The Systematic Biases In The Weak Lensing Measurement

The matter field,including the dark matter and baryon,has experienced the process of structure growth and evolution to form the structure observed at present since the very beginning of the universe.The light ray emitted from the distant galaxy is found to be deflected by the gravitational field of large structure it passed through.The corresponding shape of the distant galaxy is observed to be distorted slightly.The phenomenon is so‐called weak gravitational lensing,and the distor‐tion of the galactic shape is called cosmic shear.The distribution of the matter filed on the large scale and its history of evolution can be inferred by the statistical anal‐ysis on the shear based on the precise shape measurement.However,limited by the observational conditions,instruments,the image processing pipeline,and the shape measurement,we still need to put a lot of effort to control or even eliminate the systematic biases which,however,are currently corrected by the specific image simulation.Based on the Fourier＿Quad method and its image processing pipeline,we investigate the selection effect in shear(Chapter IV),the noise‐source coupling effect in the Fourier＿Quad method(Chapter V),and the systematic bias from the residual of point spread function(PSF)deconvolution(Chapter VI).We introduce the backgrounds in Chapter I to Chapter III and summarize the thesis in Chapter VII.We discuss the selection effects in Chapter IV.The precise shear recovery is based on the assumption that the intrinsic ellipticities of the galaxies are isotropic.The selection effects are easily caused in the stage objects detection and selection by the imposition ways which correlate to galactic shape.Therefore,the basic as‐sumption is violated due to that different galaxies have different probability to entry the sample.In Chapter IV,based on realistic image simulations and the feature of the Fourier＿Quad method that it gives a unbiased shear estimate,we test the corre‐lation between the common selection criteria and galactic shape and find the biases caused by the selection.We propose a new selection criterion to mitigate the selec‐tion effect in shear measurement.In Chapter V,we test the precision of the PDF＿SYM statistical approach in the Fourier＿Quad method and find a small amount of systematic biases which depends on the SNR and the ellipticity of PSF.A further investigation on the image compo‐nents and the algorithm reveals that the biases result from the coupling between the background noise image and the source image in the Fourier transformation.We propose to correct the bias by introducing a new noise image and rotating the PSF image.The biases are corrected to the level required by the next generation of weak lensing image surveys.We extend the lower SNR bound to lower than 10 for the Fourier＿Quad method.The correction makes it possible to make full use of the faint galaxies which account for a large portion of the real data.We investigate the effect of the residual from PSF deconvolution on the shear‐shear correlation function and the constraints on the cosmological parameters.Due to the PSF effect,the shape of the galaxy is diluted and imposed additional features from the PSF profile.The size,shape,and spectra energy distribution of PSF are supposed to be coherent and change mildly on the focal plane.Therefore,the PSF can be interpolated to the position of the galaxy by using the stars around.Due to the algorithm of star‐galaxy separation and the interpolation,it is inevitable to avoid the PSF residuals which bias the measurement of the correlation function and further cosmological parameter inference from it.We mitigate this kind of bias by using the galaxy pairs from different exposures based on that the residuals on different exposures are uncorrelated.