Bulk parameterization of air-sea fluxes: Updates and verification for the COARE algorithm

In 1996, version 2.5 of the Coupled Ocean-Atmosphere Response Experiment (COARE) bulk algorithm was published, and it has become one of the most frequently used algorithms in the air-sea interaction community. This paper describes steps taken to improve the algorithm in several ways. The number of iterations to solve for stability has been shortened from 20 to 3, and adjustments have been made to the basic profile stability functions. The scalar transfer coefficients have been redefined in terms of the mixing ratio, which is the fundamentally conserved quantity, rather than the measured water vapor mass concentration. Both the velocity and scalar roughness lengths have been changed. For the velocity roughness, the original fixed value of the Charnock parameter has been replaced by one that increases with wind speeds of between 10 and 18 m s(-1). The scalar roughness length parameterization has been simplified to fit both an early set of NOAA/Environmental Technology Laboratory (ETL) experiments and the Humidity Exchange Over the Sea (HEXOS) program. These changes slightly increase the fluxes for wind speeds exceeding 10 m s(-)1. For interested users, two simple parameterizations of the surface gravity wave influence on fluxes have been added (but not evaluated). This new version of the algorithm (COARE 3.0) was based on published results and 2777 1-h covariance flux measurements in the ETL inventory. To test it, 4439 new values from field experiments between 1997 and 1999 were added, which now dominate the database, especially in the wind speed regime beyond 10 m s(-1), where the number of observations increased from 67 to about 800. After applying various quality controls, the database was used to evaluate the algorithm in several ways. For an overall mean, the algorithm agrees with the data to within a few percent for stress and latent heat flux. The agreement is also excellent when the bulk and directly measured fluxes are averaged in bins of 10-m neutral wind speed. For a more stringent test, the average 10-m neutral transfer coefficients were computed for stress and moisture in wind speed bins, using different averaging schemes with fairly similar results. The average (mean and median) model results agreed with the measurements to within about 5% for moisture from 0 to 20 m s(-1). For stress, the covariance measurements were about 10% higher than the model at wind speeds over 15 m s(-1), while inertial-dissipation measurements agreed closely at all wind speeds. The values for stress are between 8% (for inertial dissipation) and 18% (for covariance) higher at 20 m s(-1) than two other classic results. Twenty years ago, bulk flux schemes were considered to be uncertain by about 30%; the authors find COARE 3.0 to be accurate within 5% for wind speeds of 0-10 m s(-1) and 10% for wind speeds of between 10 and 20 m s(-1).