Twinkle, Twinkle, Star No More
&ball; Physics 15, 125
New forecasting approaches could help users of ground-based telescopes predict when the atmosphere will most scramble incoming light, allowing them to better suppress the effect.
The turbulent swirl of air in Earth’s atmosphere can cause incoming starlight to flicker, much to the frustration of ground-based telescope operators. Being able to predict the properties of this optical disturbance could help observatory personnel optimize telescope design, better schedule observations, and even select future telescope locations, all of which could improve the data collected by the scientists. At the SPIE Astronomical Telescopes and Instrumentation conference in Montreal in July 2022, researchers from three different groups presented new tools that could predict atmospheric turbulence at different spatial scales up to a few hours in advance. This could help telescope operators better mitigate the effects of turbulence on telescope measurements, improving the data they collect.
Temperature gradients in the air, the topography of the Earth’s surface, and wind shear all contribute to creating atmospheric turbulence – small-scale, chaotic whirlpools of air characterized by winds that vary in speed and direction. . Variations in air density within these vortices give the atmosphere a refractive index (a parameter that determines how air deflects light) that varies rapidly in space and time. Such a very inhomogeneous refractive index distorts images of astronomical objects taken from the ground. To correct for these distortions, users can adjust the shapes of certain mirrors in telescopes to effectively “cancel” atmospheric distortions. This method works best when turbulence is low, says James Osborn, an atmospheric physicist at Durham University in the UK. If the turbulence is strong, as is the case when a large weather front is present, the data may require extensive post-processing that includes performing complex measurements in the sky to disentangle the details of the site turbulence. In extreme cases, the data may need to be discarded.
Turbulence forecasting could help avoid such data loss. Scientists could “save an hour of observing time” by better optimizing the configuration of telescope optical components using forecast data, when available, Osborn said. The three tools featured all make these predictions: one tool does so on a global scale, another aims for quick temporal measurements, and another gleans information from real-time weather parameters.
The tool developed by Osborn and his colleagues uses global weather models to convert weather data into global maps of atmospheric turbulence. The team started with atmospheric meteorological data from the European Center for Medium-Range Weather Forecasts which included vertical profiles of temperature, pressure and air humidity; and wind speed and velocity measurements taken hourly at fixed weather stations. They then incorporated this data into a model that calculated how the temperature and density of air, and therefore its index of refraction, varied as a function of the horizontal distance between two points. The team tested the model using archived observational data. By comparing their predictions to observations made over the Paranal Observatory in Chile and London for the same period, they found that the model could predict global turbulence patterns a few days in advance.
Osborn says their forecasting tool could allow for more efficient site selection for future observatories, as it would allow planners to predict conditions at different sites and then use that information to select the location where turbulence was greatest. likely to be low, or at least consistent. The tool could also help optimize observation schedules for exoplanets, for example, which require very quiet atmospheric conditions due to weak signals from exoplanets relative to their bright host stars.
While Osborn and her colleagues have chosen to create a world-scale model capable of predicting the properties of the atmosphere above any observatory, Elena Masciadri of the Italian National Institute of Astrophysics (INAF ) and his colleagues used atmospheric models that have sub-kilometer horizontal resolution and are designed to simulate conditions above specific geographic locations.
The INAF team’s original model has been operating at the Large Binocular Telescope at the Mount Graham International Observatory in Arizona since 2016, predicting local atmospheric conditions on timescales of a day or one to two hours into the future . But the team wanted to improve the accuracy of the predictions. They have now done this by combining statistical and machine learning techniques with real-time measurements.
The one- to two-hour perspective of their forecast means that the information provided by the model is relevant for optimizing data taking at the large binocular telescope, says Masciadri. She and the team now plan to develop a similar model to predict atmospheric turbulence above the Very Large Telescope in Chile.
In a third approach discussed at the meeting, Ryan Dungee from the University of Hawaii at Hilo presented a model that can identify the height of a turbulent vortex using real-time observations of the speed and direction of the vortex. wind. This information could allow observatories to take advantage of information about the spatial distribution of turbulence to improve performance. To test the model, a research team used wind speed and direction data from local weather stations and balloon launches over Maunakea Observatory in Hawaii. They then ran their model to map the altitudes of different turbulent layers based on the measured winds. They found that they could separate surface layer turbulence (up to 200 m above ground level) from free atmosphere turbulence (200 m to 2 km above ground level ) for 67% of the days they tested. Dungee says the data provided by his model could help reduce operational costs because he uses existing wind measurements to determine different turbulent layers and doesn’t need the expensive tools currently used to actively measure those layers. “Knowing which layers contain the most turbulence could lead to cheaper systems overall,” says Dungee.
The predictions and corrections provided by the three models also offer benefits beyond astronomy, the researchers say. In addition to space observations, turbulence can hamper free-space optical communications, which uses laser beams to transmit information from the ground to satellites. Turbulence can cause the beams to flicker or spread, distorting the received signal. “But if you know that the strongest turbulence is always at a certain height above the ground, you can design a system to account for that,” Osborn says. “So it all comes together.”
Rachel Berkowitz is the corresponding editor for Physics Magazine based in Vancouver, Canada.