Speaker
Description
The reproduction of most crops and wild plant species depends on interactions with animal pollinators, which are declining globally due to climate change and human activities. Understanding changes in pollinator populations and plant-pollinator interactions is crucial for predicting and mitigating biodiversity loss. Traditional methods for collecting these data are time-consuming, costly, and require specialized expertise, posing integration challenges due to varying methodologies and scales.
Camera vision methods, utilizing machine learning and computer vision, offer a promising alternative for capturing, counting, and identifying pollinators. This study aims to train and test these methods to capture pollinator and plant-pollinator interaction diversity across different biogeographic contexts. The initial focus is on prioritizing plant species for monitoring with limited camera resources, using data from the EU project SafeNet. Three monitoring options were evaluated: monitoring the most common flowering plant species, plants with diverse floral shapes, or plant genera important for pollinators across Europe. Results indicate that monitoring the most common plant species captures the highest pollinator diversity.
The second focus addresses whether automated pollinator monitoring can detect changes in pollinator diversity and interaction network structures as effectively as traditional methods across different environmental gradients. This is being tested in the EU project SEPPI with the hypothesis: H0, that network metrics are consistent across methods and taxonomic grains once rarefied for interactions, or H1, that metrics differ if taxonomic identification influences the detection of rare species and biodiversity changes.
The final focus is on developing a model to identify hoverflies, crucial pollinators with diverse morphologies and color patterns due to Hymenopteran mimicry. Currently, no AI efforts target Dipteran pollinators. We aim to train and test a model for Syrphid species identification in Germany.
This research will enhance our ability to monitor pollinators and understand plant-pollinator interactions, contributing to biodiversity conservation efforts.
| Status Group | Doctoral Researcher |
|---|
