This week I completed a two-layer budget test in Lynch Cove! I would not have accomplished this without Jilian’s support.

In this blog post, I summarize my procedure, share results for Lynch Cove, and list open questions and to-dos. Note that these results are preliminary.

Jilian’s resources

I reached out to Jilian last Friday. She has been traveling all this week, so we haven’t had a chance to meet. However, she mentioned that she had previously conducted a two-layer budget test, and she sent me her code and some procedural notes in a GitHub repo.

Figure 1 shows the example output from her two-layer budget.


Fig 1. Jilian's two-layer budget example for the whole Salish Sea.


Her code includes scripts to calculate biogeochemistry rates: respiration, photosynthesis, nitrification, air-sea gas exchange, and DO*volume.

For the exchange flow, TRAPS, and vertical exchange terms, I wrote my own code following the steps she outlined in her repo.

These resources significantly sped-up my process of completing a two-layer budget prototype— Thanks Jilian!

What is the interface depth?

My first step was to decide on the interface depth for the two-layer model. For this, I plotted annual average density and oxygen profiles in all 21 inlets for the year 2014 (Fig. 2).


Fig 2. Annual average density and oxygen profile in all 21 inlets throughout 2014. If the difference between bottom and surface density is greater than 1 kg/m3, then we consider the inlet to have two-layers. Dots indicate locations of max d/dz(DO) and d/dz(Density). These depths are labeled in the bottom left corner of each two-layer inlet.


Figure 3 shows a close-up of Lynch Cove. Conveniently, the oxycline and pycnocline line-up nicely in this inlet. For Lynch Cove, I have selected an interface depth of -6 m.


Fig 3. Close up of Lynch Cove annual average density and oxygen profile in 2014.


In the future, I will need to decide how to pick interface depth for some of the trickier inlets (like Carr Inlet).

Process notes

This section summarizes some of the key steps I took to create the Lynch Cove budget.

Exchange flow term

For the exchange flow, I used the output of tef2 sections to get the velocity and DO concentrations flowing in and out of Lynch Cove.

Per Jilian’s outlined steps, I separated the DO budget terms (DO * velocity * area) by depth into a shallow and a deep layer. Thus, the exchange flow budget terms are based on vertical depth rather than salinity coordinates.

I asssumed that positive velocities mean flow is entering Lynch Cove. Note that when I drew my section lines, I always drew them so that the terminal inlet fell on the positive side of the line (based on the sign convention described in Parker’s tef2 repo).

My code was able to reproduce the exchange flow term in Jilian’s budget (dark blue line in Fig. 4). However, my term has the opposite sign as hers. I can only assume that this difference is due to the way the section line was drawn…but I could not think of a way to verify this. If my assumption is correct, then the exchange flow term I calculated for Lynch Cove is also correct and I do not need to multiply by -1. Otherwise, my exchange flow term may have a flipped sign..


Fig 4. Rough comparison between my exchange flow term (left) and Jilian's exchange flow term (right, dark blue line) for the Salish Sea in 2014.


Thus, my exchange flow term magnitude is correct for Lynch Cove, but questions remain about its sign.

Biogeochemistry terms and air-sea gas exchange

The biogeochemistry terms and air-sea gas exchange rates seemed by far the most complex terms to calculate. Luckily, Jilian shared her code for these. All I needed to change was the interface depth and filepaths in her code.

This code took many hours to run on apogee, but it worked well.

The only issues I came across:

  1. DO consumption rates were positive, so I multiplied them by -1
  2. The code saved hourly values of DO * volume. To calculate the storage term d/dt(DO*volume), I took the time derivative of this term.

Rivers and WWTPs

There are two small rivers and one small WWTP in Lynch Cove.

I calculated DO transport rates using the river forcing files for the 2014 run, which contains daily discharge rate and daily DO concentrations.

I added all river discharge to the surface layer, and all WWTP discharge to the bottom layer.

Vertical exchange

Finally, to calculate vertical exchange, I determined the error between the storage term and all other DO budget terms:

\[Vertical\ exchange = \frac{d}{dt}\big(DO * V\big) - (consumption+exchange+photosynthesis+airsea+TRAPS)\]

Ideally, the vertical exchange terms in the surface and bottom layers should sum to zero.

Results

Figure 5 shows the inlet domain for the Lynch Cove budget analysis.


Fig 5. Lynch Cove budget domain: everything to the right of the section line.


Figure 6 shows the resulting budget. My big issue: vertical exchange is a source term for surface DO in winter, and a loss term for bottom DO in winter. This is the opposite of what I would expect.


Fig 6. Lynch Cove budget results.


Figure 7 shows the error term (sum of surface and bottom vertical exchange). The error is nonzero, but small relative to the largest terms in the DO budget.


Fig 7. Error term.


Questions and open items

  • What are the interface depths for the other 21 inlets, especially those that may not have as clear of an interface as Lynch Cove?
  • Why is the vertical exchange term the opposite of what I would expect? Is this due to the exchange flow term having the incorrect sign?
  • Is the sign of the exchange flow term correct?
    • Something else that makes me suspicious: the seasonal cycle of the exchange flow term in Lynch Cove is opposite of the seasonal cycle in the Salish Sea
  • Was it correct to multiply consumption terms by -1? Are there any other terms with a flipped sign?
  • Did I calculate the storage term correctly?
  • I used sed_sum2 instead of sed_sum, because Jilian’s code had a comment noting that sed_sum2 was the more rigorous method of calculating sediment oxygen demand.
  • How do I scale up Jilian’s biogoechemistry code to process 21 inlets?
    • We are meeting on Monday
  • Is my error term small enough?
  • How will I eventually compare 21 budgets?