|
| CODAR
Results |
| Figure
1.
This
timeseries shows the wind and water conditions at the
study site over the two-month course of our study. The
top timeseries shows wind speed, given in meters per second.
Notice that the winds are quite variable, with an average
speed of 3.1 m/s and a maximum speed of 13.6 m/s. The
bottom timeseries shows sea surface elevation, measured
in meters. Observe that we have a semidiurnal mixed tide,
with an average amplitude of 0.8 meters. |
|
| Figure
2.
First,
we wanted to determine how these two factors, wind speed
and sea surface elevation, affect radar efficacy. Recall
that the RiverSonde system simultaneously samples at thousands
of vector points throughout the coverage zone. As with
all instruments, not all of these readings are valid,
and thus CODAR uses a filtering algorithm to discard invalid
readings. Clearly we get more accurate data when the number
of valid vector readings is maximized. Thus, we use the
number of valid vector readings at a given time as a measure
of radar efficacy. This graph correlates this measure
of efficacy both with wind speed, shown in blue, and with
sea surface elevation, shown in red. We see that at low
wind speeds we have low coverage, and that coverage increases
as wind speed increases. Thus, we have best radar coverage
during moderate to strong winds. On the other hand, we
see that sea surface elevation does not affect radar coverage,
since we have essentially a uniform distribution of coverage
across varying surface elevations. Thus, radar efficacy
is independent of water depth. |
 |
| Figure
3.
Since
HF radar systems have not previously been used in inter-tidal
environments, this study sought to validate such an application.
An ADCP was deployed for a week in the same study area,
and the current velocities recorded by the RiverSonde
were compared to those recorded by the ADCP. The top graph
in this plot shows radial current velocity, in m/s, as
recorded by the RiverSonde system. There are some gaps
in the data, due to lack of wind, backscatter, and other
factors. The second graph shows velocity as recorded by
the ADCP. When we overlap these timeseries, as shown in
the third graph, we see that the RiverSonde system does
an accurate job recording velocity magnitude and phasing. |
|
| Figure
4. This
graph shows the error analysis from this radar / ADCP
comparison trial. This plot shows the correlation between
wind speed and velocity residual, namely the difference
between the velocity recorded by the radar and the velocity
recorded by the ADCP. Using the ADCP as our standard of
comparison, we see that the RiverSonde gives quite accurate
readings when we have wind speeds above 2 m/s. Once again,
we see a strong correlation between radar accuracy and
wind speed. |
| |
Download
movie showing the spatial coverage of the
RiverSonde
system during a 36-hr period HERE
|
| Movie
Description: This
movie depicts a timeseries of the spatial coverage provided
by the RiverSonde system. The movie shows an aerial view
of the entire 300-m radius coverage zone. The boardwalk
that the radar system was stationed on is shown in red.
The Atlantic Ocean is located to the right, and on flood
water flows leftward through the channels. The movie is
a series of vector maps, in which the directional arrows
show the direction of current flow. There is a graph in
the lower right hand corner that tracks the tidal phase
over this 36-hour period. |
|
|
|
 |
