Abstract
The flux of heat through the Strait of Gibraltar is known well enough that the Mediterranean Sea may be used as a climate test basin. After adjusting reported winds for changes in observing practice, the COADS for 1946 to 1988 was used together with standard heat flux formulas to estimate the long-term mean heat flux into the sea, giving 36 W m&minsu;2 more than is compatible with the Gibraltar exchange. As the estimated latent heat flux is consistent with the freshwater budget, it is suggested that standard formulas overestimate insulation in the Mediterranean. If a constant adjustment factor is used for the insulation, or for the latent heat loss, interannual variability of ±15 W m−2 is found in the total heat flux. Changes in the latent heat flux dominate, with contributions from both the humidity of the air and the saturation humidity at the temperature of the sea surface. The buoyancy flux from the sea is also examined and shows that the contributions from precipitation and runoff are important for the long-term mean, but insignificant for seasonal and interannual variability.
