Impact of assimilation of wave spectra in NEDWAM

1. Introduction

Each wave model, from straightforward Bretschneider-table to sophisticated n-th generation numerical model, depends strongly on the quality of the wind input. As to KNMI operational wave model NEDWAM wind fields are provided by the atmospheric model HIRLAM. Though this model usually performs well, the wind field will always deviate somewhat from the reality. It is not always easy to assess the quality of wave models including NEDWAM, because of the extreme sensitivity for wind input errors. This is varying depending on the growth stage of the wind waves: in young sea and/or strong wind the wave height depends linearly on the wind speed, in developed sea this dependence becomes square. In most cases, conditions are between both extremes. This sensitivity is even much greater for low frequency wave height forecasting: the presence of waves with periods greater than 10 seconds requires wind waves with significant wave height over 4 meters, which will occur at 7 Beaufort or more. As soon as these waves are present, they immediately become very high in the storm area. Because of this the variability of the low frequency wave height is very large.

 

figure 1: the North Sea, with the two northerly stations moved to the South

 

In addition to the wind as the primary factor to forcing wave models, the assimilation of observed wave data is a second supplementary option, never replacing wind forcing. Since June 2000 NEDWAM is running operationally using a system for data assimilation, developed by Aart Voorrips (1998). Verification data show that the quality of NEDWAM predicions have been improved, for short term (+6 hours) for all parameters such as wave height, but as to northerly low-frequency swell also for somewhat longer terms (+12 hours or more). From +24-hour prediction the effect is disappearing. This is related to to the time for waves with periods over 10 seconds to propagate across the North Sea from the area near the Shetland Islands to the southern North Sea, see North-Cormorant (61°NB) and Auk (56°NB) on (modified) map in figure 1.

In chapter 2 some monthly statistics are presented over period July 1999 - February 2001, taken from the UVS database (Martin Stam, KNMI). In addition some series have been added which include in part data starting from 1996.

In chapter 3 two months are compared: April 2000, before introduction of data-assimilation, and July 2000, immediately afterwards. April and July were rather similar in the North Sea area, with high wind waves with some northerly swell, sometimes (almost) just swell. Moreover, all relevant wave sensors worked well, which allowed optimal performance of data assimilation. [In August and September 2000 one or more wave buoys were missing.]

Finally, in chapter 4 the usefulness for KNMI maritime forecasters of the introduction of assimilation of wave spectra is analysed. It is concluded, that the MMD benefits from the introduction of data assimilation of wave spectra , particularly in the occurrence of northerly swell, provided that real-time directional spectra of observations in the northern part of the North Sea are available. The forecasts of low-frequency wave height in the southern North Sea have particularly improved.

 

Literature


Evert Bouws
11 June 2003