ASTR 472 · Lecture 06

Lecture 06: Aperture Synthesis and Image Formation

Radio Astronomy

Earth rotation and multiple baselines fill the uv plane imperfectly; imaging is an inverse problem with a measurable point-spread function.

Objectives and route

Connect sampled visibilities to a dirty image and dirty beam.
Explain how weighting changes resolution and sensitivity.
Identify missing short spacings and deconvolution non-uniqueness.

Route: identify what the receiver measures, derive the physical/instrument relation, calculate a scale, then test its assumptions against a diagnostic.

Terms for This Lecture

Aperture synthesis

Definition: Reconstruction of sky structure by combining Fourier samples from many baselines and times.

Why it matters here: It is needed to answer A smooth 8-arcminute H I cloud is observed with an array whose shortest baseline implies a 2-arcminute maximum recoverable scale. What map artifact or missing-flux behavior do you expect? and to interpret A finite uv sampling function multiplies the true visibility function; its Fourier transform convolves the true sky with the dirty beam.

Relationships: Use it with Largest recoverable scale and Dirty image; the lecture model is \(I_{\rm dirty}=\mathcal{F}^{-1}[S V]=I_{\rm sky}*B_{\rm dirty}\).

Boundary or common confusion: The 1.02 lambda/B_max resolution estimate is a scale, not a full synthesized-beam prediction; uv weighting, taper, flagging, and source declination matter.

Observation before interpretation

Earth rotation and multiple baselines fill the uv plane imperfectly; imaging is an inverse problem with a measurable point-spread function.

A finite uv sampling function multiplies the true visibility function; its Fourier transform convolves the true sky with the dirty beam.
Natural weighting emphasizes densely sampled short baselines; uniform weighting can improve resolution at a noise cost.
An interferometer without zero and short spacings filters broad smooth emission, even when a visually plausible deconvolution is produced.

Keep four layers distinct in your notebook: raw correlation or spectrum, calibrated product, model parameter, and physical claim.

Physical model

Let S(u,v) be the sampling function. The measured visibility is S V, and inverse Fourier transformation gives I_dirty=I_true convolved with B_dirty, where B_dirty is the inverse transform of S. Deconvolution estimates a sky consistent with sampled data and a prior/regularizer; it cannot recreate unconstrained spatial frequencies without single-dish or compact-array information.

\(I_{\rm dirty}=\mathcal{F}^{-1}[S V]=I_{\rm sky}*B_{\rm dirty}\)

Define each quantity and convention before substitution. A relation is useful only inside its stated regime.

Worked calculation

Inputs: nu=100 GHz, B_max=16 km, B_min=15 m.

synthesized FWHM scale = 0.03942 arcsec; largest-scale estimate = 24.73 arcsec

The 1.02 lambda/B_max resolution estimate is a scale, not a full synthesized-beam prediction; uv weighting, taper, flagging, and source declination matter.

Independent audit: carry SI units through the derivation, then check the limiting direction and scale.

Read the geometry or data relation

A finite uv sampling function multiplies the true visibility function; its Fourier transform convolves the true sky with the dirty beam.

Identify what is directly encoded in the figure and which inference requires an instrument model or physical assumption.

Where the inference can fail

The 1.02 lambda/B_max resolution estimate is a scale, not a full synthesized-beam prediction; uv weighting, taper, flagging, and source declination matter.

Check the frequency, velocity frame, beam definition, and calibration convention before comparing values.
Use residuals and alternate models to distinguish a weak signal from a plausible but non-unique interpretation.
Report thermal noise separately from calibration, selection, and model uncertainty.

Reasoning prompt

A smooth 8-arcminute H I cloud is observed with an array whose shortest baseline implies a 2-arcminute maximum recoverable scale. What map artifact or missing-flux behavior do you expect?

Write the predicted scaling first; then identify which measurement would falsify it.

Connection to the observing workflow

This lecture feeds into the paired lab: Baseline Geometry and Visibility Sampling. Preserve the raw-to-calibrated-to-inferred chain and carry the same units and assumptions into the analysis.

For a real observation, archive the project identifier, correlator setup, antenna configuration, calibration version, flags, beam, and data-release citation.

Synthesis

Aperture Synthesis and Image Formation is one stage in a measurement chain: celestial signal → antenna and receiver → sampled data → calibration → image/spectrum → uncertainty-aware inference.

Next: connect this result to Calibration I: Complex Gains and Bandpass.

References and scope

OpenStax Astronomy 2e §6.4 covers VLA/ALMA arrays qualitatively; ALMA Basics provides a shortest-baseline largest-angular-scale estimate.

Primary/authoritative material and provenance levels are recorded in ../reference-log.md. Numeric examples are reproducibly calculated by ../src/generate_astr472_package.py; figure geometry is original and lecture-specific.