Technische Universität Berlin

Press release | 29. September 2020 | pp

Forests in Distress

Using remote sensing images from space and artificial intelligence, two major joint research projects aimed at identifying forest damage have been launched

Our forests are true all-rounders. They are a habitat for flora and fauna, protect the climate, and are an attractive place for rest and recreation. They are also an important economic factor in rural areas. However, our forests are in distress. In the past three years, they have been exposed to extreme weather conditions. Storms, periods of extreme drought, and a massive bark beetle infestation have led to significant damage and harm to our forests. Two of the most comprehensive research projects on forest damage in Germany, “FirSt2.0” and “TreeSatAI”, are now underway at TU Berlin. The projects are being funded with a total of 2.5 million euros from the Federal Ministry of Transport and Digital Infrastructure as well as 1.2 million euros from the Federal Ministry of Education and Research.

“Climate change is causing severe problems for German forests, which provide not only a habitat and basis of life for humans and animals but also approximately one million jobs in Germany. This year alone, wildfires have torched 2700 hectares of forest. This is the worst we’ve experienced in the past 30 years,” explains Professor Dr. Birgit Kleinschmit, head of the Chair of Geoinformation in Environmental Planning, which is host to both projects at TU Berlin. She is also speaker of the project “TreeSatAI”.

© TU Berlin/FG Geoinformatik

Please note: Once you watch the video, data will be transmitted to Youtube/Google. For more information, see Google Privacy .

Bark beetles, wildfires, and storms increasingly harm forests

“However, there is also so much more affecting forests,” says Kleinschmit, who was recently appointed to the Scientific Advisory Board on Forest Policy by Federal Minister of Food and Agriculture Julia Klöckner. “The drought in 2018 and 2019 led to an explosive reproduction of the bark beetle because trees lacked water and thus the necessary mechanisms to defend themselves – We are talking about forest damage of historical proportions.”

An exact analysis of the different types of forest damage and their extent is needed before a potential strategy to adapt forests to climate change can be developed or recommendations made to the forest industry and implemented. This is where the two projects come in.

 

Vitality analysis with satellite remote sensing data and AI

“FirSt2.0” (Forest Damage inventory based on rapid Satellite technologies), which will run until June 2023, aims to develop a software product for those involved in forestry and conservation. Multi-sensor satellite remote sensing data, with in-situ measurements of ground moisture and other stress parameters as well as Artificial Intelligence (AI) and machine learning will provide a continuous examination and analysis of the state of the different types of forest damage. This service aims to provide stakeholder from forestry and nature conservation with an improved forest management that is adapted to the relevant situation. The project is coordinated by LUP – Luftbild Umwelt Planung GmbH, Potsdam. The TU Berlin Chair of Geoinformation in Environmental Planning is responsible for algorithm development within the joint project. Furthermore involved are also the Thünen Institute of Forest Ecosystems, forest owner associations, and the Bavarian Forest National Park. “This means that we are looking for methods to calculate the state of vegetation - forest damage such as declining leaf mass and wilted or prematurely falling leaves - we find on site and measure with highly sensitive technology, which we can use to derive and extrapolate the damage for a larger area, later even for the whole of Germany,” explains Dr. Michael Förster, project coordinator in the TU Berlin chair, which is funded with 560,000 euros in the project.

Citizen Science and social media data also included in forest monitoring

The joint project “TreeSatAI” (Artificial intelligence with remote sensing and multi-source geodata for infrastructure, conservation, and forest monitoring) will run until 2022 and involves two TU Berlin academic chairs – the Chair of Geoinformation in Environmental Planning, represented by Professor Birgit Kleinschmit, who also serves as speaker of the project, and the Chair of Remote Sensing Image Analysis (RSiM), represented by remote sensing and AI expert Professor Dr. Begüm Demir. Further project partners include the TU startup liveEO GmbH and the LUP GmbH. Together with the German Center for Artificial Intelligence (DFKI) and its startup Vision Impulse GmbH they will develop innovative AI methods for infrastructure, conservation, and forest monitoring. Together they will develop innovative AI methods for infrastructure, conservation, and forest monitoring. The project is particularly focused on incorporating a large amount of current satellite and drone data, information from current environmental and European databases, and social media and citizen science data into the AI algorithms to identify and monitor the current state in forests on at local, regional, and European level. 500,000 euros of the total funding will go towards TU Berlin’s contribution to the project.

According to Kleinschmit, the close cooperation with multiple small and medium-sized companies, associations, and administrations in both projects also aims to achieve an active knowledge transfer in industry and society.

www.geoinformation.tu-berlin.de/

Further information is available from

Prof. Dr.

Birgit Kleinschmit

Institute of Landscape Architecture and Environmental Planning

birgit.kleinschmit@tu-berlin.de

+49 30 314-72847