Point Spread Function (PSF)

:::info Migrated from Old Wiki This content was partially migrated from the old wiki using GitHub Copilot and should be double-checked for accuracy. :::

The PSF profile represents a single point source convolved with the instrumental PSF. It is used to model compact, unresolved sources such as quasar images in a lensed system.

Usage

psf  
  x1c        # x-coordinate of centroid (arcsec)
  x2c        # y-coordinate of centroid (arcsec)
  amplitude  # scaling of the normalised PSF grid (in counts)

Description

Unlike analytic profiles, the PSF profile does not define a parametric surface brightness function. Instead, it places a scaled copy of the provided PSF kernel (specified via esource_sub_agn_psf in the configfile) centred at (x1c, x2c) with the given amplitude.

• x1c, x2c: Position of the point source in arcseconds. These are free parameters that should be initialised close to the observed image positions (in arcsec coordinates of the science cutout).
• amplitude: A multiplicative scaling factor applied to the normalised PSF grid. Proportional to the flux of the point source.

tip

In a quad-lens system, you will typically include four psf profiles — one per lensed quasar image. Their positions can then be linked to the sources positions using the label: / link: mechanism.

:::note Subsampled PSF The esource_sub_agn_psf and its associated esource_sub_agn_psf_factor (which must be an odd integer: 3, 5, 7, …) are used by the PSF profile. The subsampled PSF provides sub-pixel accuracy when fitting bright point sources. :::