OCR Text |
Show 94 Iranga Samarasingha college of science In this research project, we analyzed high resolution spectra (R~10,000 - 80,000) to estimate the stellar parameters (effective temperature, metallicity, surface gravity) of the previously obtained low resolution spectra (R~2000, with a wavelength coverage of 385nm-920nm) taken for the Sloan Digital Sky Survey (SDSS) project, the Sloan Extension for Galactic Understanding and Exploration (SEGUE). These high reso-lution spectra have been gathered using the following observatories and instruments: Keck HIRES & ESI, Subaru HDS, HET HRS, and VLT UVES. Starting with the Subaru spectra, we computed the radial velocities for each star. In our initial method, we created a grid of 10,000 synthetic spectra varying the stellar parameters to interleave the pressure-broadened wings of H-β line of observed spectra between the grid points to find the ‘best fit' syntheses. The syntheses were generated by using the software, ‘SPECTRUM v2.76' (Gray & Corbally 1994). Then a chi-squared error analysis had been performed in order to acquire the best fit models which yield the best match, and thus the best estimates for stellar parameters. In this method, our parameter intervals were too large to obtain the solutions with desired accuracy. Also, this method consumed a large amount of time to execute the computer script. Subsequently, Monte Carlo analysis was performed in order to derive the stellar parameters more accu-rately and efficiently. In this Monte Carlo analysis, we synthesized stellar models using randomly selected stellar parameters and compared with the observed spectra. Then, a chi-squared error analysis was per-formed in order to evaluate the accuracy of the randomly picked parameters to estimate the best match-ing stellar parameters for each star. SPECTRUM is unable to model the core of the H-β line with an acceptable accuracy. Thus, in order to enhance the accuracy of the derived parameters, we decided to mask out the core of the H-β line in our comparison. Further we improved our computer script to determine the signal to noise ratio of each ob-served spectrum in order to factor it out in our chi-squared error analysis. The following figure presents the obtained results with an acceptable accuracy. Figure 1: The top panel shows the data (blue) and the best fit model (green and red) according to the Model Grid. The green line indicates the wavelength region (4850.0A to 4870.0A) which is considered for the chi-squared error comparison. The black line in the bottom panel illustrates the comparison more clearly and represents the residuals. The estimated chi-squared error value: 0.50908. Note: The core of the H-β line has been masked out. The comparison of the resulting stellar parameters between the high resolution and low resolution data will provide an insight on the accuracy of the methods used to derive the stellar atmospheric parameters from the SEGUE spectroscopic observations. These results can be applied to the significantly more distant stars, too dim to be observed via high resolution spectroscopy. References • Gray, R. O., Corbally, C. J. (1994) The Astronomical Journal, 107:742-746. • Lee, Y. S., et al. (2008) The Astronomical Journal, 136:2022-2049. BENCHMARKING SDSS-SEGUE ESTIMATED STELLAR PARAMETERS USING HIGH RESOLUTION SPECTRA OF SDSS STARS Iranga Samarasingha, (Inese I. Ivans, Tim Anderton) Department of Physics and Astronomy University of Utah UNDERGRADUATE RESEARCH ABSTRACTS Inese I. Ivans |