Exploring the Transient Optical Sky
Targets of Opportunity
TOO-1: Externally triggered observations to consider for LSST schedule interrupts include at least 4 categories of serendipitous time-variable events for which the broad-band coverage and high sensitivity and possible cadence of LSST will provide unique new capabilities. These will be primarily in response to space-based observations (for the science examples given here) but may include high priority triggers from other ground-based observatories (e.g. LIGO2 or other telescopes).
GRBs: The most obvious example are GRBs. While LSST need not (and surely can not) respond to all GRBs detected by spacecraft in the 2013 and beyond timeframe, likely new GRB capabilities (e.g. EXIST: but probably not until 2016 at earliest) will allow "smart" decisions for GRB prioritization. The GRB Epeak vs. Luminosity will allow X-ray/gamma-ray "photo-z" measurements in (near) real time, with the broad band coverage that EXIST would provide so that candidate high z (>5) would be given priority. Even with no pre-selection, GRB TOOs would not be onerous for LSST: with ~4000 sq.deg/night coverage and a full-sky GRB rate currently estimated as ~3/day/full sky, the overlap with LSST is only ~0.3/day and the total time required per TOO need only be ~10min (max; see below for Requirements). The primary science goal is to identify highest z GRBs to enable GRBs to be used as probes of cosmology and dark energy (e.g. see Lamb et al, astro-ph/0507362) and to identify, as the ultimate goal, Pop III stars as the presumed massive progenitors of the very first stellar BHs. LSST prompt imaging would enable prompt spectroscopy for redshift measurements on "real time" high priority candidates: those detected in V,R,I,Z,Y at &ge 20 within 10min of a GRB trigger and thus most likely to be high z. A crude photo-z redshift can be estimated from broad-band colors and magnitudes (Swift will have calibrated the prompt GRB to afterglow transition magnitudes relative to X-ray/gamm-ray flux ratio) and separated (partly) from unknown reddening by color-color analysis.
Fast Transients (AGN, galactic BHs, and novae): EXIST will open a new window on the variable high energy universe, generally. "Fast" transients on a range of timescales (relevant to the objects) will be discovered and could be prioritized for prompt LSST followup as TOO-1's on corresponding timescales. Each would require U,V,R,I,Z,Y imaging on cadences given below.
a)Tidal disruption of stars by dormant massive black holes in galactic nuclei are perhaps the most exciting and should be detectable by EXIST at rate ~10-30/year (Grindlay 2004). The primary science goal enabled by study of these objects is the BH content of "normal" galaxies (i.e. non-AGN) and the full understanding of the BH-Bulge mass relation and thus BH vs. galaxy formation. These would detected in EXIST data on ~1day - 1month timescales (depending on distance; out to ~200 Mpc) and so have corresponding range of followup timescales for the ~10/year accessible to LSST.
b)Galactic BHs (stellar mass) and neutron stars in low mass X-ray binaries (LMXBs) undergo ~0.5-10d outbursts due to disk instabilities. The incidence rate of short duration transients is very poorly known but may dominate; EXIST would locate these objects (to ~10-50arcsec) but "rapid" followup photometry (particularly I,Z,Y) is needed for identification. The primary science enabled is the stellar mass BH content of the Galaxy. A related objective is the incidence of 511 keV annihilation flares from accreting BHs, such as probably detected (~10h) in one BH system ("nova" Muscae) in 1991 by SIGMA/Granat.
c)Novae (nuclear burning) on white dwarfs also likely produce prompt 511 keV emission with decay times ~3h that would be detectable by EXIST and allow novae in obscured regions to be discovered and the Nova rate of the Galaxy measured. Near-IR coverage is thus most important.
TOO-2: Internally triggered TOOs are desirable for a variety of science objectives. We list only two, as examples.
Orphan GRBs: Off-axis or otherwise, gamma-ray dark GRBs are very likely, given current understanding of jets, in GRBs. Their detection and identification would allow the full understanding of GRB jets and source populations, which in turn may prove necessary for using GRBs as tracers of star formation rates out to large z. (Nakar et al 2002) have estimated that these will be optimally detected (at highest rates) at V,R ~20-22. Thus they are uniquely well suited for LSST. The "internal trigger" requirements and cadence are given below.
a) Normalized redshift distribution of observed orphan afterglows in a single snapshot (thick lines) and the integrated z distribution (thin lines). (b) Angular distribution of observed orphan afterglows in a single snapshot (thick lines) and the integrated &thetaobs distribution (thin lines). The circles depict &thetamax(zpeak, m) and the squares depict (&thetaobs). In both panels, we use F0 = 0.003 &muJy and &thetaj = 0.1. The different curves correspond to mlim = 25, zpeak = 1 (solid line); mlim = 25, zpeak = 2 (dashed line); mlim = 27, zpeak = 1 (dot-dashed line); and mlim = 27, zpeak = 2 (dotted line).
Number of observed orphan afterglows per square degree (left vertical scale) and in the entire sky (right vertical scale), in a single exposure, as a function of the limiting magnitude for detection. The thick lines are for model A with three different sets of parameters: (1) Our "canonical" normalization F0 = 0.003 &muJy, zpeak = 1, and &thetaj = 0.1 (solid line). The gray area around this line corresponds an uncertainty by a factor of 5 in this normalization. (2) Our most optimistic model with a relatively small &thetaj = 0.05 and a large F0 = 0.015 &muJy (dot-dashed line). (3) The same as our "canonical" model, except for zpeak = 2 (dotted line). The thin lines are for model B, where the solid line is for our "optimistic" parameters, while the dashed line is for our "canonical" parameters. Both models are similar for the "optimistic" parameters, while model B predicts slightly more orphan afterglows than model A for the "canonical" parameters.
Ratio of double detection of orphan afterglows to a single detection, as a function of the time delay dTobs between the two exposures. The three models are F0 = 0.003 &muJy, zpeak = 1, and mlim = 23 (dashed line), 25 (solid line), and 27 (dot-dashed line).
Targetted BH Transients Surveys: Related to TOO-1/Fast Transients-b objective above are targeted surveys for stellar mass BHs as transients in nearby local group galaxies. As LSST scans M31, M33, etc., an internal trigger on new optical/near-IR transients would trigger repeat short observations (cadence given below) for identification and study. The prime objective, again, is the stellar mass BH content of galaxies, but without the extinction and distance ambiguities of the Galactic transients.
Observational Requirements and Cadences
The TOO science programs outlined here do not require special psf or astrometric constraints; rather processing constraints. Some programs would (naturally) have seeing (psf, and airmass) constraints to execute; these would be specific to a given program.
TOO-1: The key requirement is the ability for prompt (<1min) interrupt of LSST scheduling and rapid-slew re-pointing. Given the disruption this entails, this could be reserved for only the highest priority events but it should be provided. Total time interruption for any of the TOO-1 objectives on a given night are likely to be <10-20min, with total followup rates of ~1 per 3-10days.
GRBs: The key requirement is broad-band rapid coverage: shallow (AB mag ~20) in ALL filters as rapidly as possible, followed by longer exposures in all filters (or, to achieve minimum spacing, in V, I, Z only) to AB mag ~26 if initial exposures are "blank". Real time processing is thus required on timescales of <1-3min. Total time required for any given "prompt" GRB: <20min. Followup time on succeeding nights for highest priority (e.g. high z) afterglows: ~0.5h per night, perhaps 3 observations logarithmically spaced for decay.
Fast transients: The timescales are much longer, but still require interruption of a ~1-3d schedule. Ideally schedules can be changed within ~1h to accommodate followup of these events for BH novae and nuclear novae (511 keV triggers), and ~1-10d timescales are allowed (thus no "interruption") for tidal disruption events.
TOO-2: The key requirement here is rapid, realtime processing of images. Ideally each frame is processed with very fast (on the fly) image subtraction to look for rapid events. A "real time" (i.e. within 30sec of image readout) event is processed on the fly (image subtraction from library reference frame, approximately seeing-convolved for current psf) for that filter. TOO-2 triggers interrupt subsequent programming then only for well defined thresholds for given programs (e.g. GRB Orphan Afterglows for "new" object on at least two successive exposures in each filter, etc.). Similarly for Targetted Transients Survey. Cadences for both Afterglows and Transients would be ~3 observations per night (of trigger) and ~3-10 followup nights (for highest priority transients).