Sep 09, 2021 - Smoothed kappa maps

Dick suggested that our problems with sensitivity to places where kappa is close to one and magnification diverges can be solved by smoothing the kappa maps. A natural first guess for the smoothing scale is an effective radius of the pixel. It is given by

[math]\displaystyle{ \pi \theta^2 = \frac{4\pi}{N_{pix}} }[/math]

and evaluates to

[math]\displaystyle{ \theta = (\sqrt{3}N_{side})^{-1} . }[/math]

Using the healpy smoothing routine with [math]\displaystyle{ \sigma = \theta }[/math], we pre-process the maps before lensing.

As visible from histograms, it helps to cure the tails: