We investigate how and when a nonstationary, inhomogeneous flow field with significant shear, such as that in the tropical Pacific Ocean, can be meaningfully decomposed into two distinct components: a large scale mean flow U and a turbulent flow u' that arises from the mesoscale eddy field. We propose a new methodology for making this decomposition, where appropriate, and for computing the horizontal eddy diffusivity when the decomposition is valid. Our method is based on obtaining an estimate of U(x) for each season with bi-cubic splines. The spline interpolation depends on the parameter of roughness, p, which controls the spatial resolution of the resultant field. We use simulations based on an ideal scenario (turbulence superimposed on a large-scale shear flow) to develop a metric for making a "good" choice of p. This methodology is used to make estimates of the mean flow and the diffusivity in the region 160°W-100°W, 13°S-3°S. This paper provides the practical framework required to make statistically reliable estimates of horizontal transport processes and lays the necessary foundation for extending the results from the California Current (Swenson and Niiler, 1996) to the equatorial Pacific. This work is presented in Bauer et. al. (1998).
Historical drifting buoy tracks and sea-surface temperature (SST) histories are used in conjunction with historical expendable bathythermograph (XBT) profiles to estimate the contribution to the mixed layer heat budget arising from horizontal advection and diffusion during the seasonal cycle in the cold tongue region for the period 1979-1995. We find three main results for the region [129°W-99°W, 4°S-4°N]. (a) We find that Qo, the net surface heat flux (Oberhuber, 1988), is too small to close the balance for each month from September through January. Assuming that the heat loss through the base of the mixed layer is at least 20 W m-2, we find that Qo is 80 ± 25 W m-2 too low during these five months (Figure 1). Of course, surface heat flux climatologies, numerical weather prediction surface flux estimates and satellite products exhibit large inconsistencies. Here we see an example of how the ocean measurements can help to constrain the estimates. (b) We find that both the zonal and meridional advection are weak in the spring during the peak of the warm phase in March and that both increase steadily until the peak of the cold phase in September. Although the zonal component of the advective heat flux is always less than the meridional component, it is not negligible in the latter half of the year. The annual march of the full horizontal advective flux (obtained from the Lagrangian drifters) follows the same seasonal pattern. The latter estimate includes the horizontal eddy fluxes that in this region are dominated by the effects of tropical instability waves (TIW), which induce a heat flux into the region (Hansen and Paul, 1984, JGR, 89, 10431-10440). The contribution of this process to the mixed layer heat budget is estimated as the difference between the two estimates of the horizontal advective heat flux. The TIW activity is a minimum during the spring and grows during the summer to reach a peak in activity from August to November, which is in agreement with assessments reached by other means (e.g., Enfield, 1986, JPO, 16, 1038-1054). (c) Vertical advection is dominated by the divergence of the meridional velocity throughout the year. This highlights the intimate connection between the divergent meridional flow and upwelling induced cooling. Preliminary results are presented in Swenson and Hansen (1998).
We seek to describe and diagnose the ocean-related processes which control the evolution of heat content in the mixed-layer (and therefore SST) for several regimes within the ITCZ/cold tongue region. We will also make a direct comparison with the NCEP ocean reanalysis to assess the ability of the OGCM to adequately simulate the observed processes. If this analysis suggests specific problems, we will identify them suggest improvements.
Data from satellite-tracked drifting buoys and VOS/XBT profiles for the years 1979-1995 were used to evaluate the seasonal cycle of how major oceanic processes redistribute heat in the cold tongue region of the tropical Pacific. The most active processes for the annual cycle are local heat storage and heat export by entrainment of upwelling and by mean meridional advection. Heat export by zonal advection, however, is not negligible, and meridional eddy heat fluxes associated with tropical instability waves effect a negative feedback that offsets a considerable fraction of that produced by the mean meridional advection. All of these processes mimic the essentially one cycle per year of the surface wind stress, as do those of the depths of both the bottom of the surface mixed layer and the thermocline. Because it is associated with poleward Ekman transports, upwelling, and baroclinic adjustment near the equator, the zonal wind stress component appears to be the more important. The meridional wind stress, while weaker in the annual mean, has a larger annual variation and, therefore, has equal influence on the annual variation of the scalar stress and perhaps the mixed layer thickness. The Monin-Obukov length is found to underestimate the mixed layer thickness very considerably. Finally, we produce the first estimates of the seasonal cycle of eddy heat flux convergence, which plays a significant role in the evolution of the cold tongue, and show that the eddy heat flux convergence can be quantitatively modeled as eddy diffusion with a diffusivity derived from single-particle buoy statistics. These results are reported in Swenson and Hansen (1998).
We use the extensive WOCE/TOGA data sets from drifting buoys and VOS/XBT measurements to quantify of climatological heat budget processes in the Equatorial cold tongue of the eastern tropical Pacific Ocean. One interesting application of the results is in their implication for the net surface heat flux. Results indicate that, in the cold tongue region, the heat flux climatology generated from the NCEP/NCAR Reanalysis Project is superior to other popular climatologies. In particular, the NCEP/NCAR products allow closure of the mixed layer heat budget while COADS-based climatologies do not (Figure 2). A major factor in the improvement appears to be the annual cycle of downward shortwave radiation (Figure 3). The manuscript for this work is in preparation.
Back to PACS homepage.
Page Last Modified: March 2, 1998