Science Journal | PelagicLabs

PelagicLabs integrates real-time data from 25+ buoy stations across the North Sea and Baltic, combining multiple national networks into a unified observation layer.

Data Sources

Rijkswaterstaat (RWS) - Netherlands

Stations: 8 wave buoys (Europlatform, K13, etc.)
Update frequency: 10 minutes
Parameters: Hs, Tp, direction, spectrum

Meetnet Vlaamse Banken (MVB) - Belgium

Stations: 5 buoys (Westhinder, Bol van Heist, etc.)
Update frequency: 10 minutes
Parameters: Hs, Tp, direction, water level

CEFAS WaveNet - UK

Stations: 6 North Sea buoys
Update frequency: 1 hour
Parameters: Hs, Tz, Tp, direction

Additional Networks

NDBC (US): 3 North Sea stations, hourly
DMI (Denmark): 4+ stations, 10-minute
Frost/MET Norway: 5+ stations, hourly
SMHI (Sweden): Multiple Baltic stations, hourly
Open-Meteo Marine API: Full basin grid coverage, hourly

Bathymetry Data

High-resolution depth data from EMODnet Digital Terrain Model 2024:

Resolution: 1/16 arc-minute (~115m grid)
Coverage: All European seas including North Sea
Features: Depth queries, profiles, Dogger Bank detection

Data Quality Control

Quality flags applied:
  - Range checks (Hs: 0-25m, Tp: 1-30s)
  - Temporal consistency (max Δ per hour)
  - Cross-validation with nearby stations
  - Spectral energy conservation checks

Confidence Decay

Data freshness directly impacts prediction confidence:

C_data = min(1, data_freshness / 24h)

< 1 hour: 100% confidence weight
1-6 hours: 95% confidence weight
6-12 hours: 80% confidence weight
12-24 hours: 60% confidence weight
> 24 hours: Flagged as stale

Update Pipeline

Polling interval: 5 minutes (high-frequency stations)
Processing latency: < 30 seconds from receipt
Storage: Time-series database with 90-day retention
Backup: Model-based gap filling for station outages