Did forests in Valais move upwards in the last 80 years?

Valaisan Forests moving upwards (altitude increases from top to bottom of the image) in 1946-2020 © 2024 T.-A. Nguyen

Valaisan Forests moving upwards (altitude increases from top to bottom of the image) in 1946-2020 © 2024 T.-A. Nguyen

Yes! Forests are dynamic and react to climate change. In particular, their upper limit (known as the treeline) is moving upwards due to the increase temperatures and agricultural land abandonment. In a paper out today in Remote Sensing of Environment, ECEO's PhD student Thien-Anh Nguyen shows a complete cartography in Valais since the 1940's.

Forest move: slowly, but constantly. They react to shifting land usages and to the increasing temperatures. As a consequence, the upper limit of the forest, or treeline, has been shifting up and up in the last decades. If some local studies exist documenting such search for cooler temperatures, the movement was never cartographied at a large scale. In a paper published today in Remote Sensing of Environment, a team led by ECEO grad student Thien-Anh Nguyen has cartographied the evolution of the forest cover at the treeline for the whole Valais, and since the 1940s!

Using historical images taken from airplanes, first by U.S. military pilots during the Second World War and then by Swisstopo, the team has developed a machine learning model extracting forest footprints, which can then be compared both in space and time. The images are of different resolution, colors, and quality, so wide efforts were done to make this AI model robust to such variations, including enforcing consistency of forest in time and ecologically plausible growth rates.

This map will help the team to study the relations between forest dynamics and climate change in Alpine environments.

Funding

The data used in this work were provided by swisstopo.

References

Thien-Anh Nguyen, Marc Russwurm, Gaston Lenczner, and Devis Tuia. Multi-temporal forest monitoring in the Swiss Alps with knowledge-guided deep learning. Remote Sensing of Environment, 305:114109, 2024. https://doi.org/10.1016/j.rse.2024.114109