ABSTRACT ASAR GROUND PROCESSOR VERIFICATION AND WAVE VALIDATION

ASAR GROUND PROCESSOR VERIFICATION AND WAVE VALIDATION

Y-L. Desnos(1), J. Closa(1), B. Rosich(1), A. Bellini(1), H. Laur(1), P. Meisl(2)

(1)ESA-ESRIN, Earth Observation Department, Via Galileo Galilei,

Casella Postale 64 00044 Frascati, ITALY

Email: ydesnos@esrin.esa.it, jclosa@esrin.esa.it, brosich@esrin.esa.it,

abellini@esrin.esa.it, hlaur@esrin.esa.it

(2)Mac Donald Dettwiller, 13800 Commerce Parkway,

Richmond BC, Vancouver V6V 2J3, CANADA

Email: pm@mda.ca

ABSTRACT

The European Space Agency has selected an Advanced Synthetic Aperture Radar operating at C-band (5.331 GHz) as a payload for the forthcoming ENVISAT mission. The ASAR instrument modes of operation ensure continuity of ERS SAR and feature enhanced capability in terms of coverage, range of incidence angles and polarisations. This paper presents the key features of the ASAR system. The ASAR Ground Processor is presented highlighting the concepts for its development and the performance achieved versus the ESA specifications. The selection of processing algorithm for each product is discussed based on the image quality requirements and the throughput. The processor validation using ERS data is introduced, together with examples of simulated ASAR products and their quality measurements. Finally the validation plan of the newly developed ASAR products is detailed and pre-launch validation results presented.

ASAR SYSTEM DESIGN OVERVIEW

The ASAR system design has been inherited from ERS, this is why the ENVISAT platform is yaw steered and, thanks to the provision of star sensor, the attitude control of the platform is expected to be better than ERS. ASAR will benefit from the availability of an on-board Solid State Recorder allowing up to 10 minutes recording of High Rate modes (100Mbit/s) at any point around the orbit. The DORIS orbit determination system will allow for accurate geolocation of all ASAR products produced in near real time and off-line.

The ASAR instrument design uses an active array antenna with 320 Transmit Receive modules (T/R) to produce a versatile position of the imaged swath by beam steering in elevation. Additionally, a large swath coverage (400km) is achieved using the ScanSAR technique at medium resolution (150m) or low resolution (1 km). All operating modes feature the possibility to select H or V polarisation on transmit and receive (HH or VV) and the alternating polarisation sub-modes allow to use a combination of both (HH and VV, or HH and HV, or VV and VH). Moreover, an improved wave measurement mode is offered to the users with 100 km spacing between vignettes and the capability to select the position alternating between 2 positions in any swaths (see Fig. 1).

Additionally, ASAR is equipped with a programmable digital waveform generator, which allows to optimize the radiometric quality of the products versus signal to noise ratio. As a further improvement compared to ERS, an 8 bit ADC associated with a Flexible Block Adaptive Quantization scheme to be used in 8:4, 8:3 and 8:2 compression ratios will allow for a larger dynamic range of the input signals in order to limit saturation problems. ASAR features a temperature compensation scheme to compensate for drifts observed at T/R level around the orbit. Furthermore the instrument frequency stability and datation have been improved to support interferometric applications. A detailed description of the ASAR instrument is provided in [1] [7].

The ASAR has five modes of mutually exclusive operation, which can be classified in two categories:

Fig. 1. ASAR Modes of operation

?Low data rate modes for Global Monitoring and Wave modes with an operational capability of up to 100% of the orbit.

Both modes are systematically recorded on-board and the on-board tape recorder is dumped every orbit when visible by an ESA station (KIRUNA station in Sweden) or via the Artemis Data Relay Satellite to the ESA ESRIN station in Italy.?High data rate (100Mb/s) for Image, Alternating Polarization and Wide Swath modes with operation time up to 30 minutes per orbit (including 10’ in eclipse).

The High Rate mode data are recovered according to one of the following schemes: real time transmission via an X-band link to ESA or alternative station, real time transmission via the Ka-band link using Artemis Data Relay Satellite to the ESA ESRIN station in Italy, or alternatively recorded on board on the Solid State Recorder and dumped via X- or Ka-band link when in visibility of an ESA station.

ASAR DATA PROCESSING

The development of the ASAR processor and products is based on three concepts mainly derived from the ERS experience:

?The need for users to have identical products independent of which production facility generate them.

?The preservation of the specifications of the existing ERS SAR high-resolution products (SLC, PRI, GEC) and the introduction of new products (medium resolution and browses).

?The capability to generate a large amount of products in near real time.

Following the above concepts, ESA has developed with Alcatel Space Industries and MDA an ASAR Generic Processor (PF-ASAR) able to process data from all the ASAR modes. PF-ASAR will be installed in the ESA Payload Data Handling Stations (Kiruna and ESRIN-Frascati), in the ENVISAT Processing and Archiving Centers (PACs) and in the national stations offering ESA ASAR services. The use of PF-ASAR will insure product compatibility between the different processing centers (same format and processing algorithm) and will simplify product validation.

One of the key features of PF-ASAR is the capability to process acquired data to generate medium resolution (150 m resolution, example in Fig. 2) or low resolution products (1 km resolution) and their corresponding browse images in stripline without geometric or radiometric discontinuity. The stripline image products contain data from an entire acquisition segment up to 10 minutes for Image, Alternating Polarisation or Wide Swath modes and up to a complete orbit

for the Global Monitoring mode.

PF-ASAR will be used to insure the systematic processing in near real time of all received high rate data generating medium resolution and browse products. All Wave mode or Global Monitoring mode data will be also systematically processed in near real time. Furthermore PF-ASAR will allow to process High Resolution Products from Image or Alternating Polarisation acquisitions (Precision Images, Single Look Complex or Ellipsoid Geocoded products) in near real time or off line depending on user requests. Fig. 2 shows an example of Precision Image (IMP) and Ellipsoid Geocoded Image (IMG).PF-ASAR is targeted to run on an IBM Power 3 multi-processor computer system and is designed to insure a throughput linearly proportional to the number of processing nodes available in the processing station.

Fig. 2. Example of IMM Product (left) located on the Walgreen Coast in Antarctica, IMP Product of Bathurst Island in

Canada (top right) and IMG Product of the Bay of Naples in Italy (bottom right).

Processing Algorithms

The Range Doppler algorithm is used for the image mode precision (PRI), geocoded (GEC) and complex (SLC) products. Range Doppler is used in order to meet the stringent image quality requirements while still maintaining good throughput.

A modified version of the Range Doppler algorithm [2] was developed for the Alternating Polarisarion (AP) SLC product. The algorithm was chosen such that the output product would be useful for SAR Interferometry. The algorithm is a modified phase preserving Range Doppler approach. Although the aperture covers a little more than 5 bursts, (i.e. between 2 and 3 bursts are available per target for each polarisation) the processed Doppler bandwidth is made equal to four burst bandwidths (multi-burst processing) so as to keep as much of the signal information as possible while maintaining radiometric quality. The resulting product has a modulated impulse response function (IRF) in azimuth due to a segmented azimuth spectra for each target but contains all the necessary information required for SAR Interferometry (InSAR). This modulation can be removed during InSAR post-processing using the short FFT approach [3] to select the common spectra between the two image pairs.

The Range Doppler is also used to produce the SLC imagettes for wave mode. These imagettes are used as input to the wave mode cross spectra algorithm. The cross spectra algorithm was developed by NORUT Information Technology [4]. This algorithm is an improvement over the one used for ERS since it allows the wave direction ambiguity to be resolved and compensates for the signal to disturbance ratio which is in the order of 20 dB higher. The latter also makes the product well suited for obtaining a unique solution for the underlying swell spectrum. In most of the cases, wave systems are detected and a unique solution of the wave inversion problem is achievable without using external information. PF-ASAR is currently being updated to include this new inversion algorithm developed by NORUT- IFREMER [5] and generate systematically the ocean wave spectrum Level 2 product.

The SPECAN algorithm is well suited to burst-mode data and as a result is used for most AP mode products and for WS and GM mode products. The algorithm is also used to process medium resolution products, which are produced systematically and therefore benefit from the efficiency of the SPECAN algorithm.

PF-ASAR includes a Doppler Centroid Estimator with specified accuracy of 50Hz for image and wave modes (like ERS) and 25Hz in ScanSAR modes in order to limit radiometric errors in azimuth [6].

Internal Calibration Processing on the ASAR Ground Processor

During preprocessing PF-ASAR performs, for all imaging modes, FBAQ signal decoding (from 2, 3 or 4 bits back to 8 bits) and corrects for possible downlink bit errors on the source packet annotations. Once decompressed, I/Q science data are subject to an I/Q correction (bias, differential gains, non-orthogonality). As with ERS, any necessary corrections for non-linearity may be applied in the ground processor using pre-launch instrument ADC characterization. Prior to processing, PF-ASAR also performs noise power estimation from the instrument noise measurements.

The ASAR instrument’s internal path transfer function is derived by the ASAR internal calibration scheme [7]. This is achieved by dedicated calibration signal paths to each T/R module and special calibration pulses within the instrument to make the required calibration measurements and by using these measurements to perform corrections within PF-ASAR. During normal operation in any of the ASAR measurement modes, a sequence of calibration pulses is interleaved with the normal echo pulses. These pulses characterize the active array, both on transmit and receive, on a row by row basis. As a consequence the instrument transfer function cannot be simply calculated from a few pulses, as was the case with the AMI SAR. Instead the ground processor must utilize the calibration pulses from a complete cycle through the 32 antenna rows in order to estimate the transfer function

For each of the 32 rows, the antenna and the central electronics are characterized using four types of pulses: The transmit chain characteristics of the instrument are measured by pulse P1. This pulse needs to be compensated by the pulse P1a which contains the residual contribution of the unwanted rows during P1 transmission due to imperfect cancellation errors in the module settings. The receive path of the instrument is characterized by another chirped pulse P2. It is also necessary

to characterize the central electronics independently by the use of the internal pulse P3 as the central electronics transmit and receive paths are included in both P1/P1a and P2 measurements.

In the ground processor the amplitude and phase of calibration pulses (P1, P1a, P2, P3) for each row are used to calculate the elevation gain function and the transmitted pulse replica. The elevation gain is a factor proportional to the two-way instrument gain-power product at a reference angle and it is used to measure any deviation of the instrument reference gain from its on-ground characterized value. Transmit pulse P1a is subtracted vectorially from P1 and the resulting amplitude normalized to the nominal transmit pulse value (a tx,φtx). On the other hand the amplitude and phase measurements of the receive pulse P2 are normalized to the amplitude and phase of the associated P3 measurements to remove the effects of the central electronics (a rx,φrx). All phase measurements are performed on the previously compressed pulse peaks. The overall gain is obtained combining the amplitude and phase from all 32 rows for transmit and receive with the on-ground characterized row pattern embedded in the array and a complex factor characterizing the path through the calibration loop and from the calibration coupler to the antenna face, derived from the external characterization mode of operation.

The replica of the transmitted pulse is calculated from the P1, P1a, P2 and P3 measurements from a complete row cycle (except in wave mode), the on-ground characterized row antenna patterns and the data obtained from the external characterization mode of operation over the ASAR transponders. The reconstructed replica tracks variations in all the transmit and receive circuits and is used to determine the range reference function for range compression processing. The functional block diagram of the ASAR internal calibration pulse processing on the ground processor is outlined in Fig. 3.

Antenna embedded Row Pattern

Fig. 3. PF-ASAR Internal Calibration Block Diagram

ASAR Products quality

The image quality specifications regarding the degradation introduced by the processor are summarized in Table 1. PF-ASAR meets these specifications for all products and modes with the exception of those specifications that cannot be met in GM mode due to the low time-bandwidth-product of the signal and those due to the IRF modulation in azimuth previously mentioned in the AP complex product. The modulation causes a degradation in the azimuth PSLR and ISLR as can be seen in the quality measurements performed on the simulated ASAR products from ERS raw data (table 4). However, these constrains can be overcome if measurements are performed on the IRF from individual bursts. The equivalent number of looks which has been measured during acceptance testing of the ASAR Generic Processor using pseudo-random noise is presented in Table 2.

PROCESSOR VERIFICATION ACTIVITIES

A number of activities have been undertaken in preparation for the launch of ENVISAT to verify the ASAR Generic Processor and characterize the ASAR products, developing in parallel the necessary quality analysis tools to evaluate the products and asses instrument performances.

Processing ERS data with the PF-ASAR processor

In order to verify the quality of future ASAR products with large datasets, ESA has developed a stand-alone ERS interface to the PF-ASAR system called PF-ERS. It allows to generate ASAR-like products from ERS raw data using the ASAR generic processor. Because its flexibility (e.g. direct interface with ERS raw data), the PF-ERS sub-system is an excellent tool to provide simulated ASAR products in Image Mode, Wave Mode and in Alternating Polarisation Mode (one polarisation only) where continuous ERS data is treated as a burst mode acquisition.

ERS wave mode acquisitions (up to a complete orbit) are used to verify the new ASAR wave products. All output products are generated using the PF-ASAR algorithms and follow the same product quality specifications and product formats than the future ASAR products. Table 3 shows in green the ASAR products verified with ERS data, whereas in blue appear the products tested with simulated point targets and pseudo-random noise data. Table 4 shows the quality measurements of ASAR products containing ERS data over the Flevoland transponders and rainforest scenes.

Table 1: Processor Image Quality Specification

Parameter Specification

Range and Azimuth IRF broadening

< 10% of theoretical value

PSLR degradation< 2 dB

ISLR degradation< 2dB Radiometric Error< 0.1 dB (single beam)

< 0.2 dB (ScanSAR) Absolute location accuracy< 2 pixel Geometric distortion< 0.5 pixel

Table 2: Equivalent Number of Looks

Product IM AP WS GM PRI 3.9 1.9

GEC 3.9 1.9

MR40501212 BRW807557 - 6218-21

Table 3. Validation of ASAR products with ERS data

Table 4 Quality measurements of ASAR Products generated from ERS data

PRODUCTS RANGE RESOLUTION

AZIMUTH RESOLUTION

ENL RANGE PSLR AZIMUTH PSLR ISLR LOCATION ERROR (rg/az)ASA_IMS_1P 9.7 m 5.3 m N/A -20.1 dB -27.4 dB -10.4 dB 8.5 / 33.1 m ASA_IMP_1P 15.07 m 21.6 m 3.9-20.0 dB -21.3 dB -9.49 dB 10.0 / 13.5 m ASA_IMG_1P 23.6 m 21.2 m 3.9-19.5 dB -21.7 dB -9.57 dB 10.6 / 22.5 m ASA_IMM_1P 119.7 m 145.4 m 33.0-19.3 dB -18.1 dB -5.3 dB 35.7 / 90.0 m ASA_APS_1P 8.6 m 6.5 m*N/A -13.0 dB -3.5 dB* 1.21 dB* 2.9 / 15.8 m ASA_APP_1P 26.1 m 29.9 m 1.9-20.9 dB -19.4 dB -9.79 dB 10.6 / 13.5 m ASA_APG_1P 27.6 m 28.9 m 1.9-19.8 dB -20.1 dB -9.7 dB 10.7 / 21.8 m ASA_APM_1P

119.3 m

148.5 m

34.8

-19.0 dB

-18.37

3.2 dB

35.7 / 65.5 m

*Measured on the azimuth modulated https://www.sodocs.net/doc/3e17871766.html,missioning Phase planned activities

PF-ASAR verification and product calibration activities during Commissioning Phase will be jointly carried on by ESA, one Processing Archive Center (PAC) expert representative and some PIs selected from the Envisat Cal/Val proposals under ESA coordination.

During the commissioning phase, all ASAR products will be verified and calibrated. Priorities have been established in order to concentrate first on the ASAR products that are similar to the ERS SAR products (e.g. products over swath IS2 in VV polarization). The product verification priorities are given in the table 5. Pre-launch activities to fully validate ASAR products simulated with ERS data are already on going.

ASAR products verification requires the use of dedicated transponder and rainforest scenes, as well as arbitrary data sets.The detailed activities list to be carried out is summarized in Table 6.

Processing Level

Image Mode HH or VV

Alternating Polarization Mode HH/VV HH/HV

VV/VH Wide Swath Mode HH or HV Global

Monitoring Mode

HH or VV Wave Mode HH or VV

Level 0

INSTRUMENT SOURCE PACKETS

ASA_IM__0P ASA_APH_0P ASA_APV_0P ASA_APC_0P ASA_WS__0P ASA_GM__0P

ASA_WV__0P

Level 1b MEDIUM RESOLUTION (150 m)ASA_IMM_1P

ASA_APM_1P

ASA_WSM_1P

Level 1b LOW RESOLUTION (1Km)

ASA_GMM_1P

BROWSES ASA_IM__BP ASA_AP__BP ASA_WS__BP

ASA_GM__BP

Level 1b SINGLE

LOOK COMPLEX SLC

ASA_IMS_1P ASA_APS_1P Level 1b

PRECISION IMAGE

PRI

ASA_IMP_1P ASA_APP_1P Level 1b

GEOCODED IMAGE

GEC

ASA_IMG_1P

ASA_APG_1P

Level 1b IMAGETTE

AND CROSS SPECTRA

ASA_WVI_1P Level 1b CROSS

SPECTRA ASA_WVS_1P Level 2 WAVE SPECTRA

ASA_WVW_2P

Table 5. Product Priorities during Commissioning Phase

MODE POLARISATION SWATH

Commissioning workshop (6 months after the launch)

IM VV IS2

WV VV IS2

IM VV IS1, IS3-IS7

WS VV SS1, SS2-SS5

Validation workshop (9 months after the launch)

AP HH/HV, VV/VH IS2. IS4

AP HH/HV, VV/VH IS7, IS1

AP HH/HV, VV/VH IS3, IS6, IS5

IM HH IS1-IS7

AP HH/VV IS1-IS7

WS HH

GM VV

GM HH

ASAR WAVE MODE VALIDATION

Validation of the ASAR Level 1b (cross spectra) product and the Level 2 (wave spectra retrieval) algorithm is in progress using large datasets of ERS Wave and Image Mode Single Look Complex (SLC) data collocated with in-situ measurements and numerical wave model. This has allowed to optimize the ASAR processor settings for the Cross Spectra methodology.

In-situ Data

The in-situ data consist of directional wave spectra and wind vector, and are provided by either wave models, directional buoys, platform wave radar or weather ship observations. Fig. 4 shows the location of the buoys to be used for the validation activities.

Pre-launch

In the pre-launch phase, collocated historical data and new acquisitions of ERS Image and Wave Mode data will be gathered over the in-situ stations. The products generated from Image mode will be calibrated using ERS transponders. ERS data will be processed as to equivalent ASAR products using the ASAR processing algorithm. The products will be analyzed and the processor setting will be verified and optimized.

Among the objectives of the pre-launch validation are:

?Establish the necessary calibration strategy for the Level 2 product.

?Perform a limited geophysical validation of the ASAR Wave Mode products simulated using ERS data in Wave or Image mode.

?Establish a preliminary setting of the ASAR wave level 1b and Level 2 processing algorithm using ERS data.?Establish and test the infrastructure for worldwide collocation and acquisition of the needed in-situ wind and wave data.?Evaluate availability, format and quality of in-situ observations.

?Perform a real simulation of the activities to be fully verified during the Commissioning phase.

For the Level 1b products, the ERS acquisitions processed to ASAR products will be co-located with in situ wave information which allows to simulate cross spectra. The comparison is then performed between simulated cross spectra and ERS SAR derived cross spectra.

Table 6. Activities required for the verification of the ASAR products

ACTIVITY PRODUCTS

Product Format Verification All products

Raw Data Analysis

I-Q statistics

Saturation analysis

Noise analysis

Calibration pulses analysis

Chirp Replica analysis

Timing Monitoring

Gain Droop Compensation Verification

All IM, AP, WSM, GMM, WV level 0 products

IRF analysis (based on ASAR transponders)

Image quality parameters check (resolution, peak intensity, PSLR, ISLR, point target radar cross section…)IMP, IMS, IMM, IMG, APS, APP, APG, APM, WSM,

GMM, WVI imagette

Ambiguity analysis (based on ASAR transponders)

Ambiguity location offset

Ambiguity radar cross section

Point target ambiguity ratio

IMP, IMS, IMM, APS, APP, APM, WSM, GMM Geometric analysis (based on ASAR transponders)

Localization accuracy SWST bias determination Swath width and position IMP, IMS, IMM, IMG, APS, APP, APM, WSM, GMM,

WVI imagette

Radiometric analysis (based on ASAR transponders)

Radiometric resolution

ENL estimation

Radiometric stability and accuracy NEσ0 calculation IMP, IMS, IMM, IMG, APS, APP, APG, APM, WSM,

GMM, WVI imagette

External calibration (based on ASAR transponders) Calibration constant derivation IMP, IMS, IMG, IMM, APS, APP, APG, APM, WSM,

GMM, WVI imagette

In-Flight antenna pattern monitoring Antenna pattern derivation from rainforest Calibration constant verification with rainforest σ0

IMP, IMM, APP (2 patterns per product), APM (2 patterns per product), WSM (SS1). Products with no

antenna correction

Overall instrument gain determination

Calibration pulses analysis

External characterization pulses analysis

All IM, AP, WSM, GMM, WV level 0 products

Stripline analysis

Doppler variation within slices

Doppler continuity along strips

Doppler evolution along the orbit

Radiometric continuity between slices

IMM, APM, WSM, GMM, WVI

ScanSAR specific analysis

Radiometric normalization inter-subswath

Doppler monitoring across subswaths

Beam merging

WSM, GMM

Scalloping analysis APP, APS, APM, WSM, GMM InSAR performance analysis

35-days repeat pass interferogram generation

IMS, APS

Wave specific analysis

Spectrum peak

Center of gravity

Direction and wavelength of spectrum maximum

Doppler ambiguity monitoring

WVI, WVS

Polarimetric specific analysis

Cross polarized noise level

Intensity imbalance

APP, APS, APM

Fig. 4 Location of buoys to be used for ASAR wave model cal/val

ERS data will also be processed to Level 2 products wave spectra using the newly developed ASAR inversion algorithm [5]. The wave spectra will be validated against wave model spectra. Wave height, direction and period are of particular importance within the SAR imaging domain. The wind speed and direction will be validated directly against wind information from in situ measurements.

Post-Launch

The post-launch activities will focus on the ENVISAT ASAR products (ASA_WVW_2P, ASA_WVI_1P and ASA_WVS_1P) and will consist in:

?Setting the ASAR instrument configuration in wave mode to ERS-2 like mode Swath IS2 and VV polarisation.?Verifying the processing quality of SLC imagette (quality parameters, geolocation, phase preserving test)?Calibrating the ASAR wave mode product using the ASAR mobile transponder transported to a dedicated wave-mode calibration site.

?Acquiring systematically over the wave mode validation test sites during the Commissioning phase, ENVISAT ASAR Wave Mode data

?Establish the required calibration parameters for the Level 2 algorithm using the procedure established as part of the pre-launch cal/val activity.

?Validating by comparison parameters retrieved from Level 1 and Level 2 products (SAR cross spectra, SAR ocean wave spectra, wind speed, wind direction) with the in-situ measured wind/wave parameters and wave models as described in the pre-launch validation section.

?Comparing ASAR Wave mode Level 1 Products with equivalent products generated from ERS-2 wave mode data acquisition (half an hour time difference).

Preliminary Results

In order to optimize the algorithm, the processing settings and to be able to predict the performances of the new products and algorithms, a preliminary validation has been done by NORUT-IFREMER [8]. The data used in the analysis are ERS-1 and ERS-2 Image Mode and ERS-2 Wave Mode SLC data, collocated with wave model, buoy or in-situ observations. For

Level 1b products, results (see Fig. 5) are used to establish RMS error and overall performance statistics as summarized below:

? Swell system detected, and propagation ambiguity resolved in 80% of the cases

? Peak wavelength RMS ? 50m, peak direction RMS ? 40deg.

? Mean peak wavelength = 265m. Mean azimuth cut-off wavelength = 256m.

For the Level 2 development, the preliminary results show that a consistent inversion scheme can be formulated without using any a-priori information or first guess spectrum [9]. For wave retrieval, very good results are obtained starting from ERS Image mode SLC data. For wind field retrieval, the algorithm is based on the time decorrelation and phase spectra computed from inter-look processing of SLC data. Example of ERS-2 cross spectra and corresponding phase spectra is shown in Fig.6. Note that the extracted wind direction is not aligned with the direction of the main swell wave system (white spot) detected by the SAR. Statistics from comparison with in-situ measurements from the Gullfaks C platform in the Norwegian Sea is shown in Fig. 7 which shows a good correlation between SAR and in-situ measured wind direction. The RMS error is around 35 degrees.

For wind speed retrieval, the analysis of the observed spectral azimuth cut-off caused by the wind-generated wave random motions has shown promising results. The azimuth cut-off estimator estimates the width of the ratio of the azimuth cross spectra profiles obtained at two different look separation times. Measurements carried out shows a good correlation between the time decorrelation measured using two look separation times and the in-situ wind speed. A new improved algorithm for the generation of the Level 2 product has been delivered early October 2000 for the implementation in the operational chain

CONCLUSIONS

The ASAR instrument has been completed, delivered and integrated on the ENVISAT Platform to be launched by mid 2001. It will offer several improvements compared to the ERS SAR instrument: new modes of operation, increased swath coverage, choice of polarization and incidence angle. The ASAR Generic processor, which has been successfully completed in November 1997, allows for near real time and off-line processing of all ASAR modes products and for stripline processing of medium resolution and low resolution products and their browse image. The ASAR processor will allow ESA to provide quality products to the appointed distributors and to the several hundreds selected principal investigators, contributing so to a wider range of scientific and commercial applications for SAR data. The processor has been pre-validated using ERS data to verify the product quality before launch. The validation plan for the Level 1b and Level 2 wave products has been presented with preliminary validation results using ERS data.

REFERENCES

[1] R. Torres, C. Buck, J. Guijarro, J-L. Suchail and A. Schoenberg, “The ENVISAT ASAR Instrument Verification and Characterization”, ESA CEOS SAR Workshop, October 1999 ESA-SP450.

[2] D. Stevens, F. Wong, P. Lim and Y-L. Desnos, “A Processing Algorithm for the ENVISAT Alternating Polarization Mode Single Look Complex Product”, IGARS’97, Singapore, August 1997.

[3] Frank Wong, David Stevens and Ian Cumming, “Phase Preserving Processing of ScanSAR Data with a Modified Range Doppler Algorithm”, IGARS’97, Singapore, August 1997.

[4] H. Johnsen. Norut IT, “The ENVISAT ASAR Wave Mode Cross Spectra Algorithm” CEOS SAR Workshop 3-6 February 1998, ESA-WPP 138.

[5] ENVISAT ASAR Level 2 Wave Mode Product Algorithm Specification – Software Requirements Document, Norut IT Doc. No. IT544/5-98 Ver. 2.0, July 2000.

[6] Y-L Desnos, H. Laur, P. Lim, P. Meisl, T. Gach, “The ENVISAT Advanced Synthetic Aperture Radar Processor and Data Products”, IGARSS’99, Hamburg, Germany.

[7] C. Buck, R. Torres, J-L. Suchail, M. Zink, “ASAR Instrument Calibration”, ERS- ENVISAT Symposium, Gothenburg, October 2000

[8] H. Johnsen, G. Envgen, K-A. Holda, B. Chapron, Y-L. Desnos, “Validation of Envisat ASAR Wave Mode Level 1b and Level 2 Products Using ERS SAR Data”, CEOS SAR Workshop, Toulouse, November 1999

[9] H. Johnsen, G. Engen, B. Chapron, “Envisat ASAR Wind&Wave Measurements from Level 1 product - Wind and Wave Field Retrieval Methodologies”, Norut IT, Doc. No. IT544/4-99, October9.

Fig. 5. (left). ASAR Wave mode cross spectra processed from ERS SLC together with cross spectra simulated using collocated wave mode spectra. Non-directional spectra are also shown.

Fig. 6. (bottom right). Example of ERS-2 cross-spectra and the corresponding phase spectra. The lines in the phase spectra indicate the SAR and the in-situ measured wind direction.

Fig. 7. (low right). Wind direction derived from ERS SAR SLC images versus in-situ measured wind direction.