| Summary: | Astronomical spectroscopy measures the physical properties of stellar objects using a spectrograph. The image captured by the spectrograph is subject to the dispersions and distortions. Data reduction is the process that transforms the original images from the spectrograph into final calibrated spectra. This process aims to remove or minimise the influence of the instrument, interference between neighbouring spectra, and other properties that can obfuscate the true spectra. With greater improvement of the data reduction the final spectra can be cleaner. A key component of the data reduction is the Point Spread Function (PSF) model. The PSF is the response of the instrument to a delta function of light. This thesis tested a PSF for the SAMI/AAOmega instrument on the AAT. In order to measure the improvements, several qualitative and quantitative metrics were defined. To begin with, several functions in many combinations were used to model the PSF of sparse data. It was found that the best fitting PSF was a combination of two Gaussians and two Lorentzians with offset means and combined linearly. The Gaussians were used to model the central peak while the Lorentzians accounted for large wings in the PSF. Once the optimal PSF had been selected, it was applied to calibration data. In all cases, the new PSF was found to be an improved model which reduced the systematic errors and residuals. Next the new PSF was used to reduce science data. In typical images and images with high dynamic range, the new PSF was shown to be an improvement; however, in one specific case it lead to significant errors. Several strategies were attempted in order to solve this problem, however none completely succeeded. Overall, a new PSF was found that is a better model of the SAMI/AAOmega's true PSF. While challenges exist that need to be overcome before it can be implemented in the data reduction, there are promising solutions that can be investigated.
|
|---|