The ERS-1 and Topex/Poseidon satellite altimeter are flying simultaneously. They measure the sea surface topography with different accuracies and their orbit characteristics are widely different. The joint use of the ERS-1 and Topex/Poseidon data is thus far from being trivial. Still, there is a lot to gain in combining these measurements to observe the ocean. It will be shown that, using Topex/Poseidon data, the ERS-1 orbit error can be reduced in such a way that the Topex/Poseidon and ERS-1 altimeter data become fully consistent and can be used together. The methodology used to correct for ERS-1 orbit error is first described and applied to all ERS-1 cycles (35-day repeat period) concurrent with Topex/Poseidon. Results are validated and the Sea Level Variability from ERS-1 and Topex/Poseidon is analyzed.

Variability of the Azores Current is investigated by combining hydrographic data collected in March, 1992 with ERS-1 altimeter data from January - March, 1992 and January - early April, 1994. The altimeter surface measurements contain information about both the geoid and the ocean circulation: the geoid signal dominates and is generally not sufficiently well known at the small length scales required to identify the mesoscale ocean signal. In this study we use hydrography to distinguish the mean ocean signal from the geoid and thereby determine the absolute dynamic topography (i.e. the ocean circulation signature). One leg of the ship survey lies along a ground track of ERS-1 while in its 3-day repeat missions (phases B and D). We examine the variability of the absolute surface geostrophic velocity perpendicular to the track. The observed temporal and spatial variability may be indicative of a switch in the Azores Current system from a winter situation to a summer situation.

Satellite altimetry and numerical models offer two alternative methods of studying the circulation of the remote Southern Ocean at a range of space and time scales. Repeated tracks of satellite altimeter data, from GEOSAT and ERS-1, have been used to determine characteristic length scales of oceanic mesoscale features, and the time scales of this activity, in the Agulhas Region (24-65 degree(s)S 0-60 degree(s)E). The results will be shown and discussed in terms of the local ocean dynamics and the differences in sampling between the two satellites. The Fine Resolution Antarctic Model is a numerical model of the Southern Ocean, designed to resolve mesoscale features of the circulation. A six year time series of monthly model surface dynamic heights has been sampled as if by an altimeter. These `model altimeter' data have been processed using the same methods as for real altimeter data. The spatial scales of mesoscale activity in the model have been determined using the same techniques as for the real altimeter data and the results compared in order to determine the ability of the model to recreate the mesoscale activity of the region studied.

Traditionally oceanographic features such as fronts and eddies have been monitored using satellites equipped with optical sensors such as the Advanced Very High Resolution Radiometer (AVHRR). The disadvantage of using AVHRR, however, is that it requires cloud free conditions to view the Earth's surface. Synthetic Aperture Radar (SAR) provides all weather, day/night image acquisition capabilities. Currently little is known about the effects of sea surface temperature variations on the returned SAR signal, though it is known that some variation in surface roughness is present where oceanographic features occur. Research into these phenomena is being undertaken at the Defence Research Agency (DRA) in Farnborough, U.K.. To investigate the capabilities of SAR, for ocean front detection, the Ocean Fronts Feature Analysis System (OFFAS) has been developed by Earth Observation Sciences Limited under contract to DRA. Using OFFAS, AVHRR data is used to verify and validate the SAR responses to oceanographic features. This paper is an update to that presented at Oceanology International '94, in Brighton which discussed the research and development work into the use of SAR data from the ERS-1 satellite to identify oceanographic features. Since then the modifications to the software, mentioned in that paper, have been made. OFFAS II now provides the capability to geometrically rectify the images to a regular map projection prior to the simultaneous display and manipulation of both image types. Thus allowing faster location of features of interest identified within a reference image (e.g. AVHRR), delineation of them using a tracing procedure, then the automatic redrawing of the trace on the corresponding target image (e.g. SAR). In addition the geometrically rectified SAR images can be mosaiced together, this is particularly useful where large oceanographic features span more than one image. To date, trials performed using OFFAS have indicated its value in aiding interpretation of SAR image patterns by comparison and verification with AVHRR imagery.

Infrared (IR) remote sensing from satellites is a well-proven technique for measuring sea surface temperature (SST) and for detecting and monitoring oceanographic features which have strong thermal contrast. Unfortunately, cloud cover often limits the continuity of the datasets and therefore their usefulness. There is some evidence that radar backscatter can be modified by sea surface temperature structure which raises the possibility that sensors such as synthetic aperture radar, scatterometers and altimeters could provide an all-weather complement to those operating in the IR. As a background, the results of a project which used coincident airborne radar and IR measurements of an eddy system in the Tyrrhenian Sea during October 1989 are briefly described. During a 5-day period, variations in radar backscatter of several dB occurred in a region where SST varied by 2 - 3 degree(s)C. The correlation between normalized radar cross section, sigma naught ((sigma) ₀ or sigma-0) and SST appeared to depend on the ambient wind. Unfortunately, no satellite radar data were available during the experiment, since Geosat had just failed and ERS-1 was not due for launch until 1991. Building on this work, a study has commenced in which preliminary analyses of ERS-1 altimeter data, from tracks which repeat every 3 days, have been conducted for a section of the Gulf Stream after it has separated from the US coast. The along track variation of sigma naught has been compared with contemporaneous NOAA AVHRR-2 imagery and the relationship between SST structure and sigma naught for individual passes is discussed in terms of environmental parameters such as the local wind field and ocean currents. The possibility of the interaction of environmental parameters such as waves and currents are explored and some evidence for both wave enhancement and attenuation at the north wall of the Gulf Stream is illustrated. Tentative explanations for relationships observed by the various analysis techniques are advanced and further planned work discussed.

There is currently a need for high resolution (one to ten kilometers) vector wind information in a number of oceanographic research areas. Unfortunately, existing systems that are used for extracting wind vectors can not provide such high resolution. Synthetic Aperture Radar (SAR) systems however can provide very fine resolutions (approximately equals 20 meters) and with the advent of new wide swath SAR systems such as Radarsat and ASAR can provide adequate coverage over many regions. In this paper we present an algorithm for extracting wind vector information from SAR data collected by the ERS-1 satellite and show its application to data collected during the Norwegian Continental Shelf Experiment in 1991. The algorithm estimates wind direction by examining the very low frequency clutter in the imagery caused by such phenomena as wind rows and convection cells. The wind speed is then estimated by first calculating the radar cross section of the ocean surface and then determining which wind speed is consistent with the measured wind direction and radar cross section values using existing models. Three models will be examined (CMOD4, IFREMER and a model by Wismann) and compared to the SAR data to determine their applicability, and the resulting accuracy for the estimated wind vectors will be presented.

This paper presents an analysis of altimeter wave height data, spanning a period of 9 years, provided by the Geosat, ERS-a and Topex/Poseidon satellite programs. Monthly means are derived for 2 degree latitude by 2 degree longitude bins from data for each of the three satellites, and these data sets are then intercompared with each other and against NOAA buoy wave height month means. Consistent differences are shown to exist between the various data sets, and correction factors to adjust the satellite data to the buoy measurements are suggested. These correction factors are applied to the monthly mean satellite data, which are then analyzed to investigate interannual variability since 1986. Analysis of Geosat data (for 1986- 1989) has suggested the possibility of a negative correlation between the mean winter wave heights in the NE Atlantic and the NE Pacific, and a positive correlation between the NE Atlantic values and the El Nino Southern Oscillation Index. As preliminary study with ERS-1 and TOPEX data has shown similar wave climate patterns over the period 1991-1994.

Cross-image spectra obtained by combining two sub-look SAR images are utilized in an inversion scheme for extrating the underlying ocean wave spectrum. The reasons for proposing the use of cross-image spectra instead of standard multilook spectra are twofold. First, the cross-image spectra are shown to significantly reduce the speckle noise level while preserving the spectral shape. Second, the cross-image spectra provide information about the wave propagation direction. A closed form expression of the ocean-to-SAR spectral transform and its inversion is formulated for the cross-image case. The SAR cross-image spectra are shown to yield a SNR typically 5 to 8 dB higher than the corresponding standard 3 look image spectra. The noise level is typically reduced with a factor of 2 compared to the 3 look spectra and a factor of 6 compared to the single sub-look spectra. The inherent information about the wave propagation direction are utilized in the inversion scheme to resolve the 180 degree ambiguity without using any external information. The method is demonstrated on airborne C- band SAR data.

The purpose of this paper is to understand the hydrodynamic and sedimentary processes associated to the Rhone river plume (Gulf of Lions, north-west Mediterranean sea). For this study, we use two complementary tools, satellite observations and a 3D hydrodynamic model. In the first part, remote sensing algorithms are described, especially the sea water reflectance and the total suspended matter concentration computation. In the second part, generalities about the study area are presented concerning the wind in the Gulf of Lions and the freshwater discharge variations of the Rhone river. The last part of this paper describes the study of real cases, in which satellite observations and modeling results of the river plume are compared. We were interested in two situations differing from one another in wind and outflow conditions. The shape and the extension of the computed surface sediment plume are in good agreement with observations for both situations. The study of computed oceanic parameters, the chronology of simulations allow a better understanding of satellite observations.

A mathematical model that simulates the spectral curves of remote sensing reflectance is developed. The model is compared to measurements obtained from research vessel or boat in the Baltic Sea and Estonian lakes. The model simulates the effects of light backscattering from water and suspended matter, and the effects of its absorption due to water, phytoplankton, suspended matter and yellow substance. Measured by remote sensing spectral curves are compared by multiple of spectra obtained from model calculations to find the theoretical spectrum which is closest to experimental. It is assumed that in case of coincidence of the spectral curves concentrations of optically active substances in the model correspond to real ones. Preliminary testing of the model demonstrates that this model is useful for estimation of concentration of optically active substances in the waters of the Baltic Sea and Estonian lakes.

Oil thickness is a crucial parameter in the characterization of oil spills for environmental impact. The feasibility of using active microwave sensors to measure thickness was addressed in a series of microwave scatterometer experiments performed by Simrad Marine A/S in a wave tank at the Nansen Environmental Remote Sensing Center. The thickness of the oil layer was maintained at levels similar to the thick part of an oil spill (0.1 - 1 mm). The measurements showed the capability of active microwave sensors to measure oil spill thickness when the oil type is known. In addition to thickness characterization, the experiment studied the effects of oil viscosity, incidence angle, wind speed, wind angle, microwave frequency, and polarization. The backscatter contrast was observed to be greater for lower incidence angles which indicates that the ERS-1 viewing geometry is optimum for the detection and measurement of thick oil slicks. A thickness-dependent backscatter model was developed which included the effects of oil viscosity, composite surface effects, and oil-water reflectivities. The model viscous effects saturated when the oil thickness was greater than the viscous boundary layer thickness. This explained the observed C-VV backscatter contrast saturation for low viscosity diesel oil at thicknesses greater than 0.15 mm. The model predicted contrast saturation at greater thicknesses for the higher viscosity oils. The data showed this trend but the measurements did not extend to thicknesses which tested the model completely.

One of the advantages of combining the higher resolution (approximately equals 25 meters) and re-visit times (three days) of the ERS-1 Synthetic Aperture Radar (SAR) imagery is the capability to image features and processes in the arctic that other systems can not resolve either spatially or temporally. One such process of interest is the dynamics of lead formation and closure and the subsequent growth of ice over the lead region. Much work has been done in the modelling of microwave radar cross section signatures for different stages of ice growth and ice thickness. In this paper we will present the radar cross section evolution of the lead as observed with the ERS-1 SAR. We will demonstrate how the coincident AVHRR imagery on the 24^th can be used to estimate where open water signatures are occurring in the SAR image, and then show the subsequent radar cross section changes as the leads close. We will compare the ERS- 1 radar cross section values with in situ measurement.

Satellite passive microwave sensors are the most effective means to monitor sea ice on a global scale. 18 GHz horizontal and vertical and 37 GHz vertical polarized brightness temperatures from the Scanning Multichannel Microwave Radiometer (SMMR) are compared to the 19 GHz horizontal and vertical and 37 GHz vertical polarized brightness temperatures from the Special Sensor Microwave Imager (SSMI) over the Arctic and Antarctic during the 1987 overlap period n order to merge the two time series. The Norwegian NORSEX and NASA Team multi-frequency algorithms are used on the overlapping SMMR and SSMI data sets. Sea ice extent and area are calculated and the algorithm performance is compared for both hemispheres. The NORSEX algorithm tends to give approximately 10% higher sea ice concentration values than the NASA Team algorithm.

During the last 15 years a number of case studies of the `Odden' sea ice feature in the Greenland Sea has been carried out. The `Odden' ice feature is a rapidly varying peninsula (sometimes island) of ice in the central Greenland Sea. It varies in size from tens of 7 - 800 kilometers within a few days, and both formation and decay can be this rapid. Visual and infrared images from NOAA AVHRR have been used to study this variability for many years, but due to cloud cover detailed analysis of formation and decay have seldom been possible. Examples from several years will be shown. Passive microwave observations of the thermal microwave radiation from the surface have been carried out with the NIMBUS-5 ESMR, the NIMBUS-7 SMMR and the present SMM/I. With this technique it is possible to observe rapid variability, independent of cloud cover. Examples from several years will be given. Finally, ERS-1 SAR (Synthetic Aperture Radar) images has allowed detailed studies of limited areas within the `Odden' ice feature. Examples will be given. Conclusions as to the advantages and disadvantages of the different sensors and the synergy of using all of them will be drawn.

Demand for reliable sea ice information comes from many quarters including ship routing and resource exploitation companies, weather forecasting agencies and glaciological research institution. For operational purposes, this information is typically required for local regions on short timescales. To explore this market a prototype sea ice workstation has been developed. The workstation uses data from several current earth observation sensors, combining the advantages of regional survey, all-weather capability and high-resolution imagery. The output from the workstation is an integrated sea ice chart which can be used to display combinations of ice edge, ice type, ice concentrations, ice motion vectors and sea surface temperatures. During the course of its development significant new progress in automated ice classification has been achieved together with the enhancement of existing ice motion algorithms. The quality of the sea ice information from each geophysical algorithm was assessed through validation campaigns which collected independent datasets. The results of this analysis show the ice type classification to be most accurate in identifying multi-year ice; this is probably the most critical ice category for navigational purposes. A program of end-user evaluation has also been started in which sea ice charts are supplied to operational organizations and value-added services. This will continue during 1994 and provide feedback on the use of the workstation in a semi-operational environment.

Sea ice maps are required by a diverse range of users for scientific research and operational activities. Satellite remote sensing provides opportunities for monitoring and producing sea ice maps at a range of scales, in near real time. During March 1994 ESYS Limited and the University College London Mullard Space Science Laboratory (MSSL) operated a sea ice demonstration project to supply near real time sea ice maps in the southern ocean. The sea ice information was derived from a number of data sources: DMSP SSM/I data; ERS-1 SAR and Radar Altimeter fast delivery data; NOAA AVHRR data; and PoSAT-1 imagery. The maps were supplied to three users, two involved in yacht races in the southern ocean and a ship on an oceanographic research cruise in the waters of the Princess Elizabeth Trough region of Antarctica. The demonstration was successful, supplying the users with sea ice information which they had previously not received and combining data from various sources to produce sea ice maps. The demonstration also developed operational skills within ESYS and enabled the transfer of knowledge from MSSL to ESYS.

Since early 1990 KNMI has been running an operational system to produce and distribute image-products, based on in real time received and processed NOAA-AVHRR-data. Maps of sea surface temperature and total suspended matter concentrations of the North Sea and IJsselmeer are produced with a regular frequency (minimal once a week). If necessary daily maps are produced, indicating locations of blooms of Coccolithophore algae in the North Sea, of drifting layers of Blue Algae and/or ice-cover on the water-surface of the IJsselmeer. Digital image-files on floppy-disk and color-coded hardcopies of the maps are available for the user. During the last years KNMI, in co-operation with other institutes, has put a lot of effort into the stimulation of operational use of the NOAA image-products in the marine environment.

A study is presented comparing ERS-1 SAR.FDC images and ADCP ship data in detecting mesoscale structures in the Alboran Sea during September - October, 1992. First results have shown good correlation between the shear lines of such dynamic structures in SAR images and the ship data, especially in the case of the ADCP velocity records. The two big western and eastern anticyclonic gyres, small eddies associated to the first gyre and the formation of an along current in the Algerian coast are identified in both data sets. As occur in many cases using SAR images, the wind conditions played on important role in the detection of such structures.