The Re-derivation of the NDBC Wind-Wave Algorithm

Palao, Ian M.; Gilhousen, David B.; ,

The Re-derivation of the NDBC Wind-Wave Algorithm

by Ian M. Palao, Computer Sciences Corp, John C. Stennis Space Cent, United States,
David B. Gilhousen, Computer Sciences Corp, John C. Stennis Space Cent, United States,

Document Type: Proceeding Paper

Part of: Ocean Wave Measurement and Analysis

Abstract:

The National Data Buoy Center (NDBC) currently manages a network of 57 moored buoys and approximately 7 Coastal-Marine Automated Network stations on off shore platforms which transmit wind speed, wind direction, and spectral wave energy data. As with all geophysical measurements, quality control of the data is necessary. In 1986, NDBC derived a wind-wave algorithm, based upon the relationship between the high-frequency wave energy and the windspeed, that identifies erroneous wave data. This algorithms flags wave data when the data fall beyond preset upper and lower limits for the corresponding wind speed. The algorithm performs well for most sea and atmospheric conditions, but it has limitations, especially during light winds and at stations with severe fetch limitations. Furthermore, an algorithm was needed that was more representative of the nonlinear wind/wave relationship. Therefore, NDBC initiated a study to develop a new method of quality control. Whereas the previous method compared an hourly wave energy value to linearly derived maximum and minimum values for all buoys, the new method computes a station-specific, long-term probability density distribution of acceptable energy values related to a given wind speed. This distribution is computed using atleast 6 years of hourly, buoy-measured wind speed and high-frequency wave energy observations. Current hourly pairs of wind speed and energy data are compared to the climatological `energy envelope' represented by the 1-percent contour line of the normalized, joint probability density distribution. Wind speed and wave energy observations falling outside this contour are identified as questionable. The data analyst then uses various meteorological products, such as weather maps and satellite images, to determine whether the data are in error. The result of this study clearly shows that the new method successfully models the nonlinearity of the wind/wave relationship which the previous algorithm could not. Because of its sensitivity to small changes in wave energy, the new procedure can also detect when incorrect empirical constants have been applied to the wave energy. This new approach will enable NDBC to expeditiously detect anemometer failures, wave system failures, or wave processing problems.

Return to search