AI Climate

Enlarge

In the project, remote sensing methods are utilized to investigate the effects of climate change on forests in the cross-border region of the Bavarian Forest (DE) and Šumava (CZ) National Parks. Our team investigates trends in phenology and tree species composition, while the project partners determine changes in the biomass and health status of the tree populations. Phenological changes could, for example, mean earlier sprouting of leaves as a result of warmer spring periods. In the project, long-term photo series of individual forest sections are analyzed using deep learning algorithms to determine such changes over time. By combining this with optical satellite data, it is possible to find changes over larger areas. Changes in tree species structure can be detected on large spatial and temporal scales by analyzing spectral change in optical satellite data of the Sentinel 2 and Landsat platforms. Additionally, we aim to achieve a more fine-scale spatial resolution by analyzing long-term aerial image time series. With the help of AI object recognition, individual trees can be detected in these images and their species can be determined using machine learning methods with forest inventory data as a reference. The joint output of all project partners is a map of the most endangered forest areas, as well as a methodology for climate change vulnerability, which can be utilized for informed management decisions.