The Legacy ExtraGalactic UV Survey (LEGUS)

Daniela Calzetti,, and the LEGUS Team,

The combination of UV capability, high-angular resolution, and large field of view afforded by the Hubble Space Telescope is the foundation of the Legacy ExtraGalactic UV Survey (LEGUS), GO-13364. LEGUS, a Cycle 21 Hubble Treasury program, was designed with the main goal of providing a definite characterization of the links between star formations on two fundamental scales: those of individual stars, stellar clusters, and associations on parsec scales; and of galaxy disks on kiloparsec scales (Calzetti et al. 2015b).

In order to achieve its science goal, LEGUS has obtained multi-color images of 50 nearby star-forming galaxies, in the distance range ∼3–16 Mpc. Wavelength coverage spans five bands (NUV, U, B, V, and I) by combining new WFC3 observations with archival ACS imaging data, when available. The galaxies were carefully selected to sample the full range of galaxy mass, morphology, star-formation rate (SFR), sSFR (specific SFR = SFR/mass), metallicity, internal structure (rings, bars), and interaction state found in the Local Volume where Hubble can resolve and age-date young stellar populations on parsec scales.

Many of the galaxies are well-known, iconic ones, with a wealth of additional information available in the MAST archive. The multi-color images (Figure 1) are used to secure complete inventories of the young stars, star clusters, and structures of the galaxies, together with the characterization of their ages, masses, and extinctions. For this, the ultraviolet band provides critical leverage, as it enables breaking the degeneracy between age and extinction. The goal of a complementary Cycle 22 Hubble program (GO-13773) is to obtain narrow-band imaging in the light of the Hα emission line for many of the LEGUS galaxies in order to trace the regions of recent massive star formation.

Figure 1: Details of the central regions of two LEGUS galaxies, NGC 1313 (4.39 Mpc distance, left) and NGC 5253 (3.15 Mpc distance, right). The images are color-composites using the NUV+U (blue), B+V (green), and I (red) bands. For NGC 1313, the image shows one of the spiral arms departing from the nucleus (located at the bottom-right of the image), while for the dwarf NGC 5253, the entire star-forming central region is shown. The diffuse green light close to the center of NGC 5253 is [O III] forbidden line emission emerging in the V band. The images highlight the complexity of the stellar structures in these galaxies, especially those containing young (blue) stars.

Among the specific science objectives of the project, we list: quantification of the evolution both in space and time of the clustering of star formation; discrimination among models of star cluster formation and evolution; determination of the recent formation histories of stars and clusters and their impact on the UV-star-formation-rate calibrations. We provide here a brief summary of some of the initial results.

Most stars are formed in clustered structures forming a continuous, scale-free hierarchy from parsecs to kiloparsecs (Lada & Lada 2003; Elmegreen 2003; Bressert et al. 2010). These structures are expected to arise from the self-similar distribution of a turbulence-dominated interstellar medium (Elmegreen & Efremov 1997), mediated by magnetic fields and outflow feedback (e.g., Krumholz et al. 2014). The densest peaks of the hierarchy likely survive disruption by a host of both internal and external mechanisms, and evolve as gravitationally bound star clusters.

Building upon earlier results, the LEGUS collaboration has been characterizing the properties of these structures using both stars and star clusters as tracers, and a number of techniques, including image smoothing and two-point correlation functions. The latter measures the degree of clustering of the observed data relative to a random distribution. Using ultraviolet images, which trace young star-forming structures, Elmegreen et al. (2014) found that while hierarchically distributed structures from several up to a few hundred parsecs are common in galaxies, starburst galaxies show a higher fraction of projected areas filled with star formation relative to more quiescent galaxies. The distribution of young stars is in self-similar, fractal structures from ∼20 pc to ∼2.5 kpc in the ring galaxy NGC 6503, but they diffuse across the galaxy within ∼60 Myr (Gouliermis et al. 2015). The star-cluster distribution is also within self-similar structures, with a correlation length of about 300 pc (Figure 2; Grasha et al. 2015, 2016).

When star clusters are divided according to their morphological characteristics, the compact clusters are significantly less clustered than the ‘multi-peak’ clusters (Figure 2). The three morphological categories of cluster candidates used by the LEGUS collaboration are likely to reflect physical differences as well: the compact (class 1) sources may be relaxed, massive clusters, the ‘elongated’ (class 2) sources are unrelaxed clusters, while the ‘multi-peak’ (class 3) sources are likely stellar associations. The associations, furthermore, disperse over a ∼50 Myr timescale (Adamo et al. 2016), in agreement with the clustering analysis which suggests that clusters randomize within ∼40–50 Myr (Grasha et al. 2016).

Figure 2: (top left) The cluster classification scheme adopted by the LEGUS collaboration, shown on three-color composite (U,V,I) image cut-outs (Adamo et al. 2016): Class 1—symmetric, compact cluster; Class 2—concentrated object with some degree of asymmetry; Class 3—multiple-peak system; Class 4—spurious detection (background galaxies, bright stars with extended halos, asterisms, artifacts). (bottom left) Class 1, 2, 3 (red, green, blue) cluster candidates are shown as circles, on a grayscale image of the LEGUS galaxy NGC 3344. The segment marked ‘r’ at the bottom-left of the image connects two clusters of the same class; pairs of clusters all sharing the same separation ‘r’ (in pc) are counted within each galaxy to construct the two-point correlation function. (top right) The two-point correlation function, obtained by combining cluster pairs from six different LEGUS galaxies, is shown for Class 1, 2, and 3 clusters separately, together with the cumulative function. More strongly clustered distributions display a sharper rise towards small separations. A random distribution would show as a horizontal line at log[1+ω(θ)] = 0 (Grasha et al. 2016).

Because of the wide range of galactic environments and the broad wavelength coverage it probes, LEGUS enables a host of additional science, including, but not limited to:

  • studies of supernova progenitors and their environments (Van Dyk et al. 2015);
  • investigations of the role and fate of star clusters in relation to the natal environment (Calzetti et al. 2015a);
  • analyses of cluster spatial distribution as discriminators for models of spiral galaxies (Dobbs et al. 2016);
  • tests of stellar-population models accounting for a range of astrophysical inputs, and analysis of impact on observed cluster properties (Krumholz et al. 2015; Wofford et al. 2016); and
  • studies of the physics of massive stars (Smith et al. 2016).

The first high-level data product delivery of LEGUS imaging data was performed in September 2015, and the multi-wavelength, aligned images are available at:


Adamo, A., et al. 2016, in prep.

Bressert, E., Bastian, N., Gutermuth, R., et al. 2010, MNRAS, 409, L54

Calzetti, D., Johnson, K. E., Adamo, A., et al. 2015a, ApJ, 811, 75

Calzetti, D., Lee, J. C., Sabbi, E., et al. 2015b, AJ, 149, 51

Dobbs, C., et al. 2016, in prep.

Elmegreen, B. G. 2003, in Dynamics and Evolution of Dense Stellar Systems, 25th meeting of the IAU, Joint Discussion 11, p. 34

Elmegreen, B. G., & Efremov, Y. N., 1997, ApJ, 480, 235

Elmegreen, D. M., Elmegreen, B. G., Adamo, A., et al. 2014, ApJ, 787, L15

Gouliermis, D. A., Thilker, D., Elmegreen, B. G., et al. 2015, MNRAS, 452, 3508

Grasha, K., Calzetti, D., Adamo, A., et al. 2015, ApJ, 815, 93

Grasha, K., et al. 2016, in prep.

Krumholz, M. R., Bate, M. R., Arce, H. G., et al. 2014, in Protostars and Planets VI, H. Beuther, R. S. Klessen, C. P. Dullemond, & T. Henning (eds.), University of Arizona Press, Tucson, p. 243

Krumholz, M. R., Adamo, A., Fumagalli, M., et al. 2015, ApJ, 812, 147

Lada, C. J., & Lada, E. A. 2003, ARAA, 41, 57

Smith, L. J., Crowther, P. A., Calzetti, D., & Sidoli, F. 2016, ApJ, in press (arXiv:1603.06974)

Van Dyk, S. D., Lee, J. C., Anderson, J., et al. 2015, ApJ, 806, 195

Wofford, A., Charlot, S., Bruzual, G., et al. 2016, MNRAS, 457, 4296