Todd Sikora , George Young, Nathaniel Winstead

Previous studies have demonstrated that satellite synthetic aperture radar (SAR) can be used as an accurate scatterometer, yielding wind speed fields with sub-kilometer resolution. This wind speed generation is only possible, however, if a corresponding accurate wind direction field is available. The potential sources of this wind direction information include satellite scatterometers, numerical weather prediction models, and SAR itself through analysis of the spatial patterns caused by boundary-layer wind structures. Each of these wind direction sources has shortcomings that can lead to wind speed errors in the SAR-derived wind speed field. Manual and semi-automated methods are presented for identifying and correcting numerical weather prediction model wind direction errors. The utility of this approach is demonstrated for a set of cases in which the first-guess wind direction data did not adequately portray the features seen in the SAR imagery. These situations include poorly resolved mesoscale phenomena and misplaced synoptic scale fronts and cyclones.

Measurement of Hurricane Winds using Synthetic Aperture Radars

Jochen Horstmann , Donald Thompson, Hans Graber, Frank Monaldo, Steve Iris

Several synthetic aperture radar (SAR) images of hurricanes have been acquired by the Canadian satellite RADARSAT-1 and the European satellite ENVISAT. These SAR images provide a unique opportunity to investigate the utility of SAR for estimation of hurricane winds. Using the SAR wind retrieval algorithm WiSAR, we have obtained good accuracies (rms error of 18 degrees and 1.5 ms-1) for low to moderate wind speeds using ERS, ENVISAT and RADARSAT-1 SAR data. WiSAR has been implemented at the Center for Southeastern Tropical Advanced Remote Sensing of the University of Miami to generate wind maps from SAR on an operational basis.

Derivation of the sea surface wind field from SAR is a two-step process: In the first step wind directions are extracted from wind-induced phenomena that are aligned in wind direction. The orientations of these features are derived by determining local gradients of the normalized radar cross section (NRCS) from the SAR data. Therefore, a SAR image is smoothed and reduced to resolutions of 100, 200, and 400 m. From each of these images, local directions defined by the normal to the local gradient (to within a 180 degree ambiguity), are computed. From all of the retrieved directions, only the most frequent directions in a predefined grid cell are selected. The 180 degree directional ambiguity can be removed using wind shadowing or other a priori information, e.g., atmospheric models. In the second step wind speeds are retrieved utilizing a geophysical model function (GMF) that describes the dependency of the NRCS on the near-surface wind and imaging geometry. For C-band, VV-polarization, there are a number of popular model functions, e.g., Cmod_Ifr2, Cmod4, and Cmod5. Each of these GMFs is directly applicable for wind speed retrieval from C-band VV polarized SAR images. For wind speed retrieval from C-band SAR images acquired at HH-polarization, a hybrid model function is used that consists of a GMF and a C-band polarization ratio. Cmod4 and Cmod_Ifr2 underestimate the surface wind for speeds greater than 20 ms-1 or so. The Cmod5 algorithm was specifically designed to provide better estimates of the NRCS at higher wind speeds. Differences between the GMF at low to moderate wind speeds are relatively minor. At high wind speeds (>25 m/s) however, the differences become quite significant.

The possibility of estimating hurricane winds using SAR data is demonstrated utilizing C-band SAR data acquired at both VV and HH-polarization. It will be shown that WiSAR enables one to measure wind directions as well as wind speeds at wind speeds greater than 50 ms-1. The resulting SAR-retrieved wind fields are compared to results of a numerical hurricane model as well as in situ measurements. Furthermore, the limitations of Cmod5 will be discussed with respect to hurricane conditions considering SAR as well as scatterometer measurements. Finally, the utility of SAR for estimating hurricane force winds as well as the shape and size of the hurricane eye will be discussed with respect to hurricane forecast.

No paper available for this abstract.

WiSAR - A Tool for Operational Wind Retrieval from Synthetic Aperture Radar Data

Jochen Horstmann , Wolfgang Koch

WiSAR is a methodology that enables to retrieve high resolution ocean surface wind fields from satellite borne synthetic aperture radar (SAR) data on a fully operational basis. The algorithm is suited for ocean SAR data, which were acquired at C-band of either vertical (VV) or horizontal (HH) polarization in transmission and reception. Wind directions are extracted from wind induced streaks that are visible in SAR images at scales above 200 m. These wind streaks are very well aligned with the mean surface wind direction. To extract the orientation of these streaks an algorithm based on the extraction of local gradients is used. Ocean surface wind speeds are derived from the normalized radar cross section (NRCS) and image geometry of the calibrated SAR images, together with the local SAR-retrieved wind direction. Therefore, several C-band models (CMOD IFR2, CMOD4, and CMOD5) are available, which were developed for VV polarization and are extended to HH polarization by consideration of the polarization ratio.v
Within a demonstration project WiSAR is running on an operational basis at the GKSS Research Center. WiSAR was setup to process ENVISAT ASAR data of the North Sea into ocean surface wind fields fully automated and in near real time. The resulting high resolution wind fields are made available via the internet. The system is now running operational since September 2005 and typically extracts wind fields of approximately 1200 km of ScanSAR data per day.

In this presentation we will introduce WiSAR in its manual and fully automated mode. Furthermore, we will validate WiSAR in its fully operational mode by comparison to results of the numerical model NOGAB as well as the high resolution Local Model of the German Weather Service. This will also include the comparison of the C-band models CMOD_IFR2, CMOD4, and CMOD5 as well as the comparison of polarization ratios. Finally, we will discuss the applicability of operational SAR wind retrieval and point out some limitations of the applied methodologies.

No paper available for this abstract.

Extracting Geophysical Parameters from Multi-Polarization SAR Imagery

Donald Thompson , Frank Monaldo, Jochen Horstmann

Remote sensing measurements of the ocean were first collected from space by the SeaSAT satellite in 1978 from a variety of onboard sensors including a radar altimeter, a scatterometer, and passive microwave radiometer systems. A Synthetic Aperture Radar (SAR), also on board SeaSAT, not only demonstrated the ability to measure the ocean surface wind field, but also revealed the presence of surface and internal waves, current and temperature boundaries, and shoaling bathymetry. Since SeaSAT, several other satellite SAR systems, including the European ERS-1/2 series, the Canadian RADARSAT-I, and the Advanced SAR (ASAR) on board ESA’s EnviSAT have been launched and analysis and interpretation of data from these systems have become robust and routine. In recent years, the capability of satellite SAR systems to routinely collect high-resolution measurements of the ocean surface wind vector field has been demonstrated. This capability is important in coastal areas where the high-resolution allows estimation of the surface wind vector virtually up to the coastline.

The possibility for collecting simultaneous multi-polarization SAR imagery has existed on a regular basis only since the launch of EnviSAT by the European Space Agency in early 2002. At present the (C-band) ASAR on board this satellite remains the only available source for such imagery. As a consequence, relatively few techniques have been developed to date for the utilization of such imagery in microwave remote sensing applications. However, the supply of multi-polarization imagery should soon increase dramatically with the launch in January (2006) of the Japanese ALOS (L-band) and the expected launch of the German TerraSAR-X (X-band) later this year. Both of these satellites carry multi-polarization SAR systems. Similar refinement in the interpretation and processing of data from these multi-polarization SAR sensors is required in order to take full advantage of this new technology for remote-sensing applications.

In this presentation, we will discuss our preliminary efforts to leverage previous experience to identify and validate possible new techniques that utilize multi-polarization imagery. Our primary focus to date is directed toward improving the quality of high-resolution SAR wind maps in coastal regions through the use of such imagery in the wind mapping procedure. For the initial phase of our study, we will utilize an extensive database of high-resolution RADARSAT-I (HH-pol) and EnviSAT (VV- and HH-pol) SAR imagery to characterize differences between the VV- and HH-pol backscatter cross section as a function of the surface wind vector, radar geometry, and other relevant parameters as available. We believe this procedure will eventually enable us to refine existing geophysical model functions, and perhaps also to improve the accuracy of wind retrieval using simultaneous multi-polarization imagery. Finally, it may turn out that our study of wind retrieval using multi-polarization imagery may also lead to the development of robust retrieval algorithms for other parameters of geophysical interest.

No paper available for this abstract.

A Three-Scale Ocean Bistatic Microwave Scattering Simulation Model

Xiaoqing Wang

Bistatic SAR can get ocean information more flexibly and have more potential performance compare with conventional single station SAR. With the development of the bistatic SAR technology, it will be a powerful tool for ocean remote sensing. Simulation model is very important for the study of the bistatic SAR for ocean remote sensing, in which the bistatic microwave scattering simulation model is most important.
The mechanism of Bistatic microwave scattering is somewhat similar to but much more complex than back scattering. In this paper, the ocean bistatic microwave scattering mechanism is studied and a three-scale simulation model based on IEM scattering model is given and realized. At last, some simulation results with different wave band, wind speed, bistatic parameters are given and investigated.

No paper available for this abstract.

Reducing Subjectivity in Oil Detection using SAR: The Detection Triangle

Jeff Hurley , Mark Womersley

The use of satellite-based Synthetic Aperture RADAR systems to detect oil in the marine environment has been well established for many years. Leading applications include monitoring for pollution releases (i.e., from vessels and/or oil production platforms) and observing areas for natural oil seepage.

To date, the most effective methodology for near real time analysis involves the visual scrutiny of the imagery by a trained analyst. Often using a set of defined criteria, the analyst will assign a confidence to each possible oil detection. While an attempt is made to reduce the subjectivity of the analysis, and thus improve the overall results, this is generally not possible due to differences between individuals (e.g., experience, time allotted, etc.).

Recent efforts by MDA Geospatial Services and BMT Asia Pacific have focused on developing a ‘detection triangle’ that addresses all the characteristics of the appearance of oil in a SAR marine image. Each side of the triangle represents a theme related to these characteristics. The main themes are Signature, Detectability and Morphology. Under each of these are several detailed characteristics that describe the potential oil event. This paper will examine the three themes and feature the components of each.

Current efforts are directed at reducing the subjectivity in the individual components. This is being done through identifying quantitative inputs (such as wind speed) and more clearly defining ranges for qualitative inputs (such as contextual measures). The end result is a series of inputs that result in an ‘automated’ confidence code for each detection. The analyst, while still responsible for several qualitative inputs, does not make the final decision on an overall confidence (i.e. high / medium / low). The ‘triangle’ has been derived such that the inputs of each side result in a single score, which relates to a confidence ranking. Subjectivity is reduced significantly as slight differences in single inputs would not impact the overall score or ranking. A formula, with weighting for each input, is used to derive the final scoring.

No paper available for this abstract.

Applicability of Conventional Scatterometer Model Functions to High-Resolution SAR Wind Retrievals: The Effect of Spatial Averaging

Frank Monaldo , Donald Thompson, Jochen Horstmann

High-resolution (sub 1-km) wind speed estimates from spaceborne synthetic aperture radars (SARs) are critically dependent upon the model function relating normalized radar cross section to wind speed and direction. Most investigators use some version of the CMOD4, CMOD5, or CMOD-IFR model functions. Each has its own attributes. CMOD4 tends to produce its best retrievals for wind speeds less than 20 m/s. CMOD5 has demonstrated superior high wind speed retrievals, particularly in hurricanes.
One lingering concern is that the radar cross section model functions were generally derived for conventional scatterometers with footprints 25 km in diameter or larger. These large-area measurements represent averages over smaller scale phenomena that can affect radar cross section. Such phenomena include current features, ocean surface waves and small-scale wind fluctuations. However, SAR images produce radar cross section estimates at 1-km resolution or less. Therefore, since the relationship between wind speed and direction and the estimated normalized radar cross section is non-linear, the question arises as to whether conventional model functions are truly applicable to high-resolution SAR wind speed retrievals.
To address this question, we have systematically compared Envisat ASAR wind speed retrievals averaged over a variety of areas ranging from 1 to 50 km against each other and against wind speed estimates from the Navy Operational Global Atmospheric Prediction System (NOGAPS) model. As expected, we find that the variance in the wind speed estimates increases for smaller averaging areas. Smaller averaging areas do not have average over real wind fluctuations on kilometer scales. We further find that the total estimated wind speed variability decreases for retrievals averaged over larger areas. Most importantly, there appears to be little or no systematic biases between retrievals averaged over smaller areas when compared to retrievals over larger areas. We conclude that the model functions developed for conventional scatterometers can be appropriately applied to high-resolution spaceborne SAR wind retrievals.

High Resolution Surface Winds in the Coastal Transitional Zone from SAR Satellite Imagery and MC2 Weather Prediction Model

Philippe Beaucage , Monique Bernier, Julien Choisnard, Anna Glazer, Wei Yu, Robert Benoit, Gaëtan Lafrance

The high-resolution wind fields can not be captured correctly by an interpolation of sparse observations such as from masts and buoys or from large-scale meteorological analysis. Observations distant by more than a hundred kilometres are generally poorly correlated mainly in complex region like coastal environment where important transitions between land and sea occur in topography and surface roughness. On the other hand, wind field information on a regional basis with a very good resolution is provided by Numerical Weather Prediction (NWP) model and Synthetic Aperture Radar (SAR) satellite methods.

NWP models have been used since long ago for weather forecasting applications. These complex numerical models solving the Navier-Stokes equations with different schemes for physical processes and surface properties were not intended, at first, for surface wind purposes but focused mainly on describing correctly the temperature, clouds and rain precipitations. However, the state-of-the-art atmospheric model MC2, developed by Environment Canada, has shown high wind prediction capabilities. It was used to produce the Canadian national wind atlas at a resolution of 5 km (www.windatlas.ca).

With the satellite imagery approach, many instruments derive winds from a snapshot of the sea surface roughness: altimeter, passive microwave imager, scatterometer and SAR. The benefits of these wind retrievals have already been recognized, for example the ECMWF center includes the QuikSCAT scatterometer winds (spatial resolution ~ 25 km) in their analysis. SAR scenes have the advantage of giving realistic wind speed patterns with much more details (spatial resolution ~ 400 m or better) and closer to the coastline. The SAR approach was implemented and improved by a research team of INRS (University of Québec) using RADARSAT-1 images with a geophysical model function based on CMOD-IFR2 and a polarization ratio fitted to RADARSAT-1 specificity.

This paper will focus on the characteristic of the wind fields in the coastal environment of the St. Lawrence River using the SAR and MC2 approaches. A previous study (Beaucage et al., 2006) demonstrated that these two methods used to assess the wind resource compared surprisingly well with a RMS difference of 2.07 m/s and a small bias of -0.13 m/s (SAR-MC2) at a 4 km grid-point spacing. We look at the improvement obtained when using higher resolution simulations (~ 1 km). The behaviour of SAR and MC2 surface wind speeds tend to differ as one gets close to the coastline perpendicularly. More attention is paid to the impact of wind directions (perpendicular or parallel to coastline) on the wind profile toward the coast. Finally, the influence of surface parameters such as the sea surface roughness and temperature on MC2 wind predictions is investigated in order to improve the model.

Beaucage, P., A. Glazer, W. Yu, R. Benoit, G. Lafrance and M. Bernier. 2006. Wind Assessment Using Synthetic Aperture Radar (SAR) Satellite and the Mesoscale Compressible Community (MC2) Model. European Wind Energy Conference (EWEC) Proceedings, 27 feb. - 2 march, 2006, Athens, Greece.

IBPlott - a system for operational use of space borne SAR information in the Baltic Sea

Robin Berglund , Ville Kotovirta, Ari Seina

The Baltic Sea annual transportation is about 700 million tonnes, 40% of which is transported during the winter months. During the last ten years the marine traffic has increased by 34%, and the trend is expected to continue. In the same period, however, the number of icebreakers has not increased. The smoothness of the traffic has been possible due to better ice monitoring, where the use of satellite data has become more and more important. Icebreakers need detailed ice information for their operations.

Spaceborne SAR has been used operationally by the icebreakers in the Baltic Sea since 1992. The first trials used ERS-1 data. Since 1998 Finland and Sweden have jointly acquired 100 - 150 RADARSAT scenes per ice season for ice monitoring and icebreaker operations.

The raw data is received and processed at Tromso satellite station (KSAT). The image data is then transferred using FTP and Internet to the Finnish Ice Service in Helsinki. Processing of the images consists of cutting the areas of interest from the remapped image, scaling the images to 8 bits, and then compressing the image before sending to the icebreakers. As the icebreakers do not have any specialised ice pilots, the captain and mates have to interpret the images themselves. A workstation called IBPlott has been developed by VTT to help the end-users in this task. The GIS type of workstation combines and displays all available relevant information required for making routing and assistance decisions. The information available are - satellite images, ice charts, positions and destinations of the ships moving in the area, visualisation of current speed and tracks (from the AIS system), ice forecasts, water level and weather forecasts, and most recent observations.

The challenge in the system design is to provide the end-users with relevant information from a multitude of data sources in near real time. An element called Façade has been introduced to serve as a portal for the end-user. The Façade filters, tailors and delivers only the relevant data to the end-users, and thus overcomes the main bottlenecks due to narrow bandwidth telecommunications and increased number of information products. This is emphasised when new satellites provide more images and also new imaging modes (such as dual-polarisation).

Most of the end-users in the Baltic Sea have several years of experience from SAR images. New ways of presenting SAR based products to the end-users have been piloted in research-oriented projects with funding from ESA and national funding agencies. VTT and FIMR participate in two such projects: PolarView (an ESA GSE project) and POL-ICE (a nationally funded project in cooperation with Canada). End-user feedback through questionnaires and seminars show that a rather long introduction time is needed to make the users comfortable with new products.

In the paper we present the operational system in use for making SAR images available to the end-users in near real time in the Baltic Sea combining the satellite images with other information products. The end user experiences of the system and future development ideas are shortly presented.

A simplified asymptotic theory for ocean surface electromagnetic wave scattering

Alexis Mouche , Bertrand Chapron, Nicolas Reul

A simplified theory for the prediction of asymptotic solutions for ocean surface electromagnetic wave scattering is presented. The study mostly focuses on the backscatter configuration. We compare numerical results with other existing asymptotic models, i.e. Kirchhoff, the two-scale model, the first and second order small slope approximation and the first and second order local curvature approximation. Further comparisons are made with NRCS from Ku- and C-band sensors in both HH and VV polarizations. In particular, the simplified approach reproduces the observed polarization ratio versus wind speed, wind direction and incidence angle. This approach can thus help to better understand the expected variability of the polarization ratio with the environmental conditions.

No paper available for this abstract.

Oil Spill Segmentation of SAR Images via Graph Cuts

Sonia Pelizzari , José Bioucas-Dias

This paper introduces a new Bayesian semi-supervised segmentation algorithm oriented to oil spill detection using SAR images. The data term, i.e., the density of the observed backscattered signal given the region, is modeled by a finite mixture of Gamma distributions. The prior is an M-level Markov Random Field defined on a 2D grid, enforcing local continuity in a statistical sense. The maximum a posteriori (MAP) segmentation is efficiently computed by applying recent multilevel graph-cut techniques [1], [2], namely the alpha-expansion algorithm. The parameter that controls the degree of homogeneity imposed on the scene is automatically estimated [3] and the number of used classes is optional. The proposed methodology is an extension of [4], which addressed the binary scenario and known smoothness parameter.
The effectiveness of the proposed approach is illustrated with real ERS and ENVISAT data. Its advantage over classical ICM (Iterated Conditional Modes) segmentation method is given. The method has also been applied to a MERIS image containing the same oil spill as in SAR data, as a preliminary step to investigate the potentialities of SAR/MERIS information fusion.

REFERENCES
1. Y. Boykov, O. Veksler and R. Zabih, Fast Approximate Energy Minimization via Graph Cuts, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, No.11, November 2001.
2. V. Kolmogorov and R. Zabih, What Energy Functions Can Be Minimized via Graph Cuts?, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vo.l 26, No.2, February 2004.
3. Z. Kato, J. Zerubia and M. Brrthod, Unsupervised parallel image classification using Markovian models, Pattern Recognition, Vol. 32, 591-604, 1999.
4. S. Pelizzari, Jose M. B. Dias, "Bayesian Adaptive Oil Spill Segmentation of SAR Images via Graph Cuts", on SeaSAR 2006 Proceedings.

No paper available for this abstract.

The near-real time high-resolution ice map service for the ships of the Italian Antarctic Program.

Flavio Parmiggiani

A service of Near-Real Time (NRT) high-resolution ice map delivery was set up during the two last Antarctic Campaigns of the Italian Antarctic Program, to assist ship routing in the Ross Sea. In the last campaign 2005-2006, the service became part of Polar View Project (www.polarview.org).

Two ships of the Italian Antarctic Program usually carry out field operations in the Ross Sea during the austral summer: 1) "Italica" which is the main ship of the Antarctic Program; and 2) RV "OGS Explora". "Italica" leaves New Zealand at the beginning of December to reach the Italian Base "Mario Zucchelli Station" at Terra Nova Bay (74S 165E); the ship usually encounters icefields around 68S, at the onset of the melting season, and, depending on ice conditions, its journey can last from 12 to 25 days. After the download of logistic and scientific material at the Base, the ship becomes an oceanographic research vessel and operates for more than one month in the Ross Sea. "OGS Explora", on the contrary, is a typical oceanographic vessel whose task is carrying out field work during the entire summer season.

The high-resolution NRT ice maps provided to the ships are based on Envisat/ASAR Wide Swath (WS) images and are produced by means of a procedure which consists of the following steps:

1. plan of acquisition over the area of operations notified to ESA with 2 weeks of advance notice (compulsory request by ESA);

2. ASAR WS data are acquired over Antarctica and downloaded at Kiruna Station (Sweden);

3. ASAR WS data in the standard format are available for downloading on the rolling archive of either Kiruna or ESRIN server, from 2 to 4 hours after acquisition;

4. ASAR WS data are transferred to ISAC and processed as geo-coded jpeg images;

5. processed images are transferred to the server of the Italian Antarctic Program in Rome and an "alert" is sent to the ships by e-mail;

6. processed images are available on the ship within 3 to 5 hours from acquisition time.

Examples of the products provided to the ships are presented and discussed.

SAR imaging of shallow water sand ridges aligned with tidal current

Xiaofeng Li

There have been numerous studies of shallow water bathymetric features, i.e., sand ridges or seamounts, imaged by synthetic aperture radar (SAR). Although radar signals do not penetrate water, the surface imprints of the bathymetric features can be imaged by SAR through a current-bathymetry interaction mechanism. The existing radar bathymetry imaging theories are based on the following three processes: 1) The current and bathymetry interaction generates divergence and convergence zones on the ocean surface; 2) The surface current divergence and convergence modulate the wind-generated sea surface wave spectrum; and 3) The variation of the height of short sea-surface waves induces the backscatter variations in the radar image. Most published studies use the one-dimensional radar imaging models involving the above three processes. In these models, the current divergence and convergence are constrained by the continuity equation. Therefore, for bathymetric features to be identifiable in SAR images, the models require that the flow must have a component perpendicular to the orientation of the bathymetric feature.

Although the existing radar models work well in some coastal bathymetry studies, they cannot explain some of the sand ridges observed by RADARSAT-1 SAR. In this paper, six well-known sand ridges in the Bohai Sea were shown in a RADARSAT-1 SAR image taken at 22:00:03 GMT on March 22, 2003. The sand ridges appear as finger-like quasi-linear features in the SAR image. Examining the detailed local bathymetric chart, we found that these features are indeed the sea-surface imprints of sand ridges. The water depth in this area is between 10 - 30 m. The spacing of the sand ridges is about 10 km and the length of the sand ridges is about 20 km. The same sand ridges are also visible on a MODIS true-color image. Contrary to the existing one-dimensional radar model, the tidal current in this area is roughly parallel to the major axis of these sand ridges.

Based on the shallow water equation, a conceptual 2-dimensional tidal model was applied to demonstrate the temporal variations of the current divergence and convergence which are induced by the along-sand-ridge-direction current and ridge interaction. We have experimented with different shapes of sand ridges and different phase differences between the horizontal tidal current velocity components in the X-Y directions. The results are quite persistent qualitatively. Overall, we have demonstrated a new mechanism of tidal current and bathymetry interaction that leads to current divergence and convergence on the sea surface. Therefore, this mechanism explains the fact that sand ridges parallel to tidal currents can also be imaged by SAR.

Observation of Vessel Heave with Airborne SAR

Michael Henschel , Charles Livingstone

Airborne SAR imagery may be used to extract vessel motion with three degrees of freedom. Using a novel combination of high resolution time-frequency analysis and arrayed matched filter bank processing, we have been able to demonstrate the extraction of vessel heave as well as accurate along and across track vessel velocity from Along Track Interferometric SAR data collected offshore Halifax, Nova Scotia.
By extracting heave, surge, and sway from the return signal, these quantities may be used as inputs to Doppler processing of the uncompressed SAR data -- providing more accurate focusing of the target vessel. This paper will outline the methodologies used in the information extraction and validate, with ground truth, the increase in the accuracy of the vessel velocities calculated with the derived heave motion.

Synthetic Aperture Radar for ocean current feature retrievals and surface velocity estimates

Johnny Johannessen , Vladimir Kudryavtsev, Bertrand Chapron, Fabrice Collard, Dmitry Akimov, Knut-Frode Dagestad

The Doppler shift approach and the radar imaging modelling (RIM) approach is integrated and combined with a independent wind-wave SAR retrieval tool for a stepwise retrieval of ocean current feature and surface velocity estimates.

From the RIM, SAR images are simulated provided the wind field, the 2-D surface current, the stratification in the atmospheric boundary layer (ABL) and eventual presence of film are given. In addition, RIM computes the 1-D Doppler velocity. From the observed SAR images, the wind and waves can be retrieved using SARTool, while the measured Doppler phase shift gives the 1-D surface Doppler velocity in the range direction. By an iterative procedure combining the forward simulation and the inverse estimation, RIM can be invoked with new updated fields for computation of the simulated NRCS and 1-D Doppler velocity as well as its partitioning into the contribution from wind, waves and current. This iterative loop, repeated until satisfactory comparison of simulated and observed NRCS is achieved together with the corresponding sea state and surface current fields, represents a highly new and innovative method.

Extensive testing and validation of the method is necessary to obtain a measure of the accuracies of the near surface wind-wave field and surface current derived from SAR images. Such validation studies are carried out using current measurements obtained with HF-radar systems and possibly radar measurements from the Black Sea Platform. In addition, SAR Doppler measurements for quantification of oil spill will be treated as a special test case.

Preliminary findings from this new method are promising. When fully tested and validated, it is expected to advance the use of SAR images in quantitative studies of mesoscale features such as fronts, eddies and filaments.

Radar Imaging of Current Features: A synergy of backscatter power and doppler shift

Vladimir Kudryavtsev , Bertrand Chapron, Johnny Johannsessen, Fabrice Collard, Dmitry Akimov, Knut-Frode Dagestad

Recently two new independent methods for quantitative studies of SAR imaging of current features were published in the July 2005 issue of Journal of Geophysical Research-Ocean. They are the Doppler shift approach (Chapron et al. 2005), and the radar imaging model (RIM), (Kudryavtsev et al. 2005; Johannessen et al. 2005). The study presented here combines these two approaches in a unified DopRim model. This model simulates modulation of radar cross-section and Doppler shift caused by various sea surface phenomena, like: homogeneous or spatially varying currents and near surface winds, temperature fronts and contaminants.

The role of different scattering facets (Bragg waves, specular reflected wave slopes and breakers), as well as intermediate scale waves tilting and modulating facets for different geometry of radar observation (radar wavelength, incidence angle and polarization) is analysed. It is shown that the modulation of the wind wave spectrum by surface current leads to a redistribution of the return power between the scattering facets. This, in turn, leads to modulation of the Doppler velocity making it distinct from the pure surface current. Results of these model simulations are presented for a range of surface current phenomena.

A comparison and assessment of these DopRim model simulations with available observations on background dependence of backscatter power and Doppler shift on wind speed and sea state, as well as their modulations by internal waves, surface slicks and coastal current are given. Overall the agreement is promising, although a number of uncertainties still remain suggesting that further development and refinements are needed.

No paper available for this abstract.

High resolution wind field retrievals off the Norwegian coast: Comparing ASAR observations and MM5 simulations

Knut-Frode Dagestad , Gard Hauge, Johnny Johannessen, Vincent Kerbaol, Fabrice Collard

Global weather forecasts have today a resolution of about 25 kilometres, and variations at smaller scales are thus not precisely described. More detailed forecasts are, however, possible on regional to local scales at a shorter time scale by using non-hydrostatic mesoscale numerical models. Such detailed forecasts, of wind in particular, are of large interest to the general public, and also to commercial users. A major challenge of these forecasts is to initialise the model with a good description of the initial state of wind and other meteorological parameters.

Synthetic Aperture Radars (SARs), such as the ASAR sensor onboard Envisat, are able to derive the wind field with a resolution of about 1 km. However, unlike scatterometers, a SAR cannot retrieve both the wind speed and direction. Either the wind direction must be known a priori before the wind speed can be estimated (which is most common), or vice verca. Thus we see a large potential for exploiting the synergy of SAR and mesoscale models. The models may provide the SAR with an initial value for the wind direction, and the high resolution SAR retrieved wind speed may serve the purposes of both validation and initialising of the numerical models.

We have combined the mesoscale model MM5 with ASAR images from Envisat for retrieval of high resolution wind fields for the coastal regions of Northern Norway and the Barents Sea. The SAR-retrieved and modelled wind fields are compared, to assess the quality of both the MM5-modelled and the SAR retrieved wind fields. Validation is performed against both measurements from coastal meteorological stations and offshore weather buoys. The processing chain is running in near-real-time, thus another purpose is to present the wind fields on the internet to stimulate the use of SAR-wind by operational meteorologists when making short time forecasts and nowcasts

Polar View - A Euro-Canadian Consortium Serving Polar Regions with SAR-Based Services

Thomas Puestow , Desmond Power, Charles Randell

In 2001 the European Space Agency (ESA) Ministerial Council approved a 5-year program dedicated to Global Monitoring for Environment and Security (GMES), called the Earthwatch GMES Service Element (GSE). GSE is intended to develop information services to the public sector based on exploitation of earth observation (EO) satellites. Canada is also a participant in the GSE. In 2003, ESA awarded GSE contracts, each one having a value of more than $2M. C CORE of St. John's, Newfoundland was awarded one of these contracts for their Northern View proposal. With over 30 participants from five countries, Northern View ran for approximately 2 years and successfully demonstrated the utility and market for geospatial information for northern regions. In Stage 2 of the GSE which runs from 2005 to 2008, Northern View has merged with ICEMON, another GSE activity with focus on northern environments, to form Polar View. Polar View represents an international network composed of more than 70 participants from 12 countries with the long-term objective to become a “one-stop-shop” for information required by organizations and individuals concerned with policy, development, and environmental preservation in the Earth’s polar regions. In addition to its primary focus on the Arctic and Antarctic, Polar View is also targeting areas in the mid-latitudes that are significantly affected by snow and ice. Polar View’s comprehensive portfolio of EO-based services supports the monitoring and analysis of issues related to safe shipping, water management, marine environmental security and adapting to climate change.

Evaluation of Earth Observation wind data for marine search and rescue drift forecasting

Julien Choisnard , Desmond Power, Fraser Davidson, Andry Ratsimandresy, Brian Stone, Charles Randell

The marine environment can be one of the most hostile on Earth. Its users include international commercial shipping and fishing, offshore oil and gas exploration and development, and recreational boating. In times of marine distress in Canadian waters, these all rely on the ability of the Canadian Coast Guard Search and Rescue team to respond quickly with lifesaving aid. The Canadian Search and Rescue team uses computer software, called CANSARP (CANadian Search And Rescue Program), linked to wind observations, forecasted winds and estimated historical currents to determine the search region where searchers will begin to look.

In an effort to increase the accuracy of drift prediction for search and rescue purposes, the authors have investigated the use of Earth Observation (EO) derived wind fields in drift modeling software used operationally by the Canadian Coast Guard. This investigating included both QuickSCAT and Synthetic Aperture Radar wind fields. Field campaigns were conducted off the coast of Newfoundland, Canada, during years 2004-2005, in which drift experiments were conducted with complementary EO acquisitions.

Three SAR wind retrieval approaches were used. a) an optimal inversion method (2Dvar) combining C-SAR data, the CMOD-IFR2 geophysical model function and a background numerical weather prediction wind model from CMC; b) the CMOD-IFR2 geophysical model function with an a priori wind direction from CMC model; and c) using the backscatter values corresponding to two neighboring sub-images with slight different incidence angles.

Using a scientific CANSARP simulator, SAR and scatterometers winds have been tested on the drift trajectory forecasting of common search and rescue objects, such as persons in the water or life-rafts. Our comparisons of SAR winds with scatterometer and in situ data show promising results for the optimal inversion method using a background wind model. At first glance, QuickSCAT winds do not improve the forecasted drift for any of the drifter configurations. However, an improvement is observed in short timescale drift trajectories with the use of SAR winds for most of the target configurations. The observed improvement decreases for increased duration of drift tracks, but this is to be expected since the uncertainties and errors associated drift prediction increase with time.

An ocean forecast system, the Newfoundland Operational Ocean Forecasting System (NOOFS), is being set up in order to be run in real time forecast mode using model wind fields. The ocean circulation model will generate ocean surface currents to be used with CANSARP. We present the development of the SAR derived winds data for subsequent ingestion by this ocean circulation model. In the next phase of the project, improved high-resolution sea surface currents will be generated by NOOFS forced with the EO enhanced wind fields obtained from blending SAR derived winds with the CMC model winds.

No paper available for this abstract.

SAR observations of oil spills and natural surface slicks in the Black Sea

Vladimir V. Malinovsky , Stein Sandven

Oil spill detection and natural surface slicks have been studied in the Black Sea by use of Synthetic Aperture Radar (SAR) images from ERS and ENVISAT Wideswath mode. The Black Sea is exposed to increased pollution due to growing oil and gas production and transport by tankers and pipelines. SAR images were collected for the period 1999-2005 in order to (1) investigate spatial distribution of oil spills within the Black Sea, and (2) investigate various forms of natural slicks in SAR images and comparison with other satellite and in situ data.

The spatial distribution of oil spills has been studied by first selecting SAR images where the slick signatures were clearly defined as oil spills. The criteria used for the selection were: size and shape of the slick, proximity to a ship or platform, backscatter contrast between slick area and surrounding waters and wind/SST data from other satellites. More than 400 oil spill areas were identified in 44 ERS and 28 ENVISAT Wideswath SAR images, most of them were long, narrow line features. The largest observed spill area was about 35 km2, while the majority of the spill areas (91%) were less than 3.5 km2. The distribution was concentrated along the sailing route between Bosporus and Novorossiisk, and in the western part close to oil platform and ports.

A number of cases of natural slicks were investigated, where the SAR signatures showed dark features with similar characteristics as oil spills, also called “look-alikes”. The most pronounced features were surface current structures identified by dark, curved lines showing convergence of natural slicks, depicting eddies and fronts during low wind conditions. Low wind areas near the coast and ocean fronts were also found in many images. Furthermore, internal waves were observed by their charateristic wave front and subsequent wave train identified at the surface by zones of convergence and divergence corresponding to crests and troughs of the wave train.

In September 2005 a field experiment was carried out on the south coast of Crimea to study oil spill signature in SAR images. An offshore platform was used to perform atmospheric boundary layer and sea surface measurements under moderate and weak wind conditions at the same time as SAR alternating polarisation images from ENVISAT were obtained. The SAR analysis showed that the radar contrast between the observed slicks and clean water was higher at vertical polarization than at horizontal polarization. The modulation of radar backscattering by long wind waves is about 3 times higher at vertical and 1.5-2 times higher at horizontal polarization for the slicks compared to clean water. The results will be used to develop and improve models to classify slicks and oil spills in SAR images.

The study was supported by the ESA-IAF OSCSAR project under “GMES Networking with Russian and Ukraine 2004-2005” and the National Space Agency of Ukraine.

No paper available for this abstract.

Ice drift and area flux in the Fram Strait from ENVISAT ASAR data

Stein Sandven , Kjell Kloster, Morten Stette

The Fram Strait is the main area for sea ice and liquid freshwater export from the Arctic Ocean to the Nordic Seas and the North Atlantic. The sea ice flux is a key climate parameter which is monitored by Upward-Looking Sonars on fixed moorings at 79 N in the Fram Strait. Ice drift from satellite data has been estimated using passive microwave and scatterometer data, but the resolution and seasonal coverage of these data have not been compared to similar retrievals from Wideswath SAR data. This paper investigates regular use of wideswath SAR data in order to achieve more accurate estimates of the area flux in the strait.

A sequence of ASAR Wideswath images from ENVISAT has been obtained and analysed for ice drift over a period of about two years. Starting in February 2004, NERSC has produced ice area flux profiles across 79 N using profiles of ice drift from SAR and ice concentration profiles from passive microwave data. The time interval between the images is normally 3 days. Data analysis for 2004 and 2005 shows that SAR can be used throughout the year to estimate ice drift in the Fram Strait. The mean zonal ice drift across 79 N varied between 0.10 ms-1 to 0.20 ms-1 with maximal velocities up to 0.40 ms-1. In July and August there was pratically no ice in the strait . The annual area flux was estimated to be 0. 73 x 106km2/year for 2004 and 0.66 x 106km2/year for 2005. Area fluxes retrieved from passive microwave data from 1978 to 1996 varied from 0.72 - 1.12 106km2/year (Kwok and Rothrock, 1999). The accuracy of each ice drift estimate is about 10 %. Scatterometer and passive microwave data, which are used for ice drift estimation mainly in the winter season, show general good agreement with the SAR results. Further, quantitative studies of ice drift from SAR will be conducted in order to validate ice drift from other satellite data and for model validation. The studies are supported by the EU MERSEA Integrated project, the Norwegian SatHav project and the SAR data have been provided through the ESA GMES ICEMON project.

No paper available for this abstract.

SAR-Based Oil Pollution Surveillance in Canada: Operational Challenges

Roger DeAbreu

In the fall of 2006, Environment Canada's Marine and Ice Services (Canadian Ice Service) will assume responsibility for the space-based monitoring of discharged oil pollution in Canadian coastal waters, as well as in the Gulf of St. Lawrence and Great Lakes. This represents an operationalization and expansion of the successful Integrated Satellite Tracking of Pollution (ISTOP) pioneer program led by Environment Canada with the assistance of the Canadian Space Agency, Transport Canada, and MDA Geospatial Services. The transition to formal operations was spurred by the need to enforce and respond to the recently amended Migratory Birds Convention Act and Canadian Environment Protection Act (Bill C-15), which highlighted the requirement for routine surveillance for oil pollution. ISTOP streamlines government response by guiding patrol aircraft to areas where oil spills are suspected, and by focusing any necessary prosecution and/or cleanup efforts. The presentation will describe the efforts to integrate the ISTOP program into CIS Operations. The visual and automated extraction of oil and ship targets from SAR will be described. Special attention will be given to the challenges inherent in meeting the ISTOP requirements with space-borne SAR, i.e. the routine detection of ships and oil over with wide swath radar under a range of background conditions. The impacts and opportunities presented by future SAR sensors, e.g. RADARSAT-2, RADARSAT-C, for ISTOP surveillance will be addressed as well.

Status of the Development and Implementation of a NOAA SAR Products System for Coastal and Marine Operational Applications

William Pichel , Pablo Clemente-Colon, Xiaofeng Li, Karen Friedman, Frank Monaldo, Donald Thompson, Christopher Wackerman, Christopher Jackson, Robert Beal, Hans Graber, James McGuire, Jeremy Nicoll, Erick Malaret, Ray Espiritu

After some years of applications and product research with ERS-1 and ERS-2, the Office of Research and Applications (ORA) of the National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite, Data, and Information Service (NESDIS), began a quasi-operational applications demonstration of synthetic aperture radar (SAR) products in September 1999. This Alaska SAR Demonstration (AKDEMO) initially utilized RADARSAT-1 SAR data to derive high-resolution winds, vessel positions, ice masks, and SAR backscatter imagery for use by operational agencies in Alaska, including the National Weather Service and the U.S. Coast Guard. The demonstration has now run continuously for 7 years. Active efforts to enhance the AKDEMO include: (1) derivation of products from the ENVISAT Advanced SAR (ASAR), (2) development of products from the Advanced Land Observing Satellite (ALOS) Phased-array L-band SAR (PALSAR), (3) an improved ship-detection algorithm, and (4) an experimental wave state algorithm. AKDEMO products are being used for coastal weather forecasting, education, and climatology; fisheries management and enforcement for safety of the fishing fleet and protection of fisheries and marine mammals; search and rescue; and analysis of river ice breakup and associated flooding. Steps are now being taken to develop an operational version of the product software running in the AKDEMO. The first step in this process has been to provide a portable version of the SAR wind software, called the APL/NOAA SAR Wind Retrieval System (ANSWRS). This system has been written and documented to operational standards and has been installed in ORA, at The Johns Hopkins University Applied Physics Laboratory (JHU/APL), at the Alaska Satellite Facility (ASF), and at the University of Miami Center for Southeastern Tropical Advanced Remote Sensing (CSTARS). It is planned to build upon ANSWERS to develop a fully operational SAR products system incorporating current and future automated SAR products. Outside of the AKDEMO, SAR data and products have been utilized by NOAA for hurricane wind studies, hurricane disaster response (oil spills and oil platform change detection), marine sanctuary monitoring (internal wave climatology for ecosystems research, vessel detection for enforcement of protected areas), oil spill mapping for grounded ships, ice hazards to coastal charting operations, mapping of convergence zones for marine debris detection, and ocean feature studies (upwelling, eddies, currents, river plumes). The long-term goal of the NOAA SAR operational applications program is to develop, implement, and validate a suite of fully automated SAR-derived products by the time that operational SAR satellite systems such as RADARSAT-C and Sentinel-1 are launched at the end of this decade.

Improved bayesian wind vector retrieval scheme

Vincent Kerbaol

High resolution properties of SAR imagery has truly contributed to detect, identify and analyse large scale and mesoscale atmospheric phenomena (atmospheric fronts, lows, hurricane, etc) as well as very local meteorological events (rain, breeze, relief effect, etc). In parallel, the ability to derive high resolution wind fields from SAR images using scatterometry approach has also opened new perspectives in applications where the knowledge of the fine scale spatial distribution of wind field is required (offshore wind farming, coastal sail racing competitions, etc).
Recently, research studies to define, develop and validate SAR Ocean Wind, Waves and Currents level 2 products prototypes have been initiated and supported by ESA. Characteristics of the improved wind retrieval prototype are presented here.
As a major improvement, a Maximum A Posteriori (MAP) wind vector estimator is proposed. This estimator, a sub-optimal solution of which was early introduced by Portabella et al., is optimal in a statistical (Bayesian) sense as it fully exploits the a priori knowledge on local wind field (such as obtained from numerical weather prediction -NWP- model or inferred from the SAR image itself). Baesd on this formalism, the "conventional" wind inversion scheme (which uses fixed wind direction provided by a priori wind information) is just a particular case of the general statistical estimator.
Furthermore, an FFT-based estimation method of the wind direction is proposed to complement the statistical a priori information provided by the NWP model. The wind direction is estimated from SAR signature of wind rows. Confidence level of direction estimates is characterized by a Peak Side Lobe Ratio (PSLR). Issues related to the polarization are also discussed. In particular, the choice of an appropriate wind dependent polarization ratio is proposed.
Finally, validation results obtained from the comparison between wind field retrieved from ENVISAT Wide Swath images and more than 1000 in situ buoy measurements will also be presented and discussed.

No paper available for this abstract.

TerraSAR-X and Beyond: High-Resolution Mapping of Surface Current Fields by Spaceborne Along-Track InSARs

Roland Romeiser

The fact that the upcoming German SAR satellite TerraSAR-X (launch scheduled for October 2006) will permit current measurements by along-track interferometry in an experimental "divided antenna", "split antenna", or "dual receive antenna" (DRA) mode has already been discussed in several papers. Results of a recent study indicate that the "classical" DRA mode, which uses two independent receive electronics chains in parallel, as well as a newly proposed alternative mode, which multiplexes a single receive chain between two antenna sections, will permit current measurements with a reasonable accuracy at effective spatial resolutions on the order of a few 100 m. The main disadvantage of the multiplex mode is a quite limited swath width of only 15 km (stripmap operation), whereas the classical DRA mode can exploit the full swath widths of 30 km in stripmap and 100 km in ScanSAR operation. We summarize these latest findings and provide a brief status report on the along-track InSAR implementation for TerraSAR-X.

Given the short effective baselines between 1.2 and 2.0 m and relatively high instrument noise levels of the possible along-track InSAR modes of TerraSAR-X, the predicted current measuring performance is surprisingly good and will hopefully permit a demonstration of repeated current measurements from space over some longer period in a few selected test areas. Promising applications are, for example, the moni-toring of flow patterns over changing underwater bathymetry and the monitoring of river runoff. However, the theoretical potential of the along-track InSAR technique goes far beyond the capabilities of TerraSAR-X: At longer baselines and lower instrument noise levels, velocity measurements with effective resolutions on the order of a few meters can be achieved, and the resolution of actual InSAR-derived current fields begins to suffer from superimposed "noise" of resolved surface wave motions rather than from shortcomings of the instrument itself. We discuss some fundamental relations between instrument noise level, along-track baseline, measuring accuracy, and effective spatial resolution of an along-track InSAR, and we show examples of simulated data products of improved spaceborne InSAR configurations for ocean applications, such as a proposed TerraSAR-X-like follow-on with a relatively simple antenna upgrade and a resulting effective along-track baseline of 3.6 m.

No paper available for this abstract.

Requirements for polarisation combination for ship detection in the Barents Sea

Tonje Nanette Arnesen , Øystein Helleren, Torkild Eriksen, Richard Olsen, Harm Greidanus

The Norwegian Defence has used spaceborne Synthetic Aperture Radar (SAR) operationally since 1998 to monitor Norway's more than 2 million km2 large economic and fisheries zones. The requirements for an independent capability to monitor Norwegian waters increase with the increasing shipping and fishing. This paper presents how different sources of information can be combined for surveillance in remote areas, and it identifies gaps that need to be closed to meet operational requirements.

RADARSAT-2, which will be launched in 2007, will offer large flexibility in selection of polarisation, incidence angle, and resolution. It will have VV-polarisation as well as cross-polarisation (HV and VH) channels, in addition to RADARSAT-1's HH-polarisation. This can lead to a change in preferences and strategy for open ocean surveillance. High incidence angles and HH polarisation have been preferred due to high contrast between the ship and sea. The opportunity to use cross-polarisation will probably give better possibilities for detection of ships at lower incidence angles.

The Hopendjupet-trial was carried out in the Barents Sea between April and June 2006. A total of 17 RADARSAT-1 and ENVISAT Alternating Polarisation (AP) images were used to compare the different polarisation and incidence angle combinations. Recommendations on which channels should be used are given and analyses are done to see if two channels are enough for ship detection in open sea. The trial was a collaboration between the Joint Research Centre and FFI, with support from the Coast Guard, Directorate of Fisheries, and Kongsberg Satellite Services (KSAT).

Some problems with the ENVISAT-orders were noticed. First of all, with an ordering time of two weeks (avoiding rush surcharges) it is difficult to plan the order to be sure ships are in the area. In addition, there were some cancellations and conflicting orders. This will probably not be a problem with RADARSAT-2, which will have maximum ordering time of 3 days, as well as the combination of good coverage (300 km x 300 km) and two polarisation channels.

The Hopendjupet-trial also examined the advantages of SAR in combination with Automatic Identification System (AIS) and Vessel Monitoring System (VMS). AIS and visual data were obtained from the Coast Guard, as well as VMS data from the Directorate of Fisheries. Radar satellites can detect ships not reported by AIS and VMS, while the ships detected with SAR can be identified using AIS and VMS. The trial has been useful to validate SAR, AIS, and VMS and to look at ways to improve the Norwegian Coast Guard's operative capability.

The images have also been used as a preparation for the European Sentinel-1 satellite, which will continue observations after ENVISAT. The planned polarisation combination on the satellite is VV/VH. VV is not good for ship detection due to high reflection from the sea, but the low reflection from the sea using cross-polarisation makes VH and low incidence angle a new alternative for ship detection.

Algorithm Development for Operational Satellite SAR Classification and Mapping of Great Lakes Ice Cover

George Leshkevich , Son V. Nghiem

The all-weather, day/night viewing capability of satellite Synthetic Aperture Radar (SAR) makes it a unique and valuable tool for Great Lakes ice identification and mapping provided that data analysis techniques and capability for using SAR data in an operational setting can be developed. Previous computer analysis of ERS-1 and RADARSAT ScanSAR Narrow images of the Great Lakes using a supervised (level slicing) classification technique has shown that different ice types in the ice cover can be identified and mapped and that wind speed and direction can have a strong influence on the backscatter from open water. However, for image-to-image classification, a library of backscatter signatures for different ice types is needed for use with calibrated SAR imagery. During the 1997 winter season, shipborne polarimetric backscatter measurements (0o to 60o) of Great Lakes ice types using the Jet Propulsion Laboratory (JPL) C-band scatterometer, together with surface-based ice physical characterization measurements and environmental parameters were acquired concurrently with RADARSAT-1 and ERS-2 overpass. Measured vertical-polarization (VV) backscatter values (converted to dB) for 3 ice types and calm water were applied to an 8 x 8 pixel averaged ERS-2 calibrated SAR image. However, a problem with saturation within the analog to digital converter (ADC) of the ERS-1 and ERS-2 SARs leads to a power loss resulting in an underestimation of the normalized radar cross section (NRCS). To correct for power loss, the ERS-2 image (22 March 1997) was recalibrated as described in Rosenthal et al., 1998. In addition, to account for the effects of local incidence angle, the measured (calibrated) backscatter values for the three ice types and calm open water used as "training data" were interpolated every 0.5o between incident angles 19.5o and 26.5o. These "training data" sets were then used to classify the 8 x 8 pixel averaged recalibrated image. As there were rather low power loss corrections to perform in this image, the results are similar to the first classification. Two notable differences are that 1) there is more area classified as patchy snow cover on snow ice covered black ice, and 2) a small area of open water is classified in this image as the result of the more accurate calibration and "training data" sets. When applied to a calibrated RADARSAT-1 ScanSAR Wide A (SWA) scene of the same area on the same day, the library of horizontal-polarization (HH) backscatter signatures produced classification results similar to those of the ERS-2 scene. To obtain the most accurate classification results, the RADARSAT-1 scene was classified by incidence angle of each pixel using training data for 6 ice types and open water. Moreover, the polarimetric radar backscatter measurements acquired during our 1997 Great Lakes Winter Experiment (GLAWEX 1997) reveal that multi-polarization backscatter data (such as that from ENVISAT or RADARSAT-2) for the typical snow-covered snow ice on lake ice in the Great Lakes can be used to map ice and open water without the ambiguity encountered in single polarization data due to variations in wind speed over water.

The Capability of Hurricane Wind Monitoring by SAR

Hui Shen , Will Perrie, Yijun He

Synthetic Aperture Radar (SAR) images are more and more widely used in ocean wind monitoring. It’s capability in accurate measurements for moderate ocean surface wind has been widely accepted. However, fewer studies have been done on the accuracy of SAR wind measurements in high hurricane wind conditions (>25m/s) except papers that focus on qualitative interpretation of the dynamical processes accompanying hurricane process, for example rain bands, and atmospheric boundary rolls. In this paper, we describe our newly developed hurricane wind retrieval method, which can retrieve hurricane wind vectors directly from SAR images without any external information, and the regular SAR wind direction algorithm (SWDA). Several Radarsat and Envisat SCANSAR hurricane images are analyzed. Retrieved SAR wind results are compared with in-situ measurements and QuickSCAT winds. We show that there is good potential for the capability of quantitative high wind monitoring by SAR. The challenge of more accurate wind retrieval of hurricane winds is discussed, which is the focus in our future studies.

Impact of satellite winds on marine wind estimates

Will Perrie , Weiqing Zhang, Hui Shen, Paris Vachon

Our objective is to derive high resolution wind fields and to construct high quality, basin-scale vector wind maps by combining remotely sensed RADARSAT SAR-derived wind measurements with gridded outputs from numerical weather prediction (NWP) models. The gridded NWP wind used here are fine-mesh, routine forecast wind fields, for example from the operational MSC (Meteorological Service of Canada) atmospheric model, for the NW Atlantic. While extensive data assimilation is executed in the operational simulations and forecasts, involving satellite products, no satellite wind data is presently included.
We will use data assimilation methods, to create 6-hourly maps for gridded wind fields, using wind estimates from NWP model, scatterometer wind data and SAR remotely sensed winds. Thus, an optimum combination is determined, in the sense of minimization of a cost function measuring the misfit between input data and the final wind field estimates. Verification is obtained via comparison with available operational buoy observations. Future work could consider accurate assessment of error maps, following standard objective analysis methods, correlation functions and variances for the data assimilation methodology, using all available remotely sensed data.

The Norwegian SatOcean Programme - A multi-use concept for operational marine services

Guro Dahle Strom , Per Erik Skrovseth, Terje Wahl

Satellite monitoring seems to be ideal for a country like Norway, managing large ocean areas (approx 30% of European oceans) and being located far north where polar orbiting satellites give excellent coverage. Cloud cover and winter darkness make SAR the primary source of satellite data.

The SatOcean Programme is a national marine programme to utilise Earth Observation data to support national needs. The main goal is to secure cost-effective utilisation of satellite data in support of national requirements within marine mapping, monitoring and forecasting. The programme is an initiative from Norwegian Space Centre in order to have a co-ordinated effort for governmental institutions to utilise satellite data in a cost-effective way.

Prioritised services within the programme are:
- Sovereignty and fishery control - ship detection and AIS
- Environmental monitoring - oil spill (and AIS) and water masses
- Polar and climate monitoring - sea-ice and ocean circulation
- Fish farming support - algae bloom monitoring

Acquisition and area coverage are being co-ordinated nationally, making it possible to use the same observations for different applications and services. This gives the most optimal support to all the key national services. For SAR data from Radarsat-1,-2 and Envisat a multi-use concept is developed within the programme. This is made possible due to an agreement Norway has with Canada securing data access for Radarsat-1,-2 until 2013 and the participation in the Envisat program in ESA.

Nationally funded R&D activities are also co-ordinated to support the development/improvement of SatOcean services. These R&D activities are complementary to the GMES development.

No paper available for this abstract.

Maritime Traffic Monitoring Using a Single Spaceborne SAR Sensor? Algorithms and Applicability

Gonçalo Valadão , José Bioucas-Dias

Ship detection and trajectory estimation using synthetic aperture radar (SAR) holds a great potential for maritime traffic monitoring [1]. However, the detection and estimation for moving targets in SAR is known to be a challenging problem, for it is well known that the targets appear, generally, blurred and misplaced [2]. Classically, this problem has been solved for a multi-channel configuration where stereo measurements are available [3], [4]. More recently, some single-channel based methods have been proposed, namely in [5], [6]. Therein, basically, is solved the blind angle ambiguity using a single SAR sensor, mainly by exploiting not only the phase, but also the information contained in the moving target echo amplitude. In this paper we show the performance of the basic algorithm proposed in [5], regarding to spaceborne ERS and TerraSAR-X sensors. First we present probability of detection vs. signal to clutter ratio curves as well as Cram´er-Rao low bounds curves, in order to characterize the detection and estimation procedures. Then, we illustrate the application of the algorithm to both simulated and real SAR images scenes containing moving ships. We conclude that maritime traffic monitoring using single spaceborne sensors, by using the above referred algorithm, is feasible.
References
[1] F. Meyer and S. Hinz. The feasibility of traffic monitoring with TerraSAR-X-analyses and consequences. In Proceedings of the 2004 International Geoscience and Remote Sensing Symposium-IGARSS’04, volume 2, pages 1502-1505, Anchorage, Alaska, USA, 2004.
[2] R. Raney. Synthetic aperture imaging radar and moving targets. IEEE Transactions on Aerospace and Electronic Systems, 7(3):499-505, 1971.
[3] M. Soumekh. Fourier Array Imaging. Prentice Hall, 1994.
[4] S. Barbarossa. Detection and imaging of moving objects with synthetic aperture radar. IEE Proceedings-F, 139(1):79-88, 1992.
[5] J. Bioucas-Dias and P. Marques. Multiple moving target detection and trajectory estimation using a single SAR sensor. IEEE Transactions on Aerospace and Electronic Systems, 39(2):604-624, 2003.
[6] P. Marques and J. Bioucas-Dias. Velocity estimation of fast moving targets using a single SAR sensor. IEEE Transactions on Aerospace and Electronic Systems, 41(1):75-89, 2005.

No paper available for this abstract.

Fine Scale Wind Patterns in Coastal Embayments Using RADARSAT-1

Jim Helbig , Paris Vachon, Pierre Pepin, Brad deYoung

Winds are the most important forcing mechanism of currents in the upper layer of the ocean. They consequently affect the transport and spatial distribution of buoyant passive drifters like plankton, including fish larvae and eggs. Spatial (and temporal) variations in the wind field force corresponding variations in the ocean. This consideration is especially important in coastal regions where orography induces local wind variability and bathymetry modulates currents. For management applications like surveying the number of eggs in a bay as input to the estimation of fish stock abundance, spatial and temporal variability in egg concentration can be extremely important. Consequently, oceanographic models that predict fish egg drift should be forced by realistic wind fields, i.e. those that contain small scale structure. However, small scale analyses or forecasts of wind are generally unavailable because of the coarse nature of most numerical weather model grids; moreover, in situ observation of marine winds at a spatial resolution high enough to resolve the wind is essentially infeasible. We report here on the use of RADARSAT-1 SAR imagery to ascertain spatial wind patterns in two Newfoundland bays (Trinity and Conception). Thirty two scenes (24 Wide mode and 8 ScanSAR Narrow) were captured in 2002 under a variety of wind conditions and were analysed for wind speed. A constant wind direction over the bays was assumed and was taken to be coincident with that observed at a local light station. Comparisons were made with QuikSCAT winds which, together with weather maps, provided a larger scale perspective. Wind patterns relative to out-of-bay and into-the-bay winds were derived, and an error analysis was carried out to determine the uncertainty in wind speed due to uncertainty in wind direction. These patterns are being used in models of egg drift to ascertain their importance to egg survey accuracy.

No paper available for this abstract.

Operational Utilization of SAR-Derived Winds for Forecast Operations at the Pacific Storm Prediction Centre

Laurie Neil

In a collaborative project involving the Meteorological Service of Canada (MSC) and Ottawa-based Vantage Point International (VPI) , SAR-derived wind products are being generated and utilized in near-real-time for use by marine forecasters at the Pacific Storm Prediction Centre in Vancouver. Funded in large part by the Canadian Space Agency, the MENTOR (Marine ENvironmental moniTORing) project utilizes the SAR and software engineering expertise of VPI, monitoring and modelling capabilities within MSC, as well as the meteorological and local area knowledge of prediction centre forecasters. The objective is to assess the utility of this imagery, and if appropriate establish an in-house capability for MSC to generate, manage, and display these products rapidly for operational use. This presentation describes milestones and present status of the project, and shows examples of wind images obtained to date. It also touches on the importance of training for operational users, the different purposes for which the data will be used, and the need for a continued supply of high-volume and affordable SAR data over the west coast and elsewhere for various marine applications. The need for further research to address limitations of this dataset will also be outlined.

Support for Science and Operations in Alaskan Waters

Don Atwood , Rick Guritz, Jeremy Nicoll

As the U.S. Order Desk for Radarsat-1, the Alaska Satellite Facility (ASF) has provided operational support to NOAA/NESDIS, the National Ice Center (NIC), and the Canadian Ice Service (CIS) for the past fifteen years. Upon downlink at ASF, Radarsat data is processed to Quicklook, or near-real-time, products and pushed to the Selective Active Archive, where it is further processed for wind products, sea ice analysis, and ship detection. Successful interaction between these agencies has demonstrated that SAR can be effectively used to yield important near-real-time information for decision makers. Timely data has been utilized by U.S. Coast Guard in responding to oil spills, walrus researchers requiring operational support for field campaigns, and the National Weather Service for wind speed validation. Interaction with scientists from the Bering Ecosystem Study (BEST) suggests that there is a significant need for additional information products to support ship-board research in the Bering Sea as well.
The high latitude of ASF permits short revisit times over the waters surrounding Alaska; from one to three days, dependent upon latitude. This fact, coupled with the ambitious acquisition of ScanSAR swaths, permits greater utilization of the data for near-real-time applications. Through the availability of higher level products, GIS compatible formats, and the development of new applications, SAR is beginning to be embraced by new communities of users. The intent of this presentation is to address current and planned developments of ASF near-real-time products; in particular, for the Bering Sea and the Arctic coastline of Alaska. The use of Radarsat data for the routine generation of coastal image mosaics, wind speed charts, ice motion vector plots, and optical/SAR data fusion products will be addressed.

No paper available for this abstract.

Improved Iceberg Detection with Envisat Dual Co-polarization SAR

Carl Howell , Des Power, James Youden, Kelley Dodge, Pradeep Bobby, Charles Randell

Iceberg detection has been shown to be operationally feasible using single polarization SAR. A constant false alarm rate (CFAR) which is an adaptive threshold method is widely accepted as a standard in radar target detection. This method has been successful for SAR point target detection for two reasons; targets are typically significantly brighter than ocean scatter and ocean scatter is typically some variant of normal distribution such as k, gamma or a combination thereof. Single channel CFAR methods however deteriorate under increasing ocean scatter conditions, which are predominately caused by high winds or steep incidence angles. It is also difficult to detected targets that are on the order of the SAR resolution due to their size and sometimes low intensity. These targets may be visually discernable but their assignment of target status comes at the expense of an increase in the false alarm rate.

Envisat co-polarization HH-VV data have potential for improved iceberg detection through a multivariate detection methodology. This work follows from the characterization of probability of detection (POD) vs. probability of false alarm of polarimetric ship detection as described by Liu (2005). This method follows by implementing a maximum likelihood pixel based classification on the available SAR channel data, where both ocean and targets of interest were used as training data for building class discrimination functions. Here we evaluate this approach in an operational sense, where the POD of iceberg targets are evaluated for single polarization HH, single polarization VV, and dual polarization HH/VV. We have also developed an adaptive variant, where the covariance and mean background ocean clutter are trained run time from the neighbouring or current process window. Results and analysis compare the POD for traditional CFAR methods with various combinations of multivariate discriminate functions.

No paper available for this abstract.

ocean waves retrieval in coastal zone

Fabrice Collard , Bertrand Chapron, Fabrice Ardhuin

Retrieval of surface waves parameters from SAR data has recently gained some robustness thanks to the analysis of the long time series of ERS/ENVISAT SAR Wave Mode for wave spectrum inversion. These recent improvements in the understanding of sea echo in SAR signal together with the improved removal of non-wave signatures have contributed to the ASAR level2 wave mode reprocessing. Its extension to narrow and wide swath complex images in coastal waters is presented and discussed. It includes an improved method for nonlinear imaging contribution removal before quasi-linear inversion.

Wave spectra derived from Synthetic Aperture Radar images acquired by ENVISAT in narrow and wide swath mode are compared to in situ measurements from buoys deployed in several field experiments or moored permanently.

It is shown that the extraction of swell wave parameters from the new wide swath Single Look Complex images is possible without any first guess and can bring some clear picture of wave refraction and diffraction patterns in coastal area but also within Hurricanes with an illustration of the striking wave distribution observed in several powerful Typhoons or Hurricane like Katrina that severely damaged Louisiana in September 2005.

No paper available for this abstract.

Doppler centroid anomaly extraction and interpretation.

Fabrice Collard , Bertrand Chapron, Fabrice Ardhuin

A new method for quantitative SAR imaging of ocean surface currents has been presented in the July 2005 issue of Journal of Geophysical Research-Ocean. This method is based on the analysis of Doppler centroid shift in the presence of moving ocean surface.

The two major challenges of this Doppler centroid anomaly analysis are first the estimation of the anomaly and secondly the separation of the sea state contribution and the ocean surface current contribution.

To address the first challenge, the use of precise SAR antenna misspointing form restituted attitude files is shown to be necessary together with some corrections of the bias caused by the azimuthal variations of backscatter.

The second challenge is addressed by the modelisation of the sea state effect and its assessment using 2 years of wave mode data collocated with WAM wave spectra and atmospheric model winds.

Some example are shown to illustrate both sea state and ocean surface current estimations from the careful analysis of the Doppler centroid anomaly.

No paper available for this abstract.

Role of Synthetic Aperture Radar for observing "Local Generation" of internal solitary waves in the Bay of Biscay.

José DaSilva , Adrian New, Alberto Azevedo

Large-amplitude internal solitary waves (ISWs) occurring in packets in the central Bay of Biscay are now well-documented. These waves, instead of having travelled along the thermocline from the shelf break, are instead generated locally in the central Bay by the surfacing of beams of internal tidal energy originating from the shelf break. Two main generation 'hotspots' have been identified: the northern shelf break from which ISWs travel towards the South-South-West (SSW), and the Cape Finisterre region off North-West Spain, where the ISWs travel towards the East-North-East (ENE). These are formed from large-amplitude internal tides which result from the interaction of the barotropic tide with the steep shelf-break topography. In the present paper, we investigate available satellite imagery (Synthetic Aperture Radar (ERS SAR) and ENVISAT ASAR data) to reveal the most prominent local generation 'hotspots' and the main propagation directions associated with them. It has recently been demonstrated, that the Southern wave-trains travel in slightly different directions (055 deg. T and 040 deg.T), and emanate from the Cape Finisterre region, in the Southern Bay of Biscay. The generation sites for these wave-trains lie on either side of the Ortegal Promontory (OP), an undersea "headland" projecting towards the North-West from the north-western coast of Spain (near 44N, 8.5W), and over which the barotropic tides are forced to flow. These generation sites were estimated by calculating the strength of the barotropic tidal forcing in the region, and identifying the likely propagation pathways (rays) of internal tidal (IT) energy. For each generation site, IT rays emanating from "critical" topography (where the ray slope is equal to the topographic slope) in regions of strong barotropic forcing, rise to the surface (for one site after a reflection from the sea-floor) and pass through the thermocline close to the earliest occurrences of the ISWs observed in the SAR. These rays would then produce, through nonlinear processes, the ISWs through the same "local generation" mechanism that has been used to explain the occurrence of the ISWs in the northern and central Bay. In addition to the above mentioned wave-trains, we have now identified new wave packets travelling towards the North-North-West (NNW) which also appear to be generated near OP, and are consistent with the "local generation" hypothesis (after a reflection from the sea-floor). In this paper we also provide evidence that the IT ray that propagates upwards from OP and towards the West-South-West (WSW), is capable of generating ISWs that propagate towards 235 deg. T, as predicted but not shown in our previous investigations. The methods we have used to deduce the generation sites for the SAR observed waves are expected to prove equally useful for studies in other areas of the world's oceans. The use of SAR is particularly useful for investigation of the "local generation" mechanism because, although other sensors are capable of observing ISW packets from space, it is necessary to observe simultaneously large cloud-free ocean regions, in order to validate the IT path predictions. We can thus use SAR to test if the "local generation" mechanism is more widely applicable than previously thought.

Ocean monitoring services for the Spanish coast

Celestino Gomez Cid , Juan Ignacio Cicuendez Perez

Spanish coast is very sensitive to several problems on Ocean security: illegal immigration, illegal traffic and illegal fathering. In particular Spain is suffering oils spills problems and huge problems with immigration coming from Africa where also a humanitarian problem arises. In response to this situation GMV has developed in the frame of the GMES (Global Monitoring for environment and Security) satellite based services for Spanish authorities involved in such issues. The use of SAR and innovate algorithms are being used in combination with available user data: AIS, VMS, coastal radars, and optical imagery for specific cases. The information is combined to provide monitoring and support to decision information for the users. The early results are promising and it is expected to extend such services to other users.

The algorithms extract information from the SAR image: vessel positions, estimated speed, direction and cross checked with available in-situ data. Such information is used to aware the user of the existence of the vessels and also, based on a database from intelligence, to identify suspect vessels, and perform a vessel back tracking to identify the most probable route the ship has followed.

The service provides the full chain starting from the retrieval of new available satellite data, the request of in-situ data for the specific area and timeframe using a dedicated server, the extraction of parameters, information correlation and the presentation of the results in a GIS application where the user can interact. Based on the information provided initially the user can request to process historical data to identify routes and also to change the requirements for the future satellite data acquisitions based on the new information derived and driven by the user experience.
All together will provide valuable aiding information in support of the surveillance activities of the different institutions.

No paper available for this abstract.

RADARSAT-1 observations from Davis Strait to Baffin Bay for off-shore oil seep detection.

Paul Budkewitsch , Robert McGregor, Goran Pavlic, Gordon Oakey

The investigation of possible and confirmed natural hydrocarbon seeps identified in SAR images was the focus of a study using RADARSAT-1 data in the eastern off-shore areas of Baffin Island. This work is intended to provide supportive information towards the evaluation of hydrocarbon potential from Davis Strait to Baffin Bay. With a limited data supply and a large area (~250,000 km2) of investigation, our data acquisition strategy employed a collection of three complete passes in Wide beam modes (ascending and descending passes). SeaWinds data from the QuikSCAT mission were used to estimate wind velocities close to the RADARSAT-1 data acquisitions and suitable data were analysed for low backscatter features (potential seeps). A broad range of spatial form and amplitude characteristics were found in the images and a quantitative and qualitative selection process was use to classify (rank) the slick-like features. Over 60 features were recognised, however many are suspected to have origins other than a hydrocarbon source. Few of these were detected in SAR images more than once. A good example of a persistent slick was identified in data from 2004 in the vicinity of Scott Inlet (Baffin Bay) which has been studied since at least the 1970’s. While the time window for monitoring oil-seeps with SAR in Arctic and Sub-arctic basins is short, the relative scarcity of detailed geophysical (e.g. seismic) data sources suggests that results from SAR can be used as a targeting tool for directing further evaluation and assessments of off-shore regions for their hydrocarbon potential as well as providing data for monitoring programs of active hydrocarbon seeps.

No paper available for this abstract.

Extreme wind and wave conditions observed by synthetic aperture radar on a global and regional scale

Susanne Lehner , Thomas Koenig, Antonio Reppucci, Johannes Schulz-Stellenfleth

It is well known that synthetic aperture radar (SAR) provides information on ocean winds and surface waves. SAR data are of particularly high value in extreme weather conditions where the sensor is still able to penetrate the clouds providing information on different ocean surface processes.

In this presentation some recent results on the SAR observation of extreme wind and ocean wave conditions will be summarised. Particular emphasize will be put on the investigation of tropical and extratropical cyclones in the North Atlantic performed in the framework of the project EXTROP, which is a cooperation between DLR, GKSS, University of Hamburg and the University of Miami. The study will give an overview of the capabilities of present spaeborne radar sensors and their potential to improve the forecast of extreme weather conditions.

The study is based on the use of ENVISAT wide swath, RADARSAT scansar and ERS-2 wave mode data. Wide swath and scansar data are well suited for a detailed investigation of cyclones. Several examples like, e.g. , hurricane KATRINA will be presented, demonstrating that these data provide valuable information on the two dimensional structure of the both the wind and the ocean wave field. Comparisons of the SAR observation with parametric and numerical model data will be discussed. Some limitations of standard imaging models like, e.g., CMOD5 for the use in extreme wind conditions are explained and modifications are proposed.

An analysis of a new data set of two years of ERS-2 wave mode data will be presented. The data were re-processed at DLR and comprise over 1 Millionen globally distributed high resolution SAR scenes of 10 by 5 km size. Statistics of wind and wave parameters derived using a new empirical retrieval approach (CWAVE) will be presented. The study will focus on extreme storm events in the North Atlantic and North Pacific.

Finally the study will summarize the capabilities of the new high resolution TerraSAR-X mission to be launched in October 2006 with respect to the monitoring of extreme weather conditions. The mission will provide a spatial resolution up to 1 m and has full polarimetric capabilities. The sensor is expected to provide valuable information on different air sea interaction processes like, e.g., ocean wave breaking.

Toward an Automated and Standardized SAR Coastal Environmental Toolkit

Christopher Wackerman

Under funding from NOAA/NESDIS, ONR, NAVO, and US Government, we have been developing an automated toolkit to extract coastal environmental information from SAR imagery. The goal is to have a portable system with a standardized set of environmental products that could be set up at multiple SAR imagery distribution sites. Then, instead of moving S0AR imagery around, the environmental products could be generated at the download sites and the products distributed using significantly less bandwidth.

To date, tools for estimating coastal winds, coastal waves, current fronts, ice conditions, and near-shore bathymetry have been developed and incorporated into the tool kit. A version is running operationally at NOAA/NESDIS using data from the Alaska SAR facility. Work is on-going to set up another site (perhaps at the NERSC in Norway) to demonstrate the ability to automatically generate the products at both sites and be able to access them via the internet. The hope is that eventually algorithms from other suppliers can also be integrated and their products posted on the web.

The wind , wave and near-shore bathymetry products have been validated with comparison to in situ ground truth. Root-mean-squared errors are: (1) for wind speed are 2.1 m/s; (2) for wind direction 22 degrees; (3) for significant wave height 1.1m; and (4) for near-shore bathymetry (assuming a long-dwell airborne SAR system) 1-2 m. For ice conditions, the supervised classification algorithm can classify coarse ice classes with between 70% - 90% accuracy; post-classification rules help to clean up some of the resulting errors. The current front algorithm has yet to be significantly validated.

In this talk, the long-term goal of this project will be discussed along with the status of the demonstration. The individual algorithms will be briefly described along with the validation results. Potential future directions of the project will also be addressed.

Ship Monitoring from SAR: An Automated Detection System for Ships and Oil Spills

Christopher Wackerman

Under funding from NOAA/NESDIS we have been developing an automated ship detection and oil spill detection system for spaceborne SAR imagery. The ship detection algorithm is a constant false alarms rate (CFAR) algorithm which uses a K-distribution model for the SAR ocean responses. The parameters of the K-distribution model vary spatially, and the algorithm estimates them locally from the imagery itself. The user can set the value for the probability of a false alarm to use for a given image.

The algorithm has been tested in a number of applications, most recently as part of the European DECLIMS (Detection, Classification and Identification of Marine Traffic From Space) program in its first and second benchmark tests. Based on these validations the algorithm can detect all ships that are longer than one third of the sample spacing of the image, and it can detect approximately half of the ships that are smaller (although the smallest ship in the test sets was approximately 25 m). These detection probabilities are with the false alarm rate set by the user beforehand. The algorithm also provides an estimate of ship length and width from the signature itself. This algorithm has recently been incorporated into the NOAA/NESDIS operational system for use off of the coast of Alaska.

The oil spill detection algorithm is based on work performed to estimate ice type in SAR imagery using a classification algorithm that finds the optimal linear combination of image features to separate various classes. In this case, the classes are oil and non-oil, and the set of features area characterizations of the shape and edges of dark patches in the SAR imagery. The algorithm has not been extensively validated, but has been applied to a small set of SAR images with known oil spills.

Both algorithms will be described and the validation results shown.

No paper available for this abstract.

Creating a Smart Bay - Placentia Bay, NL

Neil Cater

To address the ever-increasing supply of and demand for ocean data and information, CCMC has been a driving force behind the development and implementation of SmartBay, a demonstration project in marine information management based in Placentia Bay, Newfoundland. The area sustains a healthy inshore fishery and is a year round ice-free base for the offshore fishery. Placentia Bay contains a terminal for the passenger ferry service link to mainland Canada, an oil refinery and a large storage and transshipment port facility for Newfoundland's offshore oil production. The latter two alone account for over 500 transits of the bay by oil tankers each year. Because it also hosts an abundant and diverse marine ecosystem, it is identified as an environmentally sensitive area. The Brander-Smith Report (1990) on oil tanker safety in the wake of the Exxon Valdez disaster determined:

"the risk of spills is highest in Eastern Canada, particularly in Newfoundland. Placentia Bay is considered by many to be the most likely place in Canada for a major spill."

SmartBay creates a prototype information utility that provides tools for integrated marine management Advances in the collection and processing of ocean related data have resulted in very rapid growth in our capacity to understand our ocean environment. While collecting ocean data represents a significant investment, in and of itself it represents little value to the end user. Data only has real value when it tangibly benefits the ocean's community of interest. The requirement to turn data into application-specific information products, deliver and present these products to the mariner in an efficient and effective manner provides commercial opportunity for the private sector in terms of R&D and commercialization.

Ocean data is gathered through many means ranging from recorded human observation, to surface based instrumentation such as buoys and surveillance radars to airborne and satellite remote sensing. Initially, SmartBay will rely on surface based observation. However, the longer-term vision is that it support other layers of information. Satellite based earth observation is best suited to measurement of broad ranging ocean phenomena. Useful application layers derived from satellite earth observation include ocean colour and temperature as well as sea surface winds. The "big picture" used in applications ranging maritime security to oil spill monitoring and search and rescue operations is largely acquired through fusion of data from some or all of these sources.

We must be conscious of the fact that while Placentia Bay has been identified as the host for SmartBay, it is by no means a project that is isolated to Placentia Bay. Placentia Bay is very much attached to and influenced by the surrounding North Atlantic. Marine traffic originates from and is destined for ports elsewhere. Prevailing winds drive in pollution from the North Atlantic shipping lanes such as that from bilges of ships crossing the mouth the Bay. Storms that affect safe navigation in the bay originate outside the bay in the North Atlantic. Therefore, even if the SmartBay project were to address only the information requirements of the Placentia Bay stakeholder, it must look well beyond the confines of the bay to include the East Coast continental shelf and much of the Northwest Atlantic.

No paper available for this abstract.

Towards the Support of the St. Lawrence Global Observatory (SLGO)

Guy Aube , Yves Crevier

The governments of Canada and Quebec announced on June 1 the St. Lawrence Plan for a Sustainable Development 2005-2010. The objectives of the Plan are to identify, develop and follow-up on environmental indicators for areas in and around the St. Lawrence River. The plan covers themes such as agriculture, navigation, wetlands, pollutants, sedimentation deposits, and physiochemical properties of the water. There is a significant amount of information on the global system of the St-Lawrence being produced by various stakeholder organizations. However, there is yet to be a structure that provides easy and fluid access to this wealth of information. The CSA recognizes the importance of the creation of a SLGO and its contribution towards an IOOS international architecture. It is CSA understanding that the SLGO would contribute to the IOOS by improving the exchange of observations from in situ, air craft and EO sensors; coordinate capacity-building efforts in the Saint-Lawrence region in improving and sustaining the contribution of EO systems; support the development of EO applications and dissemination of geospatial data on past, present and future state of the Saint-Lawrence River ecosystem; etc.. Currently, the St. Lawrence community makes little use of EO data, including SAR, and geo-information tools. CSA is one of 10 federal and provincial government partners working to develop an integrated management approach for the St. Lawrence River basin. As satellite imagery is used worldwide for monitoring and managing water, land and climate change, as well as for other commercial and scientific purposes, departments and organizations of the St. Lawrence Plan partnership should opt for space-based EO technologies to meet the plan's goals more efficiently. Through the Government-related Initiatives Program (GRIP) and the EO Application Development Program (EOADP), the CSA has the capabilities to foster the development of EO applications attuned to the Saint-Lawrence River. This paper willl address: the status of monitoring and information gathering and dissemination; proposed governance structure for the SLGO; plan for the integration of EO in the SLGO for land/coastal/water base line and monitoring (science, assimilation, pre-processing, information extraction, etc.).
References:
1. Canadian Group on Earth Observations (2006) GEO Societal Benefits. 7p.
2. Canadian Space Agency (2006) Safeguarding the St. Lawrence Ecosystem. http://www.space.gc.ca/asc/eng/media/backgrounders/2006/0612.asp.
3. Douglas-Westwood (2006) Global Markets for Ocean Observation Systems. 131p.
4. Government of Canada (2006) Towards a Federal Earth Observation Strategy (FEOS). Document for consultation. 42p.
Government of Canada, Government of Quebec (2005) St. Lawrence Plan for a Sustainable Development 2005-2010. 47p.
5. Great Lakes Commission (2004) Great Lakes Observing System (GLOS) / Business Plan. 117p.
6. Group On Earth Observations (2005) The Global Earth Observation System of Systems (GEOSS) 10-Year Implementation Plan. 11p.
7. Leclerc, G. (2006) Canadian Observing Initiatives Affecting the Great Lakes. Monitoring the Great Lakes: from information to action, Rochester, NY, April 4-6 2006.
8. National Office for Integrated and Sustained Ocean Observations - Ocean.Us (2006) The first U.S. Integrated Ocean Observing System (IOOS) Development Plan. A Report of the National Ocean Leadership Counciland and The Interagency Committee on the Ocean Science and Resource Management Integration. Washington, 104p.
9. Steering Committee for the Implementation of the SLGO - SCSLGO (2006) Business Plan 2006-2009 / St. Lawrence Global Observatory (SLGO). Rimouski, 45p.

No paper available for this abstract.

Coordination for Earth Observation Marine Surveillance and Monitoring in Canada

Yves Crevier , Micheal Manore, Ross Lagrandeur

Adequate, accurate and timely information in support of marine surveillance, safety and security is critical for Canadian maritime decision makers and stakeholders to effectively deliver their services.

Earth Observation can significantly contribute to the surveillance of Canadian and international waters, ensuring an "available, reliable and affordable" source of information.

The challenges for an integrated / coordinated national maritime information management system reside in the fact that maritime issues are cross sectorial, cross disciplinary, and cross functional and has to serve the objectives of various operational and scientific stakeholders.

The overlapping use of space resources within an integrated / coordinated national maritime information management system creates a lot of collaboration opportunities among stakeholders. However, this situation also introduces potential operational conflicts inherent to the system limitations and application requirements.

Within this context, it has been suggested that CSA take the lead in the creation of a Canadian Government EO marine coordination committee to ensure the coordinated use of the space EO systems with respect to the needs and requirements of the various marine stakeholders. The committee would work towards the development of an integrated and cost-efficient use of Earth Observation resources.

The proposed presentation discusses the creation of the Earth Observation Marine Surveillance Coordination Committee - EOMSCC, identifies the principle advantages and limitations of an integrated marine surveillance system in Canada; it presents the current capabilities of existing systems in support to an operational multi-user pool; identifies opportunities for marine stakeholders to collaborate for an improved use and share of EO data and derived information; and proposes an action plan for coordination and awareness in support to an operational maritime surveillance program using satellite technology.

No paper available for this abstract.

Canadian Space Agency's Hurricane Watch program, supporting research on wind field retrieval from RADARSAT-1 images

Sonya Banal , Steve Iris, Robert Saint-Jean

Since 1999, the Canadian Space Agency is leading a project called Hurricane Watch. Its main partners are: the National Oceanic and Atmospheric Administration (NOAA), the Canadian Department of Fisheries and Oceans (DFO), MacDonald Dettwiler & Associates Ltd. Geospatial Services (MDA), the Canadian Hurricane Centre of Environment Canada and more recently the Center for Southeastern Tropical Advanced Remote Sensing (CSTARS) of the University of Miami. Since its beginning, the project has evolved from ordering RADARSAT-1 data from archives, to routine storm monitoring of the northern hemisphere cyclone prone regions (Atlantic, Pacific and Indian ocean) and dedicated last minute planning of image acquisitions. The main objective of the program has been to support research activities focused on wind field vectors extraction obtained from the analysis of SAR surface roughness information.

The 2005 Hurricane season was particularly active in terms of storm quantity and impact and a successful one for the program with a record of 72% success rate in acquiring images over hurricane's eye or edges. This paper will present some examples of Hurricane Watch acquisitions and an overview of the high value of archives acquired over the years, with a focus on the 2005 and 2006 data that were provided to researchers. We will also present the evolution of procedures used at the Mission Management Office for planning acquisitions over such geographically dynamic phenomena as cyclones. Considering the maturity of the program and the experience gained over the years by CSA Hurricane Watch team, some talks have also been initiated with partners in order to bring the program to the current level of a research-focused program to a more operational level. Finally, with the launch of RADARSAT-2 the program will be evolving, as more opportunities of scheduling acquisitions within a shorter delay will be possible.

Assisted Acquisition Planning for Maritime Surveillance with Commercial Satellite Imagery

Jeff Secker , Paris W. Vachon, Michael Robson, James Rowe

The Commercial Satellite Imagery Acquisition Planning System (CSIAPS) is a mission planning system under development by Defence R&D Canada - Ottawa (DRDC Ottawa). CSIAPS is intended to assist Collection Managers within operational centres with planning for the acquisition of Commercial Satellite Imagery (CSI). One aspect of CSIAPS is an expert system component that uses pre-defined rules to recommend sensors that can be used to ensure that the acquired imagery will permit the Image Analyst (IA) to address the specific Request for Information (RFI). The first two expert system scenarios to be addressed within CSIAPS are Maritime Surveillance and Arctic Surveillance.

Within CSIAPS, the Maritime Surveillance scenario is divided into three categories: Shipping; Sea Ice; and Oil Spill. Within the Shipping category there are three items: Detection; Classification; and Heading and Speed Estimation. Within the Sea Ice category there are two items: Iceberg Detection; and Ice Type and Ice Edge Detection. Within the Oil Spill category there are two items: Detection; and Tracking. For each item there may be one or more details (parameters) that need to be defined. For each of these items, and with knowledge of the AOI size, rules have been defined that permit the system to make a recommendation for the most appropriate satellite and sensor mode.

CSIAPS was built with a custom Graphical User Interface (GUI) that invokes algorithms in the commercial software Satellite Tool Kit (STK) for access calculations, coverage analysis and visualisations. CSIAPS is developed using Python, an open-source object-oriented programming language. The display and manipulation of raster and vector data was developed using OpenEV, an open-source library of raster and vector classes. The expert system was developed using the open-source software CLIPS. The database system for the imagery archive was developed using the open-source software MySQL.

Variational Assimilation of SAR Imagery for Near-Shore Regional Wave Estimation

David Walker

This paper describes a approach for assimilation of satellite-based SAR imagery into a near-shore spectral wave model. The basis for the approach is the SWAN near-shore wave-spectrum model (Booij et al., 1999 JGR 104, 7649) and the SAR-image spectrum is determined from the wave directional spectrum using the nonlinear mapping of Hasselmann and Hasselmann (1991, JGR 96, 10 713) and Krogstad (1992, JGR 97, 2421). The assimilation procedure is based on a variational methodology (see e.g. Bennett, 1992, Inverse Methods in Physical Oceanography, Cambridge).

In the approach, an objective function, a positive-definite measure of the difference between the predicted and observed SAR-image spectrum, is minimized by iteratively adjusting the incident-wave boundary condition for the SWAN model. The gradient of the objective function with respect to the incident-wave spectrum is calculated from the combined solution of the adjoint of the SAR model and adjoint wave model with the error between the actual SAR-spectrum and model estimate of the SAR spectrum as input. This gradient is used in a conjugate-gradient minimization procedure to determine the boundary condition which produces the best-fit between the predicted and observed SAR spectrum. It is assumed that the wave field changes slowly enough in the near-shore region that it is effectively stationary. In the objective function, the estimated boundary condition is penalized to ensure the uniqueness of the resulting estimate.

The approach is demonstrated by estimating the wave field using ERS SAR images for a 100km x 200km region off the US East coast, centered on the US Army Corps of Engineers Field Research Facility (FRF) at Duck NC. The assimilation procedure is used to estimate the wave spectrum field for a number of different SAR data sets and the results are compared to high-resolution in-situ observations of the wave directional spectrum from instruments at the FRF, and to available frequency-direction spectrum information from NDBC buoys located in the region.

Results show that assimilation of SAR-image data into the SWAN model yields good agreement for the estimation of the wave direction, significant wave height and wave period under most conditions; however, wave energy at high frequencies and large propagation angles relative to the SAR look direction is lost. This is due to well-known SAR-imaging effects.

No paper available for this abstract.

Project Polar Epsilon: Joint Space-Based Wide Area Surveillance and Support Capability

Lt. Col. Ross LaGrandeur , LCdr Andrew Samoluk

Project Polar Epsilon: Joint Space-Based Wide Area Surveillance and Support Capability, a capital project within Canada’s Defence program, will use Canada’s RADARSAT 2 as a contributing sensor and enable all-weather, day/night persistent surveillance of Canada’s Arctic region and ocean approaches.
Project Polar Epsilon is a transformational initiative to introduce space-based wide area surveillance to Canadian Government marine surveillance stakeholders. Through RADARSAT 2, Polar Epsilon will provide wide area domain awareness over Canada’s ocean approaches and Arctic region. Accordingly, much effort is being expended in algorithm development to innovate the potential of RADARSAT 2 for the sovereignty and surveillance mission. Capabilities that are being developed and which will be demonstrated during the definition phase of the project include: land surveillance of Canada’s Arctic Region via change detection techniques; ship detection; environmental sensing and ocean intelligence; direct satellite reception and processing; near-real time dissemination; and mission planning tools for satellite selection

No paper available for this abstract.

Satellite and Modeling Components of the Southern California Coastal Ocean Observing System

Ben Holt

The Southern California Coastal Ocean Observing System (SCCOOS) is a consortium of institutions that extends from Morro Bay at the southern edge of central California to northern Baja California in Mexico. The goals of SCCOOS are to streamline, coordinate, and further develop individual institutional efforts by creating an integrated, multidisciplinary coastal ocean observatory for the benefit of society. By leveraging existing infrastructure, partnerships, and private, local, state, and federal resources, SCCOOS plans to develop an operational coastal observing system to address issues in coastal water quality, marine life resources, and coastal hazards for end user communities spanning local, state, and federal governments and the public.

We describe the JPL efforts for SCCOOS which include providing a broad suite of satellite data and products and the development of an operational regional ocean model for this region. The satellite data component includes wind fields from SAR and QuikScat, optical/nearIR-derived products including sea surface temperature and chlorophyll, and sea surface heights. Except the SAR, these products are provided in near-real time. Data are accessible through a Web-based interface and can be visualized and overlaid using a MapServer interface. The 3-dimensional ROMS is a nested model with the finest grid resolution of 1 km for the entire Southern California Bight. The model assimilates various satellite products and in situ data. In the future coastal radar data will be assimilated and soon will be deriving nowcasting products for a fall experiment.

After the overview of this SCCOOS effort, this paper will focus on efforts to improve the utility of SAR wind products including product mosaicking and add land masks, validation with buoy data, and to demonstrate the value of using SAR wind products together with an assimilated blended wind products derived from MM5, QuikScat, and buoys to improve model circulation estimates.

No paper available for this abstract.

Ocean SAR Observational Requirements in GEOSS

Michael Manore

The 10-year goal of the newly-established Group on Earth Observation (GEO) is the coordination of the world's environmental monitoring activities into a coordinated, comprehensive, and sustainable 'Global Earth Observation System of Systems' (GEOSS) that will maximize the societal benefits of investments made in earth observation systems. One starting point is understanding the information requirements of the various benefit areas (e.g., improved weather prediction, reduction and mitigation of the impact of natural disasters, coastal zone management, etc.) and the corresponding observational requirements from EO systems. This presentation will provide an overview of GEO, the activities related to the definition of requirements, and the role of ocean SAR in the GEOSS.

No paper available for this abstract.

Stewardship of Coastal Waters - the Information Systems Perspective

Jerzy Graff

Stewardship is a term being used increasingly by resource industries, government agencies, and community activists to describe their philosophy of resource use. In the context of the marine environment a key enabling mechanism for effective stewardship is integrated coastal zone management or ICZM. Although ICZM is perceived to have a well defined structure and methodology content its adoption at regional and national scale over the past 10 years suggests that we are only now just beginning to get to grips with the real issues involved. Furthermore, in the context of today’s broader marine EEZ monitoring demands ICZM is only part of a widening information services infrastructure. In Europe, a European ICZM strategy document was tabled in September 2000, and it is only now, in the wake of the European Commissions Demonstration Programme on ICZM, and feedback from EU member States that a more cohesive vision of the enabling mechanism is forming; by way of the new “European Marine Strategy” document currently out for consultation. The United Kingdom itself has followed a similar path, with the government strategy on marine stewardship and conservation published in May 2002; a major review of progress “An Integrated Assessment of the State of UK Seas” published early this year, and a new national “Marine Bill” pending. An underlying feature that sits at the core of European marine stewardship evolution is a focus on geospatial information or GI and its key role in enabling the information systems infrastructure needed for effective stewardship. The role of GI and its importance in a growing Europe was prophetically set out in the 1999 document “European Spatial Development Perspective” which together with the Global Monitoring of Environment and Security initiative GMES in the same year, launched jointly by ESA and EU, has led to a wholly redefined concept of marine stewardship today reflected I believe also in Canada and USA. It is against this broader backcloth of EEZ stewardship that my presentation is set namely; focusing on the EO SAR driven information services content.

The GMES services development model provides a good starting point to view and explore the modern demands placed on EEZ stewardship and how they might be met. The GMES philosophy addresses the provision of end user information services through integration of measurement, modelling and prediction within a geospatial systems environment. A key feature of the information delivery mechanism is adoption of web services which enables centralisation of knowledge creation resources and their maintenance yet does not restrict dissemination and delivery to end users. Adopting the GMES philosophy in the marine environment poses many problems aligned mostly to web services technology issues regarding communication, marine data (including electronic charts) exchange standards, visualisation and security. Nevertheless, considerable progress has been achieved in recent years in developing new marine information services architecture and demonstrating service applications such as oil spill detection, search and rescue and coastline monitoring which highlight the importance of SAR. The ability to draw together digital data, forecasts and other information from disparate locations and centralise the complex process of knowledge generation for multi community access through a web services portal presents a major advance towards realising the true meaning of marine stewardship.

My presentation will endeavour to set the scene for marine stewardship demands today and will highlight in particular the importance of EO SAR environmental monitoring and operational forecasting as key components of an information services architecture that reflects the evolution of marine stewardship technology today.

What Is The Future for Marine SAR Exploitation

Chris Wackerman

Although the focus of this conference is on spaceborne SAR exploitation for marine and coastal regions, the future of SAR marine exploitation is expanding in a number of directions. SAR systems are being put on new types of platforms, from manned airborne systems to large unmanned systems to hand-held remote-controlled airplanes. SAR sensors themselves are evolving. Systems currently exist that can generate time-series of SAR images over the ocean (via constant rotation of the antenna) and interferometric systems have been growing additional antennas so that whole arrays can be flown over the water. Finally, advanced exploitation algorithms for SAR data are being developed that have move away from the standard approaches of estimating geophysical information directly from radar cross section patterns, and into combining SAR images with physics models to estimate larger ranges of environmental parameters.

This talk will present an overview of the current state of the art along these three dimensions of future SAR marine exploitation, and present a discussion on what the future might hold for operational SAR systems in coastal and marine regions.

No paper available for this abstract.

ESA Sentinel-1 SAR mission

Malcolm Davidson , Bjorn Rommen, Nicolas Floury, Laura Moreno Patricia, Evert Attema, Guido Levrini

The ESA Sentinels constitute the first series of operational satellites responding to the Earth Observation needs of the ESA-EU Global Monitoring for Environment and Security (GMES) program. Their mission requirements are not based on technological objectives, but rather focus on the continuity of existing services exploiting EO data and satisfying user requirements derived from GMES applications.

The ESA Sentinel-1 mission, which is scheduled to be launched in 2011, represents the SAR component of the Sentinel series. The current mission concept envisages a C-band SAR system with sigma nought, ambiguity and resolution performance very close to ERS/ENVISAT, while the revisit/coverage performance are greatly enhanced. Important applications driving the mission concept include marine (vessel detection, oil spill mapping, wind products) and sea ice mapping applications.Critical mission requirements include
- Coverage: Global coverage with daily revisit of the European Economic zone.
- Timeliness: near-realtime products in support of marine applications, weather forecasting and standard product delivery (including with satellite tasking, data acquisition, processing and delivery) in less than 24hrs.
- Processing: all data processed to level 1b/2 and disseminated in near-real time
- Resolution: single-look spatial resolution 20 x 5m (azimuth x ground range) with 220km swath for interferometric applications. Other modes are under study.
- Polarisation: Selectable dual-pol VV/HV or HH/HV
- Implementation: constellation of two satellites operating simultaneously
The final mission and sensor characteristics are currently under study. Collaboration with other SAR programs (e.g. Canadian Radarsat Constellation) is being actively pursued to enhance observations and reduce revisit time.

The presentation will discuss how the mission requirements for Sentinel-1 were derived and key characteristics of the current technical implementation. It will also discuss how this mission concept responds to GMES user requirements.

No paper available for this abstract.

RADARSAT Constellation - Changing the Paradigm of Maritime Surveillance

Guy Seguin , Andrew Eddy

Maritime surveillance covers a broad range of potential applications areas, including ice and iceberg monitoring, marine winds, oil pollution monitoring and response and ship detection. Establishing an accurate and up-to-date maritime domain awareness is critical to ensuring that users in each of these areas can achieve their disparate goals. This typically involves integrating information from a wide array of sources, including satellites, airplanes, shore radars, ships and buoys.

With the exception of meteorological information and ice monitoring, most maritime surveillance applications have made only marginal use of satellite information, particularly those tied to operational environments. RADARSAT-1 has demonstrated its value in ice monitoring and iceberg detection, and RADARSAT-2 will be much more integrated in maritime surveillance. However, delays in revisit, inability to cover wide areas on a regular basis, cost of data and insufficient redundancy are some of the main hurdles keeping space-based radar observations from becoming a backbone of maritime surveillance.

The RADARSAT Constellation, to be launched beginning in 2011, represents a paradigm shift in satellite system design, and will likely bring comparable changes to the way the maritime domain awareness picture is constituted. This paper explains the main components of the system, its maritime applications, and potential impacts on the operations of the maritime community as radar satellites play an increasingly important role in maritime domain awareness.

No paper available for this abstract.

The Interdepartmental Marine Security Working Group

Laura Hoy

The Interdepartmental Marine Security Working Group (IMSWG) is a model of an integrated national program that facilitates the collaboration of federal government departments’ and agencies’ marine security efforts.

The IMSWG is a forum to ensure integrated and collaborative initiatives. Participants act within the scope of their respective departments’ and agencies’ marine security responsibilities and programs. The IMSWG identifies and coordinates federal government actions in support of Canada’s objectives concerning public security and anti-terrorism in the maritime realm, as well as its international marine security obligations.

One of the IMSWG’s objectives is the enhancement of Canada’s marine domain awareness through the coordination of applications that would provide persistent, wide-area maritime surveillance. Through its Marine Security Coordination Fund, the IMSWG has allocated funding to projects such as the Coordinated Earth Observation Marine Security project as a means of increasing the coordination of marine security efforts across federal departments.

As interdepartmental relations play an increasingly important role in maritime domain awareness, this paper will explain the IMSWG’s purpose and function as an interdepartmental working group in the marine security realm, with emphasis on its interests and capabilities in furthering marine domain awareness.

No paper available for this abstract.

Marine Wind Analysis with the Benefit of Radarsat-1 Synthetic Aperture Radar Data

Richard Danielson , Michael Dowd, C. Harold Ritchie

A nonlinear regression approach is used to construct surface wind analyses based on synthetic aperture radar (SAR) measurements and collocated operational marine wind forecasts. Numerous Radarsat-1 SAR acquisitions along coastal regions of eastern and western North America are considered. Each analysis is iteratively constructed using a cost function that minimizes both discrepencies with the wind forecast and with the SAR acquisition. The relative importance of these discrepencies is defined by two corresponding error covariance matrices that are allowed to vary from case to case. An independent set of buoy observations is used to gauge the analyzed winds. Comparisons are first made using a conventional method of combining SAR and model data. Analyses are then assessed for various parameterizations of the SAR error covariance matrix.

Operational Ocean Management: SAR State of the art

Johnny A. Johannessen

Operational ocean management and its application of SAR data and products can be grouped according to: (i) global to regional open ocean, (ii) regional to local open shelf and coastal seas, (iii) sea ice infested polar waters, and (iv) neighbouring regional to local temporarily frozen seas. Each of these groups have distinct, but often different, requirements for both SAR observations and their spatial and temporal coverage as well as availability of near real time and off line data and products. Moreover, the availability and robustness of the SAR processing tools varies for these different application areas. In this presentation an overview in the state-of-art SAR application for operational management in Europe will be provided. It will follow the grouping specified above, capitalize on several ongoing service demonstration projects under Global Monitoring for Environment and Security (GMES), and address operational management in the context of sea state monitoring, sea ice monitoring, oil spill monitoring, ship detection and high resolution wind field monitoring. In so doing, the satellite SAR ground receiving station segments and their interfaces with thematic data centre and monitoring and forecasting centre will also be considered.

No paper available for this abstract.

From Svalbard to St. Johns: What do we bring and what can we do with it?

Johnny A. Johannessen

The Svalbard workshop in 2003 focused on answering three key questions, notably in the context of progress made since the first workshop on operational SAR coastal applications at John Hopkins APL in 1999. The questions included: a) what are the currently available tools for deriving coastal environmental products; b) what validation results have been performed; and c) what are the future directions for coastal monitoring. Upon leaving Svalbard a clearer picture regarding these questions emerged, and we agreed to target our research and development on imaging radars to provide uniform, consistent, and more sustainable products. In addition it was recommended to educate and train the user community through interactions and demonstrations on how best to utilize products from satellite SAR systems.

As we now 3 years later are arriving at the shores of St Johns refined and new capabilities have emerged thanks to continuous SAR related research that we all are attracted to. The range of users now utilizing SAR-derived products operationally are also widening and including: weather services, ferry operators, meteorological centers, sea ice centers, military organizations, maritime safety agency, dredging companies and offshore industry. These users can be grouped within several of the 9 themes of the Global Earth Observation System of Systems (GEOSS) and its European contribution Global Monitoring for Environment and Security (GMES). The presentation will highlight selected examples of recent research findings and applications of relevance to GEOSS and GMES. In so doing issues regarding strength and weaknesses of satellite SAR products will also be addressed.

No paper available for this abstract.

Creating a successful international consortium - lessons learned from 30 years of collaboration

Charles Randell

C-CORE has surveyed and thrived in the oft confused milieu that exists between applied R&D and commercialization. The corporation has achieved sustained growth over 30 years and has diversified from one domestic market to achieving 50% of its revenue internationally with a presence in numerous market sectors. As executives, we like to boast that this sort of performance requires hard work, prescience, entrepreneurial drive, and a world class team. That's true; but the vehicle, at least for C-CORE was very often collaboration. This presentation will provide a high level look at the benefits and challenges of collaboration, and a more detailed analysis of the Global Monitoring for Environment and Security program Polar View. Polar View is a collaboration of over 30 entities from 9 countries.

No paper available for this abstract.

Monitoring and surveillance for security and domain awareness - needs and requirements

Brian Whitehouse

Application of civilian spaceborne synthetic aperture radar (SAR) sensors to maritime security operations is usually viewed from the perspective of target detection. The purpose of this presentation is to introduce this subject from the perspective of ocean intelligence.

The state of the ocean influences the performance of military sensors, weapons, vessels and personnel. As a result, security operations that monitor and forecast the state of the ocean have advantage over those that do not. For example, meteorological and oceanographic information can be used to improve the target detection capabilities of spaceborne SAR sensors. Certain components of this information can be extracted from the SAR scene itself and this is giving rise to cost-effective secondary applications of scenes collected primarily for other operational purposes, such as ship, oil spill and ice detection. In Canada, this is leading to joint civilian-military applications and is blurring the distinction between SAR infrastructure designed for military and civilian operations. In both sectors, the focus is on rapid environmental assessment.

Although there will always be demand for new sensors, and although temporal resolution continues to limit the use of certain spaceborne sensors, the present limiting factor in this field is not detection or lack of practical application. It is the requirement that the information obtained by the sensors be delivered in operational time frames and formats, using available data communications infrastructure. Presently, military and civilian security operations are overwhelmed with environmental information obtained by sensors mounted on various platforms. Much of it they cannot use because it does not meet this requirement. As a result, environmental monitoring sensors that meet and often exceed engineering and scientific objectives often fail to achieve expectations of user demand within the maritime operations sector.

No paper available for this abstract.

Monitoring and Surveillance for Security and Domain Awareness: SAR State-of-the-Art

Paris Vachon

Ship detection in synthetic aperture radar ocean imagery amounts to the detection of bright targets against the ocean clutter background. The potential for ship detection depends upon many factors including the local wind and wave conditions, the observation geometry, the radar frequency and polarization, and the vessel size and type. Based upon these factors, simple models have been developed to predict the potential for ship detection. These models have contributed to understanding the ship detection performance of current SAR systems and to the design of future SAR systems.

Considerable effort has gone into the development and validation of ship detection software. A common approach is to set a pixel-based threshold according to the clutter statistics and the desired constant false alarm rate. Several pre- and post-processing steps are possible. The Detection and Classification of Marine Traffic from Space (DECLIMS) project carried out an inter-comparison of available ship detectors by developing a common benchmark data set.

As we move towards automation, such as to meet the stringent operational needs of the Polar Epsilon project, a key objective is to reduce false alarms and to cull-out known or uninteresting ships.

Polarimetry and dual-polarization image acquisition present opportunities to improve ship detection and work towards ship classification. These methods also provide the possibility of ship/iceberg discrimination.

The success of ship detection varies within an image according to the local wind conditions. Winds speeds derived from the SAR image would permit identification of regions in the image in which detection of a certain size of vessel is feasible.

It has been proposed that inter-look coherence techniques could help with the detection of smaller vessels. Unfortunately, the evidence for this remains uncertain.

Automatic Identification System (AIS), a ship-to-ship VHF transponder system intended for collision avoidance, is proving to be a valuable asset to identify ships in SAR imagery, especially as we move into the era of spaceborne AIS reception. Furthermore, AIS has permitted a heretofore unprecedented acquisition of known ship signatures, allowing development of models for ship radar cross section and target fading, which could aid in the reduction of false alarms in operational ship detection software. Attempts to carry out automatic target recognition have been demonstrated based upon SAR/AIS data sets.

State-of-the-Art in the use of SAR for Stewardship of the Coastal and Marine Environment

Gordon Staples

Stewardship of the coastal and marine environment entails the use of SAR data for a number of applications including illegal fisheries monitoring, oil pollution detection, mapping of coastal features, and detection of ocean features. In the case of illegal fisheries monitoring and oil pollution detection, SAR can provide stand-alone information, but for operational usage, SAR data are typically integrated with other data sources. For example, the combination of SAR with AIS data provides optimal information for illegal fisheries monitoring. Similarly, SAR data combined with airborne surveillance can be effectively used to counter illegal ship-bilge dumping. The objective of this presentation is assess the state-of-art in the use of SAR for coastal and marine applications. The presentation will be divided into two parts.

The first part of the presentation will provide a review of the use SAR for marine coastal and marine applications. Specifically the review will address the capabilities and limitations of the use of SAR data. The capabilities will be addressed from a perspective of whether the SAR data provides effective stand-alone information or whether the SAR data must be integrated with other data sources to meet operational requirements. The limitations will be addressed from a perspective of sensor, environmental, and orbital constraints.

The second part of this presentation will assess if upcoming SAR sensors will address any of the current limitations. For example, will radar polarimetry provide improved information, or will SAR constellations (i.e. multiple SAR sensors) improve the information due to increased re-visit? Examples will be drawn from pending SAR missions such as RADARSAT-2 and the RADARSAT Constellation Mission.

No paper available for this abstract.