Santana-Silva, Luidhy
Brazil – CBPF
GALAXY MORPHOLOGY IN THE DECam LOCAL VOLUME EXPLORATION SURVEY AND THE SOUTHERN PHOTOMETRIC LOCAL UNIVERSE SURVEY
Understanding the formation and evolution of galaxies over cosmic time requires a comprehensive analysis of their morphological features. The increasing size of large sky surveys has led to large volumes of unclassified data which turns the morphological classification into two classes (late-type, early-type) substantially difficult. In this work, we employ Convolutional Neural Networks (CNNs) to build a morphological catalog from images obtained by two complementary surveys: the DECam Local Volume Exploration survey (DELVE) and the Southern Photometric Local Universe Survey (S-PLUS). The S-PLUS survey, which counts with a combination of 7 narrow band and 5 broad band filters, allows the recovery of photometric redshift with a precision of ~0.03, reaching a depth of r ~ 21 mag. On the other hand, the DELVE survey pushes the magnitude limit down to 23.5 in g-band with a resolution of 0.2636 arcsec, which permits recovery of the faint outer parts of galaxies, especially the presence of spiral arms and companion galaxies, increasing the accuracy of the morphological classifications. Given the large coverage of the DELVE survey, to ensure a reliable training sample, we have selected 400000 galaxies with known classifications from the previous study by Dominguez et al. in 2018 which allow us to define a robust training sample of 200k galaxies classified as early-type (ET) and late-type (LT) based on the TType parameter. Remarkably, even with the use of a simple model, our classification achieves an impressive accuracy rate of 95%. Using the S-PLUS DR3 we obtain a morphological classification into ET and LT for nearly 140000 objects up to r ~ 18 mag, as well as a classification of the reliability of the stamp, i.e. the probability that the stamp contains images of a galaxy. We also extracted the Sersic index (n) using the code MFMTK for the S-PLUS i-band images and we studied its relation with the morphological classification, for different magnitude bins. We recover important relations connecting galaxy properties and their morphological features, such as the morphology-density relation and the color-magnitude relation and compare with the literature. Additionally, our ongoing efforts aim to extend the morphological classification to 200 million galaxies observed in DELVE-DR3. We aspire to create a comprehensive morphological catalog covering 20000 square degrees, reaching magnitudes up to 22. We will combine these results with S-PLUS photo-z and galaxy mass and environment determinations. The completion and release of this catalog will not only facilitate our understanding of galaxy evolution but also serve as a valuable resource for the wider scientific community.