POSTER SESSION 1
15:30–17:30, Monday, September 14
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1235 | POSTER 253
EXTENDING OCEAN COLOR INTO POLAR NIGHT: YEAR-ROUND RETRIEVAL OF PHYTOPLANKTON OPTICAL SIGNALS USING SPACEBORNE LIDAR AND DEEP LEARNING
Polar ocean color observations are fundamentally limited by low solar elevation, cloud cover, and extended polar night, resulting in substantial gaps in monitoring phytoplankton dynamics. To address this limitation, we develop a deep learning framework to retrieve phytoplankton-related optical signals from spaceborne lidar observations, enabling year-round coverage across polar oceans.
By integrating multi-mission lidar datasets from CALIOP and ICESat-2, we construct a long-term record and apply a two-branch deep learning model that combines waveform features with physically derived optical parameters. The model is trained using co-located MODIS observations and validated against independent in situ measurements, demonstrating robust performance across diverse polar conditions.
Our results show that lidar-based retrievals substantially improve winter observational coverage, capturing low-level phytoplankton optical signals under light-limited conditions where passive sensors fail. The reconstructed time series reveal coherent seasonal variability and enable the detection of persistent winter signals that are linked to subsequent phytoplankton dynamics.
This study demonstrates the potential of combining active remote sensing and machine learning to extend ocean color capabilities into polar night. The proposed framework provides a new pathway for year-round bio-optical observations and supports improved monitoring of rapidly changing polar marine environments.
Zhenhua Zhang, Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), [email protected], https://orcid.org/0000-0001-9818-5346
Peng Chen, Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), [email protected]
Siqi Zhang, Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
