🔭 astrophysics

Improvements to the NSO Farside Mapping Pipeline: Noise Reduction Updates

This paper outlines recent updates to the National Solar Observatory's Farside Mapping Pipeline that utilize noise reduction techniques to produce more accurate and consistent helioseismic maps, thereby enhancing the detection and tracking of invisible solar active regions for operational forecasting and scientific research.

Original authors: Mitchell Creelman, Kiran Jain, Niles Oien, John Britanik, Thomas M. Wentzel

Published 2026-06-04

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Mitchell Creelman, Kiran Jain, Niles Oien, John Britanik, Thomas M. Wentzel

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

Imagine the Sun as a giant, spinning orange. We can see the side facing us (the "front"), but the other side (the "back" or "farside") is hidden from our view. This is a problem because huge storms, called "active regions," can form on the back and spin around to hit us without warning. If we don't see them coming, they can disrupt our satellites, power grids, and cell phones.

For years, scientists have used a clever trick called "helioseismic holography" to peek around the corner. Instead of using a telescope to see the back, they listen to the sound waves traveling through the Sun. Think of it like being in a room with a closed door; you can't see what's behind it, but if someone is stomping their feet on the other side, the vibrations travel through the floor and walls to your feet. By analyzing these vibrations coming from the front of the Sun, scientists can build a map of what's happening on the back.

This report by the National Solar Observatory (NSO) is about fixing the "ears" of this system. The old way of listening was a bit fuzzy and full of static (noise), making it hard to hear the faint signals of storms on the farside. The team has upgraded the pipeline to make the maps clearer, sharper, and more reliable.

Here is a breakdown of their improvements using simple analogies:

1. Cleaning Up the "Static" (Noise Reduction)

Imagine trying to listen to a quiet conversation in a crowded room. The old system was like having a microphone that picked up every cough, sneeze, and clinking glass, making it hard to hear the conversation.

The "Dinner Plate" Filter: Sometimes, the cameras at the observatories get confused right after a calibration check, creating images that look like a dinner plate with ripples. The old system sometimes let these bad images through. The new system uses a smart AI (like a very picky bouncer) to spot these "dinner plate" images and throw them out before they ruin the data.
The "Bad Signal" Filter: Sometimes, the network of observatories has gaps in time (like when clouds block a site). If the data is too patchy (less than 80% full), the new system automatically discards that day's map, knowing it would be too noisy to trust.

2. Taking More Snapshots (Increased Cadence)

The old system took two big "snapshots" of the Sun's back, 24 hours apart, and mashed them together to make a 48-hour map. This was like trying to make a smooth video by only showing two frames.

The New Approach: The new system takes five snapshots, spaced only 6 hours apart. It's like switching from a slideshow to a smooth video. By averaging these five snapshots together, random noise cancels out, and the real storms stand out much more clearly.

3. Fixing the "Fading" Signal (Variable Averaging)

There was a specific problem when a storm on the back of the Sun was just about to spin around to the front (crossing the "limb" or edge). In the old system, the signal would get weaker and weaker until the storm almost disappeared right before it became visible.

The Fix: The old method averaged the data like a flat line, which smoothed out the important details at the edge. The new method uses "variable averaging." Imagine a spotlight that gets brighter exactly where the storm is, rather than dimming it down. This ensures that even as a storm is just peeking over the edge, we can still see it clearly.

4. Better Maps and New Views

The team didn't just make the sound clearer; they also changed how the maps look to make them easier to use.

Magnetic Strength Maps: They updated the math used to guess how strong the magnetic storms are. It's like upgrading from a rough estimate to a precise measurement.
New Perspectives: They created new map projections. One is like a flat map of the world (Latitude/Longitude) that is easier for computers to read. Another is a 3D "globe" view that shows the farside and the front side together, helping forecasters see the whole picture at once.

The Result

The team tested these changes on thousands of maps. The results showed that the new maps have:

Less "Static": The background noise is significantly lower.
Sharper Images: The storms look crisper and less blurry.
Better Edge Detection: Storms are no longer "lost" as they spin around the edge of the Sun; they remain visible right up until they appear on the front side.

In short, the NSO has tuned its "listening device" to hear the Sun's farside storms more clearly, giving us a better early warning system for space weather.

Technical Summary: Improvements to the NSO Farside Mapping Pipeline

Problem Statement
The National Solar Observatory (NSO) Farside Pipeline is a critical tool for space weather forecasting, utilizing helioseismic holography to detect and track solar active regions on the Sun's farside (the invisible hemisphere) via acoustic signatures observed on the front side. While the pipeline, which relies on data from the Global Oscillation Network Group (GONG), is widely used for scientific and national security applications, the quality of its seismic maps is susceptible to noise. This noise arises from various factors, including anomalous site images (e.g., "dinner plate" artifacts caused by calibration sequences), observational gaps, low duty cycles, and suboptimal averaging strategies. These issues can obscure active regions, particularly as they approach the solar limb, and reduce the reliability of the data for operational forecasting and research.

Methodology
The authors present a series of updates to the NSO Farside Pipeline, transitioning from a legacy system to a new "f6" pipeline. The improvements focus on noise reduction, calibration strategy, and data product generation:

Automated Image Quality Filtering (FQI Filter): To address the inclusion of anomalous Dopplergrams from individual GONG sites, a fully convolutional neural network (machine learning) was developed. This filter, trained on hand-labeled data, automatically identifies and excludes contaminated images (such as those with oscillatory distortions) before they are merged into the daily Dopplergram stream.
Increased Temporal Cadence: The legacy pipeline generated composite maps using two 24-hour phase shift maps separated by 24 hours (48-hour total coverage). The updated pipeline increases the temporal resolution by generating 24-hour phase shift maps at a 6-hour cadence. Composite maps are now formed by averaging up to five of these 6-hourly maps, enhancing resilience against random noise.
Duty Cycle Exclusion: A processing routine was implemented to systematically exclude 24-hour phase shift maps with a duty cycle below 80%, preventing low-quality data from contaminating the final composites.
Variable Averaging Strategy: The legacy pipeline averaged pixel values across all contributing maps, which inadvertently dampened signals near the solar limbs where data contribution varies. The new pipeline employs variable averaging, where each pixel is averaged only by the number of maps that actually contribute data to that specific pixel, preserving signal strength near the limbs.
Revised Magnetic Strength Calibration: The mathematical relationship used to derive magnetic strength from phase shifts was updated based on further data analysis, resulting in a new equation for calculating magnetic field strength ( $P_{mag}$ ) from phase shift values ( $P_{phase}$ ).
New Data Products: The pipeline now generates re-projected Latitude/Longitude maps (1° resolution) and global spherical projections. These utilize specific interpolation strategies (selected via RMSE, PSNR, and MSSIM metrics) to convert sine(latitude) projections to linear latitude.
Candidate Identification: The "find candidate" script was updated to identify active regions with higher probability thresholds (95% for automatic inclusion; 70% with consecutive detection for lower probabilities) and to provide detailed characteristics including magnetic flux and estimated return dates.

Key Results
The authors evaluated the improvements using 7,757 temporally identical maps generated by both the legacy (fqu) and updated (f6u) pipelines. The results demonstrate significant quantitative and qualitative enhancements:

Near-Limb Detection: The updated pipeline successfully mitigates the signal attenuation observed in legacy maps as active regions approach the East limb. In the new maps, active regions remain visible and trackable even as they cross the limb, whereas they often disappeared in the legacy products just before emerging on the Earth-facing side.
Image Variance: The mean image variance decreased from 0.00153 to 0.00139, with 79.8% of maps showing a reduction. The updated pipeline also established a lower variance ceiling (0.0046), indicating a dramatic reduction in the production of noisy, unusable maps.
Shannon Entropy: A measure of uncertainty and noise, entropy was reduced in 81.8% of the maps. The mean entropy dropped from 4.33 to 4.26, confirming a reduction in background noise.
Spatial Frequency: This metric, indicating pixel-to-pixel consistency, showed a reduction in 93.3% of the maps (mean reduction from -4.77 to -4.69), signifying a more consistent and higher-quality image product.

Significance and Claims
The paper claims that these updates result in more accurate, consistent, and reliable helioseismic maps. By reducing inherent noise and improving the visibility of active regions near the solar limb, the pipeline strengthens its value for both operational space weather forecasting and scientific research. The authors emphasize that these improvements allow for better continuous monitoring of solar activity without relying on expensive or temperamental space-based platforms. The work is presented as a step toward refining global helioseismic monitoring, with the updated pipeline offering a more robust tool for protecting telecommunications and national infrastructure from solar storms. The authors remain modest regarding the current state, noting that global helioseismic monitoring is a developing field and that further refinements in methodology and reliability are ongoing.

1. Cleaning Up the "Static" (Noise Reduction)

2. Taking More Snapshots (Increased Cadence)

3. Fixing the "Fading" Signal (Variable Averaging)

4. Better Maps and New Views

The Result

More like this