Main Content

TreeM Project C2: Leaf characteristics, metabolome, and herbivory

The question of the impact of abiotic factors on the health of forest tree species and biotic interactions, such as the leaf microbiome and herbivores and pathogens, is of particular importance in the face of climate change. Understanding these associations requires insights from experimental studies and approaches that allow potentially relevant mechanisms to be scaled up to variations within and between trees in the field. As part of the LOEWE-priority program (C2), we will conduct lab experiments with in-depth analyzed leaves from the field to quantify herbivory and test for effects of different metabolic and morphological as well as spectral characteristics. In addition, we will develop an automated image analysis pipeline to assess leaf traits, metabolome, and herbivory for testing experimentally supported effects in the field as well as to monitor changes through time.

AI-based image analysis workflow for Oak leaves. a) Pre-processing (thresholding and rescaling) of scanned leaves for background removal, b) Segmentation output of a Convolutional Neural Network trained with eight semantic classes, c) Extracted regions by leaf damage type, d) Extracted regions by nervure type, e) Post-processing image analysis for select geometric leaf traits. Unit are pixel values before back-transformation through the individual scaling factor.

PIs: Prof. Dr. Nina Farwig and Dr. Stefan Pinkert

Team: Dr. Andry Ny Aina Rakotomalala and Tobias Müller

Project homepage: htps://

Project funding: LOEWE-Schwerpunkt Programm - HMWK