Galaxy Zoo: Quantitative Visual Morphological Classifications for 48,000 galaxies from CANDELS (1610.03070v1)

Abstract: We present quantified visual morphologies of approximately 48,000 galaxies observed in three Hubble Space Telescope legacy fields by the Cosmic And Near-infrared Deep Extragalactic Legacy Survey (CANDELS) and classified by participants in the Galaxy Zoo project. 90% of galaxies have z < 3 and are observed in rest-frame optical wavelengths by CANDELS. Each galaxy received an average of 40 independent classifications, which we combine into detailed morphological information on galaxy features such as clumpiness, bar instabilities, spiral structure, and merger and tidal signatures. We apply a consensus-based classifier weighting method that preserves classifier independence while effectively down-weighting significantly outlying classifications. After analysing the effect of varying image depth on reported classifications, we also provide depth-corrected classifications which both preserve the information in the deepest observations and also enable the use of classifications at comparable depths across the full survey. Comparing the Galaxy Zoo classifications to previous classifications of the same galaxies shows very good agreement; for some applications the high number of independent classifications provided by Galaxy Zoo provides an advantage in selecting galaxies with a particular morphological profile, while in others the combination of Galaxy Zoo with other classifications is a more promising approach than using any one method alone. We combine the Galaxy Zoo classifications of "smooth" galaxies with parametric morphologies to select a sample of featureless disks at 1 < z < 3, which may represent a dynamically warmer progenitor population to the settled disk galaxies seen at later epochs.

Galaxy Zoo: Morphological Classifications of CANDELS Galaxies

The paper "Galaxy Zoo: Quantitative Visual Morphological Classifications for 48,000 galaxies from CANDELS" presents an extensive dataset generated through the collaborative efforts of the Galaxy Zoo project participants. The paper focuses on morphological classifications of approximately 48,000 galaxies observed by the Hubble Space Telescope (HST) as part of the Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS). The paper outlines both the methodology and findings derived from this substantial citizen science endeavor, which involved over 2,000,000 classifications contributed by over 95,000 volunteers.

Methodology

Galaxy Zoo utilizes a web-based platform to harness the collective capabilities of volunteers in classifying galaxies based on their visual morphology. The project employed a decision tree model to guide classifiers through a series of tasks, from identifying the broad appearance—such as "smooth," "featured or disk," or "star or artifact"—to more specific features such as clumpiness, spiral patterns, bars, and tidal signatures. Each galaxy received an average of 40 classifications, combining results through a consensus-based weighting method to ensure both reliability and independence in classifier votes.

To ensure the robustness of classifications, the paper describes how the team adjusted measurement parameters to account for survey depth variations and cross-validated with photometric and spectroscopic redshift data from complementary studies. This enabled correction for biases related to image depth and resolution, guaranteeing that classifications were internally consistent across diverse parameters.

Results

The paper revealed a diverse range of galaxy morphologies across multiple fields and redshift distributions, with notable findings including a considerable population of smooth disk galaxies at higher redshifts (1 ≤ z ≤ 3). These smooth disks are contrasted against featured disk galaxies prevalent at lower redshift ranges, suggesting substantial evolution in galaxy dynamics and morphology across cosmic epochs.

Implications and Future Directions

Quantitative classifications provide critical insights into galaxy structure and evolution, offering empirical data to evaluate the role of features such as clumps, spiral arms, and bar instabilities. The robust citizen science framework exemplified by Galaxy Zoo continues to drive forward the capacity to handle large volumes of astronomical data with nuanced human-driven pattern recognition far exceeding current automated capabilities.

This paper's findings concerning smooth disk galaxies at high redshifts suggest these galaxies may represent dynamically warmer progenitor populations. Such disks are less prone to instabilities like bar and spiral modes compared to their featured counterparts. This adds complexity to theories of galactic evolution, particularly regarding the transformation and settling of disk galaxies.

Future prospects include extending the Galaxy Zoo classifications to additional CANDELS fields, providing an even larger dataset for morphological exploration. Continued collaboration with professional astronomers and cosmologists will likely enhance analyses across sub-populations of the dataset, offering insights into merger frequency, disk formation, and overall cosmic structure evolution.

Overall, these morphological classifications serve as a valuable resource for ongoing research, providing a milestone in the paper of galaxy evolution through citizen-engaged science. The data, accessible and robust, afford avenues for further exploration in the interplay between observational astronomy and theoretical modeling. The project's adaptability to incorporate additional observational parameters presages its continued relevance and contribution to the astronomical community.