Technical requirements

Repeated observations. For much of the proposed GPS science, variability or proper motion is a valuable component. For the outflow work proposed above one would like to survey the same region three times over ~5 years. Such a repetition rate would be suitable for X/gamma-ray studies as well, although three passes is the bare minimum for identification of variables. Observing efficiency is reduced if the survey is split into many shallow passes, so we recommend three scans in the K filter spread over ~5 years as the best compromise, with more efficient single scans in the J and H filters.

Choice of filters. The J, H and K filters are essential for all programs since JHK imaging permits extinction correction and luminosity/distance derivation for stars whose intrinsic colours and distance are not precisely known (which will usually be the case) by dereddening in a two colour diagram. At least 2 filters are also needed to measure the temperature of stars and accretion disks.

A multipass survey in the narrow band H2 filter (2.12 micron) in a single nearby SFR is an important component of the Star Formation science case. The youngest, most embedded protostars will only be identified via detection of their molecular outflows and these outflows will also measure source age and accretion history.

UKIDSS could undertake Y band observations (see Large Area case) at ~1 micron, sampling the small peak in the spectral energy distribution of brown dwarfs near 1.05 micron and thereby providing a measurement of effective temperature. However this would not be very useful for any other galactic science outside the solar neighborhood owing to the relatively high extinction at 1 micron, so such observations should be conducted as a follow up study to determine source temperatures. The extinction of distant high luminosity main sequence stars will permit simple and reliable identification of brown dwarfs in the (J-H) versus (H-K) plane without Y band.

Numbers of YSOs. Using a Salpeter IMF with the currently accepted star formation rate of about 1 M(sun)/yr, we estimate that the galaxy contains about 2000 high mass YSOs (8-100 M(sun)) and 2 million low mass YSOs of solar mass and above, assuming pre-main sequence lifetimes of 104 yr for the high mass sources and 107 yr for the lower mass sources respectively. There are perhaps 108 YSOs of all masses, assuming there are 1011 stars in the galaxy with ages randomly distributed between zero and 1010 years. A UKIDSS GPS covering the northern half of the plane will detect 100-500 embedded high mass YSOs (assuming 50-90% are obscured by dust): an order of magnitude increase on the known population. Solar mass and above YSOs will be detected at distances of up to 5kpc to a limit of mK=18.8, depending on the amount of extinction, so the GPS will detect at least 104 of these and perhaps 105 YSOs in total. These numbers are order of magnitude extimates, which become highly uncertain at the low mass end.

Survey depth and area coverage. Large area coverage is more important scientifically than pushing right to the confusion limit: a substantial section of the plane is required to enable meaningful cross-correlation with the Astro-F and INTEGRAL surveys, and to create a useful atlas. However, obtaining the greatest practicable depth consistent with a large survey is important for all of the science cases described in Section 3.1, with the possible exception of AGB stars.

The basic logic behind our choice of area and depth is then as follows. We aim to survey to latitude ±5o, as this matches many other surveys, for example the recent MSX 4-20 micron survey, and includes most of the galactic disk at more than 2kpc distance. We then aim to survey as much longitude as possible, given a reasonable airmass restriction (Dec. > -15o) and the UKIRT declination limit (Dec. < 60o). This leads to two longitude regions defined by 15o< l < 107o and 142o< l < 230o, an area of 1800 sq. degs. Given this area, in a reasonable fraction of the total UKIDSS programme time we can afford to go approximately a factor of two longer than the standard minimum ''shallow'' depth.

Depth in K is more important than depth in J, because of reddening. It is the K-band that allows us to reach to the far side of the Galaxy. We also wish to have repeat observations, and choose to do these in K. Our final strategy is therefore to visit each position once at J and H, with twice the standard shallow depth, and three times at K, each of which being equal to the standard depth. This leads to survey limits J=20.0, H=19.1, and K=19.0.

Source confusion is potentially a limiting factor for the GPS. However UKIRT observations made with UFTI by A. Adamson have shown that confusion is not significant at K=19, at least for l > 40o. By this we mean that accurate photometry of >90% of point sources can be obtained - occasional confusion among sources at the detection limit is not important. There are complicated latitude and longitude dependent variations in confusion but these cannot be usefully quantified in advance of the survey. It will be important to test the confusion limit at 15o< l < 40o during summer 2001, in case it should be much brighter, but this is only a small part of the survey area.

The only additional requirement is the proposed H2 survey of the Taurus-Auriga-Perseus (T-A-P) region, with an area of ~300 sq. degs, defined by the presence of molecular gas. The JHK observations of this region are included in the Galactic Clusters Survey. The T-A-P region lies just below the plane, and in fact 18 sq. degs of it are inside the b=±5o band of the GPS, near l=165o and l=180o.

For the H2 survey an integration time of a 2 to 3 minutes is the minimum to detect typical outflows. The broadband K filter can be used for continuum subtraction. We aim to detect proper motions of the brighter knots, and so build up the survey in three separate passes.