Complex flow patterns were visualized using injected dyes and tracers
introduced into the upstream flow. The dye revealed the 3-D nature of the
turbulence, and it was injected using different sized syringes. One method is to
squirt a line of dye across the principle flow direction to observe how it translates
downstream. A second method is to continually introduce dye at an upstream
point so a path line forms as the dye moves downstream.
Baby powder worked well as a tracer. Quite by accident it was learned that
one brand of baby powder had small particle sizes so the flakes floated on the
surface, whereas another brand of powder had larger flakes that sank to the
bottom before moving downstream.
Large-Area Model Observations
During the week of 6-10 May 2002, engineers from the Alaska District
arrived at ERDC to examine ebb and flood general flow patterns using the
idealized flow table models. Testing commenced with the large-area model
placed on the flow table to represent the peak flood flow condition. According to
the Alaska District engineers, the large-area model
"...appeared to reproduce reasonably well known surface conditions.
The most important of these conditions are the gyres/eddies around
the Port of Anchorage, Point Woronzof, and Point MacKenzie."
It was also noted that upstream model boundaries influenced the flow patterns
downstream. Various solid objects were used to represent boundary changes
such as shifting shoals. For example, a lateral movement of Fire Island shoal an
equivalent of 914.4 m (3,000 ft) changed the downstream location of a cross-
channel current by about 1,828.8 m (6,000 ft).
Generally, the flood and ebb surface currents moved as expected with large
gyres forming in the lee of Point Woronzof, Point MacKenzie, and Cairn Point.
Figure 10 shows the large-area model during flood tide with surface flows
visualized by tracers. The closer view offered by Figure 11 shows tracers
trapped in the reduced flow area to the lee of Point MacKenzie.
Dye injection indicated significant 3-D flow structure within the large gyres
(Figure 12). An unexpected observation was formation of a strong cross-channel
current at the bottom during flood tide. This current originated to the north of
Point Woronzof and crossed the channel on a diagonal toward Port MacKenzie as
illustrated in Figure 13. The mechanism for this cross-channel flow appeared to
be flood flow separation at Point MacKenzie, which accelerated the flow at the
separation boundary resulting in a local decrease in water-surface elevation. The
cross-channel water level differential created a momentum imbalance that was
alleviated by mass flowing from the higher side to the lower side at the bottom
where resistance to cross-channel flow was least. Although this phenonomenon
had not been observed in the field, it appeared to be authenic because the
necessary flow separation was known to occur at Point MacKenzie. An
important observation was that the stepped contours in the idealized model
seemed to prevent vertical mixing between the layers.
Chapter 3 Idealized Cook Inlet Models