ASTR 420 · Lecture 14

Lecture 14: Cosmological Synthesis, Tensions, and Tests

Cosmology

Cosmology is strongest when independent epochs and messengers agree; a tension can indicate statistical fluctuation, hidden covariance, calibration systematics, or model incompleteness.

Objectives and route

Construct an evidence-weighted concordance model and compare parameter constraints without treating disagreement as an arithmetic error.
Define and distinguish multi-probe consistency, covariance, model comparison, and open questions.
Derive or audit one quantitative relation with units and assumptions.
Connect the calculation to an observational test and a limitation.

Terms for This Lecture

Joint likelihood

Definition: The combined probability of multiple datasets given shared model parameters and stated dependence assumptions.

Why it matters here: In Cosmological Synthesis, Tensions, and Tests, Joint likelihood is needed to reason about multi-probe consistency, covariance, model comparison, and open questions rather than merely label the phenomenon.

Relationship: It is interpreted alongside Covariance matrix and the lecture relation below; identify which quantities are observed, calibrated, or model-derived.

Limitation or common confusion: This summary is conditional on independent Gaussian errors; shared calibration or model covariance changes the interpretation.

\(\\chi^2=(\\mathbf d-\\mathbf m)^T C^{-1}(\\mathbf d-\\mathbf m)\)

Motivating evidence

Cosmology is strongest when independent epochs and messengers agree; a tension can indicate statistical fluctuation, hidden covariance, calibration systematics, or model incompleteness.

Cosmological claims are indirect: an observable is calibrated, placed in a model, and compared against alternatives. Keep those layers separate throughout the lecture.

multi-probe consistency, covariance, model comparison, and open questions. The axes, labels, and geometry carry the physical inference rather than serving as decoration.

Physical framework

Start from the governing relation and define its domain:

\(\\chi^2=(\\mathbf d-\\mathbf m)^T C^{-1}(\\mathbf d-\\mathbf m)\)

State the coordinate choice, units, boundary conditions, and approximation before substituting numbers. The same symbol can mean a different quantity in Newtonian, special-relativistic, and expanding-space contexts.

Read the visual

Use the figure introduced on the evidence slide as an argument: identify the measured axis, the inferred axis, the comparison model, and the parameter that controls the conclusion.

The paired lab asks you to reproduce the visual from a supplied data extract and test one modeling choice.

Worked calculation

The independent-error normalized difference is |73.04-67.4|/sqrt(1.04^2+0.5^2) = 4.88756 sigma.

This summary is conditional on independent Gaussian errors; shared calibration or model covariance changes the interpretation.

Estimate the scale first, then check dimensions and comparative direction. The displayed value is generated from shared ASTR420 constants and data.

Evidence versus inference

Raw data are not the same as calibrated observables.
A fitted parameter is conditional on the likelihood, prior, and selection function.
Independent probes can agree because the model is useful, or disagree because an assumption is hidden.

Boundary condition

Limit of the model: This summary is conditional on independent Gaussian errors; shared calibration or model covariance changes the interpretation.

Ask what happens if the dominant assumption fails. In cosmology, approximation honesty matters because distances, times, and abundances accumulate model dependence.

Active reasoning

Prompt: Combine a CMB prior, BAO distance ratio, supernova residual, and H0 comparison into a qualitative model test.

Predict the direction before calculating. Then name one observation that could discriminate between two explanations.

Connection to the seven-unit arc

Earlier tools: ASTR310 radiation, dynamics, and statistical mechanics.
Paired data activity: use the corresponding CSV and reproduce the calculation with a perturbation test.
Next step: combine this probe with an independent epoch or distance measure.

Takeaways

Construct an evidence-weighted concordance model and compare parameter constraints without treating disagreement as an arithmetic error.
A cosmological parameter is meaningful only with its model and uncertainty context.
State whether a number is live-verified, a published value awaiting spot-check, or an instructor-provided model value.
Carry the limitation forward into the lab and problem set.

Reading and provenance

OpenStax Astronomy 2e, Chapter 29, sections 29.1-29.7; embedded Figures 29.1-29.30.

OpenStax Astronomy 2e Chapter 29 embedded figure numbers were checked against the extracted text; the source index has chapter-label offsets near the Chapter 29 boundary, so the adopted PDF remains the final exercise-number authority.

Additional sources and dataset status are recorded in reference-log.md. Standard published values not re-fetched this session are level 2 and require human spot-checking.