WAMOS II: A RADAR BASED WAVE AND CURRENT MONITORING SYSTEM

Transcription

1 resented at ISOE '99 Brest May roceedings Vol. 3 WAMOS II: A RADAR BASED WAVE AND CURRENT MONITORING SYSTEM K. Reichert K. Hessner J. C. Nieto Borge 2 J. Dittmer 3 Ocean SensWare GKSS Technologiezentrum Max-lanck-Str. D-252 Geesthacht 2 uertos del Estado Antonio López 8 ES-2826 Madrid 3 Oceanographic Instruments GKSS Technologiezentrum Max-lanck-Str. D-252 Geesthacht ABSTRACT An operational wave and surface current monitoring system called WaMoS II is presented. This system is based on a conventional marine X-Band radar and provides three dimensional wave and surface current information by analysing the temporal and spatial evolution of radar backscatter from the sea surface. Measurement and data analysis take about 2 minutes so that major sea state parameters such as the unambiguous al wave number the frequency- the frequency the significant wave height the wave mean and angular spreading the wave period and lengths of the st and 2 nd spectral wave peak as well as surface current are provided in real time. The system can be operated from coastal sites moored platforms as well as from moving vessels. The possibility of unattended operation makes the WaMoS II suitable for long-term monitoring issues. The comparison of different wave parameters derived from radar and buoy measurements yields that the WaMoS II is a reliable wave and current monitoring system especially under bad weather conditions. KEY WORDS: X-band radar wave monitoring al spectra INTRODUCTION Real-time information about the sea state such as wave height and period is crucial for coastal protection as well as offshore operation management (e.g. oil platforms or ships). In the last 3 years routine sea state measurements were carried out mainly using moored wave buoys (Krogstad et al. 998). Although such devices provide reliable measurements they are easily subject to damage and loss. Therefore considerable interest has been shown lately in the use of remote sensing techniques to measure waves and surface currents. One system to remotely measure the sea state is based on a nautical X-Band radar used for traffic control and navigation purposes. In contrast to buoy measurements which are based on the analysis of the buoy motion using accelerometer and tilt sensors radar measurements of the sea state are based on the analysis of the temporal and spatial evolution of the radar backscatter information received in the near range of the radar (sea clutter). The fundamental interaction between the radar and the sea surface is assumed to be Bragg-scattering hence using a X-band radar small ripples in the centimetre range are responsible for the radar return. The longer surface gravity waves become visible in the radar images through: hydrodynamic modulation tilt modulation and shadowing (Valenzuela 978; Alpers et al 98; lant 99; Wetzel 99). The applied wave analysis which is based on the assumption of spatial and temporal homogeneity within the observed area has been validated for deep water measurements. (Young et al. 985; Ziemer and Rosenthal 987). For various deep water situations comparisons with buoy measurements have been carried through. It could be shown that the unambiguous al distribution of the ocean wave energy and the surface current can be derived from a sequence of radar images within a good accuracy (Ziemer and Günther 994; Nieto 995; Dittmer 995; Ziemer 99; Nieto et al. 998 ab). In this paper buoy and radar comparisons from a coastal application will be presented. The paper is organised as follows: in the first section the WaMoS II data capture system and the hardware requirements are presented. Data processing is discussed in the next section. The results of WaMoS II measurements are presented for a deep and a shallow water application at the end of the paper. In both cases the WaMoS II data are compared with buoy data. THE WAMOS II SYSTEM AND DATA ROCESSING WaMoS II is a hardware and software data capture system for wave and current monitoring purposes consisting of an A/D converter a C and a processing software connected to a marine X-Band radar (9.4 GHz). The minimum radar requirements for wave analysis purposes are: () A minimum antenna rotation speed of 24 rpm (antenna rotation time: RT < 2.5 s) (2) a maximum radar pulse length of 8 nsec (3) a minimum antenna length of 2.44 m. In connection with a nowadays standard radar (r.p.l. = 8 nsec RT <.7 s antenna length = 2.44m) WaMoS II obtains radar images with a range resolution of 8.5 m and an angular resolution of.9 for every radar rotation. For the standard WaMoS II analysis a sequence of 32 radar images is used so that waves in the frequency range between.25 Hz and.29 Hz can be detected. This frequency range corresponds to wave periods 3.5 s and 55.5 s. For wave measurements 99-JSC-24K. K. Reichert 5

2 $ # & C K B F F ;? > = : 8 I H J resented at ISOE '99 Brest May roceedings Vol. 3 the radar system must be operated in the near range. Typical WaMoS II operating ranges are between. km to 5 km depending mainly on the wind speed and the installation height. The minimum wind speed required for operational measurements is around 3 ms - (Nieto 997). In order to prepare the radar data for the wave analysis a sub-area of 6m 2m is extracted from the full radar image. The radar image sequence referring to the radar sub-area (xyt) is transformed into the spectral domain by using a Discrete Fourier Transformation : I ( 3 ) ( k x k y ) k k xc yc s. ( x y t) exp i( k x x k y y t) d dk x dk y k xc k yc c Here I (3) (k x k y ) is the three dimensional image k = (k x k y ) is the two dimensional wave number vector and is the angular frequency. The Nyquist limits for each spectral variable are given by k xc = / x k yc = / y and c = / t where x and y is the spatial resolution of the radar image and t corresponds to the temporal resolution of the time series. The resolution for each spectral variable is given by dk x = 2 /L x dk y = 2 /L y and d = 2 /T where L x and L y are the horizontal extensions of the sub-area and T is the duration of the measurement and therefore the length of the time series. Assuming the observed waves to be linear the location of their energy in the three dimensional I (3) (k x k y ) can be determined by applying the linear dispersions relation (Le Blond and Mysak 978): g k tanh( k d) k U (2) where g is the acceleration of gravity d the water depth and U the surface current. The surface current is obtained by minimising the distance between the position of the spectral energy in the image I (3) (k x k y ) and the theoretical position given by the dispersion relation (eq. 2) (Young et al. 985; Senet et al. 997). In order to separate the energy associated with the ocean waves from the background noise the dispersion relation is applied as a pass-band filter: ( 3) f xn ym p I ( k k ) " ( 3) I ( k xn k ym p ) if "! ' ' ' ' ' k( p d U ) k ( k xn k ym ) k( p d U ) $ % otherwise (3) where I (3) f (k x k y ( ) is the filtered three dimensional image. In order to obtain the unambiguous al the threedimensional is integrated over the positive frequency domain only (Young et al. 985; Atanassov et al. 985): c ( I 2 ) ( ( k k ) ) 2 * I 3) ( k k + ) d+ f x y f x y (). (4) To correct the effects of shadowing and tilt modulation on the wave imaging the filtered two-dimensional image I f (2) (k x k y ) is transformed into a wave F (2) (k x k y ) by applying a Modulation Transfer Function MTF (Ziemer and Rosenthal 987): ( 2) ( ) F ( k k ) - MTF( k). 2 I ( k k ) (5) x y f x y where MTF( k) / k with 2 being an empirical coefficient depending on the imaging mechanism of the sea surface. In order to estimate the significant wave height H s from navigational radar images a method developed to determine H s from synthetic aperture imagery (Alpers and Hasselmann 982) is used. This method is based on the assumption that the measured signal to noise ratio (SNR) is linearly related to H s : H A B SNR s 3 4. (6) Here A and B are empirical constants that have to be determined for each WaMoS II installation within a calibration period. In order to determine A and B the best straight line fit though a scatter plot of SNR obtained by WaMoS II versus H s obtained by a reference sensor (e.g. a buoy) is fitted by using a least-square fit. The successful application of this method for deep water applications has been shown by various authors (Ziemer and Günther 994; Nieto et al 998 ab). The names the corresponding notations definitions ranges and accuracy of further sea-state parameters that are provided by WaMoS II are summarised in table. Table : Notation definition range and accuracy of wave parameters provided by WaMoS II. Name Symbol Definition range accur. 2d wave number F (2) (k xk y) refer to equation (4) 2d frequency s frequency F (2) (f5 ) k 67 k F ( 2 ) ( k k ) f f for S( f ). 8A max( S( f )) x y hz -36 S(f) Hz ( 2 F ) ( f ) d Significant H s A < B SNR m - 2m +/- wave height % Mean period T m2 S ( f ) df (Mean zero 2 upcrossing f S( f ) df period) eak period T p / f > Mean wave eak Integrated wave spreading eak wave length (f) arctan( b( f ) / a( f )) ) p D ( f p ); f E / T (f) p ( f ) 4 2( r( f )) r( f ) a ( f ) b ( f ) ) / / p eq (2) with 9-6 m 2 - There is no limit in estimating the wave heights but up to now a H S of 2m was the highest value measured with WaMoS II. 99-JSC-24K. K. Reichert 2 5

3 b L O T W ` [ M R U a Q R Z resented at ISOE '99 Brest May roceedings Vol. 3 N p 2 / Tp and Q 2 / p k p st peak T p / f period 2) p E( f ) p p max ( E( f )) st peak wave length 2) S p eq (2) with V p 2 / Tp and Q 2 / st eak 2) X p Y ( ) p k p 2 nd peak T p2 / f period 2) p2 max ( E( f )) 9-6m³ f p -36 +/- 2 2 Z [ E( f ) p2 p2 2 nd peak wave length 2) \ p2 eq (2) with ] _ 2^ / T and 2 nd peak 2) b p Y ( ) surface current velocity surface current p2 p2 p2 2 / k p2 9-6 m³ f p2-36 +/- 2 U refer to eq. (2) -2ms - +/-.2ms- U refer to eq (2) -36 +/- 2 ) With a and b being the al Fourier coefficients 2) max refers to the first and max 2 to the second energy maximum. ³ These values indicate the typical range but they can be varied for each individual installation. DATA COMARISON Up to now WaMoS II and buoy data were compared only for deep water installations (Ziemer and Günther 994; Nieto et al. 998 ab) i.e. for areas where the water depth is larger than half of the observed wave length (deep water condition) so that the waves can be assumed unaffected by the bathimetry. In this section different sea state parameters obtained from buoy and WaMoS II for a deep and a shallow water installation are presented. Deep water Figure shows as an example the good agreement of a significant wave height (H s ) comparison between a buoy and a WaMoS II. The data were obtained at an off-shore platform (FSO Norne) located in the northern North sea with an average water depth of about 8 m. As the observed wave lengths did not exceed 5 m the deep water condition can be assumed to be satisfied. The presented data sets were obtained during a calibration phase from November 997 till January 998. The WaMoS II data are compared with buoy data gathered at the same location. During this calibration phase the WaMoS II measured every 3 hours over a period of 3 minutes continuously while the buoy delivered minutes mean values. The comparison of the two time series shows a good agreement for lower wave heights (< 2 m) as well as for higher H s values (> 3 m). Also the variations of H s with strong increases or decreases of H s exhibit similar evolution in both time series. The obtained correlation coefficient for the data sets is r =.89. This value represents an accuracy of about 9% for the WaMoS II assuming the buoy data as the truth. The obtained correlation is in the same order of what two wave rider buoys moored close to each other could reach. Fig. : Timeseries of the significant wave height (H s) gained by a buoy (solid line) and by WaMoS II (dotted line). Shallow water The data comparison presented in this section are preliminary results that were obtained within a project supported by the Bundesanstalt für Wasserbau (BAW-AK German coastal authority). In this project radar and buoy data are combined with numerical models in order to determine the wave induced forces on the Heligoland breakwater. The WaMoS II radar antenna has been installed at about 8 m above the sea surface on the island of Heligoland located in the southern North Sea. The WaMoS II analysis area was chosen about 85 m southwest of the radar antenna (see figure 2) where the water depth is about m. The wave rider buoy has been deployed approximately 2 km south of the island where the water depth drops from to 2m. The wave parameters measured by the buoy represent mean values over 3 minutes collected every 3 h. According to that the WaMoS II was set up to measure every 3 h over a period of 3 minutes (5 single measurements consisting of 32 radar images each). Figure 2 shows a WaMoS II screen plot obtained on June :4 UTC. At that time the wind speed was about 4 m/s blowing from south-west. The buoy registered a significant wave height of about m at that time. 2 These values indicate the typical range but they can be varied for each individual installation. 99-JSC-24K. K. Reichert 3 5

4 resented at ISOE '99 Brest May roceedings Vol. 3 Fig. 2: WaMoS II screen plot showing the polar radar image obtained on June :4 UTC from the WaMoS II Station Heligoland. The WaMoS II radar antenna (+) is located in the centre of the image. The dark patch in the south-east of the antenna is the island with the harbour mole. West and south-west of the antenna signatures in form of wave crest pattern are visible. The box indicates the 6mc 2m WaMoS II analysis area Aside from signatures of the island Heligoland the radar image displays surface wave signatures in form of wave crest patterns aligned almost north-southward. In order to determine the sea state parameters from such images the temporal and spatial evolution of the observed wave patterns are analysed in the selected area. Figure 3 shows the frequency- on June :4 UTC as provided by WaMoS II. The wave has a distinct energy maximum at a frequency of.3 Hz and a of 258. This corresponds to waves with a period of about 7.5 s and a wave length of about 65 m propagating east. The slightly asymmetric shape of the indicates a weak bi-modal wave field. In figure 4 the normalised one-dimensional energy S(f) (top) the mean f (f) (middle) and the al spread Spr (bottom) provided by the WaMoS II are compared with the corresponding buoy. This radar has a RT of.89 s therefore the working range is from Hz while the buoy ranges from. -.58Hz. The similarity of the two energy spectra is clearly visible; both exhibit a peak at about.3 Hz. Also the shape of the spectra show a similar behaviour that is typical for a wind sea. A steep increase at lower frequencies than the peak frequency and a slow decrease towards higher frequencies. Also the mean (Fig. 4 middle) and the al spread (fig.4 bottom) show a good agreement between the two sensors. Only where the wave energy is too low to obtain reliable al information the two sensors yield different results. SUMMARY AND CONCLUSION WaMoS II a remote sensing system based on a nautical X-band radar is presented. By analysing timeseries of radar images of the sea surface this system allows to obtain al information on the spatial and temporal evolution of sea state parameters and surface currents in real time. Significant wave height information obtained by WaMoS II have been compared with corresponding wave rider data for a deep water installation. The obtained correlation of.89 between the WaMoS II and the buoy is of the same order that two wave rider buoys deployed close to each other would reach. Furthermore a data comparison between al wave information as obtained by WaMoS II and by a wave rider buoy was presented for a shallow water coastal application. Also this comparison showed a good agreement concerning wave energy distribution and wave information. The results of these experiments demonstrate the capability of WaMoS II for al sea state measurements in deep and shallow water areas. Fig. 3: Directional frequency wave (F (2) (fd )) obtained by WaMoS II on June :4 UTC. The grey scale indicates the normalised energy. The dotted lines indicate the position of the peak frequency ( e.3 Hz) and the peak ( e 258 ). The solid lines indicate the mean frequency ( e.5 Hz) and the integrated mean ( e 56 ). 99-JSC-24K. K. Reichert 4 5

5 resented at ISOE '99 Brest May roceedings Vol. 3 Krogstad H.E. S.F. Barstow O. Haug.O. Marknussen G. Ueland and I. Rodriguez (998). SMART-8: Remotely Sensed Wave Spectra from a moored buoy roceed. Oceanology 98 Brighton. LeBlond. H. and L. A. Mysak (978). Waves in the Ocean Elsevier. Nieto J.C. (995). First experience with the use of marine radar to survey ocean wave fields roceed. of the WMO/IOC Workshop on Operational Ocean Monitoring using Surface Based Radars Geneva March 995. Nieto J.-C. (997). Análisis de Campos de Oleaje Mediante Radar de Navegación en Banda X (in Spanish) h.d. thesis at the dep. of physics of the University of Madrid. Nieto J.C. K. Reichert and J. Dittmer (998a). Use of Nautical Radar as a Wave Monitoring Instrument submitted for publication in Coastal Engineering. Nieto J.C. K. Reichert J. Dittmer and W. Rosenthal (998b). WaMoS II: A wave and current monitoring system roceed. of the COST 74 conference on al wave spectra aris 998 in press. lant W. J. (99). Bragg Scattering of Electromagnetic Waves from Air/Sea Interface Surface Waves and Fluxes Vol. II Kluwer Academic ublishers. rinted in the Netherlands. Fig. 4: top: Normalised one-dimensional energy S(f) middle: mean wave g (f) bottom: al spread Spr provided by WaMoS II (dotted line) and the buoy (solid line) obtained on June :4 UTC. Note that the frequency range of the WaMoS II is limited to.28 Hz while the frequency range of the buoy exceed.58 Hz (not shown in this figure). ACKNOWLEDGEMENT The Heligoland project is supported by the Bundesanstalt fuer Wasserbau Germany (BAW) the GKSS research center and the Technologie Transferzentrale Schleswig-Holstein. The authors thank the German Bundesamt fuer Seeschiffahrt und Hydrographie (BSH) and Statoil for kindly providing the buoy data. REFERENCES: Alpers W. and K. Hasselmann (982). Spectral Signal to Clutter and Thermal Noise roperties of Ocean Wave Imaging Synthetic Aperture Radars Int. J. Rem. Sens. Vol 3. Atanassov V. W. Rosenthal and F. Ziemer (998). Removal of Ambiguity of Two Dimensional ower Spetra Obtained by rocessing Ship Radar Images of Ocean Waves J. Geophys. Res. Vol 9. Dittmer J (995). Use of marine Radars for Real Time Wave Field Survey and Speeding up the Transmission/rocessing roceed. of the WMO/IOC Workshop on Operational Ocean Monitoring using Surface Based Radars Geneva March 995. Senet C.M. Seemann J. and F. Ziemer (997). An Iterative Technique to determine the near surface current velocity from time series of sea surface images roc. OCEANS' 97 Halifax 6-9 Oct. 997 Vol. Valenzuela G.R. (978). Theories for the interaction of electromagnetic and oceanic waves- a review Boundary-Layer Meteorology Vol 3. Wetzel L.B. (99). Electromagnetic Scattering from the Sea at Low Grazing Angles Surface Waves and Fluxes Vol. II Kluwer Academic ublishers. rinted in the Netherlands. Young I.R. Rosenthal W. and Ziemer F. (985). A Threedimensional analysis of marine radar images for the determination of ocean wave ality and surface currents J. Geophys. Res. Vol 9. Ziemer F. and W. Rosenthal (987). On the Transfer Function of a Shipborne Radar for Imaging Ocean Waves roc. IGARSS' 87 Symp. Ann Arbor Michigan May 987. Ziemer F. (99). Directional Spectra from Shipboard Navigation Radar during LEWEX Directional Ocean Wave Spectra The John Hopkins University. Ziemer F. and H. Günther (994). A system to monitor ocean wave fields roc. 2 nd Int. Conf. On Air-Sea Interaction and Meteorology and Oceanography of the Coastal Zone. Lisboa September JSC-24K. K. Reichert 5 5