## 4. Intrinsically elongated sourcesIn the method as described so far, the average galaxy is analyzed. Very accurate shear measurements require to be the average of a large number of galaxies ( for a 1- shear accuracy of 0.01), otherwise intrinsic ellipticity scatter will dominate the measurement. However, constructing is only possible if the shear and the PSF the same for all galaxies that are summed. Because of spatial variations of PSF and shear, this is usually not so. We have therefore tested the algorithm in `galaxy-by-galaxy' mode, fitting individual galaxies as PSF-smeared, intrinsically circular, sheared sources. The resulting shear estimates are then each dominated by the intrinsic galaxy shape, as in KSB; the question is how well the average of such individually PSF-corrected shear estimates can represent the actual shear without bias. Note that galaxies are not generally well-described as constant-ellipticity sources, and hence there is no guarantee that, even in principle, the shear estimates will average to the correct value. We ran our tests on various double-gaussian model galaxies, of differing axis ratios. To simulate typical galaxies, we included a round, central `bulge' component, and an outer `disk' of axis ratio between 0.1 and 1. (Simulations with different bulge axis ratios yielded essentially the same results.) These were placed at all orientations, smeared with various kinds of elliptical PSF, and analysed with the algorithm described above. The best-fit values thus derived for each galaxy were then averaged to give an estimate of the shear. As may be seen in Fig. 9, the new algorithm performs very well, essentially correcting all PSF anisotropy signal in the measured shear. By comparison, the slightly biased answer returned by the KSB algorithm is apparent as before. Residual systematics of the new method are at the level of a few tenths of a percent.
© European Southern Observatory (ESO) 1999 Online publication: December 2, 1999 |