This week I worked on some figures for the 2014-2019 simulation years. First, I resolved issues with leap years. Then I created boxplots and hypoxic volume time-series to compare hypoxic extent between years.

I would have liked to have seriously thought about hypotheses that could explain the patterns of hypoxia shown below. But between figures taking much longer to make than I hoped, creating KC drafts, and dealing with issues on klone, I simply have not had time to dedicate to hypothesis development. Rather than spending time this week discussing rushed ideas, I think it’d be best to push off our discussion of hypotheses until next week after I can give hypotheses an earnest effort.

At the end of this blog, I have also included a proposed schedule for the remainder of my time in grad school.

More details below.

DO in 2014 - 2019

Figures 1 and 2 show maps of average bottom DO and days with bottom hypoxia, respectively. They are identical to figures I shared two weeks ago, except 2013 data are omitted and 2016 data are included.

Overall, 2014 appears to have the least amount of hypoxia compared to other years.


Fig 1. Average bottom DO in Puget Sound during hypoxic season.



Fig 2. Puget Sound map of days with bottom DO < 2 mg/L.


Figure 3 shows time-series of hypoxic volume in Puget Sound for 2013 - 2019. 2013 clearly has much less hypoxia compared to other years. Even though 2014 has less hypoxia than the remaining years, it has roughly the same order of magnitude of hypoxia (not an outlier in the same way that 2013 appears to be.)

I also compared these hypoxic volume estimates to those from WA Department of Ecology. I was only able to find Ecology’s hypoxic volume estimates for 2006, so note that this is not a direct comparison. Additionally, their hypoxic volume estimates include regions in the Straits, which I have omitted from my estimates (left panel in Fig 3). Even so, Ecology’s peak hypoxic volume estimates are about 5x smaller than LiveOcean’s. That is a rather large difference. One hypothesis I have is that extreme low DO concentrations might get smoothed in SSM’s 10 vertical layers compared to LO’s 30 layers, which may be better at resolving sharp gradients in DO. Later on, it will be prudent to look at vertical profiles of DO to check my assumptions.


Fig 3. Puget Sound hypoxic volume time-series.


Lastly, I took Kate’s advice and generated some boxplots of bottom DO concentration in Southern Hood Canal. I have found that the boxplots are highly sensitive to the temporal and spatial windows that I choose. However, it does seem that in Southern Hood Canal, 2015 and 2019 tended to have lower DO compared to other years.


Fig 4. Distribution of bottom DO in Southern Hood Canal during hypoxic season.


For hypothesis development, I will mostly focus on describing the average DO patterns over our six simulation years (Fig 5).


Fig 5. Summary of low DO over 2014-2019 simulation period.


2014 anthropogenic and natural comparison

The 2014 natural simulation finally finished running early this week. I have not had time to dive into these results, but Figure 6 provides a preliminary comparison between anthropogenic and natural average bottom DO.


Fig 6. Average bottom DO in Puget Sound in natural and anthropogenic runs.


Proposed schedule

Below I’ve put together my ideal plan for the remainder of my PhD. Some key milestones include my Master’s defense at the end of this summer, quals in autumn, general exam in winter 2026, and final exam in spring 2027.

  2024 2025 2026 2027
Winter   TA for undergraduate fluids; Engage Seminar: The Science Speaker Series Prepare and complete General Exam  
Spring   CSE 412 Intro to Data Visualization with Dakota TA for something Final exam
Summer Master’s defense      
Autumn Take or audit CHIN111; Prepare for quals; Quals at end of quarter OCEAN 587 Fundamentals of Climate Change