Developing (machine learning) methods to derive ecosystem functional properties
Max Planck Institute for Biogeochemistry
Ecosystems are key mediators of global carbon-, water- or nutrient cycles. In times of global environmental change, we need to gain a profound understanding of these ecosystem functional properties. However, ecosystems differ fundamentally in their “Ecosystem Functional Properties”, for instance in their maximal photosynthetic capacity amongst others which are certainly related to their Plant Functional Traits (and their diversity). Although we have a basic understanding of these relations, it still remains a grand challenge to measure “Ecosystem Functional Properties” at large spatial scales. New remote-sensing methods may open new possibilities in this context. This PhD project will attempt to implement an automated approach to relate “Ecosystem Functional Properties” derived from in-situ observations to remote sensing information at different spatial scales. The idea is to use data -driven approaches i.e. use machine-learning techniques. Specific objectives are
Using solar induced fluorescence measurements, multi-platform hyperspectral and LiDAR information at different spatial and temporal resolution.
Upscaling local data to larger spatial scales.
Interpreting the relevance of the derived Ecosystem Functional Properties.
But the project offers a lot of space for following other ideas, e.g. the question if we can derive an index of “spectral diversity” as proxy for “Ecosystem Functional Properties”. Ultimately, we seek to link the results to simple ecological models.
The PhD student will spend a visiting period of at least 3 months at the University of Twente, Faculty of Geo-Information Science and Earth Observation under the Supervision of Prof. Dr Christiaan van der Tol..
The activity will be supervised by Dr. Miguel Mahecha.