PredictWind has made significant improvements to their weather models. 10 years of development has culminated in the release of the most accurate iteration of the PWG & PWE forecast models to date. This means that the PredictWind models now give you more forecast confidence than ever before, making your time on the water safer and more enjoyable.
To validate the improvement, we undertook an intensive verification study using real time data from 77 weather stations around the globe. For each station, 1 month of one hour interval readings were compared to the forecasts based on the new PredictWind model versus the existing model. The results of the study show on average a 12% improvement using the new PredictWind model for wind speed, and a 7% improvement in Wind Direction. PredictWind Director Jon Bilger commented, “as a company we are always looking to provide our customers with the very best forecast data, and these model updates are a big step towards achieving that goal.”
Weather modelling is a specialised science with advanced mathematical equations. The time and cost in this development is significant. Details of the update are as follows:
- Improved and more efficient model code developed over the last 10 years
- New radiation code
- 50% more vertical levels in the model
- Non-hydrostatic code
- Smaller time step of the model
- More accurate topographical and land-water mask data is now used
- Updated and more detailed land-use and soil type
- More sophisticated boundary-layer mixing scheme
- High resolution observed Sea Surface Temperature (SST) data used over water
- Initial analyses are now interpolated to each grid
- Global 50km grid forecast now used to drive the 8km domains
- Improved model settings for treatment of Tropical Cyclones on the 50km global grid
Further comment from Bilger on the exciting update, “ we are now using 15 times more computing power to produce the forecasts with this updated model, which is an added expense but well worth it. The 1km resolution forecasts are now released 30 minutes earlier than they have been to date, which is all good news for PredictWind customers”.
Details of the validation study can be seen here.