...
The Large Synoptic Survey Telescope (LSST) is a wide-field 8m telescope that will survey the southern sky over a period of 10 years. Each LSST observation (a visit) has a duration of 34 seconds consisting of 2 back-to-back 15-second exposures re-pointing to the next sky position. The LSST will implement a "scheduler" that will optimize the observing cadence against science priorities and local observing conditions. To this end the LSST project has developed an operations simulator (OpSIm) that models the temporal sequencing of visits given parameters of science proposals, constraints of hardware performance and historically based observing conditions (seeing, sky brightenss brightness and weather).
The current inputs used in the LSST OpSim for the site characteristics, seeing and weather, are based on measured data from the LSST site on Cerro Pachon and CTIO. The OpSim seeing model was based on DIMM measurements made on the LSST site from 1998 - 2006. The DIMM data were scaled for an outerscale of 30m and an 8.4m aperture following Tokovinnin et al. (199?) to form a prediction of the expected atmospheric delivered image quality that would be seen by LSST.
(add short description of current sky brightness model)
At the time of the initial OpSim development there were not sufficient data from the LSST site to develop a weather model (cloud coverage vs time). Instead we developed a weather the OpSim weather model from CTIO night logs taken from 1975 - 2005. These night logs recorded the observations of the on-site telescope operator and their assessment of net cloud cover on a scale of 0 (no clouds) to 8 (completely opaque) every 3 hours (typically 4 times per night). These logs contain no information about the spatial structure of the observed clouds and are by their nature qualitative. The current OpSim treats the cloud cover as uniform over the whole sky and derived extinction from scaling the cloud cover fraction.
The current input models However, missing from these inputs is a model for the sky brightness, transparency and cloud cover with lack the needed spatial and temporal resolution that is suitable for detailed algorithm optimization needed for the LSST scheduler.
...