invaPICS – Assessing the phenological niche of invasive species through automated image classification of citizen science photo observations
ARPHA Conference Abstracts · 17.08.2026 · 9
DOI: 10.3897/aca.9.e202212 ↗Kurz gefasst
Fachartikel in der Zeitschrift ARPHA Conference Abstracts, online erschienen am 17.08.2026. Verfasst von Yves Philippe Klinger und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Plant invasions are a major driver of global biodiversity loss, yet the role of phenology, i.e. the timing of key biological events such as budburst, flowering and fruiting, remains an underexplored aspect of invasion ecology. Citizen science photo observations offer a powerful tool to advance phenological research and can help to bridge spatial, temporal, and taxonomic gaps in existing datasets. Coupled with recent advances in computer vision and deep learning, these observations enable rapid and accurate classification of large-scale phenological data, allowing to investigate the drivers and consequences of phenological patterns in invasive species. The invaPICS project (2025–2028) aims to deepen our fundamental understanding of phenology in plant invasions by integrating photo observations with cutting-edge classification and modeling techniques. The project is structured around three objectives: First, a standardized workflow for the phenological classification of citizen science photos will be developed. Different Deep Learning model architectures will be evaluated to identify the most robust approach for classifying phenological stages, yielding a generic phenology model that can be applied across species and scales in subsequent analyses. Second, model species will be selected based on predefined ecological and phenological traits. Their observations will be classified to analyze spatio-temporal patterns of phenological events across native and invaded ranges. This analysis will also explore the environmental factors driving phenological development in both contexts, providing insights into how invasive species adapt to new conditions. Third, the project will investigate the relative phenological niche of invasive species by leveraging vegetation databases to identify co-occurring species in both native and invaded ranges. Using the classification model developed, the timing of key phenological events will be compared between invaders and co-occurring species, shedding light on potential competitive advantages or niche shifts. The findings will not only enhance our theoretical understanding of plant invasions but also inform evidence-based management strategies to mitigate their ecological impacts.
Zitieren
APA
Klinger, Y. P., Ripa, R., Lozano, V., Eckstein, R. L., Welk, E., & Kleinebecker, T. (2026). invaPICS – Assessing the phenological niche of invasive species through automated image classification of citizen science photo observations. ARPHA Conference Abstracts, 9. https://doi.org/10.3897/aca.9.e202212
Vancouver
Klinger YP, Ripa R, Lozano V, Eckstein RL, Welk E, Kleinebecker T. invaPICS – Assessing the phenological niche of invasive species through automated image classification of citizen science photo observations. ARPHA Conference Abstracts. 2026;9. doi:10.3897/aca.9.e202212
BibTeX
@article{Klinger2026,
author = {Klinger, Yves Philippe and Ripa, Ramiro and Lozano, Vanessa and Eckstein, Rolf Lutz and Welk, Erik and Kleinebecker, Till},
title = {invaPICS – Assessing the phenological niche of invasive species through automated image classification of citizen science photo observations},
journal = {ARPHA Conference Abstracts},
year = {2026},
volume = {9},
doi = {10.3897/aca.9.e202212}
}
EUVIMED