There is a newer version of the record available.

Published August 8, 2026 | Version v1.1

PINN-ECS-HAB: Physics-Informed Neural Network with Strong PDE Residuals and a Learnable Cyst-Germination Source for Forecasting Harmful Algal Blooms in the East China Sea (code )

  • 1. Zhejiang Ocean University

Description

v1.1 (corrected): missing satellite Chl-a pixels are now encoded as NaN with an explicit CHL_valid mask instead of a numeric zero fill; all experiments re-run and all tables/figures regenerated on schema_v1_1.npz; adds bilinear-resampled GOCI-II field and ships trained model checkpoints (PINN-full, ConvLSTM dense/sparse) plus the PINN-step3 ensemble prediction fields.

Files

pinn-ecs-hab-repro_code_doi_v1.1.zip

Files (22.7 MB)

Name Size
md5:fb8902ff3d10f3fa6106bab12447206e
22.7 MB Preview Download