For a network designed to support research in time-domain astronomy, all observations are "target-of-opportunity" observations. Supernovae cannot be predicted; new asteroids are discovered and must be tracked; exoplanets are better observed during transits than at other times in their orbits. For this reason, LCO users can submit requests whenever it is appropriate to do so, for example: observations of known exoplanet transits can be requested months in advance, or observations of a newly-discovered object can be requested for "tonight." LCO's scheduling software, which produces an observing schedule that's optimized over the entire network, runs whenever a new observing request is received. The time for the software to produce a new schedule, from several thousand active requests, is typically less than 10 minutes, and the typical timescale of the lag between the moment an observing request is submitted and the moment the observation can be started at the telescope is less than 15 minutes.
For some research programs (e.g. gamma-ray burst follow-up), LCO's typical response time is not fast enough, and observations can be requested in rapid-response mode, instead. When an observing request is tagged as "rapid response," it circumvents the normal scheduling process and gets dispatched to the appropriate telescope immediately. The observation currently in-progress at that telescope is aborted. If a rapid-response observation cannot be executed immediately, it will be executed as soon as possible, up to a limit of 6 hours.
Tests of the rapid-response mode have demonstrated that the time between request submission and execution is sometimes as short as 2 minutes. The median rapid response time is 6 minutes. Because rapid-response observations disrupt normal network operation, the time made available in this mode is limited. PIs of research programs seeking an allocation of rapid-response observations must justify the need for this observing mode in their (TAC-reviewed) proposals.
Observations in time-critical mode do not disrupt normal scheduling. They are submitted to the scheduler with high priority, so that they have a good chance of getting scheduled at a specific time. Time-critical mode is reserved for observations that must be made at relatively tightly constrained times that rarely occur. Some examples are observations of predictable (but infrequent) transits, observations simultaneous with other observatories, and follow-up observations of rare events that evolve quickly. Like rapid-response mode, there is limited time available in time-critical mode, and requests for time-critical observations must be justified in proposals.