line

With g = 1 and b = 0.6 nm known, what z follows from y = 600.6 nm for the hypothetical 500 nm rest feature?

Subtract the additive wavelength indication offset: 600.6 − 0.6 = 600 nm.
Form the ratio 600/500 = 1.2.
Subtract one.

z = 0.2 under the supplied identification and calibration model; no speed or distance has been inferred.

Test a different multiplicative calibration

Can g = 1.001 reproduce the same fitted slope m = 1.2 with a different z?

The model constrains m = g(1 + z).
Set 1 + z = 1.2/1.001.
Obtain z ≈ 0.198801199 and retain b = 0.6 nm.

Yes. This alternative predicts all three same indications, so the line pattern alone does not independently determine g and z.

Try the reasoning

May an observed wavelength in nm be divided directly by a rest wavelength entered numerically in angstroms?
Show the answer and reasoning
No.
Convert to compatible units and wavelength conventions before forming the dimensionless ratio.
Does a plausible identification of one noisy peak uniquely establish a redshift?
Show the answer and reasoning
No.
Alternative features, blends and artefacts may fit; use the broader pattern and quality information.
Does the toy’s agreeing third line separate the multiplicative scale g from 1 + z?
Show the answer and reasoning
No.
Every line contains their same product under the declared affine mapping.
Does z by itself specify one universal astronomical distance?
Show the answer and reasoning
No.
Distance inference needs physical and geometric assumptions and a declared distance measure.

What to carry forward

Define redshift from identified, consistently calibrated wavelength evidence.
Separate pattern agreement from instrumental scale identification and shared uncertainty.
State additional physical assumptions before converting a measured shift into motion, distance or cosmic history.

References

Teaching resources and research supporting this lesson’s explanations and examples.