EGU24-22189, updated on 11 Mar 2024
https://doi.org/10.5194/egusphere-egu24-22189
EGU General Assembly 2024
© Author(s) 2024. This work is distributed under
the Creative Commons Attribution 4.0 License.

AI-Driven Classification of X-band Radar Dual-Polarimetric Coherence Data for Mapping and Monitoring Penguin Colonies on Zavodovski Island, South Sandwich Islands

Nicole Richter1, Malte Schade1, Oliver Cartus2, and Tom Hart3
Nicole Richter et al.
  • 1RWTH Aachen University, Germany
  • 2GAMMA Remote Sensing, Switzerland
  • 3Oxford Brookes University, UK

Synthetic Aperture Radar (SAR) remote sensing has become indispensable for monitoring natural and anthropogenic hazards. Here we present a novel application of SAR technology in the ecological domain. Focusing on Zavodovski Island, home to the world's largest penguin colony, we address the challenges posed by limited optical satellite observations due to persistent cloud cover and rare ground surveys due to the extreme remoteness of the study site.

Our proposed approach involves the application of an AI-driven classification algorithm to high spatiotemporal resolution X-band radar interferometric coherence data. Despite the inherent limitations of optical observations, we showcase the potential of dual-polarimetric, high-resolution SpotLight SAR in measuring phenology and its effectiveness in detecting, mapping, and monitoring penguin colonies on Zavodovski Island. The study leverages temporally dense SAR data, utilizing the TerraSAR-X and PAZ satellite systems, and spatially high-resolution data gathered during two field campaigns in February and December 2023.

This research not only highlights the innovative use of SAR in ecological monitoring but also underscores the broader applicability of SAR technology in diverse domains. By contributing to the understanding of penguin colony dynamics, our study exemplifies the transformative impact of SAR remote sensing on ecological health indices. This contribution demonstrates the capabilities of SAR technology in addressing unique challenges and expanding its utility beyond traditional hazard applications.

How to cite: Richter, N., Schade, M., Cartus, O., and Hart, T.: AI-Driven Classification of X-band Radar Dual-Polarimetric Coherence Data for Mapping and Monitoring Penguin Colonies on Zavodovski Island, South Sandwich Islands, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22189, https://doi.org/10.5194/egusphere-egu24-22189, 2024.