Skip to main content
Sandbox Physics

Weak-Lensing Mass Map

Each galaxy is distorted only a little. How do thousands of them add up to a map of invisible mass? Mass in front of distant galaxies stretches their images slightly, always around the mass rather than toward it. One galaxy is too randomly shaped to show this, but averaging thousands cancels their own shapes and leaves the common stretch. Turning that stretch pattern back into mass gives a map, except for a uniform layer of mass, which leaves shapes unchanged.

Drag invisible dark-matter halos and their lens planes. Watch background galaxy shapes line up around them, then rebuild the mass map from those shapes alone.

Interactive modelWeak-Lensing Mass Map
Einstein radius: from shapes / true24.0′′ / 21.7′′24.0^{\prime\prime}\,/\,21.7^{\prime\prime}
Mass-map peak over noise11.111.1
Lensing efficiency of the selected plane0.6830.683
Cross-shear null test0.156 σ0.156\,\sigma
Mass map from shapesSynthetic catalog · E mode
Null map: should be noiseSynthetic catalog · B mode
Stretch around the selected haloSynthetic catalog · noise-free truth
Which planes the sources feelCalculated · flat cosmology
How the maps and the fit are made

Shapes are averaged in 0.375 arcminute cells, rotated back to convergence in Fourier space and smoothed with a Gaussian. The E map holds the lensing signal; the B map is what lensing cannot make, so it shows the noise and the edge errors. Both share one colour scale. Rings mark the true halo positions, dashed when the halo sits behind the sources.

The fit treats the selected halo as isolated, with the same core and truncation radius as the truth, and uses annuli from 0.75 to 6 arcminutes. Neighbouring halos and a mass sheet are not in the fit model, which is why A hidden mass sheet fools it.

Thousands of slightly stretched galaxies weigh the mass in front of them

BackgroundLight from a distant galaxy bends as it passes mass on the way to us. Weak lensing stretches each image by a few percent, tangentially around the mass, while galaxies are intrinsically elliptical by about a quarter.

Why it mattersMost of the mass in galaxy clusters and in the cosmic web does not shine. Lensing responds to all mass, luminous or dark, so averaged galaxy shapes are one of the few direct ways to map it.

Focus question
If a second cluster sits behind the source galaxies, how much does it change their shapes?
One-sentence intuition
Shapes measure the reduced shear, and the distance ratio between lens and source sets how strongly each plane counts. The mass map is fixed only up to a uniform sheet.

Core mathematical model

Convergence and shear of a halo

κSIS(θ)=γSIS(θ)=θE2θ,θE=4πσv2c2DlsDs\kappa_{\rm SIS}(\theta)=\gamma_{\rm SIS}(\theta)=\frac{\theta_{\rm E}}{2\theta},\qquad \theta_{\rm E}=4\pi\frac{\sigma_v^2}{c^2}\frac{D_{\rm ls}}{D_{\rm s}}

For a singular isothermal sphere the convergence and the tangential shear are equal and fall off as one over radius. The Lab softens the centre with a core and cuts the halo off far out.

What a galaxy shape measures

g=γ1−κ,ϵ=ϵs+g1+g∗ϵs,⟨ϵ⟩≃gg=\frac{\gamma}{1-\kappa},\qquad \epsilon=\frac{\epsilon^{\rm s}+g}{1+g^{*}\epsilon^{\rm s}},\qquad \langle\epsilon\rangle\simeq g

Each observed ellipticity is the intrinsic one transformed by the reduced shear. Intrinsic orientations are random, so the average over many galaxies estimates the reduced shear.

Shear back to mass

κ^E+iκ^B=(ℓ12−ℓ22)−2i ℓ1ℓ2ℓ2 γ^(ℓ),κ→λκ+(1−λ) leaves g unchanged\hat\kappa_E+i\hat\kappa_B=\frac{(\ell_1^2-\ell_2^2)-2i\,\ell_1\ell_2}{\ell^2}\,\hat\gamma(\boldsymbol\ell),\qquad \kappa\to\lambda\kappa+(1-\lambda)\ \text{leaves}\ g\ \text{unchanged}

In Fourier space the shear is the convergence rotated by twice the wave-vector angle, so the inverse rotation recovers it. The imaginary part should vanish for lensing; the zero wave vector, a uniform sheet, is lost.

Common difficulties

Galaxies do not point at the mass

Typical misconceptionLensing stretches background galaxies along lines pointing toward the cluster.

Better mental modelImages are stretched tangentially, along circles around the mass. Turn on Shapes without scatter and look at the galaxies just outside a halo ring.

A flat map does not mean no mass

Typical misconceptionWhere the reconstructed map is near zero there is no mass.

Better mental modelThe inversion sets the mean of the field to zero, and a uniform sheet of mass changes no shape. Mass behind the sources is also invisible to them.

Run the experiment

  1. 01

    Take the scatter away

    Choose Shapes without scatter and compare the galaxies near halo A with those far from both halos.

    What to observe: Near the halo every galaxy is stretched along a circle around it; the fitted Einstein radius matches the true one, and the B-mode map is almost blank.
  2. 02

    Move the sources in front of a lens

    From Two lens planes, drag the source plane in the side view from redshift 1.2 down past halo B at redshift 0.6.

    What to observe: Halo B fades from the mass map and its photon kinks disappear: once the sources are in front of it, the distance ratio is zero.
  3. 03

    Hide a mass sheet

    Choose A hidden mass sheet and compare its maps and shapes with Two lens planes.

    What to observe: Shapes and the reconstructed map are identical, yet the halo's true Einstein radius dropped by a quarter and violet glow covers the field. Only the drawn sizes grow, because magnification does change.

Related experiments and articles

What this page links to directly

See every article series on the content map →

Weak gravitational lensing and mass maps · Reading path

From galaxy shapes to a map of mass

Articles on this path that use this experiment, followed by the other experiments they suggest.

  1. ArticleHow can galaxies that are barely stretched map dark matter?
  2. ArticleConvergence, shear and what a galaxy shape measures
  3. ArticleTurning a shear field into a mass map
  4. ArticleShear correlation functions and redshift tomography
  5. ArticleRecompute shear correlations and tomography from a simulation
  6. ArticleHow to read the DES Year 3 lensing and clustering analysis
Something broken, confusing, or missing on this page?Tell us
What kind of feedback?

Sent with this page’s address, its Lab ID, the site language and your device type. Please leave out personal details. Feedback is kept for up to 12 months.Privacy policy