Measurements of Ocean Wave Spectra and Surface Currents Dennis Trizna Imaging Science Research, Inc. dennis @ isr-sensing.com
Presentation Outline: Introduction: Standard Marine Radar vs. Single Image of Waves: ~Tilt Modulation (SMR) vs. Orbital Wave Velocity (CMR) Summed Images of Waves: Enhanced Roughness (SMR) vs. Mean Current (CMR) 3D-FFT analysis: MTF scales Radar/Wave Spectra (SMR) vs. Direct measure (CMR) Analysis approach: image spectra to wave spectra, Standard Products Modulation Transfer Function (MTF) multi-parameter variable (Wind/Wave Dir) Wind speed/direction effects may affect MTF CMR is only affected by Wind dependence when lack of winds smooths surface Coherent marine radar approach:similar analyses (3D-FFT, Sums) Image sums of velocity field can provide radial current maps at 3-m resolution 3D-FFT analysis provides a direct measure of orbital wave speed, DWS Comparison of COR COH, shows issues with magnetron frequency drift Affects accuracy of 3D-FFT analysis, requires filtering
Introduction to the Standard Marine Radar Transmit phase random pulse-to-pulse Sum Several, N Gain ~N Signal is Video (I 2 + Q 2 ) Log-Magnitude Amplitude recorded Radar Wave Images due ~ long-wave slope Video to Wv=Ht scale is empirical - MTF Transmit phase is fixed pulse-to-pulse Sum Several, N Gain ~N 2 Signal is I and Q Phase is Φ = ARCTAN(I/Q) δφ = Φ j+1 Φ j ~ Radial Comp. Orb. Vel. Orb-Wave Vel to Wv-Ht Spectra is direct
Prototype Layout: Koden / Si-Tex Radar Pedestal maintained Above Plane SMA-Con components Project to combine all elements Single Package Below Solid State Amp, 5 W Cooling system on Prototype 10-min On / 20 Min off ops
New Sub-assembly board: Previous components (Top & Bottom Plate)=> on a single board, ~ 4 x 7 HPA on bottom now part of Transceiver Sub-assembly 5 watts output Serves as Pre-amp for larger power amplifier for future (10-25 watts?) New design with bias currents turned on during pulse only reduces heating
Pulse Compression Example: Reference Function used as Raw Data 50-80 MHz FM Chirp In-Phase/Quadrature P-C Output, Sine-Cosine, Point target equivalent For 30-MHz chirp, get 33-ns pulse echo (vs. 80-ns for marine radar) Wider bandwidth/shorter pulse is possible, if RF filter widened (worse SNR?) Smaller area/weaker-echo + noise Increase due to wider bandwidth
Coherent Radar Magnitude / Radial Velocity Polar Display Range-Az Echo Intensity Bright is strong echo area Wave crests give strongest echo Shallow Wave Breaking strongest Deep waves breaking less active Range-Az Radial Velocity Dark - Receding crest Orb-Wave-Vel Bright - Approach crest Orb-W-Vel Wave-Ht ~ Orb-Wave-Vel Shallow crests faster -8 to +8 m/s for 2 KHz PRF View To shore View To shore View To sea View To sea
11-14 November 2009 Storm @ Duck NC IDA Passage Left - Video Magnitude signal Right - Phase-difference or Doppler shift signal +180 deg Phase diff between pulses @ δt=500 µs = + 16 m/s radial velocity Phase difference signal longer range coverage than video
11 Nov 09 13:00 2.1-m Hrms 13 Nov 09 11:00 4.2-m Hrms
Cartesian Transformed Radial Velocity Maps 11 Nov 11 - Cartesian Transformation of Radial Velocity Image Shoreline lies vertically Recede Doppler is darkest Approach Doppler is whitest Shoaling waves show higher Doppler shift than deep water Set 64x64 window for 3D-FFTs for directional wave spectra Use data directly for wave height maps in real-time forecasting apps Doppler sense lines up with wave direction - dominated by orbital wave motions
Standard Summary Product *12 of 32 Kx-Ky spectra, 1 of 512 PPI images, 512-sum Image, 26-day frequency spectrum History - All elements available as ASCII/Binary files for analysis
Recent COHrad Marine Radar results 11-13 Nov 2009 Storm, Duck NC site - 2-pulse Doppler Image Windows
COH Marine Radar results 11-13 Nov 2009 Storm, Duck NC test site Frequency Spectrum derived from Previous Figure Doppler 3D-FFT Spectral Peaks-512 rotations, 11.6-min ***** No MTF Scaling Required Direct Measure *****
Summary of Preliminary Wave-height Comparisons Good Agreement for 2.5 4.3 m Hrms Fit parameters not yet calculated Worst Outliers due to 64 rotations (2.5 min) vs. 512/10min later Full week of data to be analyzed to determine error bars
COH -512 Summed Images - Mean Intensity (L), Doppler (R) Surface current is dominated by wind drift and Bragg scatterer velocity - Care must be taken in separating out Orbital wave velocity for current measure Wave Height is determined by modulation about the mean ** Blue dots on right are stationary buoy echoes - Mean Doppler is 0 Wind Direction
Mean Surface Current Comparisons at FRF Radar pixel chosen over AWAC acoustic Doppler sensor Two dates analyzed, compare well with in-situ Full set to be analyzed and error bars determined
COH - COR Comparisons of Intensity Imagery COH Xmit pulse is 33 ns compressed vs. COR 80 ns COH higher range resolution Will affect shortest waves in spectrum that can be imaged Some pulse-to-pulse COR dithering gives added broadening to COR images Otherwise amplitude results are similar
COH - COR Comparisons of Radial Velocity Imagery COR radial velocity striations due to frequency drift of magnetron Creates radial streaks that appear in Cartesian transformed images
Keelung Coastal Current Measurement Demo Test Courtesy: Prof. Dong-Jiing Doong, Current Map COR Radar geometry ebb tide flow Keelung islet Keelung sill flood tide flow National Taiwan Ocean University
Bathymetry and the locations of COR radar and the ADCP -30-40 -25-20 -15 ADCP -10 COR ADCP: N25 o.09.232,,e121 o 46.742,
Range-vs.-Azimuth Intensity Map ADCP Rip Current??
Range-vs.-Azimuth Radial-Current Map Range-vs.-Azimuth Radial-Velocity Map, cm/s ADCP Cosine-Squared-Azimuth dependence if spatially uniform Rip Current??
Courtesy: Prof. Dong-Jiing Doong: Range of Currents Expected Keelung tide from Oct. 15 to 18 30 cm/s range as seen with radar predicted Speed (m/s) observed
Keelung Current Measure Test Demo COR Radar used as test of principle of current mapping Same post-processing tools Passes export control as standard marine radar Range-limited compared to COH, phase noise COH radar will provide higher quality radial velocity data No radial velocity striations as COR has Emissions testing to be done, Export control paperwork Directional wave spectra without MTF empiricism No environmental dependence System will run continuously for 1- month period
Future Plans Two Coherent Radar prototypes have been built 1 st was shipped to Un Michigan Lake Michigan rip current studies Microwave transceiver sub-assembly Being completed by contractor, due for delivery 1 November Lab and field testing to be conducted over next quarter Delivery of systems in 2 nd quarter 2011? Export control application, Emissions testing must be completed Turn-key system, with software options: Directional Wave spectra Surface Currents Vector Surface current maps using a pair of system - a future goal Bathymetry over coverage of radar on 50-m pixel grid in Beta testing
Summary New s tested Fully Coherent Radar built using Marine Radar Pedestal and Antenna New Sub-assembly transceiver system available 4 th -Q 2010 Coherent-on-Receive modification to Standard Koden 25 kw radar, IF output Place holder until COH product line is full steam Orbital-Wave Radial Velocity measured using 2-pulse phase differencing Directional Wave Height spectra a direct measure with no MTF needed Mean currents at FRF AWAC comparison test points Future: Current mapping over select area using 2 ~orthogonal look radars Bathymetry maps using Shallow-Water Dispersion Rule inversion Real time product at: http://www.frf.usace.army.mil/radar/frfdisplay.jpg
Range-vs.-Azimuth Radial-Current Map Range-vs.-Azimuth Radial-Velocity Map, cm/s ADCP Cosine-Squared-Azimuth dependence if spatially uniform
Range-vs.-Azimuth Intensity Map ADCP