Wälder aus dem All: Die Umgebung von Darmstadt aus Sicht der europäischen Copernicus-Sentinel-Satelliten (Foto: ©ESA)

Risikoerkennung

Satelliten als Retter der Wälder

Worldwide, forests are the habitat for most amphibian, bird and mammal species, but also for around 300 million people. A further 1.6 billion people earn their living from forests. Forests have numerous ecological functions: They regulate the water balance, protect against erosion, store carbon – the list goes on. The benefits of this ecosystem are estimated at over 100 trillion euros per year. This is more than the entire global gross domestic product.

Although forests can adapt to long-term changes in environmental conditions, they are vulnerable to sudden disturbances such as heavy pest infestations, fire and storms. Climate change is increasing these risks: Insect pests such as the bark beetle are multiplying more, drought favours forest fires and severe storms are becoming more frequent.

Early action can minimise risks and reduce the economic and ecological damage. In the EU project Satellites for Wilderness Inspection and Forest Threat Tracking, SWIFFT for short, scientists at L3S are working with European partners to develop a holistic forest monitoring system that recognises impending dangers with the help of satellite images. SWIFFT aims to provide forest managers with low-cost, simple and effective remote sensing tools that are supported by powerful machine learning models. The identified risks are clearly labelled and displayed on overview maps.

The satellite data is provided by Copernicus, the European Union’s earth observation programme. The large amounts of data and the rise of artificial intelligence make amazing things possible: “Thanks to openly accessible satellite data in combination with the latest AI algorithms, we are able to regularly analyse the entire forests of Europe in a short space of time,” says Sven Ysker, who works on assessing the risk of forest fires at L3S.

The new tools are tested under real conditions and optimised in a targeted manner. This is why several end users from the forestry industry are also involved in the project, including the French Groupe Coopération Forestière, an association of forestry cooperatives that manage 2.1 million hectares of private forest. Rīgas meži is also involved. The municipal company manages forest land, parks and gardens in the Latvian capital Riga.

As soon as the overview maps are finalised and the areas with windthrow damage, insect infestation and fire risk are shown in detail, you can get started: Forest managers can use the tools to recognise risks, intervene in good time and allocate resources efficiently. The system is expected to monitor and protect up to 39.6 million hectares of forest by 2030. This could save the forestry industry up to half a billion euros in monitoring costs.

„Mit den nachhaltigen, effektiven und kostengünstigen Waldbewirtschaftungsinstrumenten von SWIFTT wird Europa in der Lage sein, den Klimawandel effektiver zu bekämpfen und die Artenvielfalt durch gesündere Wälder zu bewahren“, sagt Ysker.

Vorgestellte Projekte
Kontakt

Prof. Dr. Jörn Ostermann

Jörn Ostermann ist Mitglied des L3S und leitet das Institut für Informationsverarbeitung der LUH. Er forscht auf dem Gebiet der Signalverarbeitung mit Fokus auf Video-, Sequenzierungs- und Audiodaten.

Dipl.-Ing. Sven Ysker

Sven Ysker arbeitet im Bereich Machine Learning mit dem Forschungsschwerpunkt der Risikoabschätzung. Er beschäftigt sich zur Zeit mit der Waldbrandrisikovorhersage mittels künstlicher Intelligenz.