CFHT UM2025 - Presentation Details
Abstract Details
Title: Improving Observing Efficiency at CFHT: Predictive Models for Image Quality and Sky Background
Presenter: Sébastien Fabbro
Abstract:
Optimizing observing time at CFHT is essential for maximizing scientific output. We present refined data-driven models that forecast IQ and sky background brightness by analyzing CFHT's extensive archival data. Our methodology applies statistical techniques to a comprehensive dataset spanning two decades of operations, incorporating MegaCam and WIRCam observations, weather telemetry, and solar activity records. Our models predict delivered IQ with 0.07" precision and identify optimal dome vent configurations that can reduce observing time by ~12% for target signal-to-noise ratios. Building upon our previously demonstrated IQ improvement framework, we now present new sky background brightness predictions with better than 10% accuracy, significantly enhancing dark time utilization. We identify the primary physical drivers affecting these parameters and discuss how these predictive capabilities enable dynamic queue scheduling adjustments and real-time optimization of observatory parameters. This presentation covers our methodology, model performance, and implementation strategy, highlighting how this approach requires minimal resources while leveraging existing data infrastructure to measurably improve telescope efficiency.