Previous Article Next Article Table of Contents
Gemini Operations: A Queue Scheduling Simulation (1Dec94) (from USGP, NOAO Newsletter No. 40, 1 December 1994) At the recent meetings of the US Gemini Science Advisory Committee and the Gemini Science Committee, the scientific case for non-traditional observing modes was explored. As input to this discussion, a simulation program has been developed, allowing an assessment of the relative advantages of queue scheduled observing and "classical" observing. Although the simulation requires a number of assumptions and simplifications, it demonstrates some of the gains that queue scheduling produces for certain types of observations. The Gemini model includes six instrument configurations, each with different requirements for environmental conditions (phase of moon, transparency, water vapor, and seeing). The table below lists these requirements. Note that 10 um imaging is limited by telescope emissivity, rather than the atmosphere. The Fiber Feed programs are spectroscopy carried out by connecting a small number of fibers from the Gemini/North focal plane to the CFHT high resolution spectrograph. The MOS-high resolution category signifies high spatial resolution spectroscopy for which the best seeing is required. The MOS-low resolution observations are background limited, but are not necessarily aimed at achieving the highest spatial resolution. Each program consists of 24 exposures requiring one hour each under median conditions. Each night is eight hours long. Each program is given a scientific grade, and the distribution of grades mimics the top 25% of a gaussian distribution between 1 and 5. The simulation runs through a semester twice with the same programs, first by assigning a randomly chosen program three nights at the telescope and then by selecting a program each night based on the current conditions. It is assumed that the environmental conditions are constant throughout an entire night. The transparency, water vapor, and seeing are drawn from distributions measured at Mauna Kea. In order to achieve some statistical accuracy, the simulation is run for 100 semesters. The chart below shows the average success of queue and classically scheduled observations as a function of proposal grade. Note that it is the percentage of observations completed which is shown. The average number of observations in each bin is printed above the chart. Water Program % of Time Moon Skies vapor Seeing 10 um Imaging 15 any photometric any not sensitive 5 um Spect. 10 any spectroscopic low not sensitive 2 um Imaging 25 dark spectroscopic any background or grey limited Fiber Feed 10 dark spectroscopic any not sensitive or grey MOS-high res. 20 dark spectroscopic any requires 20% best MOS-low res. 20 dark spectroscopic any background limited The results of the simulation for the total set of observations is as follows (Listed in the Classical and Queue columns are the number of observations completed/the number of possible observations and the corresponding percentage): Classical Queue Total Observations 619/1440 (43.0%) 972/1440 (67.5%) Programs Completed 13.2/60 (22.0%) 39.5/60 (65.8%) 10 um Imaging 106/224 (47.3%) 169/224 (75.5%) 5 um Spectroscopy 68/133 (51.0%) 130/133 (97.7%) 2 um Imaging 191/354 (54.0%) 272/354 (76.8%) Fiber Feed 76/150 (50.8%) 111/150 (73.9%) MOS-High Resolution 29/289 (10.2%) 83/289 (28.6%) MOS-Low Resolution 148/289 (51.2%) 207/289 (71.7%) [Figure not included] Todd Boroson
Previous Article Next Article Table of Contents