Table Of Contents

This Page

Potential density anomaly from APDRC-ARGO monthly mean product

We construct the anomaly of potential density σ observed by the ARGO float Hawaii5145 using the APDRC-ARGO monthly mean product. The APDRC-ARGO product successfully capture the seasonal variability as well as the large-scale change due to the eastward drift of the float (Fig. 1). The anomaly is plotted in Fig. 2. σ and its anomaly are compared to nitrate in Fig. 3. The anomaly computed from the APDRC-ARGO product is close to be simply σ detrended in time. This works because the float motion is mostly eastward but that would not be the case with more complicated trajectory, in which case the definition of the anomaly with the APDRC-ARGO product would become necessary.

The correlation between σ and nitrate is only 0.4-0.5, while that between the σ anomaly and nitrate reaches 0.70-0.75. The high correlation confirms that most of the variability in nitrate at a certain level is captured by the variability in the depth of the isopycnal surfaces.

Thus, can we use the anomaly in σ as a proxy for nitrate anomalies? Yes and no. Yes, because, as demonstrated above, the anomaly in σ captures most of the nitrate variability. No, because the events in nitrate and in σ anomaly differ in amplitude.

Also, it is not clear what is the relationship between σ and nitrate. If the only mechanism was horizontal stirring and contours of constant nitrate were parallel to contours of constant potential density, then there would be a one-to-one relationship between σ and nitrate, valid everywhere. In reality, the contours are not parallel –for instance, as we have seen in the domain studied here, isopycnal surfaces shallow eastward while contours of constant nitrate stay relatively horizontal. In this case, for the same potential density anomaly, different amplitude in nitrate anomaly can occur, the larger one will occur at the location at which the vertical gradient of nitrate is the larger. The relationship is also likely to be complicated by submesoscale processes such as frontogenesis; thus two strained filaments of same density anomaly can have different nitrate concentration if in one of them, but not the other, submesoscale processes have injected nitrate at some point in its past. All these mechanisms can also explain why the correlation between the σ anomaly and nitrate is below one.


Figure 1: σ (a) as measured by the ARGO float Hawaii5145 and (b) along the float trajectory but in the APDRC-ARGO monthly mean product. See APDRC_ARGO_monthly_mean_script.m in RESEARCH/PROJECTS/MARINE_BIOLOGY/SUBMESOSCALE_PROCESSES/FSLE/analysis/Johnson_etal_09.


Figure 2: Anomaly in σ between (a) and (b) of Fig. 1.


Figure 3: 150-200 m vertical averaged (a) σ and its anomaly (to which a constant has been added for comparison with σ) and (b) nitrate.