Urban tree health assessment by using bi-temporal airborne laser scanning

Urban trees are key components of cities’ green infrastructure and sustainable urban planning. Healthy trees provide essential ecosystem services, including air purification, temperature regulation, and biodiversity enhancement. However, trees in urban environments face many stressors such as heat, air pollution, road salt, and limited growing space. Many of these stressors are expected to become more pronounced under climate change. Therefore, cities need efficient methods to assess the health of their trees.  

Remote sensing techniques, such as multitemporal airborne laser scanning (ALS), provide detailed three-dimensional information on tree structure. However, most research has focused on forests, while urban trees have not been studied to a similar extent. In this study, we investigate the potential of multitemporal ALS data to assess urban tree health. By analyzing changes in tree height and crown growth over time, tree health can be inferred using physiological principles, as trees under stress photosynthesize and grow less efficiently than those growing under favorable conditions.  

We used ALS data collected across the entire city of Helsinki during the summers of 2015 and 2024. For each tree, height and canopy area growth were calculated over a nine-year period using a traditional watershed segmentation method, and growth indices were then calculated for each tree by size class and species. ALS-derived tree metrics were integrated with an open geospatial tree register containing information on more than 55,000 urban trees, including diameter at breast height and species. Field reference data from 1,119 visually assessed trees were used to evaluate the accuracy of the ALS-based tree health estimates. Relationships between ALS-derived tree growth metrics and field-based health scores were analyzed using correlation analysis and statistical modelling to assess method performance.  

The results indicate a strong correspondence between ALS-derived growth indices and field-based reference data. Our model performed particularly well in identifying declined trees, with especially strong performance for young and mid-sized trees. Together, these findings demonstrate the potential of ALS data for assessing urban tree health and supporting practical, evidence-based urban planning and decision-making.