WRF

Expedition WRF weather simulations

Expedition WRF is best thought of as a system to add small scale thermal and land effects to the GFS forecast.

It is aimed at sailing locations where good high resolution model data isn't otherwise available (eg NZ).

Various areas are also available on demand for regattas, such as the Americas Cup, Sail GP, Olympics and Mediterranean sailing venues.

Data

MSLP, rain, 10m wind and gusts, 20m wind, 2m temperature and dew point, CAPE & total cloud cover. Others available on request.

Updated every 6 hours (4x daily), but up to 3 hourly for special events. Update times can be checked in Expedition before downloading.

Resolution

1/12° (0.083°, < 5nm), 1 hour increments to 72 hours

1/36° (0.028°, < 2nm), 30 minute increments to 36 hours

1/108° (0.009°, 1/2nm), 30 minute increments to 24 hours

Implementation

Expedition WRF uses GFS and other surface datasets for initialisation and boundary conditions. The 1/36° domains are similar in horizontal and vertical resolution to the HRRR.

Expedition WRF uses 45 to 60 levels, depending on the domain. WRF can be run with 35 or even fewer levels, but more is generally better.

Nesting inside GFS has two main benefits:

Merging WRF and GFS data is seamless.

Timeliness and skill level of the forecast is improved. We don't have NOAA's resources to run WRF or MPAS on a global domain, at least to the same level as the GFS.

Notes

WRF can over-estimate 10m winds, especially in stable conditions or where there is a lot of wind shear. From experience this implementation does especially well in sea breeze and thermal scenarios.

For example, Expedition WRF was very good during the 2021 Americas Cup and exceptionally good for sea breeze situations. Users rated it as the best of the available models. Similarly, it does very well in the Mediteranean.

What it isn't

It isn't a replacement for the national mesoscale models such as Access, Arome, Harmonie, HIRLAM, HRDPS, HRRR, ICON-DE and UM-UKV on the Exp grib server.

It also doesn't replace the global models for larger scale areas.