Weak gravitational lensing is a powerful technique to measure the large-scale distribution of gravitating matter in the universe (both dark and luminous) from careful measurement of the distorting effects of the light from distant galaxies passing intervening mass distributions. This approach has become a key focus in observational study of dark matter, and with upcoming new surveys of the sky (e.g. the University of Hawai'i PanSTARRS project and the Canada-France-Hawai'i Telescope legacy Survey) it will be possible to make measurements of cosmological parameters from a direct probe of the dark matter component of the universe.
My current interest is in the methods we use to make the shape measurements and redshift (cosmological distance) measurements of galaxies. The distortion effect on an individual galaxy is extremely small (percent, or less) and because we cannot know the intrinsic shapes of individual galaxies, only their statistical distribution, the signal can only be measured by averaging over large numbers of galaxies. Such a measurement must be free from systematic error to very high accuracy in order to avoid biasing the cosmological measurements, and at Oxford Tom Kitching and I, working with former graduate student Catherine Heymans, Ludo van Waerbeke and Alan Heavens, have developed a new Bayesian method of measuring galaxy shape in the context of faint weak-lensing surveys. With former graduate student Edd Edmondson we have also applied Bayesian methodology to the problem of measuring redshifts when only limited information is available.
See also: lensfit - galaxy shape measurement for weak lensing surveys