Heiko Dankert

Detection of wave groups in SAR images and radar image sequences

The properties of individual wave groups in space and time utilizing synthetic aperture radar (SAR) images and nautical radar image sequences are studied. This is possible by the quantitative measurement and analysis of wave groups both spatially and spatio-temporally. The SAR, with its high spatial resolution and large coverage, offers a unique opportunity to study and derive wave groups. In addition to SAR images, nautical radar image sequences allow the investigation of wave groups in space and time and, therefore, the measurement of parameters such as the group velocity. The detection of wave groups is based on the determination of the envelope function, which was first adopted for one-dimensional (1-D) time series by Longuet-Higgins. The method is extended from 1-D to spatial and spatio-temporal dimensions to derive wave groups in images and image sequences. To test the algorithm, wave groups are derived from SAR images and two radar image sequences, recorded at locations in deep and shallow water. It is demonstrated that the algorithm can be employed for the determination of both location and size of wave groups from radar images. Investigating the detected wave groups in radar image sequences additionally allows the measurement of the spatial and temporal development of wave groups and their extension and phase velocities. Comparison of measured wave group velocities in shallow and deep water gives a deviation of the average value from the group velocities resulting from linear wave theory and shows a clear oscillation of the group velocities in two dimensions.

A Marine-Radar Wind Sensor

A method, called WiRAR, is developed to measure the wind vector using a marine X-band radar as sensor. WiRAR extracts local wind directions from wind induced streaks, which are visible in radar images at scales above 50 m. It is shown that the streaks are very well aligned with the mean surface wind directions. Wind speeds are derived with WiRAR from the normalized radar cross section (NRCS), by parametrization of its dependency on the wind vector, which was performed by training of a Neural Network. The dependency of the NRCS on sea state and atmospheric parameters, such as air-sea temperatures and humidity, were studied with respect to further improvement of WiRAR. Therefore, sea state parameters are extracted from radar-image sequences by derivation of the Signal-to-Noise Ratio (SNR) and wave phase speed at the spectral peak cp. The SNR is directly related to the significant wave height Hs- Recently, the research platform FINO-I has been set-up in the German Bight. This platform provides various environmental data, such as wind measurements at different heights of up to 100 m for studying the atmospheric boundary layer, as well as air-sea temperatures, humidity, and other meteorological and oceanographical parameters. WiRAR is applied to radar-image sequences acquired by a marine X-band radar aboard FINO-I. The derived wind vectors are compared to wind measurements at the platform. The comparison of wind directions resulted in a correlation coefficient of 0.99 with a standard deviation of 12.8deg and for wind speeds with a correlation coefficient of 0.99 with a standard deviation of 0.41 ms-1, respectively. In contrast to traditional offshore wind sensors, the retrieval of the wind vector from the backscatter of the ocean surface makes the system independent of the sensors motion and installation height and reduces the effects due to platform induced blockage and turbulence effects.

Wind- and wave-field measurements using marine X-band radar-image sequences

This paper describes two algorithms for the retrieval of high-resolution wind and wave fields from radar-image sequences acquired by a marine X-band radar. The wind-field retrieval algorithm consists of two parts. In the first part, wind directions are extracted from wind-induced streaks, which are approximately in line with the mean surface wind direction. The methodology is based on the retrieval of local gradients from the mean radar backscatter image and assumes the surface wind direction to be oriented normal to the local gradient. In the second part, wind speeds are derived from the mean radar cross section. Therefore, the dependence of the radar backscatter on the wind vector and imaging geometry has to be determined. Such a relationship is developed by using neural networks (NNs). For the verification of the algorithm, wind directions and speeds from nearly 3300 radar-image sequences are compared to in situ data from a colocated wind sensor. The wave retrieval algorithm is based on a methodology that, for the first time, enables the inversion of marine radar-image sequences to an elevation-map time series of the ocean surface without prior calibration of the acquisition system, and therefore, independent of external sensors. The retrieved ocean-surface elevation maps are validated by comparison of the resulting radar-derived significant wave heights, with the significant wave heights acquired from three colocated in situ sensors. It is shown that the accuracy of the radar-retrieved significant wave height is consistent with the accuracy of the in situ sensors.

Ocean wind fields retrieved from radar-image sequences

An algorithm is presented for retrieving wind vectors from radar-image sequences acquired by a standard nautical radar near at near grazing incidence. The radar operates at X-band (9.5 GHz) with horizontal and vertical polarization in transmit and receive. The algorithm consists of two parts, one for wind direction and another for wind speed retrieval. Wind directions are locally extracted from wind induced streaks, which are approximately in line with the mean wind direction. The algorithm assumes wind direction as normal to the gradient of the amplitude image, which is approximated by finite differences over an appropriate length. The resulting wind direction is taken as normal to the retrieved local gradients. Wind speeds are derived from the radar cross section, by parameterization of its dependency on the wind vector using a neural network. The algorithm was tested and validated using data from a radar mounted in the North Sea. The applicability of nautical radars for wind retrieval is shown for both tower based and ship borne (moving) instruments.

Retrieval of surface-current fields and bathymetries using radar-image sequences

An algorithm for retrieving high-resolution surface-current fields and bathymetries using nautical radar-image sequences is presented. The image sequences are acquired by a common marine X-Band radar. The algorithm retrieves 3d wave-number frequency spectra for each pixel in the analyzed area. With the local 3d spectra maps of hydrographic parameters, wave spectra with its integral spectral parameters, e.g. the local significant wave height, are determined. Thereby areas of about 2 km in radius are covered by the radar. The spatial distribution of these parameters gives a better understanding of the processes near the shoreline for better application of coastal or river training measures. Especially in harbor areas and regions with coastal structures, where the wave fields are highly inhomogeneous due to shallow-water effects like refraction and shoaling as well as building effects like diffraction, is the potential of the developed method. The method is also applicable on moving vessels to get permanently information about local currents, changing water depths in near-shore regions around the ship, especially when the ship is approaching or leaving a harbor.

Ocean surface winds retrieved from marine radar-image sequences

A new method for wind-field retrieval with spatially and temporally high-resolution using marine radar-image sequences is presented. The method is based on analyzing the movement of wind gusts, which become visible in radar image sequences after filtering. In contrast to previous methods, this new technique requires no calibration phase of the radar system. The retrieved wind directions are compared to wind directions of the recently developed method, were wind directions are extracted from wind induced streaks that are orientated in wind direction. Wind speeds are derived from the backscatter of temporal integrated radar-image sequences using a empirical model function, which was parameterized by training a Neural Network. The different methods are applied to radar image sequences acquired by a marine X-band radar mounted aboard an offshore platform in the North Sea. The radar derived winds from more than 1300 radar-image sequences are compared to in-situ wind data measured at the platform. In contrast to traditional offshore wind sensors, the retrieval of the wind field from the backscatter of the ocean surface makes the system independent of the sensors motion and installation height and reduces the effects due to platform induced blockage and turbulence effects

Ocean winds retrieved from X-band radar-image sequences

A new method for retrieving wind speeds and directions using nautical radar-image sequences is presented. The method consists of two parts, one for wind direction and another for wind speed retrieval. Wind directions are locally extracted from wind induced streaks, which are approximately in line with the mean wind direction. The algorithm assumes wind direction as normal to the local gradients of the amplitude image. Wind speeds are derived from the radar cross section, by parameterization of its dependency on the wind vector, which is performed by training of a neural network. For verification of the method the wind direction and speed from nearly 1400 radar-image sequences are compared to in situ data from a wind sensor. The accuracy and limitations of the method are discussed. A second new method is introduced, which enables to retrieve spatial and temporal wind fields from radar-image sequences. Thereby the wind streaks are available in space and time. The local velocity and direction of the wind pattern of each point in the investigated area is determined using tensor-based techniques. This method has the advantage that no calibration of the radar images or training of a neural network is necessary.

Ocean wave groupiness from ERS-1/2 and ENVISAT imagettes

In this study a global data set of reprocessed synthetic aperture radar (SAR) data acquired by the European Remote Sensing satellite ERS-2 is used to study ocean wave grouping using wavelet based methods. For more than a decade the ERS-1/2 satellites have continuously recorded SAR images of the ocean surface. Operating in wave mode both instruments have acquired about 1400 imagettes of 10/spl times/5 km. size (every 200 km along the orbit) each day, which allows to study ocean waves on a global basis. Only coarsely gridded SAR image spectra are available as official wave mode products from the European Space Agency (ESA). As the full image information is required for the present study about 3 weeks of ERS-2 SAR wave mode raw data were reprocessed to 34000 complex SAR images using the BSAR processor from the German Aerospace Center (DLR). ENVISAT satellite, which was successfully launched on February 28, 2002, will provide almost 3000 imagettes a day due to its higher sampling rate (every 100 km). Applying a wavelet edge detection method on the SAR-amplitude-density image and using a region growing approach for the edgefree areas allows examinations of the wave groupiness on a single image. These examinations include group size and number of large groups. The wavelet coefficient as a measure for edge strength is correlated to both wave height and steepness. The wavelet method is compared with an alternative approach, which is based on the classical Hilbert-transform technique. For the latter method the actual sea surface elevation field has to be known. Therefore a quasilinear inversion scheme is used which estimates the surface elevation from complex SAR data.

Estimation of Friction Velocity Using Tower Based Marine Radars

The friction velocity is estimated from image sequences of a marine Radar, which operates at grazing incidence with X-band at horizontal polarization in transmit and receive. Therefore, radar image sequences are analyzed in space and time. The direction of the friction velocity is extracted from streak like features visible in the image resulting from the temporal integrated radar image sequence. The orientation of these streaks are determined by derivation of local gradients of the radar images. The magnitude of the friction velocity is derived from the measured normalized radar cross section by a geophysical model function (GMF), which is parameterized by training of a Neural Network. For further improvement of the GMF the radar retrieved signal to noise ratio, which is strongly related to the significant wave height, is taken into account. The methodology is validated at FINO-I, a research platform in the North Sea, were various meteorological and oceanographical parameters are measured on an operational basis. The radar retrieved friction velocities are compared to in-situ wind directions as well as to the friction velocities estimated from in situ measurements using the TOGA COARE formulation. The comparison resulted in a standard deviation of 13deg for wind direction and 0.41 ms-1 for the magnitude of the friction velocity. In contrast to traditional measurements the retrieval of friction velocity from marine radars is free of platform induced effects, e.g., turbulence, and can be used from moving platforms.

Wind measurements at FINO-I using marine radar-image sequences

A marine X-band radar operating at grazing incidence and horizontal polarization in transmit and receive is used to analyze the backscatter of the ocean surface in space and time. This resulted in a method, called WiRAR, for retrieving the wind vector from radar-image sequences. The method extracts local wind directions from wind induced streaks, which are visible in radar images at scales above 50 m. It is shown that the streaks are very well aligned with the mean surface wind directions. Wind speeds are derived from the RCS, by parametrization of its dependency on the wind vector, which was performed by training of a Neural Network. Recently, the research platform FINO-I has been set-up in the German Bight. This platform provides various environmental data, such as wind measurements at different heights of up to 100 m for studying the atmospheric boundary layer, as well as air-sea temperatures, humidity, and other meteorological parameters. FINO-I provides, for the first time, a broad data base for detailed investigation and validation of our methods. WiRAR is applied to radar-image sequences acquired by a marine X-band radar aboard FINOI. Thereby different atmospheric parameters, like temperatures and humidity, were studied regarding a further improvement of WiRAR. Using this extended version of WiRAR the marine radar system continuously derives winds from radar data sets. These are compared to the in-situ wind data measured at the platform. In contrast to traditional offshore wind sensors, the retrieval of the wind vector from the backscatter of the ocean surface makes the system independent of the sensors motion and installation height and reduces the effects due to platform induced blockage and turbulence effects.