Speaker
Description
Astronomical seeing refers to the clarity and sharpness of celestial observations, governed largely by atmospheric turbulence influenced by dust, humidity, wind, and temperature fluctuations. Poor seeing conditions distort images, degrade measurement accuracy, and lead to inefficient use of valuable telescope time. These atmospheric fluctuations also critically affect geodetic systems such as Satellite Laser Ranging (SLR) and Lunar Laser Ranging (LLR), which require stable, well characterized atmospheric paths to achieve millimeter-level precision. Traditionally, seeing is measured using instruments such as Differential Image Motion Monitors (DIMMs), which while accurate, are costly and operationally demanding, limiting their accessibility across many astronomical and geodetic sites.
This study investigates a cost-effective, data-driven alternative for estimating and forecasting seeing by integrating all-sky camera imagery and Cloud Sensor measurements collected at the SARAO Hartebeesthoek site. The project is currently in an early development phase and has completed the initial CRISP-DM (Cross-Industry Standard Process for Data Mining) stages, including business understanding, data understanding, and data preparation. Exploratory analysis reveals meaningful atmospheric features relevant to both astronomical seeing and the stability of laser-ranging signals. Initial baseline models such as Multi-Layer Perceptron (MLP), Random Forest, XGBoost, and Long Short-Term Memory (LSTM) networks demonstrate promising capability in capturing nonlinear and temporal relationships between environmental conditions and observed seeing.
Although model refinement and validation are ongoing, anticipated outcomes include real-time or short-term seeing predictions that improve observational scheduling, reduce wasted operator time, and enhance data quality for both astronomical and geodetic applications. By leveraging existing, lower-cost instrumentation, this framework offers a scalable and practical solution that can be extended to additional observatory sites across Africa.
This project highlights the potential of machine learning to strengthen observational and geodetic capabilities while promoting cost-effective innovation across African facilities.
Keywords: Astronomical Seeing, Atmospheric Turbulence, Machine Learning, Data-Driven Prediction, SLR/LLR, Geodetic Applications
| Stream | Science or Engineering |
|---|