A wide-field WFCAM survey would advance our understanding of star formation in three main areas: (1) the star formation process; (2) the distribution of star formation regions (SFRs) throughout the galaxy and spatial variation in their properties; (3) the shape of the IMF. Some of this science overlaps with the Galactic Clusters Survey, which is targeted toward nearby star clusters outside the galactic plane.
The star formation process. A consensus is beginning to develop concerning the main evolutionary stages of low and high mass star formation but the picture becomes very uncertain for the youngest objects. These are initially visible only at far-IR and radio wavelengths, but later become bright near-IR sources as the birth cloud dissipates. High mass YSOs swiftly settle on to the Main Sequence and blow away their circumstellar material. Their presence in a cluster can profoundly influence the evolution of lower mass YSOs by triggering more star formation or by photoionising the envelopes of low mass systems which are still in the process of forming. This is seen in Orion and M16. Hence there are probably many modes of star formation and it is desirable to obtain high quality images of a large number of SFRs to see all the various modes at as many different stages as possible. The embedded, rapidly accreting phases of star formation are the least understood. We estimate that the GPS will detect 100-500 embedded high mass YSOs (8-100 M(sun)), at least 104 YSOs of solar mass and perhaps 105 YSOs integrated over all masses (see Technical requirements). This is an order of magnitude increase on the known populations.
UKIDSS can answer the following key questions by surveying several hundred SFRs:
Questions (1) and (3) can be addressed by JHK imaging. The evolutionary stages will be characterised by luminosity, extinction, K band excesses, circumstellar nebulae and variability. With 104 low mass sources the sample could be divided into e.g. ten mass and ten age bins yielding a statistically reliable sample of 100 sources per bin. The existing mid-IR galactic plane survey of the Midcourse Space Experiment, MSX (Egan & Price 1996) will be very useful for detecting and quantifying weak IR excesses, at least in nearer clusters which were spatially resolved. A key bonus of a very large survey will be the detection of 10 to 100 FU Ori systems during the onset of an outburst. FU Ori systems are YSOs in the rare state of very rapid accretion and associated outflow activity, with a duty cycle of ~1% of the pre-main sequence period. This unsteady accretion behaviour is believed by some leading researchers (e.g. Hartmann & Kenyon 1996) to be ubiquitous among YSOs at some level, although the level of contrast between high and low accretion states may have some variation. FU Ori stars are young, at least partially embedded systems, which undergo a large increase in brightness (6 magnitudes for FU Ori) with spectral signatures of strong accretion. This bright state persists for years or decades. Only eight such systems are known, none of which have undergone an outburst recently enough for the onset to be observed with modern IR instrumention. The rise in flux would easily be detected with a few epochs of data over a period of ~5 years. The pre- and post-outburst fluxes will quantify the frequency and magnitude of these events. A multipass survey is also essential to identify YSOs via variability in T Tauri stars and Herbig Ae/Be stars. This will tend to identify the older, optically visible YSOs which lack clear IR excesses. Only ~1/3 of YSOs have detectable K band excesses (Kaas 1999) and these cannot be measured if the source is obscured at J band.
A handful of parsec-scale ''macrojets'' have very recently been discovered in nearby SFRs, subtending significant fractions of a degree on the sky and with dynamical ages implying a very early origin. A 2.12 micron H2 survey will detect a statistically useful number of macrojets, measure their proper motions and dynamical ages. The assembled statistics will answer Question (2). H2-K difference maps will distinguish unresolved knots from stars and reflection nebulae from line emission nebulae. Crucially, the maps will detect the youngest, most embedded (Class 0) systems with no apparent source via stellar jets, providing key data for Question (1). Knots and working surfaces in the jets will also trace the ejection/accretion history of YSOs over a period of thousands of years and the interaction between jets and the interstellar medium (ISM). The proper motion information will also illustrate the dynamics of each jet. Do macro-jets accelerate, decelerate, or exhibit periodic velocity variations? Such information is important for jet models and for interpreting the influence the large-scale ISM has on macro-jets, but requires a statistically significant sample of measurements. The detection of more FU Ori systems will shed light on the relationship between steady/unsteady accretion and stellar jets, i.e. Question (3). An H2 survey would become time consuming and ineffective at large distances (kpc) so we propose a complete JHK+H2 survey of the Taurus-Auriga-Perseus molecular cloud complex, since this is the nearest SFR in the northern hemisphere. The region lies on the galactic plane but extends some way below it. The region has an area of 300 sq. degs, defined by the presence of molecular gas, e.g. Ungerechts & Thaddeus (1986). It includes all the well known dense Lynds subclouds, the lower density gas between them and the nearby high mass IC348 star forming region. This will provide an unbiased survey of outflows in SFRs, probing Class 0, Class I and Class II systems. Since this region will also be observed at JHK in the Galactic Clusters survey, there is no need for narrow band continuum observations. Radio continuum follow-up with SCUBA/SCUBA-2 and CO observations with HARP at the JCMT are expect to begin in the 2004-2005 time frame as soon as the first scan is complete. We would expect to detect several dozen macrojets in this complex but the number cannot be well quantified. The JHK component is included in the Galactic Clusters-IMF case. The survey would be followed up by observations of the continuum sources which power the jets using SCUBA/SCUBA-2 and by large scale CO(J=2-1) mapping of the molecular outflows with HARP on the JCMT.
Distribution of star formation and the IMF. The distribution of SFRs in the galaxy is at present inferred primarily from low resolution CS and IRAS mapping. We wish to learn how the density of SFRs varies with distance from the galactic centre and in different spiral arms and also study spatial variations in the IMF. At present there is evidence that star formation efficiency is higher at the solar circle than in the inner galaxy (Bronfman et al. 2000; Casassus et al. 2000) but this information depends upon the heating of dust and gas by young OB stars for IRAS detection of dense clouds emitting in the CS(J=2-1) line. Since low mass YSOs are brighter than their main sequence counterparts, UKIDSS will have the sensitivity to study the IMF down to brown dwarf masses in the nearer spiral arms and down to a few tenths of a solar mass throughout the galaxy (except along lines of sight which suffer very high extinction). These data will complement the more sensitive UKIDSS Cluster Survey by studying a much larger number of clusters in a wider range of environments, and will therefore measure how much environmental dependence exists in the IMF. The distances to clusters will be determined from the colours of the brightest members (see below), and independently by radio follow up. The MSX mid-IR satellite located several hundred SFRs, most of which will be accessible to WFCAM. The coadded IMFs of all the clusters will yield the first near-IR measurement of the galactic star formation rate, which is presently a highly uncertain quantity.