This week I finished two-layer volume and DO budgets in 5 terminal inlets. They are listed below, along with their average Aug/Sep bottom DO concentrations in parentheses:

  • Lynch Cove (0.0 mg/L)
  • Penn Cove (1.8 mg/L)
  • Case Inlet (1.9 mg/L)
  • Carr Inlet (4.0 mg/L)
  • Budd Inlet (5.8 mg/L)

I also began constructing bar charts to compare the dominant processes influencing bottom DO in all of the inlets. It seems like the exchange flow and vertical exchange are always the largest terms. Perhaps the differences between inlets are truly the smaller terms, such as respiration.

More details below.

Interface depths

Following our discussion last week, I selected interface depths that were about 1/3 the total depth. The exceptions to this is Budd Inlet, in which I used 1/2 the total depth instead of 1/3 because this inlet is very shallow.

The resulting interface depths are:

Inlet Depth at moor Interface Depth
Lynch Cove 20 m 6 m
Penn Cove 20 m 6 m
Case Inlet 25 m 8 m
Carr Inlet 53 m 18 m
Budd Inlet 12 m 6 m

Two-layer volume budgets

Before re-attempting two-layer DO budgets, I started by creating two-layer volume budgets in each of the inlets. In these figures, I have included both TEF and Eulerian results.

TEF and Eulerian have same sign

Lynch Cove and Budd Inlet appear to have a persistent exchange flow throughout the year, with inflow always entering the bottom layer (Fig. 1 and Fig. 2)


Fig 1. Lynch Cove two-layer volume budget.



Fig 2. Budd Inlet two-layer volume budget.


TEF and Eulerian sometimes have different sign

However, the other inlets all appear to have exchange flow reversals throughout the year, with more inflow coming in through the surface layer than the bottom layer (Fig. 3, Fig. 4, and Fig. 5).


Fig 3. Penn Cove two-layer volume budget.



Fig 4. Case Inlet two-layer volume budget.



Fig 5. Carr Inlet two-layer volume budget.


For the rest of this analysis, I am showing results using TEF Qin and Qout. However, I also have results for the Eulerian version as well. In inlets with exchange flow reversals, I wonder whether the Eulerian versions would be more appropriate.

Two-layer DO budgets

These budgets were constructed using the TEF exchange flow method: assuming all of Qin goes into the bottom layer, and all of Qout exits from the top layer. I have also created these same two-layer DO budgets using Eulerian horizontal advection, and can share those figures upon request.

I also refactored Jilian’s bio processing code to process all inlets simultaneously. Even though the script still takes 6+ hours to run, I only needed to run it once total (rather than once per inlet, which was my greatest concern).

Figures 6 through 10 show the two-layer DO budget time series in all 5 inlets.


Fig 6. Lynch Cove two-layer DO budget.



Fig 7. Penn Cove two-layer DO budget.



Fig 8. Case Inlet two-layer DO budget.



Fig 9. Carr Inlet two-layer DO budget.



Fig 10. Budd Inlet two-layer DO budget.


I also note that Penn Cove has a seasonal error term that is more negative during winter. Though the magnitude of this error term is small, it seems almost comparable to some of the smaller budget terms.

The Case Inlet error term also has a strange positive blip near the beginning of August. I haven’t looked into this in depth, but I only see this blip in the TEF budget, and not the Eulerian version of the budget.

Budget bar charts

Using the two-layer TEF budgets, I created bar charts of the average DO transport rates over the time intervals of one year (first row of Fig. 11), spring (second row of Fig. 11), and summer (third row of Fig. 11).

These rate terms are volume-averaged, meaning that I divided the volume-integrated rates by the average volume of each layer over the same time period.

I have also listed the inlets in ascending order of average Aug/Sep bottom DO.


Fig 11. Bar charts of volume-averaged DO transport rates in 5 terminal inlets.


I was hoping that these bar charts would show a clear difference between the hypoxic and oxygenated inlets, but alas, that is not the case.

How do we continue this analysis?

After zooming into the bar charts, it seems that respration is the next largest term behind the exchange flow and vertical exchange. I am hypothesizing that the ratio of respiration to DO sources is larger in hypoxic inlets compared to oxygenated inlets. Furthermore, I hypothesize that this large ratio of respiration to DO sources begins earlier in the year in hypoxic inlets compared to oxygenated inlets.

To test this hypothesis, I plotted a time series of the ratio of respiration to the source terms. The source terms include TRAPS, photosynthesis, and a residual term, which is the sum of the exchange flow and vertical exchange terms.

\[DO\ usage\ ratio = \frac{Respiration}{TRAPS + photosynthesis + (TEF + vertical\ exchange)}\]

I am not showing the resulting figure, as it didn’t tell me much… Rather, I’d like to use this opportunity to discuss other options for analysis.

Bar charts with combined horizontal and vertical exchange

Figure 12 shows another budget bar chart, but this time using the net sum of the TEF exchange and the vertical exchange, which I’ve labeled “residual.”


Fig 12. Bar charts of volume-averaged DO transport rates in 5 terminal inlets, with horizontal and vertical exchange combined in one term.


Next steps

  • Start writing again!
  • Continue analyzing budgets. What makes hypoxic inlets different than oxygenated inlets??
  • Investigate sensitivity to depth?
    • I could start by remaking volume budgets, which would take less time than the DO budgets.
    • Look at TEF interface depth compared to salinity profile over time at all 5 inlets
  • Extend prior scatter plot analysis through 2019 (dots are average of 2014-2019, and error bars to indicate standard deviation)