2014MNRAS.438.2732D

Identifying galaxy clustering at high redshift (i.e. z > 1) is essential to our understanding of the current cosmological model. However, at increasing redshift, clusters evolve considerably in star formation activity and so are less likely to be identified using the widely used red-sequence method. Here we assess the viability of instead identifying high-redshift clustering using actively star-forming galaxies (submillimetre galaxies, SMGs, associated with overdensities of BzKs/LBGs). We perform both a 2D and 3D clustering analysis to determine whether or not true (3D) clustering can be identified where only 2D data are available. As expected, we find that 2D clustering signals are weak at best and inferred results are method dependent. In our 3D analysis, we identify 12 SMGs associated with an overdensity of galaxies coincident both spatially and in redshift - just 8 per cent of SMGs with known redshifts in our sample. Where an SMG in our target fields lacks a known redshift, their sight line is no more likely to display clustering than blank sky fields; prior redshift information for the SMG is required to identify a true clustering signal. We find that the strength of clustering in the volume around typical SMGs, while identifiable, is not exceptional. However, we identify a small number of highly clustered regions, all associated with an SMG. The most notable of these, surrounding LESS J033336.8-274401, potentially contains an SMG, a quasi stellar object (QSO) and 36 star-forming galaxies (a >20σ overdensity) all at z ∼ 1.8. This region is highly likely to represent an actively star-forming cluster and illustrates the success of using star-forming galaxies to select sites of early clustering. Given the increasing number of deep fields with large volumes of spectroscopy, or high quality and reliable photometric redshifts, this opens a new avenue for cluster identification in the young Universe.