Before supercomputers, a structural model helped scientists predict currents in Puget Sound

One of the first working models of Puget Sound was a scaled-down concrete reproduction, with actual water running through channels, around islands and into bays, inlets, and harbors. Motors, pumps and timing gears are part of an elaborate mechanism that replicates tides and river flows in the still-functioning model.

A physcial model of Puget Sound shown without water.
A physical model of Puget Sound still operates within the Oceanography Building at the University of Washington. The model, shown here without water, can replicate tidal flows, freshwater currents and estaurine mixing. Photo: Kathy Newell, UW Oceanography

Built in 1950 in the Oceanography Building at the University of Washington, the model has been used to predict the dispersion of oil following spills, to characterize bacterial plumes from simulated sewage overflows, and to help site new sewage-treatment plants, among other projects. In 1988, the model was even used to deduce the trajectory of a human body that floated for two days after a young daredevil jumped off the Tacoma Narrows Bridge.

Kathy Newell, who has been involved with the model for 40 years, was an undergraduate in the UW oceanography program in 1982, when she was assigned a class project to study currents and eddies from Commencement Bay through the Tacoma Narrows and around Vashon Island. She used the 12-foot-long model to help her design the project, which included releasing and tracking actual floats on the waters of Puget Sound.

“When I first encountered the model, I noticed a loud humming from the motors that ran the tidal engine,” Newell recalled. “I was impressed with the detail, all of the contours along the shore. I thought it was really cool.”

The Puget Sound model being constructed.
The Puget Sound model, built in the 1950s, took its shape from upside-down forms that represented water-filled regions. The deepest areas appear as peaks in this photo. John Lincoln, upper right, oversaw the initial construction. Photo: UW

The model covers an area from Admiralty Inlet south to Olympia, including the main basin of Puget Sound as well as Hood Canal. It was built by carefully constructing upside-down wooden forms to represent the bathymetry of Puget Sound, the deepest water forming peaks in the layout. The concrete was poured around the forms, one section at a time. Then the sections were turned over, the forms removed, and the pieces joined together to create the full structure.

A common experiment through the years has been to drop a bit of dye into the water and watch the movement of color. Dispersion is quicker near rivers and narrow channels but slower in the upper reaches of dead-end bays. Thanks to the miniaturized landscape, time is speeded up. An hour in the real world becomes about three seconds in the model.

Originally, tidal elevations depended on a plunger immersed in saltwater and connected to an elaborate gear system, which replicated solar and lunar cycles. Later, a computer was used to direct the depth of the plunger based on those cycles. Rivers are represented with copper and glass tubes attached to variable pumps that can alter the flow of freshwater coming into Puget Sound.

The effects of flowing water — including friction and surface tension — cannot be scaled down as easily in the model as one can shrink the landscape. To help maintain realism, the vertical elevations in the model have been exaggerated by 35 times compared to horizontal distances. For example, a horizontal mile in real life is 1.8 inches in the model, yet that same 1.8 inches of vertical distance in the model equates to about 172 feet (.03 miles) in the real world.

As an oceanography student, Newell was never far from the model, built in a nondescript storage room on the first floor of the Oceanography Building. She went on to complete her bachelor’s degree in biological oceanography. After summer field work with NOAA, she joined the staff in 1982 as an oceanographer and has remained there since.

Today, as a research science engineer, one of her duties is to oversee the operation and maintenance of the aging Puget Sound model, which requires ongoing painting and structural repairs. The model remains in active use, helping students and others understand the movement of freshwater and saltwater throughout Puget Sound.

“There is something special about a physical model — just being able to stand over the model and orient yourself to all the basins,” Newell said.


Forms being poured to build the Puget Sound model.
The forms were poured one section at a time to produce the land contours. Photo: UW

When it comes to simulating the natural world, computer modeling has largely displaced physical models in the realm of science. Today, computer models not only provide mathematical answers to questions, they can even be designed to show animated movies to help visualize water flow and many other natural functions. Still, in an animated landscape, a person cannot reach out and touch the land and water, nor walk to the end of Puget Sound in a few short steps.

One of the most unusual applications of the physical Puget Sound model came in 1988, after a 22-year-old man jumped off the Tacoma Narrows Bridge. The man was David Mygatt, an Army colonel assigned to Fort Lewis. Friends with him on the fateful day of Friday, Feb. 5, told police that David was an experienced bridge jumper who decided to take on the 221-foot jump and then swim to the Tacoma side of the waterway, where he would be picked up. He never made it to the planned rendezvous.

 A body was spotted by a fisherman later that day near Fox Island south of the bridge, but there were no further reports of the body until it washed up two days later at Seattle’s Alki Point, some 20 miles north of the bridge. An autopsy revealed broken bones and internal injuries.


View of the Tacoma Narrows Bridge in 1967
Tacoma Narrows Bridge before the addition of a second span. Photo: reivax via Wikimedia Commons

At the time of the jump, the winds were calm, and rapid currents would have pushed the body to the south, not the north. Investigators, using the Puget Sound model, set the model for the reported date and time of the jump, 3 a.m. on Feb. 5, 1988, then they floated a tiny bead in the water that was moving through the Tacoma Narrows. They watched the bead’s movement through the water to replicate how a floating body could have ended up so far north. The reported jump time of 3 a.m. was largely confirmed by the model — even though the body first moved about two miles south of the bridge (near Fox Island) before turning to the north.

Had the jump occurred a half hour earlier, at 2:30 a.m., it never would have gone anywhere north of the bridge, according to a report written by Curtis Ebbesmeyer, a UW oceanographer, and William Haglund of the King County medical examiner’s office. Had the jump occurred a half hour later, at 3:30 a.m., it could have been caught in two eddies near the north end of Vashon Island rather than drifting around Alki Point, where it finally came to rest.

“By comparing the bead counts (every two hours) with times when the body was observed, the approximate time of the jump was inferred to be a fraction of an hour earlier than given in the police reports,” or about 2:45 a.m., according to findings published in the Journal of Forensic Sciences.


The Puget Sound model in its present-day surroundings.
The Puget Sound model in its present-day surroundings. Photo: Kathy Newell, UW Oceanography

While the Puget Sound model has been used for many things, the investigators in this experiment said they were able to prove that modeling can be useful to law enforcement when looking for bodies, limiting a search area, or determining where a body went into the water based on its final resting spot.

Videos about the Puget Sound model produced by Richard Strickland for Puget Sound Institute in cooperation with the UW School of Oceanography:

This article is part of a series focused on different models and their uses within the Puget Sound ecosystem. The project is jointly sponsored by King County and the Puget Sound Institute.


Other articles in the series:

Six things that people should know about ecosystem modeling and virtual experiments

Health of killer whales examined through Bayesian network modeling and informed predictions

Researchers use a qualitative network model to test ways to boost production at shellfish farms

Quantitative models, including Ecopath, take food web studies to a higher level of analysis

Prey and predators create varying life-or-death conditions for salmon, as shown with Atlantis model

About the Author: 
Christopher Dunagan is a senior writer at the Puget Sound Institute.