A Bottom-Up and Top-Down Approach to Cloud Detection

Weather monitoring has always been an element of observatory operations. For a robotic telescope there is the added complication that software needs to understand the ever changing atmospheric observing conditions in order to respond in real time, continuously balancing the schedule for both facility calibrations (i.e., standard stars) and targeted observations according to the TAC-assigned science priorities. For the Liverpool Telescope, in the past year we have been testing a new multi-threaded approach. We have long operated a single-element, integrated-all-sky, 10 mu m bolometer on site. To this we have added real-time photometric monitoring of field stars around the science target and analysis of publicly accessible weather satellite images. This gives us three estimates of any night's photometricity; two ground-based looking up through the cloud (optical and thermal IR) and one satellite-based looking down at the observatory. We present a comparison of the results from the different methods and share our experiences selecting between the complementary data sets to support real-time observing decisions.