The Journal
Lab notes, validation write-ups, and methodology updates. We show our work. Proof over slogans.
Architecture: Physics-Informed Hybrid ML Models
How we combine physics-based wave propagation with XGBoost and Random Forest to achieve RMSE < 0.15m while maintaining interpretability.
Audit: Validating North Sea Gadoid Behavior
Comparing PelagicLabs algorithmic outputs against 15 years of ICES catch data. Result: 91% alignment.
Methodology: Deep Water Wave Physics & Swell Propagation
The dispersion relation, group velocity, and energy flux equations driving our swell arrival predictions.
Study: Dogger Bank Swell Attenuation Model
How a 25-30m sandbank filters Atlantic swells before reaching Belgian beaches, and our depth-ratio attenuation algorithm.
Methodology: Thermal Stratification Models
How we calculate the oxy-thermal squeeze for freshwater predators using remote sensing and local sensors.
Methodology: Spectral Partition Analysis
Multi-modal sea state decomposition: separating primary swell, secondary swell, and wind sea for quality predictions.
Audit: Species-Specific Thermal Response Validation
Gaussian thermal preference indices calibrated against peer-reviewed ecology data for 12 target species.
Study: Solunar Period Correlation Analysis
84% correlation between predicted major periods and reported catch times across 2,400 logged sessions.
Study: North Sea Swell Pattern Classification
Four distinct swell archetypes: Scottish Low, Skagerrak Wrap, Atlantic Filtered, and Local Wind Sea, and their surf potential.
Methodology: Freshwater Prey Behavior Database
23 prey species with regionally-calibrated activity windows, depth patterns, and seasonal emergence timings.
Technical: Real-Time Buoy Network Integration
25+ stations across RWS, MVB, CEFAS WaveNet, DMI, and SMHI, with 10-minute to hourly update frequencies.
New validation reports and methodology updates published regularly.
Questions about our methods? science@pelagiclabs.io