This research was sponsored by the NSF Hydrology Program and the US Geological Survey.
INTRODUCTIONOver the last several years, we have developed two types of non-contact microwave surface velocity sensors for river applications. One, called RiverRad, is a pulsed Doppler radar that is capable of measuring the horizontal distribution of both horizontal components of surface velocities across rivers. This radar has been deployed in tests with the USGS Hydro21 Committee both on riverbanks and on helicopters. A second, simpler system called RiverScat has been developed as a cheaper alternative to RiverRad. An individual RiverScat probe will measure only a single component of horizontal surface velocity at a location determined by the antenna's small footprint on the water surface. An array of RiverScat probes has been deployed for several months on the Cowlitz River bridge at Castle Rock, WA in collaboration with the USGS and has been successful at producing time series of surface velocities during that period. We believe that RiverScat will operate from helicopters as well as RiverRad and with much less expense and effort. A hand-held version of RiverScat is planned for the future.
BASIC PRINCIPLES RiverRad and RiverScat both obtain their signal from surface roughness present on the river; no artificial targets are necessary. This surface roughness backscatters according to the well established principles of Bragg scattering in which the backscatter is a resonant phenomenon that selectively comes from a short surface wave of a particular wavelength, the Bragg wave (Plant, 1990).
Figure 1. Diagram of the RiverRad scattering geometry showing many range bins in two look directions.. RiverScat's geometry corresponds to a single range cell from a single antenna of RiverRad.
Figure 1 shows two antenna beams directed at slight angles, θ, from the perpendicular toward the upstream and downstream directions, the preferred RiverRad configuration. U indicates the component of velocity along the river while V indicates the cross stream velocity. Range bins are indicated by the lines within the footprints of the antennas. If the mean Doppler shifts of the return to the upstream and downstream beams, fu and fd, can be determined, then the velocity components at each range bin are given by U = λ(fu - fd)/(4 sin θ) V = λ(fu + fd)/(4 sin θ) where λ is the microwave length. The difficult part of the measurement is determining accurately the values of fu and fd since the Doppler spectra of microwaves backscattered from rough water surfaces are not simple.
RIVERRAD MEASUREMENTS FROM RIVERBANKS Figure 2 shows RiverRad in the laboratory and in the field.
Figure 2. Left - RiverRad in the laboratory in a single antenna configuration. Center - RiverRad deployed on the San Jouquin River near Vernalis, CA in a two-antenna configuration. Right - RiverRad deployed on the Skagit River near Mt. Vernon, WA in a single antenna configuration. RiverRad was operated on the San Jouquin River in California in a fixed position in a shed with two antennas mounted on top of the shed. On the Skagit River near Mt. Vernon, WA, we deployed RiverRad mounted in a van. A ground penetrating radar was operated from the cableway seen behind RiverRad in the figure. Results of these experiments have been reported by Costa et.al., GRL, 27, 2000. With RiverRad mounted in the van, we were able to move around the country and measure surface velocities on several rivers in a single day. Figure 3 shows the consistency of results obtained with RiverRad on the San Jouquin River while Figure 4 compares RiverRad measurements with measurements of surface velocity using a sonar on a floating "Boogie Board" (BoogieDopp).
Figure 3. RiverRad measurements on the San Jouquin River during several days in April, 2002 showing the consistency of the measurements on a given day and the changes from day to day.
Figure 4. Comparisons of RiverRad surface velocity measurements with those from a sonar floating on a "Boogie Board", the BoogieDopp.
RIVERRAD MEASUREMENTS FROM HELICOPTERS RiverRad was deployed on a helicopter on several occasions in 2000 and 2001. Figure 5 shows it mounted in a Jet Ranger helicopter and in flight over the Cowlitz River near Castle Rock,WA .
Figure 5. RiverRad (in an earlier configuration) mounted on a Jet Ranger helicopter. Left - the electronics in one passenger seat. Right - In flight over the Cowlitz River near Castle Rock, WA. The two long white objects just above the helicopter skids are RiverRad antennas in this configuration; a second pair was mounted on the opposite side. The large white box just below the cockpit is a USGS ground penetrating antenna.
Figure 6. Comparison with RiverRad surface velocity measurements (circles with error bars) with in situ measurements with a current meter (asterisks) and an acoustic doppler current profiler (ADCP, X) Figure 6 shows the results obtained with RiverRad on the helicopter. Details of these experiments are available in Melcher, et.al., GRL, 29(22), 2002. The ranging capabilities of RiverRad were not necessary in these experiments and in the future we plan to deploy the smaller, simpler RiverScat for helicopter measurements.
THE RIVERSCAT ARRAY AT CASTLE ROCK To date, RiverScat has only been deployed as an array on the Cowlitz River bridge at Castle Rock, WA. Figure 7 shows a close-up of one RiverScat unit and a view of 4 of the 8 units that were mounted on the bridge.
Figure 7. Left - Close-up of one unit of the RiverScat array mounted on the Cowlitz River bridge at Castle Rock, WA. Right - Four of the eight units that are presently mounted on the bridge.
This array at Castle Rock has now produced several months of river surface velocities. A four month time series of these velocities compared with the stage measured by the USGS is shown in Figure 8.
Figure 8. Time series of surface velocities (black) from the eight RiverScat units mounted on the Cowlitz River bridge at Castle Rock, WA. The red line is the USGS stage in feet with 32 feet subtracted. By assuming that the mean velocity at each probe site is approximately 0.85 times the measured surface velocity, we have computed discharge from our measurements and a single measurement of river depth with a sounding rod. The results are compared with the standard USGS rating curve for the site in Figure 9.
Figure 9. Discharge versus stage from microwave surface velocity measurements (dashed) compared with the standard USGS rating curve for the site (solid).
We have developed another, more compact version of RiverScat for use on helicopters and for deployment at other bridge sites. This version can be run on rechargeable batteries and can accommodate up to four individual units. Results are viewed on a laptop computer. A photo of this version of RiverScat is shown in Figure 10.
Figure 10. A more compact version of RiverScat for use on bridges and helicopters.
In the near future, we hope to improve both RiverRad and RiverScat further to make them more compact and amenable to running on solar panels. We hope to have a tripod-mounted version of RiverScat available in the near future. We can arrange to provide both RiverRad and RiverScat to interested parties. For more information, contact Bill Plant at billpop@earthlink.net.
|
