SED Templates Calibration Based On Perturbation Algorithm

Distance is one of the most important parameter to study celestial objects.For those distant extragalactic objects,we can calculate their distance by measuring redshifts.The accuracy of redshift measurement is very important to cosmological study.Photometric redshift is an efficient way to obtain redshift.With the development of techniques,photometric redshift has a considerable high precision.Machine learning and SED template fitting are two widely used methods to estimate photometric redshift.The accuracy of photometric redshift derived by template fitting method is largely depend on how well the SED template applied fitting the catalogue samples.There are two types of SED templates,empirical templates based on observation and synthetic template derived by theoretical models.However,empirical templates may have troubles in representing the color of galaxies due to the limitation of observation samples which will cause influence on the accuracy of photometric redshift.CSST is a space based survey which is designed for high-accurate cosmological study.The telescope will be launched in high-galactic latitude region,providing optical imaging data of seven bands of NUV,u,g,r,i,z,y.The observation area is about 17,500 square degrees.It is expected to provide billions of photometric data and hundreds of millions of spectrums in ten years.In this paper,in the goal of improving the photometric redshift accuracy,we use optimized perturbation algorithm on the existing 4 different SED templates,by selecting 30 k galaxies with high signal to noise ratio in the CSST mock catalogue.It shows that the optimized templates fit the catalogue better.The accuracy of photometric redshift is improved as the fraction of outliers in the derived photometric redshifts is reduced by up to 64%,scatter up to 44%,and bias up to98%,when compared to estimates using a standard set of spectral templates.In addition,the perturbation algorithm can reconstruct the high resolution features of SEDs by using broad band photometric data.Also,by changing the color of SED template,the perturbation algorithm can partly solve the completeness problem of SED templates.