1 of 8

WL Focus Group Update

👻 Ian Harrison 👻

SKAO Cosmology SWG Meeting, Goerlitz

18 June 2025

Slides at: http://bit.ly/ianh_skacosmo25

2 of 8

SKA Weak Lensing Reminder…

  • Measure growth of structure through lensing effect on galaxy shapes
  • Require > 1 arcmin-2�resolved galaxies at z~1
  • Require to measure shapes with very high accuracy
    • Hard with interferometers
  • 3.6 sigma detection in FIRST
    • Chang Refregier, Helfand 2004
  • Upper limits from SuperCLASS
    • Harrison et al 2020
  • SKAO-mid will be first radio telescope with necessary resolution, sensitivity(?)

2

Slides at: http://bit.ly/ianh_skacosmo25

3 of 8

AA* is Not Good

  • Simply cut T-RECS according to Srmsand PSF size
  • Normalise to specs of:�2.7 gal arcmin-2 at�0.55 arcsec PSF and�3 uJy depth�from AA4 forecast
  • Never hit 1.0 gal arcmin-2�threshold no matter how�deep you go
  • Maybe something like 75km baselines you could still get a detection within a smaller area?

Source 1

Source 7

Threshold for weak lensing to work 👆

Depth [uJy]

Number of resolvable�galaxies [arcmin-2]

Slides at: http://bit.ly/ianh_skacosmo25

4 of 8

Methods – DeepShape

  • Paper from Nice group here
  • Deep learning approach to shape measurement for interferometer Weak Lensing
  • Benchmarks against previous image- and uv-plane methods
  • …working on running on SuperCLASS DR1 and DR2 data

Slides at: http://bit.ly/ianh_skacosmo25

5 of 8

Data – LOFAR ILT

  • Paper from Leiden group under review
  • Use 0.3” International LOFAR (ILT) images x HSC optical images
  • Detect correlation between radio and optical position angles 𝛼
  • Forecast WL signal of 1.8σ for 128 hours and 6.8σ for 3200 hours with ILT

Slides at: http://bit.ly/ianh_skacosmo25

6 of 8

Forecasts – Shear x CMB Lensing

  • Forecast x-correlation signal with CMB Lensing from Simons Observatory here
  • For SKA1 get useful constraints on source redshift distribution
    • cf Alonso et al�~10x worse from galaxies x CMB lensing
  • For SKA2 observable x-corr more constraining than Euclid x-corr

Slides at: http://bit.ly/ianh_skacosmo25

7 of 8

Preparation – UKSRC Demonstrator Case

  • Access to UKSRC resources to demonstrate science case
  • e.g. imaging / analysis of SuperCLASS data
    • Have 2022 submission to work from
  • Due 27th June
  • Help gladly welcomed!

Slides at: http://bit.ly/ianh_skacosmo25

8 of 8

Others…

  • Ongoing CosmoStat+ TOSCA project
    • WP1: “From visibilities to galaxy shapes”
  • Works on Kinematic lensing
    • https://arxiv.org/abs/2201.00739 and many others by UArizona group
  • Work on SuperCLASS DR2
    • Images awaiting some final calibration
    • Methods re-implemented in modern python
    • Participation of DeepShape group
  • Work on self calibration in SKA1 x Optical Stage IV surveys
    • Led by Roberto Ingrao
    • Bugs fixed, draft being written

Slides at: http://bit.ly/ianh_skacosmo25