I think I’m beginning to make a breakthrough! More details below.

The main story

This week I want to begin with a figure from the last blog post. Figure 1 shows the time series results of the two-layer DO budget in Lynch Cove. Panel (d) is a new addition since last week, which shows relevant DO concentrations.


Fig 1. Lynch Cove two-layer DO budget.


What we observed last week is that the two largest terms in the DO budget, for all five inlets, are the exchange flow term and the vertical flux term. In the bottom layer, the vertical flux is the largest sink of DO, and the exchange flow is the largest source of DO. (This is important! We’re going to circle back to this point!)

Because these terms were so large, and largely balance each other, we combined them into a single term called “recirculation.” This recirculation term represents the physical processes that influence DO.

Last week, I also created seasonal bar charts to compare the magnitude of different terms in the DO budget. This week I improved these bar charts, trying to consolidate information to only what is most important. Figure 2 shows the summary DO budget chart for the bottom layer in all 5 inlets.

The columns represent different time intervals over which I averaged. On the far left is the Annual average, followed by two-month averages.

Panel (a) shows relevant DO concentrations, including the average deep layer concentration, the absolute minimum concentration recorded in the inlet within the time interval, and the average inflowing concentration via TEF exchange. The dashed line is at the hypoxic threshold (2 mg/L)

Panel (b) shows the time-averaged DO transport rates, which are the raw (volume-integrated) values from the budget time series (e.g. Figure 1).

Panel (c) is showing the same DO budget rates as in panel (b), except that the rates have been averaged by the volume of the bottom layer. These rates now tell us, on average, how much the DO concentration is changing per day in one cubic meter of water.

Panel (d) is showing the average time rate of change of DO in the inlet. These values are effectively the sum of the sources and sinks in panel (c). Positive values mean that DO is increasing, and negative values mean that DO is decreasing.


Fig 2. Summary of bottom layer DO budgets in all 5 inlets.


So, what did I learn from this figure? From panel (c), I learned that per unit volume, Budd inlet has the largest volume-averaged source and sink terms. Budd Inlet is the shallowest inlet of the five, and it is rather narrow too, so it is reasonable that there’s just a lot going on in its tiny volume. Despite having such large rate terms per unit volume, the sources and sinks still mostly balance each other. Budd Inlet also has the highest DO concentration out of all of the other inlets [panel (a)].

Penn Cove, Case Inlet, and Carr Inlet all have similar average DO concentrations to each other [panel (a)], though Carr Inlet does not have the extreme hypoxic lows that we observe in Penn Cove and Case Inlet.

Compared to all other inlets, Lynch Cove has lower DO concentrations throughout the year. It is the only inlet in which the entire bottom layer becomes, on average, hypoxic during the summer. However, the DO budget term rates for Lynch Cove aren’t much different than the other inlets [panels (c) and (d)].

With the exception of Budd Inlet which tends to have smaller rate terms, it seems like during spring and summer (when we are losing DO), the magnitude of d/dt(DO) is similar across all of the inlets. The annual average d/dt(DO) is also generally 0, so the inlets aren’t losing or accumulating DO over the year.

Therefore, the local processes influencing DO don’t seem to be the key as to why Lynch Cove is so much more hypoxic than all of the other inlets. Instead, the baseline (annual average) seems to be what sets up how low DO gets during the summer.

To visualize this, Figure 3 shows a time series of average bottom layer DO concentration for all of the inlets. I have subtracted the annual mean DO concentration from each inlet, so they are all centered about zero.


Fig 3. Bottom layer DO time series of all 5 inlets, with annual mean subtracted out.


In this figure, we can confirm that Budd Inlet has the least amount of deviation from its annual mean value compared to all other inlets. Additionally, we observe that Lynch Cove has comparable seasonal variability in its average bottom layer DO concentration to Penn Cove, Case Inlet, and Carr Inlet. This evidence thus corroborates that a low baseline DO is what makes Lynch Cove become so much more hypoxic than the other inlets, and not a larger magnitude of local oxygen consumption processes.

So, what sets the baseline DO concentration? If we recall my comments on Figure 1, the inflowing exchange flow is the largest source term of DO to the bottom layer. And in Figure 2, panel (a), the plus symbols show that the inflowing DO concentration to Lynch Cove is lower than the inflowing DO concentration to any other inlet. Therefore, Lynch Cove appears to simply have a lower baseline concentration of DO, so it does not take a uniquely large amount of DO consumption for the inlet to become hypoxic.

Revisiting bottom layer budget of all inlets

Finally, I am showing all of the bottom layer DO budget time series in all 5 inlets in Figure 4. These rates have all been normalized by the layer volume. What I observe is that oxygen consumption is mostly balanced by recirculation processes in Budd Inlet, and there is very little change to the storage term. In the other four inlets, the storage term is mostly dominated by changes in the recirculation term, especially during the fall and winter. This evidence again suggests that for inlets with lower DO (i.e. Lynch Cove, Penn Cove, Case Inlet, and Carr Inlet), physical mechanisms are the most dominant processes in the bottom layer oxygen budget.


Fig 4. Bottom layer DO budget in all 5 inlets. Rate terms have already been volume-averaged.


Summary

  • Budd Inlet, the most oxygenated inlet, has a high baseline DO concentration, and relatively small seasonal variability from its annual mean DO concentration
  • Lynch Cove, Penn Cove, Carr Inlet, and Case Inlet all experience a similar amount of local oxygen depletion and replenishment throughout the year. However, Lynch Cove has a lower annual mean DO concentration, so it becomes the most hypoxic out of all of these inlets. Thus, baseline DO concentrations are important in establishing whether an inlet will become hypoxic during the summer or remain more oxygenated.
  • Baseline DO concentrations are heavily influenced by the DO concentration of the incoming exchange flow.
    • Local processes aren’t the main story! Larger-scale circulation is!

Other thoughts

  • Is this conclusion obvious?? This is the only major conclusion I’ve been able to draw so far, but it feels a bit underwhelming.
  • Another question/hypothesis:
    • Why is the inflowing exchange flow at Lynch Cove so much lower in DO concentration than all of the other inlets?
    • My hypothesis: Hood Canal is very long, so the incoming exchange flow must travel a long distance for a long time before reaching Lynch Cove. All of this travel time means more time for oxygen to be respired. Furthermore, tidal currents are weak in Hood Canal, so the bottom branch of the exchange flow may not be well ventilated. In contrast, inflowing water to the inlets in South Sound may have higher DO because of potential ventilation and mixing through Tacoma Narrows. Similarly, Penn Cove is closer to the Admiralty Inlet which has strong mixing.
      • Another thought: The inflowing water to the bottom layer of Lynch Cove is almost analogous to the outflowing water of a shorter terminal inlet, like Penn Cove.
      • Perhaps I should consider “inlet length” or “inlet aspect ratio” as a characteristic in my scatter plots (I think Alex may have mentioned this idea before, and it slipped my mind until now). But Lynch Cove is tricky, since it’s the tail end of Hood Canal. What would I use as the length in this case?
  • I have started to extract TEF data for 2017, so I can re-calculate the two-layer DO budgets in a more “normal” DO year. I plan to extract budget information from all 21 inlets (since I refactored the bio extraction code such that more inlets will not significantly add to the processing time).
    • Do I observe the same behavior in 2017 as I did in 2014? If I expand the analysis to more inlets, will I continue to find that oxygenated inlets have little seasonal variability (like Budd Inlet) and hypoxic inlets tend to have larger seasonal variability and lower baseline DO, or will we learn something else?
  • The timing of the seasonal DO cycle in Lynch Cove seems to be slightly lagged compared to the other inlets (I saw this more profoundly in the upper layer average DO concentrations, which are not shown here). What is driving this time lag?