Table Of Contents

This Page

12.06.10: Temperature salinity and potential density during WH01a and WH02a

We plot here the temperature, salinity and potential density σ referenced at the surface during WH01a (Figs. 1 to 3) and WH02a (Figs. 4 to 6). The initial temporal resolution of the data is at least 10 minutes for WH01a and 3 minutes for WH02a but 1-day averages are shown here to eliminate the tides and high-frequency internal waves. Notice the high-frequency “upwelling” events after day 150 in WH01a (Fig. 3) and around day 200 (Fig. 6) that are reminiscent of the events measured by ARGO floats (see this note for instance). Notice also that these events occur during the Spring when stratification in the upper layers is weak. Can we see the same relationship in the ARGO data?

Because the quantities shown are 1-day averages, there is no risk of internal waves or period less than 12 hours to be aliased in the time series. Furthermore, because the apparent period of the events of upwelled density surfaces is at least 10 days and the lowest frequency of internal waves is about 1.3 days at these latitudes, internal waves cannot explain these events.


Figure 1: Temperature during WH01a. All quantities are 1-day averages.


Figure 2: Salinity during WH01a.


Figure 3: σ during WH01a.


Figure 4: Temperature during WH02a.


Figure 5: Salinity during WH02a.


Figure 6: σ during WH02a.

Internal wave signal can, nonetheless, be aliased in the 5-day time series of ARGO float. For instance, the ARGO float observations near station PAPA in the eastern subarctic Pacific Ocean show high-frequency events of upwelled density and nitrate contours (see Figs. 12 and 13 in this note) that are likely to be due to internal waves, known to be vigourous at this location.

To estimate the contribution of the internal wave field to the ARGO float observations in the subtropical gyre and to know how well the events of upwelled density contours are sampled by the ARGO measurements, we compare the 10-min averaged σ to its 5-day subsampled time series and 1-day averaged time series (Figs. 7 and 8). We see that the 5-day subsampled time series recovers the low-frequency (periods larger than 1 day) component of the signal; in particular, the internal wave field does not seem to be aliased in the ARGO observations. The lack of regularity and the large amplitude of the events weaken also the hypothesis that internal waves are the cause of the events. The ARGO-like time series, however, can either miss an event or underestimate its amplitude.


Figure 7: σ during WH01a: (a) 10-min averages, (b) 5-day subsamples and (c) 1-day averages.


Figure 8: As in Fig. 7 but during WH02a.