🔭 astrophysics

A Singular Value Decomposition Framework for Jovian Radio Emissions from Parker Solar Probe

This paper presents a novel Singular Value Decomposition framework coupled with dual-stage noise filtering and Eigenfaces analysis to successfully isolate faint Jovian radio emissions from the Parker Solar Probe's noisy solar wind data, thereby validating the spacecraft's potential as a long-term distant observatory of Jupiter's magnetosphere.

Original authors: Evan Wille, Zack Li, Marc Pulupa, Leon V. E. Koopmans, Philippe Zarka, Stuart Bale

Published 2026-08-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Evan Wille, Zack Li, Marc Pulupa, Leon V. E. Koopmans, Philippe Zarka, Stuart Bale

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

Jupiter is a world that never sleeps, even in the dark of space. While the planet is famous for its swirling clouds and massive storms, it is also a relentless broadcaster of radio waves. These signals, invisible to the human ear, are generated deep within the planet's magnetic field, a vast, invisible cage of energy that traps charged particles. This magnetic cage is constantly interacting with Jupiter's smallest moon, Io, a volcanic world that acts like a giant generator, pushing electrical currents through the magnetic field and triggering bursts of radio energy. For decades, scientists have listened to these broadcasts to understand how Jupiter's magnetic engine works and how it connects to its moon. However, listening from Earth or even from nearby spacecraft has always been difficult. The signals are faint, and the space between the observer and Jupiter is filled with a chaotic roar of solar wind and static from the spacecraft itself, often drowning out the very whispers of the planet researchers are trying to hear.

For the last several years, the Parker Solar Probe has been racing closer to the Sun than any spacecraft before it, designed to study our star up close. But as it speeds through the solar system, it has also been passing by Jupiter, offering a unique, if distant, vantage point. The probe carries sensitive instruments capable of hearing radio waves, but for a long time, the data it collected from Jupiter was considered too noisy to be useful. The spacecraft's own electronics and the constant stream of particles from the Sun created a wall of interference that made it nearly impossible to separate the faint Jovian signals from the background chaos. Researchers knew the signals were there, buried under layers of static, but they lacked a way to dig them out without getting lost in the noise.

A new study by a team of astronomers has changed that. They developed a mathematical method to clean up the Parker Solar Probe's data, effectively turning the spacecraft into a long-term radio observatory for Jupiter. The team did not try to filter out the noise by looking for specific shapes or patterns that they expected to see. Instead, they treated the entire dataset as a complex puzzle. They took the raw radio recordings, which span six years of observations, and rearranged them based on the rotation of Jupiter and the orbit of Io. By aligning the data to these cosmic clocks, the signals from Jupiter began to line up in a repeating pattern, while the random noise from the solar wind and the spacecraft remained scattered and disorganized.

Once the data was organized this way, the researchers applied a powerful mathematical tool known as singular value decomposition. In simple terms, this process breaks down the complex data into its most essential building blocks. It separates the strong, repeating structures—the true radio signals from Jupiter—from the weak, random jitters of the background noise. The method acts like a sieve that lets the structured signals pass through while holding back the chaos. The result was a dramatic improvement in clarity. The team found that their method boosted the strength of the Jovian signals relative to the background noise by an average of 1.54 decibels, with some individual bursts becoming more than three decibels clearer. This might sound like a small number, but in the world of radio astronomy, it is the difference between hearing a whisper and understanding the words.

The cleaned data revealed the true nature of Jupiter's radio emissions with striking detail. The researchers were able to map out exactly when and where these radio bursts occur, confirming that they are tightly linked to the position of Io. The signals appear in specific arcs and lanes that repeat with the moon's orbit, a pattern that matches what scientists have seen from other missions but had never been able to see so clearly from such a distant vantage point. The study also showed that the radio waves come in two distinct types of polarization, or "handedness," which correspond to different parts of Jupiter's magnetic field. One type of signal is strongest when Io is at a specific point in its orbit, while the other peaks at a different time, creating a complex but predictable rhythm that the new method captured perfectly.

To prove that their method was working, the team compared their results with data from the Juno spacecraft, which orbits much closer to Jupiter. While Juno's instruments were often blocked by their own electronic interference in the frequency range where Jupiter's hectometric radio waves are strongest, the Parker Solar Probe's data, once cleaned, showed the same radio bursts clearly. By matching the timing of the signals seen by both spacecraft, the researchers confirmed that the waves were indeed coming from Jupiter and traveling across space to reach the probe. The time delay between the two observations matched the travel time of light across the distance between the two spacecraft, providing a solid, independent confirmation that the signals were real and not just an artifact of the cleaning process.

This work demonstrates that spacecraft not designed for planetary radio astronomy can still serve as powerful observatories if the right tools are used. The method developed by the team is automated and does not require a human to look at every piece of data and decide what is a signal and what is noise. It can be applied to the vast archives of data collected by the Parker Solar Probe and potentially to other missions in the future. By turning a noisy, chaotic dataset into a clear picture of Jupiter's radio activity, the researchers have opened a new window into the dynamics of the Jovian system. They have shown that even from a distance, and through a wall of interference, the radio voice of Jupiter can be heard, revealing the intricate dance between a planet and its moon in a way that was previously impossible.

Technical Summary: A Singular Value Decomposition Framework for Jovian Radio Emissions from Parker Solar Probe

Problem Statement
Since its 2018 launch, the Parker Solar Probe (PSP) has served as a long-term, distant observer of the Jovian system, recording high-resolution radio and plasma wave data during 21 perihelion encounters. However, extracting faint Jovian decametric (DAM) and hectometric (HOM) emissions from this dataset presents a significant analytical challenge. The data is routinely saturated by quasi-thermal plasma noise, spacecraft instrumental interference, and spectrally wide solar bursts (Types II and III). Traditional extraction methods, which rely on close-proximity spatial constraints or localized direction-finding, are inefficient and prone to subjective bias when applied to a distant, non-dedicated platform like PSP. Furthermore, standard Stokes $V$ polarization data is only available during high-rate telemetry windows (approx. two days per month), precluding continuous long-term monitoring.

Methodology
The authors present a generalized, empirical pipeline designed to blindly separate structured planetary signals from a stochastic heliospheric background. The methodology proceeds through four primary stages:

Phase-Folding and Data Formatting:
Raw dynamic spectra are converted into a periodic, phase-folded reference frame relative to Jupiter's System III longitude ( $\lambda_{III}$ ) and Io's orbital phase ( $\Phi_{Io}$ ). Using NASA SPICE ephemerides, timestamps are corrected for light travel time and projected onto these rotating coordinates. This restructuring transforms temporal data into repeatable, periodic arrays, establishing a baseline for noise reduction without temporal averaging.
Deterministic Noise Filtering:
- Quasi-Thermal Noise Whitening: A dynamic noise-percentile function evaluates the median and 75th percentile of the positive half of the data frequency-by-frequency. This generates a frequency-dependent multiplier to flatten the background noise continuum, ensuring the local electron plasma frequency limit is not misclassified as transient bursts.
- Solar Burst Mitigation: An automated thresholding algorithm detects spectrally wide transients (solar bursts) by identifying sustained increases in standard deviation across frequency channels while the mean amplitude remains near zero. These corrupted intervals are removed and replaced via median inpainting scaled to the local noise floor.
Singular Value Decomposition (SVD) Extraction:
The filtered spectrograms are decomposed using SVD ( $M = U\Sigma V^T$ ). The separation of signal from noise is governed by the distribution of singular values, which exhibit a steep decline followed by a "noise shoulder."
- Rank Selection: A deterministic rank parameter $k$ is established by retaining singular values up to an inflection point.
- Entropy Thresholding: To objectively select modes, the Shannon entropy of the normalized spectral basis vectors is calculated. Structured Jovian emissions concentrate power in a subset of channels (low entropy), whereas stochastic noise distributes power uniformly (high entropy). Modes are retained only if their entropy $S_i$ falls below a threshold $S_{lim} = 3.0$ nats.
- Reconstruction: The matrix is reconstructed using only the retained low-entropy modes, effectively isolating the Jovian signal.
Eigenfaces for Global Trends:
A secondary matrix factorization, adapted from the "Eigenfaces" technique used in facial recognition, is applied to the entire archive. Individual spectrogram matrices are flattened into vectors and stacked to derive a dominant global structural mode (the "Eigenface"). A normalized correlation coefficient $C$ (cosine similarity) is computed for every spectrogram against this basis, serving as a scalar detection metric for long-term morphological trends and archival screening.

Key Contributions and Results

Signal-to-Noise Ratio (SNR) Improvement: The pipeline demonstrates empirical SNR improvements of $1.54 \pm 0.11$ dB on average, with individual events reaching gains up to $3.32$ dB.
Validation of Signal Separation:
- Autocorrelation: The SVD-reconstructed matrix exhibits a lag-1 autocorrelation of $r_{svd} = 0.797 \pm 0.003$ , confirming spatially correlated structure. Conversely, the residual matrix yields $r_{res} = 0.040 \pm 0.017$ , consistent with uncorrelated white noise.
- Cross-Polarimetric Validation: The extracted $C_{XY}$ cross-correlation data (which serves as a proxy for Stokes $V$ ) shows a Pearson correlation of $r = 0.783$ with independently processed Stokes $V$ spectrograms during high-rate windows. Control experiments applying SVD directly to Stokes $V$ data yielded significantly lower performance ( $r_{svd,V} = 0.32$ and an SNR improvement of $0.64$ dB), confirming that the extraction advantage stems from the specific $C_{XY}$ signal chain rather than the SVD algorithm itself.
Geometric Independence: The pipeline successfully recovers known physical features without imposing geometric constraints during extraction. This includes:
- Concentrated Io-phase peaks at $\Phi_{Io} \approx 100^\circ$ and $260^\circ$ .
- A suppression trough near $\Phi_{Io} \approx 150^\circ$ .
- Systematic hemispheric asymmetry in longitude distributions (bimodal for Left-Hand Circularly Polarized emissions, flatter for Right-Hand).
Complementarity with Juno: The framework successfully isolates HOM transients in frequency bands (150 kHz–3.5 MHz) where Juno-Waves data is contaminated by instrumental interference. Coincident detections between PSP and Juno align with the geometrically predicted light-travel-time delay ( $\sim 36.84$ minutes), providing independent confirmation of the Jovian origin of the extracted signals.

Significance and Claims
The paper establishes the viability of PSP as a long-term, non-dedicated Jovian radio observatory. By utilizing a robust, automated mathematical framework based on SVD and phase-folding, the authors demonstrate that faint planetary signals can be extracted from high-noise heliospheric environments without relying on localized spatial constraints or manual curation.

The authors modestly frame the "Eigenface" metric not as a definitive pixel-level classifier, but as a computationally efficient, probabilistic pre-screen that enriches candidate pools by approximately 30-fold. The primary significance lies in providing an open-source, generalized methodology that can be retroactively applied to other heliospheric archival datasets or integrated into future planetary missions to objectively identify transient signals in complex noise environments. The work confirms that distant spacecraft, when equipped with sensitive instrumentation and processed through rigorous mathematical frameworks, can yield scientifically valuable planetary radio astronomy data.

More like this