U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications
Up to now, the individual contributions of Q, Q, Q and Q to the mixed layer temperature change have been estimated and no single term has been found which accounts for the observed variations of Q. Estimation of the remaining terms in equation (1), meridional advection and diffusion, require off-equatorial information. Such data are not available continuously throughout the study period ( Figure 1) because of deployment schedules and instrument failures. Coverage is best in boreal spring of each year and these records are discussed in the next section. First, however, we consider how well the sum of the terms available throughout the entire record can balance the heating.
Figure 6e shows the comparison of the sum of the surface heat flux, penetrative radiation, zonal advection, entrainment and vertical diffusion (Q = Q + Q + Q + Q) and the observed mixed layer heating (Q). Over the entire 29-month record the correlation coefficient between the two series was 0.71; limiting the record to 24 months in 1986-87 increases the correlation to 0.80. Both of these correlations are significant at the 95% level. The linear regression coefficient (1.1) between the two complete series indicates that the fluctuations in Q are about 10% larger than those observed in surface heating. The offset between the two series showed that the mean surface heating was about 30 W m larger than Q. Thus an additional heating source (perhaps meridional diffusion) is required to balance these terms.
Observed and computed heating agreed reasonably well in 1986. During the initial development of warm conditions in the eastern Pacific in August through November 1986 [MH], the mixed layer was relatively deep and was warming by up to nearly 100 W m. This heat was provided by the net solar radiation (which was enhanced at that time by the semiannual cycle in Q and the reduced Q associated with the deep mixed layer), and variations in all the oceanic terms. Changes in latent heat flux do not appear to contribute to this warming. The large increase in heating in mid-September 1986 was associated with reduced oceanic cooling as the zonal advection was near zero, vertical entrainment was reduced, and vertical diffusion was near its minimum.
The phase of the seasonal cycle from February through July 1987 was also rather well represented by the computed heat flux terms. Ignoring the rapid rise in January, the general spring warming in 1987 required about 50 W m in February. This heat was provided by the net solar radiation and reduced vertical diffusion. Reduced zonal advective cooling is nearly balanced by increased entrainment. The subsequent cooling in May-July appears to result primarily from a large increase in vertical turbulent diffusion out of the mixed layer. This diffusion and increased entrainment is sufficient to counteract the increased mixed layer heating associated with zonal advection.
Relatively large (>50 W m) discrepancies between the surface heating and the heat flux occur throughout the record although the phase of the two terms generally agree. In January 1987 the observed mixed layer heating was nearly 60 W m larger than the estimated heat flux into the mixed layer. In May 1987 the estimated cooling was nearly 100 W m too large. The period March-May 1988 also remains anomalous. At that time changes in heating and heat flux were out of phase. As the mixed layer cooled by about 10 W m (and SST dropped nearly 8°C), the heat flux was trying to warm the ocean by up to 50 W m. This period is considered in the context of boreal spring 1986 and 1987 in the next section.
Go back to previous section or forward to next section
PMEL Outstanding Papers
PMEL Publications Search