U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications

The upper ocean heat balance in the western equatorial Pacific warm pool during September-December 1992

Meghan F. Cronin and Michael J. McPhaden

Pacific Marine Environmental Laboratory, NOAA, Seattle, Washington

Journal of Geophysical Research, 102(C4), 8533-8553 (1997)
This paper is not subject to U.S. copyright. Published in 1997 by the American Geophysical Union.

7. Discussion

On diurnal timescales and to a large extent even on longer timescales, the warm pool's SST variability is controlled by variability in the insolation and latent heat loss. During wind bursts, reduced insolation and increased latent heat loss tend to cool the upper ocean temperature, while during quiescent periods (periods of suppressed atmospheric convection), increased insolation and reduced latent heat loss tend to warm the upper ocean temperature. An important result of this analysis, however, is that horizontal heat advection can also be a dominant process controlling SST variability.

Horizontal heat advection requires a horizontal current oriented along a temperature gradient. While the long-term mean surface currents and temperature gradients in this region are small, wind forcing due to the easterly trades, modulated by westerly wind bursts, can cause large O(50 cm s) equatorially trapped surface currents which can be either westward or eastward depending upon the forcing. The time-longitude plot of the equatorial Pacific SST [Reynolds et al., 1994] for January 1991 through December 1994 (Figure 14) indicates that at 0°, 156°E the zonal temperature gradient is generally positive (warmer water to the east, cooler water to the west). A simple scale analysis can determine how large the temperature gradient must be to produce substantial heat advection: throughout the entire 4-year period, the weekly SST at 0°, 156°E never varied by more than 1.5°C, although during warming and cooling events SST varied by up to ±1°C month. For a 50 cm s zonal current to produce a 1°C SST change in 1 month by zonal advection, an SST gradient of 1°C in 1300 km or approximately 12° of longitude would be required. Temperature gradients of this magnitude are not uncommon in the warm pool, as can be seen in Figure 14 (e.g., September-October 1991, September-October 1992, and August-October 1993). Additionally, although not resolved by the weekly 1° × 1° SST fields, sharp temperature fronts, such as those observed during COARE, could also result in significant SST variability due to advection. Thus the warm pool surface currents and temperature gradients are potentially large enough to make horizontal heat advection an order one process in the time-varying heat balance.

 

fig14sm.gif (16799 bytes)

Figure 14. Time-longitude plot of the Reynolds et al. [1994] SST from January 1991 through December 1994, along the equator (average of 0.5°S and 0.5°N) from 140°E to 110°W. The study period is highlighted by a line along the left edge of the time axis. Likewise the study region is highlighted by line through the contoured time series at 0°, 156°E. The CI is 1°C and 0.5°C for SST values less than and greater than 28°C, respectively. The 28.5°C contour (dark line) indicates the boundary of the warm pool.

 

Using 4 years of data from 250 surface drifters in the western equatorial Pacific, Ralph et al. [1997] show that while advection by the long-term mean current is negligible, long-term mean advection is not. On average, eastward surface currents tend to be correlated with an increased positive temperature gradient, and thus the variable currents and temperature gradients result in a net long-term heat advection which is a cooling process in the warm pool. The correlation in this case is due to westerly wind bursts which tend both to produce eastward surface jets and to increase the temperature gradient by cooling the SST in the western portion of the warm pool (where wind speeds and, presumably, evaporative cooling associated with wind bursts are larger). It should be noted, however, that the statistical range of variability in their analysis also includes westward warm advection events such as we observed in October 1992. During this event, the westward South Equatorial Current and a positive temperature gradient combined to produce substantial warming at 0°, 156°E. Instantaneously as well as climatologically, heat advection events are important in the warm pool's heat balance.

The dependence of entrainment cooling on the depth of the thermocline causes a further complication to the simple one-dimensional heat balance. On the basis of the turbulent energy budget and the empirical three-dimensional heat balance (Figure 11), the strength of entrainment cooling depends upon the location of the reservoir of cold water relative to the source of turbulence. When the thermocline is shallow, as it was in September and early October 1992, turbulent entrainment mixing can be an important cooling process in the upper ocean heat balance. The 2.5-year time series of temperature as a function of depth at 0°, 156°E (Figure 15) shows that from February through October 1992 and again from July through October 1993 the thermocline was extremely shallow. One might expect that during these times entrainment cooling was potentially large. In contrast, throughout the COARE IOP (November 1992 to February 1993), the thermocline was below 70 m and one might expect that entrainment cooling was reduced during this period. Indeed, Anderson et al. [1996] show that for much of the COARE IOP the upper ocean temperature tendency was in near-balance with the surface heat flux rate (see their Figure A1).

 

fig15sm.gif (8398 bytes)

Figure 15. Time series of the subsurface temperature profile at 0°, 156°E. Data have been low passed with a 29-day triangular filter. The CIs are the same as in Figure 4a.

 

While the deepening of the pycnocline in October 1992 occurred as a dynamical response to the October 1992 westerly wind burst, as can be seen in Figure 15, not all westerly wind bursts have this response. The differences are due to the fact that pycnocline responses to wind burst forcing in general depend not only on local Ekman convergence but also on the excitation of equatorial waves, most notably Kelvin and Rossby waves. Previous studies have shown that the mix of local wind forcing and remote forcing at a particular location in the warm pool can lead to a variety of responses (upwelling, downwelling, or more complicated evolutions) [e.g., McPhaden et al., 1990, 1992; Delcroix et al., 1993]. The details of the response depend on the space/time structure of the wind forcing, the location of the observational array relative to the forcing region, and oceanic initial conditions on which that forcing operates.

Upper ocean stratification can also be affected by high levels of precipitation in the warm pool region. In particular Lukas and Lindstrom [1991] hypothesized that heavy precipitation in the western Pacific can lead to the formation of a barrier layer, i.e., a salt-stratified layer between the bottom of the surface isopycnal layer and the top of the thermocline. Using CTD data from the western equatorial Pacific, Ando and McPhaden [1997] found that during COARE the barrier layer was on average about 10 m thick. With the vertical resolution of the moored time series data (Table 1), this thickness is at the margins of detection. Nevertheless, during a period of persistent rain in the second week of December, short-lived, shallow barrier layers greater than 10 m thick were observed in the 0°, 156°E moored data. As can be seen in Figures 4a and 7, during these events the stability of the surface freshwater stratification was at times enough to support temperature inversions such that the SST was cooler than subsurface temperatures. Similar short-lived, shallow barrier layers were observed at 2°S, 156°E during the COARE IOP [Anderson et al., 1996; Smyth et al., 1996; Huyer et al., 1997]. Using a one-dimensional mixed layer model, Anderson et al. [1996] show that the SST cooling and subsequent temperature inversions associated with these shallow barrier layers are due in part to the increased loss of penetrating shortwave radiation across the base of the mixed layer. For barrier layers deeper than the shortwave extinction depth (~25 m), reduced entrainment mixing should result in warmer SSTs, consistent with the Lukas and Lindstrom [1991] hypothesis. A more complete understanding of the climatic impacts of barrier layer variations will require analyses of ocean-atmosphere interactions involving the long-term, large-scale hydrological cycle in and over the ocean, work which is currently underway.


Return to previous section or go to next section

PMEL Outstanding Papers

PMEL Publications Search

PMEL Homepage