Applicant

Dr. Ellen Oldenburg
Heinrich-Heine-Universität Düsseldorf
Department Biologie
Institut für Quantitative und Theoretische Biologie
Universitätsstraße 1
40225 Düsseldorf

Project description

Polar marine ecosystems face rapid environmental change, yet interactions between zooplankton and microbial communities, critical for food-web dynamics and carbon cycling, remain poorly understood across trophic levels. This project addresses this gap through comparative, cross-hemispheric analysis of plankton communities in the Arctic and Antarctic, integrating advanced image-based zooplankton monitoring with established microbial time-series data. The project comprises three interconnected work packages. WP1 develops and validates a deep-learning classifier (DeepUVP) for automated zooplankton identification from Underwater Vision Profiler (UVP) images. Building on our DeepLOKI framework, we train classification models using annotated datasets from Antarctic (AWI, BAS) sources, generating continuous, depth-resolved time series of zooplankton abundance from moored UVP deployments in both polar regions, the first standardized, cross-polar zooplankton dataset at high temporal resolution. WP2 integrates these zooplankton data with long-term microbial observations to reconstruct cross-trophic ecological networks. Using co-occurrence analysis, Convergent Cross Mapping (CCM) for causal inference, and Energy Landscape Analysis (ELA), we identify functional community modules, directional predator-prey interactions, and environmental tipping points. This approach establishes causal relationships between taxa and links community dynamics to physical drivers such as light availability, temperature, and sea-ice cover. The analysis reveals whether bottom-up or top-down controls dominate polar planktonic systems and identifies keystone taxa governing seasonal transitions. WP3 conducts comprehensive comparative synthesis of Arctic and Antarctic plankton communities. Applying identical network and stability analyses to both polar regions, we determine whether polar ecosystems operate under shared ecological principles driven by solar forcing or follow region-specific trajectories. Key questions include: Do both poles exhibit analogous seasonal "winter reset" dynamics? Are functional modules structurally similar despite taxonomic differences? How do stability landscapes and resilience differ between polar systems? The project leverages established collaborations with AWI, BAS, and the SPP network, ensuring access to rare, high-quality polar datasets. Our analytical toolkit, previously validated on Arctic microbial communities, is extended to include zooplankton and cross-trophic interactions, representing a major methodological advance. Expected outcomes include validated deep-learning tools for UVP image analysis, directed food-web networks revealing causal ecological interactions, and the first quantitative, bipolar comparison of plankton community stability and function. These insights will enhance predictive understanding of polar ecosystem responses to environmental change and contribute foundational knowledge to the SPP goals.

DFG Programme Infrastructure Priority Programmes

Term since 2026