Where will that water be tomorrow?
by Charlie N. Barron, Jr.
Where will that water
be tomorrow? This simple question reasonably summarizes the problem I addressed
in my dissertation. It is probably one of the first oceanographic riddles
ever addressed by early sailors or philosophers who never imagined a place
like Texas A&M. I narrowed the problem to a prediction of surface motion
on the Texas continental shelf for periods of four days or less. Applications
of this prediction might include forecasting the movement of an oil spill
or directing helicopters in a rescue at sea. I limited my prediction resources
to those which would likely be available at short notice, estimates of surface
winds and remote observations of the sea surface. My goal was to develop
a simple model which used these data to provide useful predictions of surface
flow for a few days.
I began my oceanographic research in the fall of 1988 working as an undergraduate
research assistant with Dr. Drew Vastano. Investigations using drifter observations
and velocity estimates from paired infrared satellite images led to a Honors
Program University Undergraduate Fellowship project and culminated in my
master's thesis in May 1992. Support for much of my graduate academic work
was provided by an Office of Naval Research Graduate Fellowship and a NASA/Texas
Space Grant supplemental fellowship. The remainder of support was derived
from my position as a research assistant on Dr. Vastano's TEXFLOW project.
[51K] Four-day trajectories of actual
(white) and simulated (purple) drifters near Galveston in January 1993.
Triangles mark positions after four days and circles mark positions at 00:00
The observations and techniques I learned from my thesis project guided
me in my dissertation work. I developed a circulation model which included
wind stress, bottom friction, realistic bottom topography, and realistic
coastline geometry, all within a coordinate system that transformed the
irregular boundaries into a rectangular grid. Influences such as tides,
density variations, and alongshore pressure variations were neglected.
My research led me to try some new approaches to old problems. I found that
accurate assessment of initial conditions is necessary to develop reasonably
reliable flow predictions, so I created a procedure to define initial conditions
based on satellite-derived velocity estimates. Another refinement involved
the treatment of segments of the model domain boundary which fall over open
water rather than along the coastline. These segments form artificial boundaries
which should allow energy and mass to cross freely without altering the
response of the interior. In practice, a truly open boundary is difficult
to implement numerically. I developed a new method for simulating open boundaries
on the edge of the model domain and found it to be superior in some aspects
to previous techniques. Finally, my initial model formulation assumed that
motion would be nearly geostrophically balanced; that is, the pressure gradient
would balance accelerations caused by the Coriolis effect, the effect of
the rotation of the earth. Observations revealed flow events which deviated
significantly from this balance, however, so I investigated ways to account
for ageostrophic effects.
Freely drifting buoys provided observations for evaluating model simulations.
These drifters were drogued to approximately three meters of depth and were
equipped with transmitters so they can be tracked by a satellite while they
move with the ocean currents. I compared the actual trajectories with trajectories
simulated using the model flow predictions. On average, the real and simulated
trajectories agreed to within ten kilometers after two days and within fifteen
to thirty kilometers after four days. The results show that this simple
model can provide useful circulation predictions over short intervals.
Note: Charlie graduated in December 1994 and holds a Naval Research Laboartory/Joint
Oceanographic Institutions postdoctoral fellowship at Stennis Space Center,
Mississippi. He continues to research circulation on the Texas-Louisiana
shelf in order to better define limits on predictability imposed by the
chaotic nature of real circulation and the approximations of numerical modeling.
[Next | Previous
| Contents | Home]
Oceanography, Texas A&M University
Updated July 24, 1995