Galaxy Zoo for Astronomers

This page aims to give background to the scientific motivation behind the project for a more technical audience.

You may be looking for data - public Galaxy Zoo data is available in a range of formats here.

Alternatively, you can browse a list of papers published by the team here.

Why is galaxy morphology important?

To first order, the morphology of a galaxy is a tracer of the orbital dynamics of the stars in it, but it also carries an imprint of the processes driving star formation and nuclear activity in galaxies. Visual morphology produces classifications which are strongly correlated with other, physical parameters. To give a single example, the presence of multiple nuclei and extended tidal features indicates that the dominant mechanism driving star formation is an ongoing merger. Equally, the absence of such features implies that the evolution of the galaxy may be being driven by slower ('secular') processes.

Traditionally morphology has been derived either by visual inspection of galaxy images (e.g. Hubble 1926, de Vaucouleurs 1991, and more recently e.g. Nair and Abraham 2010 or via morphological parameters, such as Concentration, Asymmetry, Clumpiness, M20, the Gini coefficient, etc (e.g. Conselice 2003, Lotz et al. 2008). Strictly speaking, these parameters are morphological 'proxies', each with its own attendant biases, which are typically checked and calibrated against visual inspection. A visual approach is generally more resistant to changing signal-to-noise and resolution in images (e.g. Lisker 2008), making it an ideal method for determining galaxy morphology. Nevertheless, morphological parameters have been valuable for classifying large survey-scale datasets, for which visual inspection by individuals (or small groups of researchers) can be prohibitively time-consuming.

Classifications of Large Survey Datasets Using Galaxy Zoo

Galaxy Zoo (Lintott et al. 2008, 2011) pioneered a novel method for performing large-scale visual classifications of survey datasets. Using more than half a million members of the general public, the project has classified – via direct visual inspection - the entire Sloan Digital Sky Survey spectroscopic sample and all existing Hubble Space Telescope surveys (around 1.5 million galaxies in total). With more than 40 classifications per object, Galaxy Zoo provides both a visual classification and an associated uncertainty (which is challenging to estimate if there are only a few human classifiers). The classifications themselves have been demonstrated to be of comparable accuracy to those derived by expert astronomers (see Lintott et al. 2008).

Galaxy Zoo Science Highlights

The Galaxy Zoo science programme has contributed to a diverse set of topics, largely focused on the nearby and intermediate-redshift Universe. Some recent highlights include the largest studies of galaxy mergers (Darg et al. 2010), tidal dwarf galaxies (Kaviraj et al. 2012), dust lanes in early-type galaxies (Kaviraj et al. 2012) and bars in disc galaxies (Masters et al. 2011, 2012) in the nearby Universe to date. One of the unique aspects of Galaxy Zoo over automated morphological measurements is the possibility of serendipitous discoveries (often aided by volunteer led discussion on the Galaxy Zoo Forum). These have included the discovery of 'green peas' (a class of compact extremely star-forming galaxies in the local Universe; Cardamone et al. 2009) and perhaps most famously "Hanny's Voorwerp" (Lintott et al. 2009) along with a survey of similar AGN-ionised gas clouds (Keel et al. 2012). The availability of a large sample of galaxies with both color and morphological information has led to the important realisation that color, not morphology, is most strongly correlated with environment (Bamford et al. 2009; Skibba et al. 2009), leading to intriguing subclasses of galaxies like red spiral galaxies (Masters et al. 2010) and blue ellipticals (Schawinski et al. 2009).

The Many Phases of Galaxy Zoo

The current site has incorporated images from a variety of different telescopes, including SDSS, HST, UKIRT, VST and the CTIO 4-m Blanco telescope. The SDSS images are from DR8 and the Southern Galactic cap, which increases the sample size of galaxies in the local universe by 40%. The HST images are from CANDELS, the largest HST Treasury Program. This survey is designed to take advantage of the advent of Wide Field Camera 3 (WFC3), which is rapidly opening up a new window into galaxies at z > 1 – the first 50% of the lifetime of the Universe. Previous HST imaging at these epochs largely sampled the rest-frame ultraviolet, since the available survey instruments (e.g. ACS) operated in the optical wavelengths. However, the near-infrared capabilities of the WFC3 (a factor of 20 better than NICMOS) are providing us with unprecedented rest-frame optical data of galaxies at z > 1.

CANDELS is using the WFC3 near-infrared filters to image 800 arcmin2 in established HST legacy fields (e.g. GOODS, COSMOS). The near-infrared WFC3 images are particularly important because morphological analysis is best performed in rest-frame optical wavelengths, which trace the underlying stellar population of the galaxy rather than just the UV-bright star-forming regions.

The combination of the SDSS and CANDELS Galaxy Zoo samples will offer a formidable tool for answering significant open questions that demand a morphological analysis. For example, at what epochs was the Hubble sequence established? How and when were the primordial spheroids formed? What was the relative role of major mergers and secular processes in driving star formation and black hole growth in the early Universe?

The UKIRT images are from the Large Area Survey (LAS) as part of UKIDSS, which imaged 4000 sq. deg. of the sky overlapping with the SDSS fields. We selected all galaxies classified in Galaxy Zoo 2 that had high-quality UKIDSS imaging, totaling about 70,000 galaxies. The images shown on the site are a colour-composite of the Y, J, and K-band images (where the J-band uses dithered observations to improve the angular resolution in that channel). These images will allow us to trace morphology as a function of wavelength; for example, whether the observed bar fraction increases for galaxies in the rest-frame infrared.

Galaxy Zoo also includes a small set of SDSS images that have been artificially processed to simulate the observed effects of redshift. This set used a morphologically diverse sample of galaxies from Galaxy Zoo 2 and were processed using the FERENGI code (Barden, Jahnke, & Hausler 2008) out to redshifts of z=1. Results from these images were used as calibration for galaxies in the GZ: Hubble data, since it allowed us to assess resolution and brightness-dependent biases independent of true galaxy evolution.

In early 2015, three new sets of data were put into Galaxy Zoo. Two of them are from Hubble surveys for which we already have GZ morphological classifications --- GOODS and CANDELS. The new images explore how changing the depth of data (a surface brightness limit) affects the morphological measurements. The new CANDELS images are from 2-epoch data (which is shallower than the 5-epoch data already cataloged), and the GOODS images come from the full-depth images, supplementing the shallow observations already completed in GZH. Finally, we added roughly 1000 images each of galaxies from SDSS in their monochrome ugriz filters, which will allow a more careful measurement of morphology as a function of wavelength.

The Dark Energy Camera Legacy Survey (DECaLS) is a public imaging survey intended to supplement spectroscopic measurements being collected by various iterations of the SDSS. Using the Dark Energy Camera mounted on the 4-m Blanco telescope at the CTIO in Chile, the DECaLS team is imaging 6700 square degrees of the sky overlapping with the SDSS footprints, including Stripe 82. Images are taken in the g, r, and z bands and have significantly better angular resolution and point-source sensitivity than the SDSS images. DECaLS images combined with Galaxy Zoo morphologies will be used for several science goals, including measurement of the Hubble sequence at lower luminosities, detection of tidal tails and minor mergers, and serendipitous discovery of rare and unusual objects. DECaLS images were put into Galaxy Zoo beginning in mid-2015.

Galaxy Zoo is also working with state-of-the-art simulated images of galaxies in order to test the physical models that go into these simulations and assess whether the reproduced morphologies match those that are seen in the Universe. The Illustris simulation is a massive cosmological simulation that traces both dark matter and baryons via realistic physical models over evolving conditions down to the Universe at its present day. Galaxy Zoo is classifying simulated images of galaxies formed in the Illustris simulations, which appear as if they were observed with the SDSS telescope located at a distance of 223 Mpc (redshift z=0.05). The galaxies appear at a variety of angles and against several sets of real sky backgrounds, allowing for a direct comparison to the morphologies observed in previous projects such as Galaxy Zoo 2. Illustris data were put into Galaxy Zoo beginning in mid-2015.

Another new source of images is provided by the Galaxy And Mass Assembly Survey. GAMA combines multi-wavelength data from many ground and space-based survey facilities in order to study the structure of galaxies and dark matter on scales from thousands to millions of light years. In particular, GAMA has added valuable distance information for nearly quarter of a million galaxies via a large redshift survey on the AAT. This enables the environment of galaxies, the groups and large scale structure in which they live, to be very well characterised. GAMA has recently benefitted from images provided by the Kilo-Degree Survey (KiDS) on ESO's VST. These dramatically improve on SDSS, the previous source of GAMA's optical imaging. The better resolution and depth will allow us to study fainter structures in many more galaxies, helping us to further understand the links between environment and morphological transformations. GAMA KiDS images were first added into Galaxy Zoo at the end of 2016.

The purpose of the Galaxy Zoo project is to answer a variety of scientific questions, to prepare the ground for morphological work using future instruments like the JWST, and to produce samples of morphologically selected high-redshift galaxies for follow-up using instruments like the extremely large telescopes and ALMA.