Doctoral position (m/f/d) in Hybrid Forest Modeling

Doctoral position (m/f/d) in Hybrid Forest Modeling

NEU

Teilzeit

Forschung und Lehre

The CRC 1537 - ECOSENSE invites applications for a doctoral researcher (m/f/d) position at the Chair of Biometry and Environmental System Analysis Doctoral position (m/f/d) in Hybrid Forest Modeling

Bewerbungsfrist: 15. April 2026

Veröffentlichungsdatum: 23. März 2026

Eintrittstermin: 1. Juli 2026

Arbeitsumfang: Part-time position (75 %)

Kennziffer: 00004901

Beschreibung

Who we are:

In ECOSENSE we explore new ways to quantify ecosystem processes and stress impact of climate change by developing novel sensor principles and measurement methods ( ). We established a comprehensive infrastructure and sensor network in a nearby forest including three canopy access towers.

For our second phase starting this July we are seeking 15 doctoral researchers from environmental and engineering sciences. ECOSENSE offers highly interdisciplinary research and training in a cooperative and vivid working environment.

The project C4 aims to identify hot moments and hot spots of forest dynamics within ECOSENSE. We focus on combining all measurements with process and empirical modeling by calibrating models to measured physiological process variables, environmental drivers, and ecosystem productivity. Furthermore, we intend to evaluate existing process models and improve them using our novel high-resolution measurements of many processes.

The position is located in the group of Biometry and Environmental System Analysis (Prof. Carsten Dormann), which closely collaborates with the groups for Biogeochemical Systems Modelling (Prof. Rene Orth) and Sensor-based Geoinformatics (Prof. Teja Kattenborn) in this project.

Your Task:

Within ECOSENSE, you will combine process knowledge, empirical data, and machine learning to improve dynamic forest modeling for prediction and inference. You will contribute to the development of a novel hybrid modeling framework (Forest Informed Neural Networks, FINN) by developing, training, and testing a hybrid forest growth model based on ECOSENSE data (Photosynthesis, Sap Flow, GPP, Volatile Organic Compounds (VOCs), soil variables, remote sensing, tree growth, etc.). Your main objective will be to balance process based understanding and empirical models (in particular Deep Neural Networks) to develop a model of tree growth across multiple scales. In addition, you will integrate your model in FINN (Pichler & Käber, 2025) to better understand the underlying mechanisms within the context of a dynamic forest model and compare it with the other models (e.g. QUINCY). We expect from our candidate the preparation of scientific publications of our results and methodological advancements.

For implementing the Deep Neural Networks, you will use the R-Package cito, which provides an user-friendly interface and only requires very little knowledge of Machine Learning. However, we expect a high level of motivation to familiarize with Machine Learning and ecological process modeling.

Your qualifications and contributions to ECOSENSE:

· You hold a Masters degree in environmental sciences, forest science, biology, computer sciences, or any other field related to modeling, machine learning, or ecology. If already available, please include your M.Sc. thesis in your application.

· Basic knowledge of tree physiology and plant ecology.

· Familiarity with fundamental concepts of forest dynamics such as succession trajectories and demographic processes.

· Some experience of process modeling, ideally with forest models, and model evaluation.