2023 Shortlist

Outer Space, our garbage dump

Entry type: Single project

Country/area: Germany

Publishing organisation: Zeit Online

Organisation size: Big

Publication date: 2022-10-18

Language: German

Authors: Paul Blickle, Robert Gast, Nicolás Pablo Grone, Andreas Loos, Axel Rudolph, Julius Tröger, Benja Zehr

Biography:

Robert Gast is an astrophysicist turned journalist, he is the space reporter of ZEIT ONLINE.
Dr. Andreas Loos has a background in journalism and mathematics.

Project description:

Space debris poses a threat to both satellites and future space missions. The littering of space with defunct satellites and old rockets has accelerated dramatically these last years, mainly due to space launches from the private sector, above all Elon Musk’s Starlink project. Experts warn that soon a chain reaction could unfold, where fragments from collisions make parts of Earth’s orbit dangerous to use. With the help of interactive storytelling and an exclusive data analysis, our project shows how big the problem already is, while also describing what is at stake and what could be done to improve the situation.

Impact reached:

For the first time, we are making the entire extent of the space debris problem visible with this project. Comparable projects often only show actual satellites and not even the smallest pieces of space debris that have been created by collisions, for example. We therefore turned to the European Space Agency (ESA) and obtained the data set that ESA itself uses to produce its forecasts. It contains all space debris collisions of the past, as well as sophisticated simulations for the time evolution of the debris cloud.

Techniques/technologies used:

For the transformation of Kepler to Euclidean coordinates, we used a slightly modified version of the ESA tools Orekit and Orekit Labs (for python bindings). Moreover, we developed a simple physical model for debris particles (ignoring e.g. atmospheric friction) in Orekit to simulate the particle orbits for short time spans. This was necessary to classify the particles according to their orbit types (GEO, LEO, and so on). Since the dataset contained several million rows, we additionally experimented for visualization with an aggregation of particles to a cubic space grid (density) using a self developed search algorithm.

For the representation of the collision, we used 3D models that corresponded to the actual satellites of the 2009 crash. The representation of the Earth in Lower Earth Orbit was realized using an HDRI file. The moving particles showing the distribution of the scrap metal were simulated using X-Particles. For the static satellites and scrap parts we placed light points at the correct position data of the satellites. Detailed 3D models and especially visualizations with hundreds of thousands of points, as used in this project, are much too large to download, let alone animate performantly on cell phones. Therefore, we rendered the data directly into the animation software Cinema 4D and Redshift. This video in turn was then saved in many thousands of frames and displayed on scroll events. This turned out to be the smoothest playback method – especially on older, less powerful smartphone devices.

Context about the project:

The data is more accurate and complete than many publicly available catalogs of space debris, which often include positions only for objects that can be tracked by radar. However, this is not possible for debris less than ten centimeters in diameter. Here, only simulations of past collisions and explosions can estimate how much debris was created in each case and how it has spread in orbit over time. Esa uses a model called Master (Meteroid And Space Debris Terrestrial Environment Reference) for this purpose. In the dataset we used, more than 500 space debris events from the past were tracked and combined with data from measuring stations, according to Esa. The computational effort for this is considerable, but in the end a fairly accurate picture of the space debris situation emerges.

When working with large datasets it might be useful to have mathematical and computer science knowledge in the team, as we at ZEIT ONLINE do.

What can other journalists learn from this project?

A topic is more nuanced than you might think at first glance, which can pose challenges when showing a visualization of your data. Many media reports about space debris paint a sort or caricature of the problem: A cloud of debris so thick that no rocket can pass through it anymore. In fact, this will probably never happen, as space is really really big. Still, debris is a problem, but more in the sense that it increases the risk from negligible to noticeable: Over time, more and more satellites (and space stations) will be hit by debris, causing them to fail, making spaceflight more expensive and unreliable.
To the broad public, however, this will probably only become noticeable in a few decades, the experts we interviewed emphasized, while preventive action is required today. The problem is therefore reminiscent of climate change: We have to change our course of action now in order to avoid severe problems in the future.

This was a challenge when developing the visualization of the junk cloud we show in the first part of the story. We wanted it to a) give the right impression while b) still creating a strong visual effect. In the end, we decided to show just those objects from our data set that surpass a specific size, while discussing the balance between alarmism and the long winded nature of the problem openly our story.

Project links:

https://www.zeit.de/wissen/2022-10/weltraumschrott-satelliten-raumfahrt-gefahr-forschung