In this paper the different results of estimated air-sea heat fluxes by different computational schemes of bulk formula are researched. The varying results may bring another source of the uncertainties of flux climato...In this paper the different results of estimated air-sea heat fluxes by different computational schemes of bulk formula are researched. The varying results may bring another source of the uncertainties of flux climatology. In addition to the classical scheme and the sampling scheme, a revised sampling scheme named semi-sampling scheme is proposed in this paper. In this revised version, the monthly means of (qs- qa)W and (Ts- Ta)W are calculated from each simultaneous measurement as that in the sampling method. But the transfer coefficients in this version are defined based on monthly means of W and instead of those in each measurement in order to use the files of monthly mean variables and in COADS (Woodruff et al. 1987).The results of a comparison study show that the covariance between (Ts - Ta )' and W' and that betWeen (qs- qa)' and W/ are less than 2 W / m2 and 5 W / m2 respectively in most areas, which are about one order of magnitude less than the standard deviation of the fluxes. Therefore the difference betWeen tWo schemes is not statishcally significant on average. However, in the regions off coast of Asian and North American continents, the contribuhon of covariance between (qs - qa )' and W' can be as large as 15 W / m2, which is about one third to a half of the standard deviation. Since there are files of monthly mean (qs - qa)W and (Ts- Ta)W for each year in the COADS. we propose to use these files directly in calculation of climatology. It is not necessary to use every individual report. The flux climatology calculated directly from the long-term mean values of basic meteorological elements shows an significant difference from that calculated from the mean of each year's flux, particularly for the latent heat flux which is on the order of 10 to 15 W / m2, about 3-4 times of the difference between semi-sampling and classical schemes and about one third of the standard deviation. Therefore it indicates that the time nit6rval of months is perhaps a better one in the calculation of flux climatology. It is not recommended to use the long--term means of the met6orological variables.展开更多
文摘In this paper the different results of estimated air-sea heat fluxes by different computational schemes of bulk formula are researched. The varying results may bring another source of the uncertainties of flux climatology. In addition to the classical scheme and the sampling scheme, a revised sampling scheme named semi-sampling scheme is proposed in this paper. In this revised version, the monthly means of (qs- qa)W and (Ts- Ta)W are calculated from each simultaneous measurement as that in the sampling method. But the transfer coefficients in this version are defined based on monthly means of W and instead of those in each measurement in order to use the files of monthly mean variables and in COADS (Woodruff et al. 1987).The results of a comparison study show that the covariance between (Ts - Ta )' and W' and that betWeen (qs- qa)' and W/ are less than 2 W / m2 and 5 W / m2 respectively in most areas, which are about one order of magnitude less than the standard deviation of the fluxes. Therefore the difference betWeen tWo schemes is not statishcally significant on average. However, in the regions off coast of Asian and North American continents, the contribuhon of covariance between (qs - qa )' and W' can be as large as 15 W / m2, which is about one third to a half of the standard deviation. Since there are files of monthly mean (qs - qa)W and (Ts- Ta)W for each year in the COADS. we propose to use these files directly in calculation of climatology. It is not necessary to use every individual report. The flux climatology calculated directly from the long-term mean values of basic meteorological elements shows an significant difference from that calculated from the mean of each year's flux, particularly for the latent heat flux which is on the order of 10 to 15 W / m2, about 3-4 times of the difference between semi-sampling and classical schemes and about one third of the standard deviation. Therefore it indicates that the time nit6rval of months is perhaps a better one in the calculation of flux climatology. It is not recommended to use the long--term means of the met6orological variables.
