Step 1.3: GLEAM + In-Filled Point Sources

[steven-murray]

1.3: Gleam + In-Filled Point Source

This test is an extension of #47 which adds synthetically-generated point sources to the existing GLEAM catalogue where it does not have coverage. The point of this isolated step is simply to ensure that the recovered power spectrum has high dynamic range (the same test as for 1.0, 1.1 and 1.2).

The in-filled point sources should be uniformly distributed on the sky, and drawn from dN/dS down to the minimum flux density of GLEAM.

• Simulation Component: GLEAM + In-Filled Point-Sources
• Simulators: vis_cpu
• Pipeline Components: pspec
• Depends on: Nothing.

Why this test is required

To increase the range of LSTs that we can simulate in the end-to-end for H1C IDR3 and beyond, we need to be able to fill in point sources in places that GLEAM doesn't cover. This tests whether we can do that reliably.

Summary

A brief step-by-step description of the proposed test follows:

• Generate point source catalog from dN/dS in GLEAM missing regions
• Combine GLEAM with synthetic catalog
• Run vis_cpu on the catalog
• Compute EoR visibilities with vis_cpu
• Use pspec to generate the power spectrum in some LSTs affected by the missing regions
• Check that the dynamic range is large enough to detect EoR, and that EoR is unbiased.

Simulation Details

Two simulations are required: the GLEAM+In-Filled, and the EoR. These should be done at the specs required of the end-to-end, so they can be re-used.

Criteria for Success

• Dynamic range of FG > XX
• EoR PS unbiased to ~1% for k > XX.

[bhazelton]

Where is the code that generates the gleam + filled in sources going to live? There's an abandoned PR on pyradiosky for this, wondering whether that's going to be picked back up or if it's going somewhere else.

[steven-murray]

@bhazelton That's a good question. pyradiosky seems like a good place. Why was the PR abandoned, btw?

[bhazelton]

I'm not sure, it was started by Samir and I've given him feedback several times but there's been no new commits since March. He did comment on the PR in July to say he was planning to keep working on it but I've heard nothing since.