An audiovisual representation of a geomagnetic excursion for public engagement

This methodological communication paper details the creation and public reception of an audiovisual representation of the last geomagnetic excursion. Utilizing a global model of the Earth’s magnetic field, we developed a visualization and sonification designed to engage a broad audience with this extreme event in Earth’s history. The work reached over one million people online and sparked positive engagement, demonstrating the power of artistic interpretation in communicating complex scientific concepts. This project highlights the potential of innovative media formats for public outreach and science communication in the Earth sciences.

1 Introduction

The global magnetic field is a crucial part of the Earth system. It is used for navigation by animals and humans and it protects us against harmful radiation from space. It is also a window into the Earth’s interior (Merrill et al., 1996). Magnetic fields cannot be sensed by humans directly. Today, a global network of dedicated observatories and satellites provides continuous, high fidelity measurements of the field. These measurements are used to observe fundamental processes, such as the flow in the core (Gillet et al., 2010), currents in the atmosphere (Alken and Egbert, 2025), and even tiny signals generated in the ocean (Saynisch-Wagner et al., 2023). The magnetic field of the distant past is recorded in lavas and sediments (Merrill et al., 1996). These records reveal that the magnetic poles have switched position multiple times in the Earth’s history (Figure 1a). This can also be seen from the magnetization of the ocean floor (Figure 1b). Magnetic reversals, when the poles switch, are extreme events in the evolution of the Earth’s magnetic field. There is another class of extreme events recorded in sediments: geomagnetic excursions. The mechanism behind the two types of events is likely related to motions in the fluid outer core, but is not yet fully understood. Sometimes excursions are considered aborted reversals (e.g., Buffett, 2024). During reversals and excursions the magnetic field becomes very weak for a period of time. However, in an excursion the field returns to the original polarity after the event. This can be seen from the evolution of the axial dipole term - the strongest component of the field, which is parallel to the rotation axis, see Figure 2.

Figure 1

Figure 1. Magnetic polarity changes in the Earth’s history: (a) Global polarity timescale from Ogg (2020). White stripes correspond to “normal” polarity, when the field is in the same configuration as it is today, blue stripes indicate reversed polarity. (b) Records of magnetic polarity changes in the magnetized crust at the ocean floor, here shown for the Southern Pacific Ocean. The data is taken from the World Digital Magnetic Anomaly Map (Choi et al., 2025). Different polarities are indicated by red and blue stripes in the image.

Figure 2

Figure 2. Axial dipole during the Matuyama-Brunhes reversal (from the model GGFMB by Mahgoub et al. (2023); left panel) and the Laschamps excursion (from the LSMOD.2 model by Korte et al. (2019); right panel). During the reversal, the dipole changes its sign and remains with the opposite sign. For the excursion, the dipole returns to the original polarity.

In this paper, we describe how we created a visual and sonic representation of the last geomagnetic excursion - the Laschamps event. The aim of this representation was to spark interest in this fascinating aspect of Earth physics and to showcase one of the most extreme events in its history to a general audience. We use the concept of magnetic field lines to highlight how different the configuration of the magnetic field in an extreme state is from its stable, dipolar appearance we see today and in recent history. Adding sound to the visual representation creates an intuitive, corporeal, and immersive experience. The combined audiovisual work was shared via the YouTube channel ESA Extras, associated to the European Space Agency (ESA): https://www.youtube.com/watch?v=6Tc7XI0iUYU.

This manuscript is intended as a methodological communication contribution. In the following Section 2, we briefly explain how global models of magnetic excursions are built. One of those models is used as a back-bone of our audiovisual representation. In Sections 3 and 4, we lay out how the visualization and sonification of the model were created, respectively. Section 5 gives detailed insights into how the piece was received, by looking at available statistics. We draw conclusions in Section 6 and give an outlook in Section 7.

2 Magnetic field models for excursions

The magnetic field of the past is recorded by two distinct mechanisms: On one hand, material that is heated beyond its Curie temperature “freezes” the magnetic field when it cools down again. For geomagnetic excursions, these are lavas from volcanic eruptions. On the other hand, tiny magnetic particles record the magnetic field when they accumulate at the ocean floor and at the bottom of lakes. While lavas provide point-records in time and space, the sediments provide continuous time series of magnetic field evolution at single locations. These time series can reveal magnetic excursions (Nowaczyk et al., 2012). Multiple records can be combined to form a global picture of the magnetic field during the event.

For the most recent excursion, the Laschamps event ca. 40,000 years ago, several such models exist (e.g., Korte et al., 2019; Panovska et al., 2018; Panovska et al., 2021). All of these models are constructed by regularized least-squares inversion. The magnetic field is parameterized by a set of coefficients, the spherical harmonics, that evolve in time. These coefficients are inverted from the data by an iterative algorithm (Bloxham and Jackson, 1992). The algorithm interpolates the data globally, and aims to balance physically motivated smoothness constraints and fit to the data. The result can be used to infer the magnetic field at the Earth’s surface, in its interior down to the core-mantle boundary and in the near-Earth space environment. Both visualization and sonification make use of this property.

3 Visualization

3.1 Field lines

Because magnetic fields cannot be sensed by the human body, we rely on representations to interpret and understand them. A common way of representing vector fields (like the magnetic field) are field lines. These are lines that follow the field direction. Where the field is stronger, the lines are closer to each other. In a typical school experiment, iron filings are distributed in a glass container and a bar magnet is placed such that the filings orient themselves along the magnetic field (Figure 3a). The emerging pattern is precisely that of the magnetic field lines. Depictions of the Earth’s magnetic field also frequently show its dipolar field by means of field lines (Figure 3b). People are somewhat used to these images, so showing the Earth’s magnetic field in a similar way, but in a vastly different shape during an excursion helps people understand the level of magnetic chaos going on.

Figure 3

Figure 3. Examples for typical encounters of field lines in teaching: (a) Experiments with iron filings show the dipolar pattern of a bar magnet. (b) The Earth’s magnetic field lines for the year 2025. Because it is largely dominated by a dipole, the field lines form a pattern similar to the iron filings.

Generating field lines for a known vector field is relatively straight forward. Given the field vector $\mathit{F}$ in cartesian coordinates at a specific location in space $\mathit{x}$, the idea is to move in tiny steps along the field directions. Starting at a point ${\mathit{x}}_i$, the next point along the line ${\mathit{x}}_{i+1}$ is given by

${\mathit{x}}_{i+1}={\mathit{x}}_i+ϵ\cdot \widehat{\mathit{f}}\left({\mathit{x}}_i\right),$

where $ϵ$ is a small number and $\widehat{\mathit{f}}({\mathit{x}}_i)=\mathit{F}({\mathit{x}}_i)/|\mathit{F}({\mathit{x}}_i)|$ is a vector of unit length that points along the field direction. Connecting these points gives a field line. Starting multiple lines at several points in space will provide a visual representation of the field.

3.2 Field lines of the Earth’s magnetic field

Magnetic field models can be used to access the field vector in space. Because they do not include a description of the dynamo process in the core, the field models for the Laschamps event are valid only above the core-mantle boundary. We are interested in field lines in the near-Earth space around the globe, so this is sufficient for our purposes. We used the software pymagglobal (Schanner et al., 2020) to evaluate the GGF100k model (Panovska et al., 2018) and access the field vector. Starting at random locations on the globe, Runge-Kutta integration (a slightly more complex algorithm than the one outlined above) was used to calculate the lines of the magnetic field. The result is a collection of lines in three dimensional space.

3.3 3D visualization and composition

The three dimensional line data was imported into the visualization tool ParaView (Ahrens et al., 2005). The lines were rendered as tubes, with color depicting the radial part of the field (the part that points directly towards or away from the Earth). A sphere was added at the center of the scene, to represent the Earth. The surface color corresponds again to the radial part of the field. For artistic purposes, coastlines and exaggerated topography (NOAA National Centers for Environmental Information, 2022) were added to the sphere. For the background, we used randomly placed white dots on a black rectangle, to give a star-like impression. We chose the perspective from which the complex field during the Laschamps excursion looks most interesting (Figure 4). For example, some field lines show a clearly three dimensional shape, in stark contrast to the typical field lines of a dipolar field, which lie in planes.

Figure 4

Figure 4. Three dimensional rendering output of magnetic field lines during the Laschamps event, ca. 40,000 years ago.

Employing these techniques, we rendered 1,200 images for the timespan covering the Laschamps excursion, 42,500 to 39,500 years ago. We combined these images into an 80 s long video, using a framerate of 15 frames/second or 2.5 years/frame. To make the beginning of the video more interesting, we added a zoom-out animation. To give a sense of geological time, we also added a counter in the corner of the video.

4 Sonification

4.1 The original piece

The sonification was originally made for a specific venue in Copenhagen, Denmark, and conceived as a multi-channel audio only installation. It was later adapted and incorporated into the Laschamps visualization. However, since the fundamentals of the sonification are the same, we will give a description of how this original piece was conceived and produced.

The original project was called The Sounds of Earth’s Magnetic Field and ran from 24-30 October 2022 on a unique 32-channel sound system called The Sound Wells, which is dug into the ground of a 15,000 m² public square called Bent Fabricius-Bjerres Plads (formerly Solbjerg Plads; Figure 5a). The Sound Wells features 32 weatherproof speakers placed in shafts and shielded by perforated well covers. Furthermore, it is a 32-channel system so that the output of each speaker can be controlled individually, resulting in a totally immersive experience for the listener.

Figure 5

Figure 5. Speaker layout in the original sonification piece The Sounds of Earth’s Magnetic Field. (a) Speaker layout of the Sound Wells on Bent Fabricius-Bjerres Plads (formerly Solbjerg Plads) in Copenhagen, Denmark. (b) Speaker layout imposed on world map, to generate individual sonifications for the speakers at Bent Fabricius-Bjerres Plads.

The Sounds of Earth’s Magnetic Field was an artistic auditory representation of the planet’s magnetic field and its movements over the past 100,000 years at 32 different locations across the globe. These locations were selected by imposing a map of the world onto a map of the speaker layout on the square (Figure 5b) and then determining the longitude and latitude of each speaker’s location. Thus, when the listener walked for instance north across the square, they walked towards the arctic region of the auditory representation. This was an important part of the project, since not only is Earth’s magnetic field in constant change, but it also varies from one location on the planet to the next. Therefore, multiple speakers is a perfect way of representing local variations. Furthermore, with the speakers embedded in the ground, projecting sound upwards, the listening experience had a canny affinity with the magnetic field protruding from the Earth’s core out into space. This connection was in fact what inspired Klaus Nielsen to start this project.

The main protagonist of the piece is the core field generated by oceans of molten iron in Earth’s outer core, and this is the part that was adapted for the Laschamps visualization. In the original project, however, an additional layer was added representing the effect on the magnetic field by charged particles from the Sun. Data derived from ground observatories in Greenland of a so-called substorm event on 3 November 2021 was used to create a series of high-pitched crackling, static noises. This part was produced by Nikolai Linden-Vörnle, at that time a student at DTU. Admittedly, the two elements of the project had no temporal or spatial connection. They were combined as an artistic liberty to create a sonically dynamic result.

4.2 Sonification as a method

There are two main approaches to sonification as a method for representing data: Direct audification and parameter mapping. In the first, data is directly converted and transposed into an audible format. In the second approach, data is used to control sound, for instance the pitch, volume, or timbre of an instrument. Direct audification tends to output noisy results that are relatively similar to each other: static, crackling noises, or low frequency hums. A classic example of direct audification is the Geiger counter, wherein audible clicks represent the number of ionization events detected by the instrument. Parameter mapping covers the vast and diverse field of sonifications wherein data is not directly audified. Any sonification in which the sound sources are audio recordings or electronic or acoustic instruments applies parameter mapping. This was the method applied in the magnetic field sonification.

4.3 Data selection and processing

Data was extracted from the above-mentioned GGF100k model at 32 locations on the planet corresponding to the speaker layout of the public square. A combination of different parameters was chosen: Field intensity and the three components (theta, phi, radial) at both the Earth’s surface and the core mantle boundary, which lies roughly 2800 km below the surface. In addition, the rate of change for the same parameters was used. The piece was designed to last for 5 min and so a sampling rate of one data point per 100 years was chosen, which gave a sufficiently smooth resolution of 3.33 data points per second.

To use the data for musical purposes it was converted into the industry standard protocol MIDI (Music Instrument Digital Interface), which is an open format used by all electronic music devices. Most MIDI parameters have a resolution of 128 points, but the pitch bend function has a resolution of 16,384 points which makes it ideal for a smooth conversion. The data time series were offset and scaled to the range 0–16,383, converted into MIDI files (.mid) using the open source Python tool, py_midicsv (Wedde, 2024), and imported into a Digital Audio Workstation (DAW).

4.4 Conceptualizing the soundscape and audio processing

Sonifications of space data have become very popular, and to some extent the outlines of a style or genre have already emerged which leans towards either ambient electronic music or the symphonic film score. For the sonification of Earth’s magnetic field it was a deliberate choice to break with this mold.

The goal was to give the audience a feeling of listening to a living planet in constant flux, and so a blend of different field recordings was chosen and processed to convey an organic yet otherworldly soundscape. Recordings of rocks tumbling, wood creaking and breaking, and mudslides were combined with more randomly chosen sounds, for instance a table being dragged across the floor (an homage to the sound designers of Star Wars who amongst other sources used this sound for the growl of Chewbacca), a turtle eating lettuce, an anaconda gliding across a riverbed. All recordings were processed in the DAW using different techniques such as time-stretching, pitch shifting, and filtering so that each element was difficult to recognize, but the overall organic feel was retained.

The data time series were then used to control different playback parameters in such a way that the dynamics of the piece were directly linked to the data. Most noticeable is the magnetic field strength which was mapped to the overall volume: when the magnetic field strength goes high, the volume increases.

4.5 Reach and longevity of the original piece

The piece was picked up by ESA’s communications team for Earth Observation and published on the official ESA SoundCloud account where it quickly amassed over 1 million plays. This triggered extensive interest from other media, leading to interviews with Radio Dubai and BBC among others. The piece was brought to the attention of UNOOSA - the United Nations Office for Outer Space Affairs - who wished to include the project in their report on the use of sonification as method of inclusion (United Nations Office for Outer Space Affairs, 2023b) as well as invite Klaus Nielsen to set up a smaller 16-channel version of the installation at the UN building in Vienna during the 2023 general assembly of COPUOS (The Committee on the Peaceful Uses of Outer Space; United Nations Office for Outer Space Affairs, 2023a). This installation was made possible through the generous sponsorship of Genelec, a Finnish speaker company, who provided 16 speakers, and further from ESA’s communications team. Furthermore, the piece formed the background for a musical performance by Klaus Nielsen at the 2023 general assemblies of IUGG (International Union of Geodesy and Geophysics) and EGU (European Geosciences Union).

4.6 Merging sonification and the Laschamps visualization

The Laschamps visualization described in section 3 focuses on an event already present in the original sonification. But where the original piece represented 100,000 years in 5 min, the visualization represents the 3,000-year period from 42,500 - 39,500 years ago in 80 s. A sampling rate of one data point per 1,875 years was chosen which gave a resolution of 20 data points per second. Furthermore, where the original piece used data from 32 different locations, the Laschamps version relies primarily on global data. In addition, the declination, inclination, and field intensity at two random locations was used and panned out left and right in the stereo field.

This new data set was applied to the same audio recordings as in the original pieces. This way the movements in the audio correspond to the visualization. However, the audio tracks contained certain percussive elements (sounds of branches breaking or rocks hitting each other). Since the timings of these elements were not controlled by data, they needed to be manually aligned to the occurrences of field lines in the visualization using a DAW (here Ableton) that also supports video editing.

5 Analysis of reach and audience

In this section, we will analyze the reach of the video across traditional and social media, including YouTube, and the demographics of the audience.

YouTube is a global platform, accessed by over 100 million daily users, across over 100 countries and available in at least 80 languages. ESA’s main YouTube account, at the time of writing, has over 1.25 million subscribers. There, the organization publishes a variety of videos to share news and information, covering everything from new satellite launches to new scientific results gleaned from satellite observation of Earth and space. ESA’s most popular video, from 2015, features the Soyuz undocking, reentry and landing and has so far reached over 18 million viewers. Another widely viewed video, garnering 6.1 million views, features the astronaut Samantha Cristoforetti cooking on the international space station. Recently ESA’s YouTube team noticed a gap for extra content that might not be associated with ESA’s major focus campaigns, but being relevant for ESA satellite missions and campaigns would have wider value to both scientists and the public. For this reason, a secondary account was opened, called ESA Extras. It joins the ranks of other ESA channels catered to more specific audiences, such as the ESA Space Science Hub and ESA Education.

The sonification and animation was therefore published to ESA Extras, as well as being uploaded to ESA’s main website with an embedded link back to YouTube. It was also posted on ESA’s Earth Observation Instagram channel as a reel. As of October 2025, the video had over one million views on YouTube, making it the channel’s most popular video to date, and has accrued over 60,000 views on Instagram.

5.1 Who watched the video?

We performed an analysis of statistics gathered both from YouTube as well as ESA’s media monitoring team, covering a period of 4 months directly after the video was uploaded between 10 October 2024 and 10 February 2025. The media monitoring data comes from Aitastic, focusing on more traditional media such as online news websites, and Talkwalker for social media listening of platforms such as Facebook, Instagram, LinkedIn, YouTube, and X. We also used information from SimilarWeb to compile statistics on website audience demographics.

In this time, the video was viewed 508,711 times on YouTube, totalling 8,300 h of watch time. The average duration of a watch was 58 s, which is 72% of the total time of the video. This indicates that most users watched most of the content. The top five locations of viewers were the US (42%), Croatia (5.5%), Canada (5.4%), Egypt (4.3%) and Greece (3.9%), but the video reached a total of 137 countries (see Figure 6a). YouTube’s statistics suggest that 69.3% of viewers were male, while 30.7% were female. Just 0.1% of viewers were in the 13-17 age category, with 2.7% in 18–24, 10.9% in 25–34, 22.2% in 35–44, 21.7% in 45–54, 20.5% in 55–64, and 21.8% in 65+ (see Figure 6b).

Figure 6

Figure 6. View statistics from YouTube. (a) Spatial statistics. Countries with no data are shown in gray. Note that the data is shown on a logarithmic scale. (b) Views across different age categories. Only 0.1% of the views are in the category 13–17 years.

5.2 How did people find the video?

Just 2.5% of views came directly from YouTube, with the remaining 97.5% coming from external sources. According to YouTube, the Australian science news website ScienceAlert.com was the top referrer, responsible for 204,017 views - 41% of the total - thanks to an article titled “Earth’s Flipping Magnetic Field Heard as Sound Is an Unforgettable Horror”, in which the video hosted on the ESA Extras YouTube account was embedded, meaning users could watch the video after reading an introductory article about it. The next biggest referrer was Yahoo.com (6.4%), likely from a link to that ScienceAlert article, followed by the Croatian news website Index.hr (5.5%), the US-based Futurism.com (4.3%), Facebook (3.9%), Lebanon24.com (2.7%), the Greek website Enikos.gr (2.3%), Ampproject.org (2.0%), another Croatian website Klik.hr (1.5%) and the Bosnia-Herzegovinian website Nezavisne.com (1.5%), complete the top ten. See also Table 1, for some links to media articles, covering our work.

Table 1

Table 1. Links to selected media articles covering our work.

According to media monitoring data from Aitastic, the predicted reach of that ScienceAlert article in the period of 10 October 2025 to 10 February 2025 was 348,466. That represents a conversion rate - the percentage of views based on the predicted reach - of 58.5%. In this analysis, ScienceAlert was the traditional news outlet with only the 16th highest potential reach. The top two websites, responsible for 2.75 million and 2.4 million potential reach, respectively, were USA Today and the Daily Mail. These websites did not provide many conversions, because in both cases the video was uploaded directly to their platforms. The predicted reach overall of the animation and sonification was 17.7 million via traditional online media (such as news websites and blogs) and 38.3 million for social media (platforms such as Facebook, LinkedIn and X). If we compare YouTube’s top 50 referring websites with the list of websites obtained by media monitoring, only 21 of them match, suggesting the potential reach figure would be greater. If we compare those 21 matches, on average the conversion rate is 8.84%, which would grant an extra 212,000 views for the video via the Daily Mail website, for example. However, the actual views recorded are a more robust statistic. Of the potential reach of 17.7 million from traditional media, 408,602 views were gleaned - a conversion rate of 2.3% - compared to just 23,241 from social media’s much larger potential audience of 38.3 million - giving a conversion rate of 0.06%. A very large majority of social media reach came from a single Facebook channel belonging to Hashem Al-Ghaili, who has over 34 million followers, followed by various ESA X accounts and ESA Extras on YouTube which make up 87% of the rest of social media views. Where engagements are concerned - a metric that tracks things like likes, retweets, shares, etc., on social media - the standout was Doctor Fisión, whose reach of 294,000 achieved over 35,000 engagements on a YouTube video he posted.

The media monitoring and social listening metrics only record the country from which articles or social media posts were posted, rather than the location of the readership. They also miss some of the sources, including the major articles from Croatia and Greece. But, of the articles that were tracked, we can see that for traditional media the United States was the country with the largest potential reach, with 38 articles reaching 5.6 million people - 31.7% of the potential audience. The country with the second greatest reach was the UK, with six articles reaching a potential audience of 2.58 million (14.7% of the total), followed by Indonesia (8 articles, 2.19 million, 12.4%), Colombia (3 articles, 0.94 million, 5.3%), and Germany (5 articles, 0.72 million, 4.0%). If we split by region, Europe had the largest share of the potential reach, with 32.5%, followed by North America (31.7%), South America (18.0%), Asia (15.5%), Australasia (2.1%), and Africa (0.25%). For social media the data is massively skewed by the single Facebook channel previously mentioned, therefore giving the US 89% of the audience. The social listening metrics also point to an audience that is 66.2% male and 33.8% female, and 79.5% of the audience on social media is under the age of 35.

5.3 The challenges of understanding online audience demographics

Access to a global audience via the internet is a huge opportunity. However, internet outreach does not guarantee equal access to all potential audiences, and there are challenges in understanding those audiences.

It is worth addressing the latter point as it relates to the statistics presented above. YouTube belongs to Google, which can track the demographics (such as age, location and gender) of viewers via several means. One is direct user-provided information via accounts that are signed in, for which there is no guarantee of accurate information (users may knowingly provide false demographic information, or use a VPN), and many of which may be bots (Lazer and Radford, 2017; Cesare et al., 2018). Other methods include third-party cookies, which rely on inferences based on online activity (accuracy is often above 80% when detecting gender from “digital traces” (Hinds and Joinson, 2018) but that still leaves around a fifth of people being misgendered (Fosch-Villaronga et al., 2021)) as well as Android and iOS identifiers. If none of these options applies, then the user’s demographics remain unknown, so in the best-case scenario the numbers presented here represent merely an estimate of the age and gender of an incomplete (only two genders are considered, for example,) subset of the viewership.

It should be noted as well that all online audience analytics should be taken with a pinch of salt, as there are large discrepancies between numbers reported by different platforms even regarding audience interactions such as total visits or unique visitors (Jansen et al., 2022).

5.4 Traditional versus social media

What is very interesting is how many conversions happened from traditional versus social media, compared to the potential reach. Social media clearly has the potential to reach many millions in a single post, and an individual post can garner tens of thousands of likes, comments and shares. In the specific case of this video though, the more successful conversions - people watching and engaging with the video itself, rather than a post referencing it - came from more traditional news websites. However, this is just one video, and cannot be taken as evidence of a wider trend or used to support a communication strategy, and the data is incomplete.

5.5 Audience sentiment

Unfortunately, the demographic data can only tell us so much about the audience reached. We do not know what the readers, watchers, and listeners took from their experience in terms of things learned or remembered. We cannot follow or take part in a conversation like we can with a face-to-face event. Comments for the video are turned off, which means we cannot analyze responses there for an indication. The 22 comments on the video shared on Instagram might offer some idea, though this number is not at all representative of the total number of people who watched the video. Using a manual classification, of the 22 comments 12 were positive (54.5%), 3 were negative (13.6%), and 6 were neutral (27.3%).

One of the more common ways to describe the sonification in the title of articles was as an “unforgettable horror”, or similar sentiments (this is likely a result of the first sonification mentioned in this paper, which was released around Halloween with a title of a similar nature on the ESA website). Whether or not that affects what the viewer of the content takes away, or whether or not that matters, is a subject for a different analysis and difficult to ascertain here. The use of phrases such as “unforgettable horror” potentially lands in the clickbait category of titles, which can elicit different responses in readers ranging from curiosity to mistrust, depending on the actual outcome of engaging with that content (Jung et al., 2022). The fact that users watched most of the video, on average, suggests that as well as the video being very good, this is a case of positive nudging - a successful way to introduce the topic to a broad audience, leading to a satisfactory outcome in terms of engagement.

6 Conclusion

The analysis shows a great reception of the audiovisual piece we created. Nobody expected that within a year of releasing the video, it would reach more than one million people. A central conclusion is therefore that creative and non-conventional approaches to science representation can significantly increase the reach, especially among a general audience of non-scientists. We encourage our colleagues to experiment and collaborate with artists. Additionally, as our analysis shows, the audience was broadly international. However, at least on YouTube, we did not reach the younger age groups as well as the older ones (Figure 6B). While we published our work also on Instagram, next time it may be worth to consider platforms that have a younger audience in general, such as TikTok. Finally, a lesson learned is that the initial presentation and framing of the work can strongly influence how the work is represented by the media. When released via ESA’s website, the title of the initial sonification was “The scary sound of Earth’s magnetic field”. This led to other articles reporting with a similar framing (see Table 1), for example, “The Sound of Earth’s Flipping Magnetic Field Is an Unnerving Horror” (ScienceAlert), “The Sound of Earth’s Magnetic Flip Is Seriously Unsettling” (VICE) or “The Creepy Sounds the Earth Made When Its Magnetic Field Flipped Will Haunt Your Dreams” (Futurism).

7 Outlook

Before closing with an outlook, we would like to share two personal anecdotes related to this work: One is about a colleague who told us he saw a story about turning the magnetic field into sound in a major German news outlet. The other one is about a friend who found that the video had been used by a science YouTuber, to illustrate new findings about the Laschamps excursion. Even though both the colleague and the friend are close, they were not aware initially that we created the video and it was great to find that it reached them independently - without us telling them.

Motivated by the success of the initial piece, we worked on a more artistic follow-up. An audiovisual interpretation of the last geomagnetic reversal, the Matuyama-Brunhes reversal, premiered at the general assembly of the EGU 2025 (Nielsen et al., 2025). It was also released on the ESA Extras YouTube channel: https://www.youtube.com/watch?v=_blivWRpp80. While for the Laschamps excursion, we worked independently on visualization and sonification, for the follow-up piece, we used a more intensive feedback loop to shape video and sound together. Overall, the collaborative approach between scientists and artists to create something in between was a great experience. It was always interesting to exchange creative ideas and balance them with the aspiration of scientific accuracy.

Sonifications and visualizations have been used in other research fields as well, from climate sciences (Nagai, 2025; Hawkins, 2018) to gravitational physics (Ubach and Espuny, 2024; James et al., 2015). Our work adds to the general experience, that artistic interpretation can convey scientific findings and capture the attention of a general non-expert audience to a far better extent than text and graphs.

Data availability statement

Publicly available datasets were analyzed in this study. This data can be found here: git.gfz.de/sec23/korte/pymagglobal and git.gfz.de/sec23/korte/pymagglobal-3d-fieldlines.

Author contributions

MAS: Conceptualization, Data curation, Methodology, Visualization, Writing – original draft, Writing – review and editing. KN: Conceptualization, Data curation, Methodology, Writing – original draft, Writing – review and editing. PB: Data curation, Formal Analysis, Writing – original draft, Writing – review and editing. SP: Data curation, Methodology, Validation, Writing – review and editing. GK: Conceptualization, Supervision, Validation, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. MAS acknowledges funding by the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft), grant 388291411. MAS and SP acknowledge funding by the Helmholtz association within the Changing Earth InnoPool project “SOLVe”.

Acknowledgements

The following persons were instrumental in bringing the sonification project to fruition: Chris Finlay, Clemens Kloss, Marie Wigger, and Mikkel Otzen (DTU Space), Ingeborg Okkels and Finn Markwardt (The Sound Wells), Honora Rider, Chiara Forin, Romina Persi, and Matt Taylor (ESA).

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Ahrens, J. P., Geveci, B., and Law, C. C. (2005). “Paraview: an end-user tool for large-data visualization,” in The visualization handbook. Editors C. D. Hansen, and C. R. Johnson (Academic Press/Elsevier), 717–731.

CrossRef Full Text | Google Scholar

Alken, P., and Egbert, G. D. (2025). Modeling toroidal currents in the ionospheric electrojet regions. J. Geophys. Res. Space Phys. 130, e2024JA033080. doi:10.1029/2024JA033080

CrossRef Full Text | Google Scholar

Bloxham, J., and Jackson, A. (1992). Time-dependent mapping of the magnetic field at the core-mantle boundary. J. Geophys. Res. Solid Earth 97, 19537–19563. doi:10.1029/92JB01591

CrossRef Full Text | Google Scholar

Buffett, B. A. (2024). Excursions, reversals, and secular variation: different expressions of a common mechanism? Geochem. Geophys. Geosystems 25, e2024GC011604. doi:10.1029/2024GC011604

CrossRef Full Text | Google Scholar

Cesare, N., Lee, H., McCormick, T., Spiro, E., and Zagheni, E. (2018). Promises and pitfalls of using digital traces for demographic research. Demography 55, 1979–1999. doi:10.1007/s13524-018-0715-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Choi, Y., Dyment, J., Lesur, V., Garcia Reyes, M., Ishihara, T., Litvinova, T., et al. (2025). World digital magnetic anomaly map version 2.2.

Google Scholar

Fosch-Villaronga, E., Poulsen, A., Søraa, R., and Custers, B. (2021). A little bird told me your gender: gender inferences in social media. Inf. Process. and Manag. 58, 102541. doi:10.1016/j.ipm.2021.102541

CrossRef Full Text | Google Scholar

Gillet, N., Jault, D., Canet, E., and A, F. (2010). Fast torsional waves and strong magnetic field within the Earth’s core. Nature 465, 74–77. doi:10.1038/nature09010

PubMed Abstract | CrossRef Full Text | Google Scholar

Hawkins, E. (2018). The warming stripes. Available online at: https://www.climate-lab-book.ac.uk/2018/warming-stripes/(Accessed September 12, 2025).

Google Scholar

Hinds, J., and Joinson, A. N. (2018). What demographic attributes do our digital footprints reveal? a systematic review. PLOS ONE 13, e0207112. doi:10.1371/journal.pone.0207112

PubMed Abstract | CrossRef Full Text | Google Scholar

James, O., von Tunzelmann, E., Franklin, P., and Thorne, K. S. (2015). Visualizing interstellar’s wormhole. Am. J. Phys. 83, 486–499. doi:10.1119/1.4916949

CrossRef Full Text | Google Scholar

Jansen, B. J., Jung, S.-g., and Salminen, J. (2022). Measuring user interactions with websites: a comparison of two industry standard analytics approaches using data of 86 websites. PLOS ONE 17, e0268212. doi:10.1371/journal.pone.0268212

PubMed Abstract | CrossRef Full Text | Google Scholar

Jung, A.-K., Stieglitz, S., Kissmer, T., Mirbabaie, M., and Kroll, T. (2022). Click me.! the influence of clickbait on user engagement in social media and the role of digital nudging. PLOS ONE 17, e0266743. doi:10.1371/journal.pone.0266743

PubMed Abstract | CrossRef Full Text | Google Scholar

Korte, M., Brown, M. C., Panovska, S., and Wardinski, I. (2019). Robust characteristics of the laschamp and mono lake geomagnetic excursions: results from global field models. Front. Earth Sci. 7, 7–2019. doi:10.3389/feart.2019.00086

CrossRef Full Text | Google Scholar

Lazer, D., and Radford, J. (2017). Data ex machina: introduction to big data. Annu. Rev. Sociol. 43, 19–39. doi:10.1146/annurev-soc-060116-053457

CrossRef Full Text | Google Scholar

Mahgoub, A. N., Korte, M., and Panovska, S. (2023). Global geomagnetic field evolution from 900 to 700 ka including the matuyama-brunhes reversal. J. Geophys. Res. Solid Earth 128, e2023JB026593. doi:10.1029/2023JB026593

CrossRef Full Text | Google Scholar

Merrill, R. T., McElhinny, M. W., and McFadden, P. L. (1996). The magnetic field of the Earth: paleo-magnetism, the core, and the deep mantle. San Diego: Academic Press.

Google Scholar

Nagai, H. (2025). Polar geosonif-i: a python-based comprehensive web tool for geo-data sonification in polar regions doi:10.5194/egusphere-egu24-14238

CrossRef Full Text | Google Scholar

Nielsen, K., Schanner, M., Mahgoub, A., Panovska, S., and Kervalishvili, G. (2025). Sound reversal - hearing earth’s magnetic field switch poles 780k years ago doi:10.5194/egusphere-egu25-3279

CrossRef Full Text | Google Scholar

NOAA National Centers for Environmental Information (2022). Etopo 2022 15 arc-second global relief model. doi:10.25921/FD45-GT74

CrossRef Full Text | Google Scholar

Nowaczyk, N., Arz, H., Frank, U., Kind, J., and Plessen, B. (2012). Dynamics of the laschamp geomagnetic excursion from black sea sediments. Earth Planet. Sci. Lett. 351 (352), 54–69. doi:10.1016/j.epsl.2012.06.050

CrossRef Full Text | Google Scholar

Ogg, J. (2020). “Chapter 5 - geomagnetic polarity time scale,” in Geologic Time scale 2020. Editors F. M. Gradstein, J. G. Ogg, M. D. Schmitz, and G. M. Ogg (Elsevier), 159–192. doi:10.1016/B978-0-12-824360-2.00005-X

CrossRef Full Text | Google Scholar

Panovska, S., Constable, C. G., and Korte, M. (2018). Extending global continuous geomagnetic field reconstructions on timescales beyond human civilization. Geochem. Geophys. Geosystems 19, 4757–4772. doi:10.1029/2018GC007966

CrossRef Full Text | Google Scholar

Panovska, S., Korte, M., Liu, J., and Nowaczyk, N. (2021). Global evolution and dynamics of the geomagnetic field in the 15–70 kyr period based on selected paleomagnetic sediment records. J. Geophys. Res. Solid Earth 126, e2021JB022681. doi:10.1029/2021JB022681

CrossRef Full Text | Google Scholar

Saynisch-Wagner, J., Baerenzung, J., Hornschild, A., and Thomas, M. (2023). Tidal transports from satellite observations of earth’s magnetic field. Sci. Rep. 13, 13302. doi:10.1038/s41598-023-40448-3

PubMed Abstract | CrossRef Full Text | Google Scholar

Schanner, M. A., Mauerberger, S., and Korte, M. (2020). Pymagglobal - python interface for global geomagnetic field models, GFZ Data Serv. doi:10.5880/GFZ.2.3.2020.005

CrossRef Full Text | Google Scholar

Ubach, H., and Espuny, J. (2024). Sonification of gravitationally lensed gravitational waves/sonificació de l’efecte de lent gravitatòria en ones gravitacionals

Google Scholar

United Nations Office for Outer Space Affairs (2023a). “Notes from the underground: the music of earth’s magnetic field,” in An Interview with Klaus Nielsen, Musician and Project Manager at the Technical University of Denmark. Available online at: https://www.unoosa.org/documents/pdf/Space4PersonswithDisabilites/The_Music_of_Earth_Magnetic_Field.pdf.

Google Scholar

United Nations Office for Outer Space Affairs (2023b). “Sonification: a tool for research, outreach and inclusion in space sciences. A special report by the space for persons with disabilities project”. Available online at: https://www.unoosa.org/documents/pdf/Space4PersonswithDisabilites/UNOOSA_Special_Report_on_Sonification_2023.pdf.

Google Scholar

Wedde, T. (2024). py_midicsv—v4.1.2. Available online at: https://github.com/timwedde/py_midicsv.

Google Scholar