Salish Sea Model analysis at Puget Sound Institute

This repository contains a collection of post-processing files designed to create tables, graphics, and movies from Salish Sea Model output. The work was funded by King County and the coding platform was designed and developed by Rachael D. Mueller (unless otherwise specified) under the supervision of Stefano Mazzilli and Joel Baker. Code development also included collaborations with Ben Roberts, Su Kyong Yun, and Marielle Larson. See contributors for a full overview of this collaboritive effort.

The focus of this project was to better understand how nitrogen in rivers and waste water treatment plants affects oxygen levels in the Salish Sea, as delineated by regions established by the Washington State Department of Ecology and shown below.

My job was to develop the coding platform and analyses to address questions of how these different regions respond to changes in nutrient inputs from rivers and wastewater treatment plants. It was the first time that I've taken more of a back seat in modeling work. The model setup and scenarios were determined before I was hired on, and my computation skills were wanted more than my oceanographic and data analysis knowledge. My main contribution was to make the magic happen in an organized and coherent structure that supports the project goals of enhancing transparency as well as access to knowledge.

My accomplishments include:

1. Reducing the turn-around time on post-processing model output from weeks to days.
2. Strategically developing code in order to produce a well-organized and efficient system while meeting an urgent need that was established prior to my being hired.
3. Identifying the mathematical equivalent of a series of logic statements that were used in a "rounding method" to quantify deviations in dissolved oxygen (used in a regulatory standard) and simplifying the calculation of noncompliance with a mathematically equivalent, albeit more simple, approach.
4. Developing information on marine conditions. Click here for some examples of graphics and links to code.
5. Creating a method to batch process runs and synthesize requested information via:
   1. spreadsheets,
   2. graphics, and
   3. movies...or....
   4. all of the above.

The system that I developed stategically uses a .yaml configuration files to organizing file paths and run information as well as using a shapefile as the source of information for each model node (e.g. depth, region name, etc). I also introduced SLURM arrays in bash scripts to help facilitate batch processing of multiple runs. This system relies on files that are version-controlled on two different repositories, both of which are only visible/accessible by project collaborators (at the time of this posting):

1. https://github.com/UW-PSI/SalishSeaModel-grid
2. https://github.com/UW-PSI/SalishSeaModel-spreadsheets

These repositories ought to be cloned at the same directory level as this repository, using the repository names above (especially for the grid repository). They were all tagged with version-1 to benchmark completion of this project. Results from this work have been incorporated into three reports (one each for the SOG/Bellingham, Whidbey, and Main regions) as well as workshop presentations that have knowledge-shared with around 400 participants.

Please note: I have tried to address the places where urgent need won over my senses of better/preferred coding practices, but I probably missed some spots. I also had plans to further streamline this process but had to leave this project "as is." Please contact me if you see something wonky.

A special thanks to Ben Roberts for sharing information and his ssmhist2netcdf code for post-processing SalishSeaModel output. Thanks as well to Su Kyong Yun (PNNL) for her help and guidance in developing this code. Lastly, I very much appreciated the support and guidance of Marielle Larson in crafting the narrative for this project.

License and Copyright

Please see the LICENSE file for details.

All contents in this repository copyright (c) 2022 by University of Washington and created by Rachael D. Mueller (unless otherwise specified) in collaboration with the Project Contributors at the University of Washington's Puget Sound Institute. This work was funded by King County.

An overview of most of the products created by these scripts can be found in Table of Contents.