Oceanum Logo



Oceanum Baltic Sea ECMWF Wave Forecast

February 2025

   
Model SWAN 41.31
Forecast horizon 7 days
Spatial resolution 0.05 degree (~5 km)
Temporal resolution 1 hourly
Region 9E - 30.3E, 53.8N - 66N
Forcings ECMWF winds, sea ice, and Oceanum spectra
Update frequency 12-hourly

Dataset description

The Baltic Sea wave forecast dataset provides operational wave predictions across the entire Baltic Sea basin, including the Gulf of Bothnia, Gulf of Finland, Gulf of Riga, and the Danish Straits (Figure 1). The domain encompasses the coastal waters of Sweden, Finland, Estonia, Latvia, Lithuania, Poland, Germany, and Denmark. Wave forecasts are produced using the SWAN (Simulating WAves Nearshore) third-generation spectral wave model, with a 7-day forecast horizon updated every 12 hours (00, 12 UTC).

Wind forcing is provided by ECMWF IFS Open Data global atmospheric model. Sea ice concentration is prescribed from GFS to account for wave attenuation in ice-covered regions during winter months. Spectral boundary conditions are supplied by the Oceanum Western Europe ECMWF wave forecast. Bathymetry is derived from the GEBCO 2024 global bathymetric grid.

The modelling setup employs the ST6 source term parameterisations. Spectra are discretised into 36 directional bins and 43 frequency bins, covering a frequency range from 0.037 to 1.98 Hz with 10% logarithmic increments, extending to higher frequencies to capture the short-period wind waves characteristic of the enclosed Baltic Sea. The model features a regular grid with a 5 km (0.05 degree) resolution.

The dataset provides hourly forecast estimates for key ocean wave parameters (Table 2) including spectral quantities integrated over the full spectrum and for spectral partitions. Partitions are defined from an 8-second split (sea/swell) and from the Watershed method, which identifies one wind-forced partition and up to two swell partitions. Forecasts are archived for 30 days, and frequency-direction wave spectra are available at 1610 sites distributed across the domain. Nowcast datasets are also available, constructed by retaining the most recent data from each forecast cycle to provide a continuous near-real-time historical record.

Figure 1

Figure 1. Mean significant wave height from the Baltic Sea hindcast domain (used for forecast validation). The locations of 2D spectra hourly output are shown by the black dots.


Validation

The wave model physics and calibration have been validated against satellite altimeter observations for the corresponding hindcast domain. Validation results are available through the Oceanum Hindcast Satellite Validation App, which provides density scatter plots, quantile comparisons, and statistical metrics for the Baltic Sea region.


Data description

Table 1. Data description.

Field Value
Title Oceanum Baltic Sea wave forecast
Institution Oceanum
Access Oceanum Datamesh
Source SWAN 41.31A
Source terms ST6
Forecast horizon 7 days
Update frequency 12-hourly (ECMWF)
Archive period 30 days
Temporal resolution 1 hourly
Spatial coverage [9E, 53.8N, 30.3E, 66N] at 0.05 degree
Spectra output sites 1610
Frequency discretisation 43 frequencies between 0.037 - 1.98 Hz at 10% logarithmic increments
Direction resolution 10 deg
Bathymetry GEBCO 2024 Grid
Winds ECMWF IFS
Sea ice NOAA GFS
Boundary Oceanum Western Europe ECMWF wave forecast

Linked Datamesh datasources

ECMWF-forced (12-hourly updates):

Nowcasts (continuous near-real-time archive):


Gridded output parameters

Integrated wave parameters are stored hourly over the domain at the native model resolution. Table 2 describes the gridded output parameters.

Table 2. Gridded output parameters.

Variable names link to the corresponding NERC Vocabulary Server standard name where available. All parameters are defined on the time, latitude and longitude coordinates.

Variable Long Name Units
depth depth below sea surface m
dpm mean direction at the spectral peak of wind and swell waves degree
dpmsea mean direction at the spectral peak of wind waves below 8 seconds period degree
dpmswe mean direction at the spectral peak of swell waves above 8 seconds period degree
dspr directional spreading of wind and swell waves degree
fspr normalised width of the frequency spectrum of wind and swell waves -
hs significant height of wind and swell waves m
hsea significant height of wind waves under 8 seconds period m
hswe significant height of swell waves above 8 seconds period m
icec sea ice cover -
pdir0 mean direction of wind waves (partition 0) degree
pdir1 mean direction of primary swell waves (partition 1) degree
pdir2 mean direction of secondary swell waves (partition 2) degree
pdspr0 directional spreading of wind waves (partition 0) degree
pdspr1 directional spreading of primary swell waves (partition 1) degree
pdspr2 directional spreading of secondary swell waves (partition 2) degree
phs0 significant height of wind waves (partition 0) m
phs1 significant height of primary swell waves (partition 1) m
phs2 significant height of secondary swell waves (partition 2) m
ptp0 peak period of wind waves (partition 0) s
ptp1 peak period of primary swell waves (partition 1) s
ptp2 peak period of secondary swell waves (partition 2) s
pwlen0 mean wavelength of wind waves (partition 0) m
pwlen1 mean wavelength of primary swell waves (partition 1) m
pwlen2 mean wavelength of secondary swell waves (partition 2) m
tm01 mean absolute wave period of wind and swell waves from the first frequency moment s
tm02 mean absolute wave period of wind and swell waves from the second frequency moment s
tps smooth relative peak wave period of wind and swell waves s
tpssea smooth relative peak wave period of wind waves below 8 seconds period s
tpsswe smooth relative peak wave period of swell waves above 8 seconds period s
xwnd eastward component of wind velocity m/s
ywnd northward component of wind velocity m/s

Spectra output

Frequency-direction wave spectra are stored hourly at the spectra output sites within the domain. Table 3 describes the spectra output variables, using the exact variable names served by Datamesh.

Table 3. Spectra output variables.

Variable names link to the corresponding NERC Vocabulary Server standard name where available. Spectra are defined on the time, site, freq and dir coordinates; lon and lat are per-site data variables giving each site’s location.

Variable Long Name Units
efth sea surface wave variance spectral density m² s / deg
dpt water depth m
wspd wind speed m/s
wdir wind direction degree
lat latitude degrees_north
lon longitude degrees_east

www.oceanum.science