How have ecosystems changed over time and how can nature be better protected? To answer these questions, more than a million satellite images from former spy programs could be used, a study shows. The once secret recordings have been released for decades, but have so far hardly been used for environmental research. Together with modern technology, the spy images could reveal historical influences of humans and climate change.
During the Cold War, US spy satellites captured a large number of images starting in the 1960s. The analogue black and white images show the earth from above and capture both inhabited and uninhabited areas. Many of these originally secret images were released to the public starting in 1996 and have since become available worldwide. Researchers are already using the old military images in many disciplines, for example in archaeology, glaciology or civil engineering, to supplement modern civilian remote sensing data. Civilian images often only exist for selected areas and only since the 1980s, and in high resolution only from the 2000s. However, the historical images have so far hardly been used in ecological research and nature conservation.
What can be seen from the satellite images?
A team led by Catalina Munteanu from the University of Freiburg now wants to advance the use of the previously secret satellite images in environmental research. To do this, the researchers first evaluated the metadata of more than a million images from the four declassified historical US spy satellite programs to find out how comprehensive the information they contained was. The analysis showed that, in contrast to more modern remote sensing data, the old images cover almost the entire earth and are available for all seasons, mostly even with high resolution. The team then examined which ecological studies had already used images from these spy satellites. From this they derived possible future areas of application.
The spy images are therefore well suited for historical comparisons and research into legacy issues. The recordings could, for example, be used to better understand how individual habitats have changed over the last few decades and how humans or climate change have influenced the ecosystems. This also provides more precise conclusions about the species that lived there before and now. “For example, the Black Forest underwent intensive human use after the Second World War, which dramatically changed the composition of the forest and the functioning of the ecosystem and is likely still affecting the ecosystem's ability to cope with contemporary challenges, such as climate change “, explains Munteanu.
Further examples include extensive deforestation in Romania or changed river courses in Albania, which become clear when compared with the historical images. However, if you only compare more modern satellite images, these early changes are no longer visible. Thanks to modern image processing and analysis, the old satellite images could now offer better insights into the long-term changes in ecosystems, species populations or human influence on the environment since the 1960s, according to the researchers. Knowing what ecosystems once looked like could help avoid false conclusions about their current state and also help restore and better protect them in the future.
Hurdles to image use
However, difficulties must also be overcome in order to effectively use the images from spy satellites in ecological research, warn Munteanu and her colleagues. Among other things, access to the data is limited because it is time-consuming and expensive to process and correct the analog images so that they can be evaluated digitally. There are currently no uniform workflows within the scientific community. “Our work is therefore a call for interdisciplinary collaboration to solve these problems and unlock the full potential of these incredible data treasures,” says Munteanu. The researchers hope that artificial intelligence could also help with the actual evaluation of the images in the future.
Source: Catalina Munteanu (University of Freiburg) et al., BioScience, doi: 10.1093/biosci/biae002