Validation of the Operational Wave Model WAM and SWAN-2009

The significant wave height (Hs) from the operational wave models at met.no WAM and SWAN, is validated against EnviSat Radar Altimeter (RA-2) and in-situ buoy observations for the year 2009. WAM is run at 50km, 10km and 4km resolution (WAM50, WAM10 and WAM4) and is forced with 10m surface winds from the numerical weather prediction model Hirlam to produce a 66 hour forecast. SWAN is a wave model employed at met.no to forecast waves in coastal regions and is run operationally for Trondheimsleia and Karmøy. It is forced with 10m winds from UM to produce a 36 hour forecast with a grid spacing of 500m. We find that the agreement between observed and modeled Hs is very good for all WAM models with a correlation exceeding 0.95. However, for higher waves, Hs is overestimated in WAM. When comparing WAM10 and WAM50 we find that the behavior of the two models are quite similar, but WAM10 performs better than WAM50. When increasing the resolution from 10km to 4km, small improvements are shown. This may be due to the fact that the available buoys are located offshore where the advantage of WAM4 can not be seen. When comparing WAM4 and SWAN for the coastal areas at one observation site at Karmøy, we find that SWAN performs better than WAM4. SWAN is run on a finer resolution than WAM4 and have better physics to solve waves in shallow waters. Secondly, three cases during January and October, with high winds and high sea state has been evaluated for WAM50 and WAM10. We find that the storms are well predicted in both models, but often with an overor undershoot in Hs, depending on the observation site.