Pacific Toy Spill Fuels Ocean Current Pathways Research

Earth in Space Vol. 7, No. 2, October 1994, pp.7-9, 14. © 1994 American Geophysical Union. Permission is hereby granted to journalists to use this material so long as credit is given, and to teachers to use this material in classrooms.

When several containers of children's bathtub toys spilled over a ship's side and were released into the Pacific Ocean, who would've thought a concerted research project to study the ocean's currents would be the result?

Curtis C. Ebbesmeyer, Evans-Hamilton, Inc., Seattle, Wash.; W. James Ingraham Jr., National Marine Fisheries Service, National Oceanic and Atmospheric Administration, Seattle, Wash.

On January 10, 1992, a container vessel crossing the North Pacific Ocean along the great circle route from Hong Kong to Tacoma, Wash., encountered severe storm conditions near the International Date Line. At 44.7°N, 178.1°E, a dozen forty-foot (13.3 m) containers were washed overboard, one of which held about 29,000 plastic bathtub toys (see Figure 1). As they fell overboard, some of the steel cargo containers may have been torn open by the vessel's stays, or they may have been ruptured by collisions with other containers.

Fig. 1. Four types of spilled bathtub toys: blue turtle, yellow duck, red beaver, and green frog. The toys were released from containers that were washed off the side of a ship during a severe storm in January 1992.

Ten months after the spill, plastic toys began washing up on beaches near Sitka, Alaska. According to computer simulations the toys drifted toward the southeast Alaska coast, past both the site where about 61,000 Nike shoes spilled from a boat in 1990 and past Ocean Weather Station Papa, where many drift bottles have been released.

This spill, like that of the Nike shoes, provides new data for research on ocean current pathways because it was far larger than the usual planned instantaneous release of 500-1000 drift bottles. Given the typical 2% recovery rate of objects from the mid-Pacific Ocean, the drift pattern documented here is based on hundreds of recoveries compared with 10 or so drift bottles. The 6- to 12-cm toys manufactured in China were packaged in a plastic housing glued to a cardboard backing. To check on the probable time for release of the toys, sample packages obtained from retail stores were soaked in a bucket of seawater; after a day in water, the glue deteriorated, releasing the toys. Therefore, shortly after the container fell overboard, it is possible that up to 29,000 toys were released into the Pacific Ocean.

The beached toys were traced to one container through the following evidence: according to company records, the toys were shipped in a single container; the vessel's log showed that this container was spilled at the stated date and position; and all four types of toys spilled were found by beachcombers. Unfortunately, the container itself was not found.

Beachcomber Reports Help Track Drift Pattern

To obtain reports of toys found on beaches, articles and advertisements were placed in a newspaper widely distributed near the toys' first landfall. A beachcomber found the first landfall, consisting of six toys, at ~57.1°N, 135.7°W on November 16, 1992. A second beachcomber found 20 toys at ~57.8°N, 136.4°W on November 28, 1992. Unfortunately, the reports were not specific enough to provide a breakdown of the type of toys recovered at each location. Without this information we were unable to differentiate the windage – the effect of the wind – on each type of toy. From other beachcombers, we learned that approximately 400 toys, positively identified from photographs, were found between November 1992 and August 1993, along approximately 850 km of shore between the city of Cordova and Coronation Island, bordering the eastern Gulf of Alaska. To expand our search for beached toys, an article was placed in the monthly newsletter sent to Canadian lighthouse keepers from the Queen Charlotte Islands southward to Vancouver Island. We also presented our findings at a beachcombers' fair held in Washington state. None of the toys were reported south of Coronation Island, suggesting that once the toys reached the vicinity of Sitka, they drifted north.

Model Is Used To Simulate Drift

To simulate the toys' drift, we used a computer program, a numerical ocean-wide model that covers the subarctic and subtropic Pacific Ocean from 20° to 66°N with a 90-km grid. It combines long-term mean surface geostrophic currents, which move toward the right, with empirically generated surface-mixed layer currents, which are functions of wind speed and direction. These two measurements form daily data from which trajectories – possible flow paths – are interpolated. Wind speed and direction were derived from daily sea level pressure data collected between 1946 and 1993.

The basic equations for the model were established by tuning the Gulf of Alaska portion of the model so that the model trajectories for September 21 to December 31, 1978, matched the trajectory of a satellite-tracked drifter for the same dates and starting locations. Further tuning enabled the model to simulate trajectories of drifting objects with windage. By matching simulated trajectory endpoints with recovery dates and locations, the model showed that the drift bottles drifting eastward from 48 °N, 165°W in March 1962 and Nike shoes drifting eastward from 48 °N, 161°W in May 1990 exhibited similar characteristics over their approximate 9-month interval.

Simulations Show Toys Took Early Northeastward Turn

To account for the larger windage on the toys, the functions for current speed and angle of deflection were adjusted to determine which model trajectory ended near Sitka, Alaska, at the time of the first reported landfall. Increasing the basic speed coefficient by 50% and subtracting 10 degrees from the basic angle of deflection function caused the end of the trajectory on November 16 to vary 6° in latitude. Holding the angle constant and varying the basic speed coefficient from a 40% to 50% increase caused the latitude of landfall to vary by approximately 1 degree of latitude. Among these combinations, a unique pair, equivalent to increasing the speed coefficient by 50% and decreasing the angle of deflection function by 5 degrees, provided a close match between the modeled and reported landfall. Modeled trajectories for 1992 tuned to high windage drift for toys and tuned to normal drift for satellite-tracked drifters were computed. Although the directions of both trajectories were similar, the trajectory modified for windage made a landfall during the autumn, while the unmodified trajectory at that time was still about 900 km offshore and didn't reach the coast until 6 months later. The separation was considerably larger after 2 years of drift.

For perspective, the 1992 toy spill trajectory was compared with trajectories from other years by running the model with the same high-windage coefficients and starting from the same location on January 10 of each year between 1946 and 1993. From the spill location, the eastward flow appears to be orderly and repeatable for about 6 months in a well-defined drift channel that gradually widened until a turning point occurred (see Figure 2, near 45°N, 145°W).

Fig. 2. Site where 29,000 children's bathtub toy animals washed overboard on January 10, 1990 (T), and dates and locations where 400 toys were discovered by beachcombers. Trajectory of the toys, simulated with a computer model, passed near the site (N) where 61,280 Nike shoes were spilled and near Ocean Weather Station Papa (P) where 33,869 bottles were released. Thick line with circles on January 1 and July 7, 1992, shows our best estimate for the toys' trajectory; by using a 50% increase in the windage coefficient and a 5° decrease in the angle of deflection function, the track arrived near Sitka, Alaska, at the time of first recoveries, November 16, 1992. Thin line with triangles every 6 months shows the slower trajectory for objects without significant windage. Coded symbols as follows: black square, location of reported toy recoveries; large triangle, location of low-windage objects on November 16, 1992; small symbols, locations on January 1 and July 7 for 1992, 1993, and 1994.

The 1992 trajectory, however, branched northeastward unusually early at 45&176;N, 162&176;W. Beyond the turning point, most of the trajectories turned northward into the Gulf of Alaska, southward along the California coast, or continued straight to the coast. After 10.3 months of drift in 1992, the latitude of the toys' landfall was farther north than 77% of the other simulated trajectories. The large symbols in the Figure 3 show where the toys would have been during other years at this time.

Fig. 3. Five trajectories selected from the set of 48 simulations of the same toy spill each year between 1946 and 1993 on January 10 showing the range of spatial variability on November 16, 1992 (first recovery date; large symbol), and their destinations after 2 years (double symbol; see codes at lower left); 1959 (squares), when the toys would have travelled in a loop around the Gulf of Alaska gyre; 1992 (circles), when the toys after reaching the Sitka area would go westward then northward through Unimak Pass into the Bering Sea; 1961 (inverted triangles), the most southerly trajectory; 1984 (diamonds), when the toys would have looped back to the northeast in an area of increased residence time; and 1990 (triangles), when the toys would have travelled the farthest westward north of Hawaii.

For trajectories extending out to 2 years, five major types of trajectories were evident (see Figure 3). To represent the northward trajectories, 1959 was selected because of the well- defined loop around the Gulf of Alaska gyre and 1992 because it was the only trajectory penetrating the Bering Sea. For southward trajectories, 1990 passed farthest north of Hawaii, 1961 reached the farthest south, and 1984 looped northeast of Hawaii, illustrating an area of increased residence time.

The Toys' Trajectories: Some Likely Scenarios

Many types of man-made objects drift on the surface of the North Pacific Ocean. The movement of surface drifters is influenced to varying degrees by currents, winds, and waves. As they drift, these objects are constrained to the sea surface and thus cannot follow the motion of water parcels that also move vertically. Though understanding of the toy dynamics is incomplete, we believe that they illustrate some aspects of the behavior of objects that float exposed to winds and waves in the 4 cm above the sea surface.

Surface drifters frequently have organisms attached to them, such as corals on glass floats. Depending on the associated windage, the objects and organisms will be transported to various locations in the North Pacific Ocean and adjoining water bodies. Our calculations suggest that the fate of long-term drifting objects is not only sensitive to the great interannual variations, but also that the variation increases with the amount of surface area subject to windage. Here we provide a quantitative estimate for the amount of displacement difference between objects that float mostly submerged – Nike shoes – and those that float with considerable area above the water's surface – the plastic toys.

Our best estimate for the toys' trajectory is that it passed close to the Nike shoe spill site, as well as to Ocean Weather Station Papa, where drift bottles have been released for many years. By comparing the recovery rates of the toys and shoes with those of drift bottles, we learned that the containers must have opened, spilling most of their contents.

The total number of toys reported, ~400, is about 1.4% of the 29,000 lost overboard. Historical drift bottle releases made near the toy spill are listed in the Table 1. All totaled, 16 bottles were recovered from 663 releases for a recovery rate of 2.4%.

Table 1. Drift Bottle Recoveries From Releases Made Near the Toy Spill (44.7°N, 178.1°E).
DateLatitudeLongitude ReleasedNumber RecoveredNumber
May 28, 196044.97°N175.02°W2313
March 1, 196245.00°N175.00°W1206
March 2. 196245.00°N175.00°W1204
February 22, 196644.97°N172.02°E961
February 23, 196645.00°N169.00°E962
Total  66316

In the 1990 Nike shoe spill, only four of five containers opened, releasing up to 61,280 shoes. At least 1600 shoes were recovered. We estimate that at least 2.6% of the shoes spilled were recovered, a value close to that for drift bottle recoveries. Anecdotal evidence suggests that many additional toys were recovered but not reported. Since the recovery rate of toys, shoes, and bottles is approximately equal, it appears that most of the toys and shoes escaped from their containers, except for one container of shoes that sank.

To determine the long-term fate of the toys, the computer simulation was continued beyond the first landfall, showing that the toys traveled counterclockwise around the Gulf of Alaska, and through passes west of the Alaska peninsula. By January 1994, the toys had arrived in the southeast Bering Sea and presumably were trapped in the seasonal sea ice.

By the following spring during breakup, given the prevailing currents, it is reasonable to assume that the toys eventually would be transported northwestward by the Bering Slope Current, then turn north and head for the Bering Strait along the Anadyr coast, or south in the Kamchatka Current toward the Oyashio Current. The toys passing through Bering Strait into the Arctic Ocean would be carried to Point Barrow and from there north of Siberia with the Arctic pack ice, eventually reaching the North Atlantic Ocean. The toys turning south might merge eventually with the Kuroshio Current and be carried past the spill site.

Drifts from the area north of Bering Strait in the Chukchi Sea have been reported from several locations in the North Atlantic: wreckage from The Jeannette frozen in the pack ice near Hearld Island in November 1879 was found 5 years later on the southwestern coast of Greenland; three drift casks containing messages released near Point Barrow in 1899 and 1900 were recovered 6-8 years later on the northern coasts of Norway and Iceland and the southwest coast of Greenland; several bottles released in the vicinity of Nome, Alaska, were found approximately 10 years later in Iceland, Ireland, and Norway; and a drift bottle, released June 26, 1979, in the Bering Strait was found in western Scotland 7 years later on July 6, 1986.

The drift southward from the Bering Slope Current has been traced by a number of drift bottles and satellite-tracked drifting buoys, some of which were caught in the Kuroshio, and one buoy that subsequently headed southeast from Japan. Furthermore, a survival suit containing the victim's skeletal remains lost in the Bering Sea was found 3 years later in Hawaii. Therefore after a number of years we also expect some of the toys to be dispersed around the North Pacific Ocean, the Arctic Ocean, and the northern North Atlantic Ocean.

Given the substantial release of 29,000 toy animals, we anticipate that by the year 2000 a few toys will have been transported to many oceanic locations in the Northern Hemisphere. We ask your help in reporting any subsequent finds of the toys or other interesting drifting objects.

Source: Eos, Vol. 75, p.425, September 13, 1994.

Return to Science and Society
Return to Starting Point