Downward longwave and shortwave radiation fluxes were measured for 13 days in May 1988 in the equatorial west Pacific; estimates of net radiation fluxes from these data were checked against a net radiometer over 9 days. The pyrgeometer observations were corrected to match the net radiometer measurements at night, when the latter is substantially more accurate. These measurements were compared with various empirical formulae. Three widely used empirical formulae for net longwave radiation out of the ocean consistently underestimated our observations by up to 21¿5 W m-2; however, the Brunt-Budyko formula had no systematic bias. Two empirical formulae for the daily mean net shortwave radiation entering the ocean both overestimated the observations by up to 17 W m-2, although the scatter of daily estimates from observation was about ¿30 W m-2. Routine ship's meteorological observations were used to estimate turbulent heat fluxes, with four empirical formulae. The 18-day means differed by up to 28 W m-2, due to different treatment of the increase in bulk transfer coefficient at low wind speed.

The Liu et al. (1979) algorithm gave the highest fluxes, and agreed well with direct measurements of turbulent fluxes reported elsewhere. A further systematic error in the measurement of air-sea temperature difference accounted for about 11 W m-2. These systematic errors all tend to reduce the net heat flux into the ocean; their sum is as much as 60--75 W m-2 for some choices of empirical formulae. This tends to confirm the hypothesis of Godfrey and Lindstrom (1989) that ocean mixing and advection are too weak to carry a substantial heat flux away from the mixed layer of this region. Temperature profiles collected near 4¿S, 150¿E showed a clear warming from May 8 to May 20 in the top 20 m, equivalent to a heat storage rate of about 30 W m-2. Directly measured radiation fluxes were available for 10 days of this period of calm, sunny weather; turbulent heat fluxes were estimated for the same period with the Liu et al. (1979) algorithm. The net heat flux into the water estimated in this way averaged 38 W m-2. While the heat flux measurements were made as much as 400 km from the site of the heat storage observations, our observations provide some support, at least in calm weather, for the hypothesis that the net heat flux into this region is only of order 10 W m-2. ¿1991 American Geophysical Union