Global Models
ECMWF
SPIRE
UKMO
PWE
PWG
GFS
- #1 for accuracy & global resolution 9km
- #1 for land-based forecasts
Regional Models
PWE1km
PWG1km
UKMO2km
NAM
HRRR
AROME
- #1 Coastal resolution of 1km
- #1 Accuracy out to 36 hrs
Outstanding accuracy
Our exclusive PWE & PWG forecasts are generated with proven technology and the highest resolution on the web.
Read about who uses our models to get an insight from the experts.
Powerful marine tools
In addition to forecasts, we offer the most powerful marine tools on the market to make your time on the water simpler, safer and more efficient.
Weather Alerts
Get notified when conditions are perfect
Departure Planning
Let us find you the best date to depart
Live Weather
Over 20,000 live observations
Weather Routing
Optimal routes every time
The world’s best global models
We offer the leading global models so you can have complete confidence in the forecast.
ECMWF
The European Centre for Medium-Range Weather Forecasts model is highly regarded by meteorologists around the world. ECMWF’s HRES model consistently rates as the top global weather model and offers a record breaking 9km resolution. This data has a very high acquisition cost, and is not available to most weather providers.
- Resolution: 9km
- Range: 10 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
SPIRE
Spire’s satellite constellation uses radio frequency sensors to create a 3D profile of atmospheric temperature, pressure, and humidity over 10,000 times per day. Spire outputs a global resolution of 12km with an advantage in forecast accuracy for remote locations.
- Resolution: 12km
- Range: 10 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
UKMO
The UK Meteorological Office UKMO model, is used by many Met Offices and weather companies around the world. UKMO has a 15km resolution model output and studies show that it is on par with ECMWF model for offshore forecasts, and slightly behind the ECMWF & Spire models for land-based forecasts.
- Resolution: 15km
- Range: 7 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
GFS
The Global Forecast System model from NCEP is the most widely used data source as it is free of cost. Whilst it is a good model, it has a lower accuracy score than the models above. GFS runs at a 13km resolution, but is only output at 27km globally.
- Resolution: 25km
- Range: 10 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
PWE
PredictWind creates two exclusive global models. The first is PWE, which uses the ECMWF global initial conditions for the model run. It offers exceptional short term accuracy.
- Resolution: 8km
- Range: 7 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
PWG
PWG is PredictWind’s second exclusive global model, which uses the NCEP (GFS) global initial conditions for the model run. As with PWE, the PWG model offers exceptional short term accuracy.
- Resolution: 8km
- Range: 7 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
Why only six global models?
Our goal is to provide the very best data, while not overwhelming you with every available weather model. Many weather sites try to impress with a long list of models but we choose quality over quantity.
For example, the CMC (GEM), Arpege and NavGeM models are good, but are not as acurate as our 6 models. Similarly, the GEFS (Ensemble GFS) model has value but only if you are looking at forecasts beyond 10 days.
High-resolution regional models
Our regional models provide short range, high resolution forecasts for selected regions around the world. Find out what resolutions are available in your area by viewing our Global Coverage
PWG 1km
Our exclusive PWG high resolution model uses the NCEP (GFS) global initial conditions for the model run. PWG offers exceptional short term accuracy and a record breaking resolution of 1km, capable of displaying land and thermal effects.
View the Evaluation Report
- Resolution: 1km
- Range: 1.5 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
PWE 1km
Our exclusive PWE high resolution model uses the ECMWF global initial conditions for the model run. PWE offers exceptional short term accuracy and a record breaking resolution of 1km, capable of displaying land and thermal effects.
View the Evaluation Report
- Resolution: 1km
- Range: 1.5 Days
- Time step: 1 Hour
- Update frequency: 12 Hourly
UKMO 2km
The UK Meteorological Office UKV or UKMO 2km model is a new high resolution model covering the entire UK and Ireland. It offers excellent short range accuracy and is detailed enough to display geographical effects, such as funnelling, thermals and wind shadows
- Resolution: 2km
- Range: 5 days
- Time step: Hourly for days 1 & 2, then 3 hourly
- Update frequency: Hourly for 0-12h , 6 hourly for 13-48h, 12 hourly for 48h+
HRRR
HRRR by NOAA is a real-time model covering the USA with a resolution of 3km. The model is updated hourly and is supplimented with radar data assimilation every 15 minutes, refining the forecast for even greater accuracy.
- Resolution: 3km
- Range: 18 hours
- Time step: 1 Hour
- Update frequency: Hourly
AROME
Arome by Meteo-France was designed to improve short range forecasts of severe events across Europe, such as severe storms, fog and urban waves. It outputs a resolution of 1.25km and is highly regarded by top racing navigators as it beats the ECMWF forecast.
- Resolution: 1.25km
- Range: 42 Hours
- Time step: 1 Hour
- Update frequency: 00z, 03z, 06z, 12z, 18z
NAM
NAM is one of NOAA’s major weather models, covering most of North America. NAM is a mesoscale model, capable of modelling land and other features at a higher resolution than global models, leading to improved forecast accuracy.
- Resolution: 5km
- Range: 84 Hours
- Time step: 1 Hour
- Update frequency: 6 Hourly
Wave Models
WaveWatch III
The PredictWind wind forecasts, PWG and PWE, are used to drive the WaveWatch III wave model which produces an accurate 50km resolution wave forecast.
WAM
The WAM model from ECMWF is run at an incredible 14 km resolution globally, and is currently rated the best wave model from any national weather centre.
Hour by hour forecasts
Most marine forecasts give you a generalised forecast in the morning or the afternoon. We show wind and weather updates every hour, with accurate, highly detailed weather maps for your entire local area.
Hour by hour forecasts
Most marine forecasts give you a generalised forecast in the morning or the afternoon. We show wind and weather updates every hour, with accurate, highly detailed weather maps for your entire local area.
Our Process
1
Condition Files Entered in Model
The initial conditions files are entered into the model. The files can be comparable to a 'photographic' snapshot that contains the current state of the Earth's atmosphere. PredictWind uses two sources for these files (from the USA and Canada) so you have a comparative result to give you confidence in the forecast.
2
50km Forecasts Generated
Imagine dividing the atmosphere into 50km x 50km square grids over your local area. These initial conditions files step forward in time to create a 5 day forecast. For each 50km x 50km square region the model uses a complex equation to calculate the meteorological fields such as wind, temperature, humidity, pressure, rain and their interaction.
3
8km Forecasts Generated
This process is then repeated for the 8km model run, though it is 'nudged' by the 50km model run.
4
1km Forecasts Generated
Finally the 1km resolution model is completed, which is 'nudged' by the 8km model run. Significant computing resources are required to run the model in this configuration for all the PredictWind locations around the world.
5
Maps & Graphs Generated
Wind forecast maps and graphs and text are generated for the 20,000 worldwide forecast locations.
6
Repeat Every 12 Hours
This process is repeated automatically every 12hrs for both initial sources (i.e. 4 forecasts every day). There is no historical data involved, so each 12hr forecast is independent from the previous.
Model Physics
All weather models use many complex equations and different physics to calculate meteorological data such as wind speed, wind direction, temperature, pressure, clouds and rain. The way that the equations are encoded into computers and the set-up parameters may also differ.
PredictWind uses the open source Conformal Cubic Atmospheric Model (CCAM) to generate PWG/PWE weather forecasts based on different initial conditions and domains that are
PredictWind uses the open source Conformal Cubic Atmospheric Model (CCAM) to generate PWG/PWE weather forecasts based on different initial conditions and domains that are
tailored by PredictWind. Unlike most atmospheric models, CCAM employs a variable resolution grid which is focused with higher resolution over an area of interest (down to 1 km) but also has lower resolution elsewhere to still simulate large-scale atmospheric behaviour. This approach helps to better use computational resources as well as avoiding errors or inconsistencies at simulation boundaries which are associated with more traditional downscaling methods.