by Refiz Duro (Austrian Institute of Technology), Hanns Kirchmeir (E.C.O. Institut für Ökologie Jungmeier), Anita Zolles (Bundesforschungs- und Ausbildungszentrum für Wald, Naturgefahren und Landschaft) and Günther Bronner (Umweltdata)

Changing climatic circumstances have a significant impact on forests: besides higher temperatures, more intense and frequent storms and drought spells affect forest growth. To see what the future is bringing, and to be able to deal with forest conservation and management, it is necessary to answer the question “how quickly do trees grow in an environment of climate change?” We take on a challenge to answer this question by integrating state-of-the-art data collection and AI-based methods.

Forests are the largest terrestrial sinks for carbon and some of the richest biological areas on Earth. Climate change is, however, affecting forests through increasing temperatures, changing precipitation patterns and the growing number of biotic and abiotic disturbances [1]. Saving forest ecosystems is thus one of the key measures to mitigate climate change and save biodiversity. To maintain and improve forest biodiversity and forest resilience to climate change, updated forest policies and forest management strategies are being developed and implemented in adaptive forest management. They all require up-to-date data of the forest conditions including the vitality and health of trees, as well as the ongoing tree growth (i.e., carbon sequestration).

Initially, tree and forest growth have been assessed mainly for economic reasons in order to build forest yield tables as simple “growth models” and a basis for improved forest management and taxation. Only within the past 50 years, improved tree and forest growth models have become available (e.g., [2]). They do not, however, allow for consideration of instantaneous changes in growth due to climatic extremes (e.g., drought, heat) and the changes of phenological patterns (i.e., length vegetation season).

Nowadays, new and more accurate measurement equipment has become available and new tree and forest characteristics have become the focus of forest and environmental science. The latest measurement equipment allows, for example, to assess intraday variation of tree radial growth (automatic dendrometers), to assess the water status of trees during a day (sap flow). This is used to characterise the tree and crown architecture including their habitat and microhabitat functions with terrestrial laser scanning in highest precision, or to assess tree health and vitality with airborne and satellite imaging / laser scanning. These new measurement techniques provide a huge amount of quantitative forest data, but their correct analysis and interpretation strongly requires advanced analytics to fully utilise the obtained data, achievable nowadays with the advanced probabilistic and machine learning approaches (e.g., neural networks).

In this sense, the AI4Tress project [L1] funded by the Austrian Federal Ministry for Climate Protection, Environment, Energy, Mobility, Innovation and Technology and the Austrian Research Promotion Agency’s (FFG) “AI for Green” initiative, has taken on the challenge of developing a predictive AI-based climate-sensitive tree-growth model, exploiting the availability of the data, and applying the optimal machine learning strategies. Multidisciplinary experts from forestry to data science come from 6 Austrian project partners: AIT Austrian Institute of Technology GmbH (coordinator, AIT), Bundesforschungs- und Ausbildungszentrum für Wald, Naturgefahren und Landschaft (BWF),  Umweltdata GmbH (UWD), Know-Center GmbH Research Center for Data-Driven Business & Big Data Analytics (KC), GeoVille Informationssysteme und Datenverarbeitung GmbH (GeoVille) and E.C.O. Institut für Ökologie Jungmeier GmbH (E.C.O.).

Figure 1: Top row: the basic concept of the AI4Trees project targeting development of a single tree growth model integrating climatic extremes and phenological patterns change. Bottom row: Data collection through TLS (left) providing a 3D image of the surroundings (middle), and dendrometer measuring tree growth (right).
Figure 1: Top row: the basic concept of the AI4Trees project targeting development of a single tree growth model integrating climatic extremes and phenological patterns change. Bottom row: Data collection through TLS (left) providing a 3D image of the surroundings (middle), and dendrometer measuring tree growth (right).

The project’s first phase (started April 2022) has been focused on collecting high-quality, in-field data directly from forest sites selected from the Europe-wide forest monitoring program (ICP-Forests), which has been providing high-quality data on the vitality and adaptability of trees, nutrient cycles, critical load rates, and water balance [3]. Leveraging these high-quality datasets, statements can be made about climate change, air pollution, biodiversity, and the overall state of single tree stands. The project partner BFW maintains six ICP-Forests core measurement sites, where for each site, there are 10 trees equipped with automatic dendrometers, delivering tree diameter data at every 60 (or even 15) minute intervals.

Dendrometers are basically highly sensitive measuring instruments that detect minuscule changes in a tree diameter (resolution of 1–5 micrometres), making them excellent to assess intraday variation of tree radial growth, which could be due to  instantaneous changes in growth due to environmental extremes over short time spans (e.g., heat or cold waves). Furthermore, terrestrial laser scanning (TLS) technique delivers single tree point-clouds not only allowing extraction of traditional tree features like diameters at different heights, tree height and crown dimensions, but also providing the possibility of statistical approaches for calculation of various metrics, e.g., point-cloud percentiles along the tree height and competition between neighbouring trees. Repeating TLS measurements on a regular basis allows to estimate tree growth within a short period (1–3 years) in point-cloud data, which is still a challenge. Some measurement runs have already been performed and the collected TLS data has been processed, and together with the dendrometer data give an excellent basis for the forthcoming activities targeting development of an explainable and predictive AI-based climate-sensitive single tree growth model.

The successful outcome of the project has the potential to empower response to minimise potentially harmful climate change impacts on modern societies in line with the UN Sustainable Development Goals.


[1] R. Seidl, M.-J. Schelhaas, W. Rammer, and P. J. Verkerk, “Increasing forest disturbances in Europe and their impact on carbon storage,” in Nat. Clim. Change, vol. 4, no. 9, pp. 806–810, Sep. 2014.
[2] H. Hasenauer, “Concepts Within Tree Growth Modeling,” in Sustainable Forest Management: Growth Models for Europe, H. Hasenauer, Ed. Berlin, Heidelberg: Springer, 2006, pp. 3–17.
[3] M. Lorenz et al., “Forest Condition in Europe,” Federal Research Centre for Forestry and Forest Products (BFH), United Nations Economic Commission for Europe Convention on Long-Range Transboundary Air Pollution, 2004.

Please contact:
Refiz Duro
AIT Austrian Institute of Technology GmbH, Austria
This email address is being protected from spambots. You need JavaScript enabled to view it.


Next issue: October 2024
Special theme:
Software Security
Call for the next issue
Image ERCIM News 132
This issue in pdf


Image ERCIM News 132 epub
This issue in ePub format

Get the latest issue to your desktop
RSS Feed