Warning: This writeup was done to help organize
my thoughts prior to the presentation.
The actual presentation may have been quite different!
...and I've just heard that it was accepted pending minor revisions. Comments, suggestions, questions would be very much appreciated.
A barrier layer is the layer between the base of a shallow mixed layer and the base of the isothermal layer ~near the top of the thermocline. Thus, the CTD cast on the left has no bl, while the cast on the right shows a barrier layer that is ~45 m thick. These are the casts made famous by Lukas and Lindstrom in 1991. As discussed by Lukas and Lindstrom, this layer acts as a barrier to entrainment of cold thermocline water into the surface, and therefore has major consequences on the western equatorial Pacific warm pool heat balance. Also, as discussed by the previous speaker, barrier layers cause the mixed layer to be shallower than would be expected from the thermocline depth and this shallow mixed layer can trap momentum and cause wind forced surface currents to be shallow and fast.
In this analysis I use daily averaged temperature and salinity data from a mooring at 0, 165E to compute barrier layer thickness, where the isothermal layer depth and mixed layer depth are defined respectively in terms of a temperature step of 0.25C from the surface and an equivalent density step.
Before we look at these time series, let's review the barrier layer formation and erosion mechanisms.
BL can form through....
2. Tilting. If there is a pre-existing horizontal salinity gradient within an isothermal layer, then advection within this layer by a vertically sheared horizontal flow can cause the horizontal salinity gradient to tilt into a vertical stratificiation. -- again, I show the resulting barrier layer in green.
3. Barrier layers can also be advected into a region,
4. and likewise, a pre-existing barrier layer can grow through vertical stretching.
Likewise, BL erosion processes include:
2. slumping -- which is similar to the tilting process. However in this case the vertically sheared flow tilts a horizontal gradient into a gravitationally unstable stratification, which then generates overturning and turbulent mixing.
3. barrier layers can be advected out of a region, and,
4. can be thinned through compression.
The WWB that I will be looking at is the November 1989 WWB. For those of you who know your WWB, this is the one described by McPhaden et al. (1992).
Each of these panels shows the time-longitude section from September 1989 through May 1990, along the equator from 145E to 165W, -- Reynolds' SST, Delcroix et al. SSS, TAO zonal winds, and Xie and Arkin Rainfall. The Color pixel time series is the barrier layer thickness at 0, 165E.
Within a day of the November 1989 WWB onset, a ~100 m thick BL formed. This is Huge. The thickest barrier layers that Delcroix described earlier was just 35 m. This barrier layer here is 100 m thick and after 5 months was still ~ 35 m thick.
There are a couple of things special about this WWB. This WWB occured after an extended period of strong easterlies associated with the 1988 La Nina. Because of this recent cold event, the eastern-edge of the warm/fresh pool was near 165E. As typically occurs during WWB, the warm pool cooled, reducing the SST zonal gradient. However, the rainfall, if anything, increased the salinity gradient associated with the eastern edge of the freshpool.
Also notice how in response to the Nov 1989 WWB, the eastern edge of the warm/fresh pool moved eastward towards its normal position. We strongly suspect this was due to zonal advection of the front.
Could the formation of this very thick barrier layer be due to advective processes -- tilting of this zonal salinity gradient? Or is it due to the rainfall associated with the WWB? During this event there was a total accumulation of over 300 mm.
OK, let's look at the mooring data.
These panels show time series as measured by the TAO current-meter buoy at 0, 165E.
We see all the features typical of a WWB response.
As the zonal winds (shown in panel a) switched from being easterly to being westerly, zonal currents (shown in panel b) switched from being ~40 cm/s westward to almost 100 cm/s eastward. This eastward acceleration of the surface waters caused a large shear above the top of the thermocline (shown as the blue line).
Likewise, the thermocline deepened due to Ekman convergence. In many WWB responses, the deepening is often attributed to both Ekman convergence and increased mixing. This is not the case here though because in fact the easterlies had been quite strong before the WWB, so there wasn't really an increase in the wind speed. Also, the mixed layer was quite shallow because of this intense freshening near the surface. What caused this 1 psu freshening? and why didn't it extend to the top of the thermocline?
Rainfall cannot account for it, since accumulation of 325 mm, distributed over 40 m, could only cause of freshening of ~0.3 psu.
This sheared eastward flow coincident with barrier layer formation strongly implicates the zonal tilting process. This sheared flow could also be caused by the trapping of momentum above the barrier layer, via a positive feedback -- First rainfall generates weak stratification, which generates weak shear, which tilts the horizontal gradient into vertical stratification, which then traps more momentum and generates stronger sheared flow, which further tilts the gradient... etc.
However, before we say our job is done, look at this meridional shear. The coincidence of the meridional shear with surface flow from the north and flow from the south, strongly implicates tilting in the meridional plane. But was there a meridional Salinity gradient of the correct sign and strength?
This shows a CTD section made along 165E about 1 week prior to the November 1989 WWB. Notice that there is 34.5 psu water at ~ 4N. So yes, it appears that meridional tilting was as large as or larger than the zonal tilting term producing the thick barrier layer.
* Tilting mechanism requires shear. A positive feedback can develop between the formation of surface stratification through sheared flow advection, and the formation of shear through surface trapping.
* WWB advective processes can cause thick BL to form, particularly if there is a pre-existing zonal and/or meridional salinity gradient.
* Because strong trades are typically associated with high SSS on the equator and relatively large SSS gradients, it may be that the first WWB following an extended period of strong trades is most effective at forming barrier layers.
Meghan F. Cronin
Pacific Marine Environmental Laboratory
7600 Sand Point Way NE
Seattle, WA 98115 USA
Meghan Cronin's Home Page
DOC | NOAA | OAR | PMEL