Prospects and pitfalls of gravitational lensing in large supernova surveys

Aims. To investigate the effect of gravitational lensing of supernovae in large ongoing surveys. Methods. We simulate the effect of gravitational lensing magnification on individual supernovae using observational data input from two large supernova surveys. To estimate the magnification due to matter in the foreground, we simulate galaxy catalogs and compute the magnification along individual lines of sight using the multiple lens plane algorithm. The dark matter haloes of the galaxies are modelled as gravitational lenses using singular isothermal sphere or Navarro-Frenk-White profiles. Scaling laws between luminosity and mass, provided by Faber-Jackson and Tully-Fisher relations, are used to estimate the masses of the haloes. Results. While our simulations show that the SDSSII supernova survey is marginally affected by gravitational lensing, we find that the effect will be measurable in the SNLS survey that probes higher redshifts. Our simulations show that the probability to measure a significant (3 sigma) correlation between the Hubble diagram residuals and the calculated lensing magnification is greater than or similar to 95% in the SNLS data. Moreover, with this data it should be possible to constrain the normalisation of the masses of the lensing galaxy haloes at the 1 sigma and 2 sigma confidence level with similar to 30% and similar to 60% accuracy, respectively.