Estimating effective wind speed from Gemini Planet Imager's adaptive optics data using covariance maps

Abstract

The Earth's turbulent atmosphere results in speckled and blurred images of astronomical objects when observed by ground based visible and near-infrared telescopes. Adaptive optics (AO) systems are employed to reduce these atmospheric effects by using wavefront sensors (WFS) and deformable mirrors. Some AO systems are not fast enough to correct for strong, fast, high turbulence wind layers leading to the wind butterfly effect, or wind-driven halo, reducing contrast capabilities in coronagraphic images. Estimating the effective wind speed of the atmosphere allows us to calculate the atmospheric coherence time. This is not only an important parameter to understand for site characterization but could be used to help remove the wind butterfly in post processing. Here we present a method for estimating the atmospheric effective wind speed from spatio-temporal covariance maps generated from pseudo open-loop (POL) WFS data. POL WFS data is used as it aims to reconstruct the full wavefront information when operating in closed-loop. The covariance maps show how different atmospheric turbulent layers traverse the telescope. Our method successfully recovered the effective wind speed from simulated WFS data generated with the soapy python library. The simulated atmospheric turbulence profiles consist of two turbulent layers of ranging strengths and velocities. The method has also been applied to Gemini Planet Imager (GPI) AO WFS data. This gives insight into how the effective wind speed can affect the wind-driven halo seen in the AO image point spread function. In this paper, we will present results from simulated and GPI WFS data.