🔭 astrophysics

windsoCC: reconstructing the wind-driven halo in MagAO-X images using wavefront sensor telemetry

This paper introduces windsoCC, a novel post-processing workflow that utilizes high-cadence wavefront sensor telemetry from the MagAO-X instrument to reconstruct and remove the wind-driven halo artifact, thereby significantly improving the recovery of extended astrophysical objects like the disk around HR 4796A.

Original authors: Jay K. Kueny, Jared R. Males, Alycia J. Weinberger, Laird M. Close, Joseph D. Long, Joshua Liberman, Sebastiaan Haffert, Eden McEwen, Maggie Y. Kautz, Olivier Guyon, Logan Pearce, Parker T. Johnson, K

Published 2026-08-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jay K. Kueny, Jared R. Males, Alycia J. Weinberger, Laird M. Close, Joseph D. Long, Joshua Liberman, Sebastiaan Haffert, Eden McEwen, Maggie Y. Kautz, Olivier Guyon, Logan Pearce, Parker T. Johnson, Katie Twitchell, Jialin Li, Alex Hedglen, Avalon Gower, Warren Foster, Jhen Lumbres, Lauren Schatz

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Invisible Wind That Blurs the Stars

Imagine trying to take a crystal-clear photo of a tiny, glowing firefly sitting next to a blindingly bright spotlight. In the world of astronomy, this is the ultimate challenge: spotting faint planets or dusty disks swirling around distant stars. To do this, astronomers use giant telescopes equipped with a superpower called "adaptive optics" (AO). Think of AO as a magical, self-correcting camera lens that can instantly reshape itself to cancel out the twinkling and blurring caused by Earth's atmosphere. Without it, the atmosphere acts like a wavy window, turning sharp points of light into fuzzy blobs.

However, even with this magic lens, there is a sneaky ghost that haunts these images. Because the atmosphere is constantly moving, the telescope's correction system is always just a tiny fraction of a second behind the wind. This delay creates a strange, ghostly smear of starlight that stretches across the image, looking like a halo or a comet tail. Scientists call this the "wind-driven halo." It's a persistent noise that can hide the very objects astronomers are trying to find, especially if those objects are large and diffuse, like a ring of dust around a star. The big question has always been: how do we remove this ghost without accidentally erasing the real treasure we are looking for?

Chasing the Wind with a Digital Detective

This paper introduces a clever new method called windsoCC (which sounds like a wind chime, but is actually a software pipeline) that acts like a digital detective to solve this mystery. The team, led by Jay Kueny and colleagues using the MagAO-X instrument on the 6.5-meter Magellan-Clay telescope, realized that the telescope's own "eyes" (the wavefront sensors) were already watching the wind, even if the scientists weren't paying attention to the data in the right way.

Here is how the magic works, step-by-step:

1. The Frozen Flow Theory
The scientists started with a simple idea called the "frozen-flow" hypothesis. Imagine the atmosphere is made of invisible, transparent sheets of wind moving across the sky. As these sheets blow, they carry tiny ripples of turbulence with them. Because the wind moves at a steady speed, these ripples look like they are "frozen" in place but sliding across the telescope's view. If you take a picture of these ripples and then take another picture a split-second later, the ripples will have shifted slightly. By measuring exactly how much they moved, you can figure out the wind's speed and direction.

2. The WindsoCC Pipeline
The team built a computer program named windsoCC to do this math automatically. They fed it thousands of images from the telescope's wavefront sensor, which were taken at a super-fast speed of up to 3,600 times per second.

The Trick: The program took a snapshot of the wind ripples and compared it to a snapshot taken a tiny bit later. It used a technique called "cross-correlation," which is like sliding one puzzle piece over another to see where the patterns match best.
The Result: When the patterns matched, a bright "peak" appeared on the computer screen. By watching how these peaks moved over time, the software could draw a map of the wind, telling them exactly how fast the wind was blowing and which way it was heading for different layers of the atmosphere.

3. Reconstructing the Ghost
Once they knew the wind's speed and direction, the team used this information to build a mathematical model of the "wind-driven halo." They created a fake version of the ghostly smear that matched the real wind conditions perfectly. It's like knowing exactly how a shadow will fall based on the position of the sun; they could predict exactly where the noise would appear in the final photo.

4. The Big Test: HR 4796A
To see if this worked, they tested it on a famous star system called HR 4796A, which is surrounded by a beautiful, giant ring of dust.

The Problem: In the original images, the wind-driven halo was so bright and messy that it almost completely hid the details of the dust ring.
The Solution: Instead of using aggressive filters that might chop off parts of the real ring, they used their new wind model to subtract the ghostly halo directly from the raw images before doing any other processing.
The Outcome: The results were dramatic. When they removed the wind-driven halo using their new method, the dust ring around HR 4796A popped out with incredible clarity. The images showed the ring's structure much better than before, without needing to use harsh filters that usually damage the image.

Why This Matters

The paper doesn't just suggest that this might work; they actually demonstrated it with real data from the MagAO-X telescope. They showed that by listening to the telescope's own "heartbeat" (the wavefront sensor data), they could predict and remove a major source of noise that has been a headache for astronomers for a long time.

The key takeaway is that you don't always need to fight the noise with a sledgehammer (aggressive filtering). Sometimes, if you understand the wind that causes the noise, you can simply sweep it away, leaving the beautiful, faint details of the universe perfectly intact. This method turns the telescope's own data into a superpower, allowing us to see the "fireflies" next to the "spotlights" with a clarity we haven't had before.

Technical Summary: windsoCC – Reconstructing the Wind-Driven Halo in MagAO-X Images

Problem Statement
The wind-driven halo (WDH) is a persistent, low-spatial-frequency noise artifact inherent to adaptive optics (AO) systems, caused by servo-lag errors. Unlike static speckles, the WDH is dynamic, changing in intensity and position angle throughout an observation. This poses a significant challenge for high-contrast imaging, particularly for extended astrophysical objects like circumstellar disks. Standard post-processing techniques, such as spatial filtering or aggressive algorithms like Karhunen-Loève Image Projection (KLIP), often fail to remove the WDH without simultaneously filtering out the smooth, diffuse features of the target object. Consequently, the WDH frequently remains the limiting source of noise in high-contrast images, obscuring the innermost regions of circumstellar environments.

Methodology
The authors present windsoCC, a novel data processing pipeline designed to reconstruct and remove the WDH using wavefront sensor (WFS) telemetry. The methodology relies on the following core components:

Data Source: The pipeline utilizes high-cadence (up to 3.6 kHz) closed-loop pyramid WFS telemetry from the MagAO-X instrument on the 6.5-meter Magellan-Clay telescope.
Wind Vector Extraction: Based on Taylor's "frozen-flow" hypothesis, the pipeline applies cross-correlation (CC) operations to the WFS pupil images. By correlating pupil images with themselves across a range of temporal delays, the system identifies moving peaks in the CC response maps. These peaks encode the speed and direction of turbulent atmospheric layers.
- The pipeline partitions WFS data into 12-second mini-datasets.
- It employs a segmented, overlapping CC method (analogous to Welch's method) to maximize signal-to-noise, utilizing a maximum correlation time delay ( $\Delta t_{max}$ ) of 2 seconds within these blocks to capture slow-moving ground layers.
- Matched filtering (MF) and unsharp masking are applied to enhance peak detection.
- A "tripwire" search region is used to dynamically reject non-physical noise peaks, retaining only those that traverse the region linearly and radially outward.
- The Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm is used to classify discrete turbulent layers from the extracted wind vectors.
Forward Modeling and Removal:
- The wind vectors derived from windsoCC inform the priors for a parametric WDH model.
- The model uses a 2D analytical function incorporating a Gaussian envelope, a "pinch" parameter for lobe opening angles, and a radial flaring parameter.
- This parametric model is forward-modeled through the KLIP algorithm using the DiskFM tool (part of the pyKLIP package) via a Markov Chain Monte-Carlo (MCMC) framework.
- Once the best-fit model is determined, the individual WDH components are extracted and subtracted from the raw science frames before final PSF subtraction, preserving the signal of extended objects.

Key Contributions

windsoCC Pipeline: The development of an automated pipeline that cross-correlates closed-loop WFS telemetry to empirically map atmospheric wind vectors during post-processing.
Parametric WDH Reconstruction: A method to translate wind telemetry into a physical model of the WDH footprint, allowing for the reconstruction of the artifact's morphology based on atmospheric conditions.
Pre-Processing Subtraction: A workflow that removes the WDH from individual science frames prior to standard PSF subtraction, avoiding the need for overly aggressive spatial filtering that degrades extended object signals.

Results
The authors applied this workflow to archival MagAO-X observations of the circumstellar disk around HR 4796A taken on two separate nights (March 10 and March 13, 2023) in $i'$ and $r'$ bands.

Wind Vector Validation: The wind vectors extracted by windsoCC were cross-referenced with historical weather data from the Copernicus ERA5 reanalysis dataset. The extracted vectors aligned well with the multi-layer atmospheric profiles above Las Campanas Observatory at the time of observation.
Image Recovery: By subtracting the reconstructed WDH components from individual frames before KLIP processing, the authors achieved a dramatic improvement in object recovery. The resulting images of the HR 4796A disk showed clear, preserved features at visible wavelengths ( $r'$ -band at 615 nm and $i'$ -band at 762 nm) without the artifacts typically introduced by aggressive filtering.
Modeling: The forward modeling successfully fit the WDH using combinations of 4 to 5 distinct models, corresponding to the turbulent layers identified by the pipeline.

Significance
The paper claims that this work demonstrates the critical value of retaining and utilizing WFS telemetry for advanced post-processing. By bridging the gap between closed-loop wavefront error tracking and science image reconstruction, the authors show that the WDH can be effectively mitigated without sacrificing the signal of extended astrophysical objects. This approach establishes closed-loop WFS telemetry as an essential asset for maximizing contrast in high-contrast imaging, particularly for the study of circumstellar disks and other extended structures where traditional filtering methods are insufficient. The work also notes that detailed forward modeling and pipeline architecture details have been submitted for peer review in the Journal of Astronomical Telescopes, Instruments, and Systems (JATIS).

The Invisible Wind That Blurs the Stars

Chasing the Wind with a Digital Detective

Why This Matters

More like this