RA2 conclusions 2

Slide1: 

RA-2/MWR Cross-Calibration and Products Validation Orbit Validation Objectives, Approach, Results and Recommendations Summary J. Benveniste, M.P. Milagro, B. Greco, P. Féménias et al. P. Vincent, B.Duesmann (OVT)

Presentation Overview: 

Presentation Overview Cross-Calibration and Validation Objectives Approach Activity so far and results Algorithm Verification Product Validation Cross-calibration on ERS-2 RA/MWR MWR calibration and Validation results Orbit Validation Team results Conclusion Recommendations RA-2 MWR DORIS

Objectives of the ENVISAT Validation Workshop: 

Objectives of the ENVISAT Validation Workshop Recall of the In-flight calibration of the instruments Recall of the Verification of level 1b data processing (engineering parameters) -- Review the evolution Review the verification of the Level 2 product algorithms (geophysical parameters) Review the consistency of the level 2 data products Provide an error estimate of the Level 2 products Formulate conclusions and recommendations to be delivered at the final plenary session.

In One Slide... : 

In One Slide... From the RA-2/MWR Cross-Calibration and Validation Team

Slide5: 

Sea Level Anomaly in Oct 2002

Objectives of the Cross-Calibration and Validation: 

Objectives of the Cross-Calibration and Validation Cross-calibrated data products:- Determine relative biases between ENVISAT RA-2 and ERS-2 RA, at the mm level for the range, or/and other altimeters (T-P, GFO, Jason). Rationale: to ensure ERS-2/ENVISAT data products consistency and unify the two mission-long time series with high precision. Validated data products:- Geophysical validation of the processing models and the data products. Rationale: to authorize distribution of validated data products to all users.

Cross-Calibration and Validation Approach (1): 

Cross-Calibration and Validation Approach (1) (i) Cross-calibration of the RA-2 main geophysical parameters against ERS-2 (and/or TP, GFO, Jason):- Global comparison at cross-overs and along the collinear tracks of millions of globally distributed data (significant error reduction) Global statistical comparison against ERS-2, Models, In-situ data (ii) Preparation of long-term drifts detection:- Global tide gauge network for RA-2 range monitoring

Cross-Calibration and Validation Approach (2): 

Cross-Calibration and Validation Approach (2) (iii) Product validation Verify with real data the consistency of the product package (document, media, format and actual dataset) Quantify the inherent validity and accuracy of the range and altitude wave height, wind speed measurements and the geophysical corrections. Estimation of pseudo-time tag errors by minimization of crossover height differences over ocean. The data products are used by the whole CCVTeam for the cross-calibration activities: This will enhance the quality and thoroughness of the validation.

Cross-calibration objectives for RA-2 & MWR data: 

Cross-calibration objectives for RA-2 & MWR data

RA-2/MWR CCVT Activity So Far : 

RA-2/MWR CCVT Activity So Far Quick look data inspection at Switch-on (night of 12 March) In house verification of reference processors versus NRT processor (IPF) Level 2 Algorithm Verification w/ ESLs Verification and tuning for IPF and F-PAC Support to Instrument and Level 1B verification Pseudo-time tag bias, range bias and sigma0 bias investigations MWR Instrument and Level 1B verification Met offices validation of NRT data circulation. Preliminary cross-calibration with one cycle of data

Algorithm Verification: 

Algorithm Verification Level 1b/2 data from the IPF verified through comparison with real data Test Data Sets from the ref. processors: 3 IPF algorithm updates since launch Only very minor discrepancies present in current IPF Level 2 algorithms verification and products validation activities performed by the ESLs: Ocean retracker and geophysical corrections validated, ice2 retracker verification preliminary, more ongoing OCOG/sea-ice retrackers and LIP algorithms verified

IPF status: 

IPF status IPF version (4.53) operative at Kiruna since 27 Nov. previous v. 4.52 operative since mid August Remaining discrepancies: Very minor differences in few output values from reference ocean retracking default ocean ranges when ref. value was non-default negative SWH values over non-ocean surfaces No impact on product quality as there are no ocean waves on land!

CMA status : 

CMA status Version of CMA used for the cycle of data provided to the CCVT members (4 Oct- 16 Nov.): V5.4 New version (V5.5) to be operative in CMA by mid December. This new version solves some known (minor) bugs in v.5.4

Range Noise Estimation: 

Range Noise Estimation Range noise estimate is: 8 cm rms at 18 Hz 1.8 cm rms at 1 Hz for SWH < 2 m

Range Standard Deviation: 

Range Standard Deviation Ku S Ku peak at 8 cm (1.8 cm) S peak at 27 cm (5.2 cm) This confirms that RA-2 is indeed a low noise instrument

Range Standard Deviation: 

Range Standard Deviation Estimated on the ocean at cross-over of tracks CLS: Pseudo Time Tag Bias : -0.3 ms DEOS:

Pseudo-Time Tag Bias: 

Pseudo-Time Tag Bias Estimated on the ocean at cross-over of tracks DEOS: -0.5 ms CLS: -0.3 ms MSSL: -0.4 ms NCL U -0.4 ms +/- 0.5 ms

Cross-Over Analysis: 

Cross-Over Analysis Estimated sea surface height RMS on the ocean at cross-over of tracks CLS: 8.98 cm with orbit error correction: 6.25 cm NCL U: 9.5 cm MSSL : 8.5 cm --> Confirms low noise from Envisat Altimetry (on top of ocean natural variability)

Ionospheric Correction: 

Ionospheric Correction Impact of the ionospheric correction on the estimated sea surface height RMS on the ocean at cross-over of tracks Confirms RA-2 doing better!

Estimated Relative Range Bias: 

Estimated Relative Range Bias DEOS: -40.4 cm / ERS-2 CLS: -40.7 cm / CLS01 MSS OSU: -43.1 cm / OSU 95 MSS OSU: -39.95 cm / Great-Lakes Tide gauges

Significant Wave Height: 

Significant Wave Height The special design of RA-2 with two additional digital filters located on the waveform leading edge to improve its slope estimation is giving excellent results in picking up the smaller waves. The SWH algorithm has been further improved to estimate smaller waves

Significant Wave Height: 

Significant Wave Height

ENVISAT Performance Coming Off Land: 

ENVISAT Performance Coming Off Land Envisat RA2 gives good values for SWH and U10 < 2s when leaving land close to 90°.

SWH Regression Against Buoys: 

SWH Regression Against Buoys 20 NDBC buoys (< 30mins , < 50 km) N = 30 ODR Correction: “True” Hs = 1.0579 EnvHs - 0.1438 95% conf. Intervals: Gradient 0.9701 - 1.1456 Intercept -0.3513 - 0.0636 SWH (Ku) not significantly different from buoy “truth”. 1 months data: error margins ± 9%, ± 20 cm

SWH in Near Real Time: 

SWH in Near Real Time Ready to be used by met offices !

Rain Is Under Control: 

Rain Is Under Control

Rain Flag Performance: 

Rain Flag Performance

Passive Sigma0 Calibration: 

Passive Sigma0 Calibration Calibration experiment methodology validated on small data sample Improved method to complement the sigma0 absolute calibration transponder experiment by providing the inter-frequency bias Need more data to reach final results

Altitude in FDGDR/FDMAR: 

Altitude in FDGDR/FDMAR * 30 meter Orbit error * 4 cy/rev * Jumps Use DORIS Navigator instead (30 cm)!

NRT Sea Surface Height dominated by orbit error: 

NRT Sea Surface Height dominated by orbit error Inclusion of the full rate DORIS Navigator on FDMAR product is fundamental to real time oceanography

Conclusion (1/2): 

Conclusion (1/2) Ref. processors fully verified and optimised Only “ice2” verification is preliminary, more ongoing IPF verified, since last update (27 Nov) F-PAC data processing verification perfect match for the outputs from ice1/ice2/sea-ice retrackers between F-PAC and IPF Very minor differences existing wrt NRT data (ocean retracker over non-ocean surfaces, flags, ..) No show-stopper for the Cross-Cal activities

Conclusion (2/2): 

Conclusion (2/2) The preliminary1 performance estimation of RA-2/MWR yields very good results Low Range noise estimated over low waveheight area of the ocean: at 1 Hz: 1.8 cm for Ku-band, 5.2 cm at 1Hz for S-band SWH compares well with ERS-2 and Models: 30 cm rms RA-2 measures low waveheights much better than ERS (a feature!) Algorithm has been further improved Wind speed rms only 1.4 m/s Sigma0 cross-cal w/ERS-2 on more data Range preliminary cross-calibration on ERS-2 : bias = - 40 cm 1The RA-2/MWR Cross-Calibration and Validation Team has worked with four months of FDMAR and (only) one cycle of IGDRs, GDRs, SGDRs data (need three).

Next Steps: 

Next Steps Future upgrade of L2 Auxiliary files Sigma0 cross-calibrated on ERS-2 The next step will be to introduce the Absolute calibrated sigma0 and upgrade geophysical models using sigma0 (wind, sea state bias, MWR tropospheric correction) Future upgrade of CMA to account for latest algorithms modifications on L2 ref. processors and IPF| Finalise Validation activity on 3 cycles of data Perform Cross-calibration on 3 cycles of data

MWR Calibration and Validation Conclusions: 

MWR Calibration and Validation Conclusions The MWR Brightness Temperatures are well calibrated for the two channel, as expected The Level 2 Wet Tropospheric correction is in very good agreement with the ERS-2 MWR one. Dh bias < 5 mm. Envisat MWR already appears to be in better agreement with the radiosoundings, when compared to ERS-2 MWR MWR 36 Ghz channel may require anyway some stabilization due to aging and thermal cycling of the components Current drift from the BOM of 0.25 counts/K No specific on-board problem has been identified.

MWR Recommendations (1/2): 

MWR Recommendations (1/2) Implementation of the new MWR Side Lobes corrections within Level 1B (Improvement > 1cm in specific zones) Algorithm validation on-going, to be completed by mid-February 2003. Specifications ready by the end the 2002 Priority for the MWR level 2 algorithm to have a RA-2 Sigma0 in line with ERS-2, in a first step. ‘MWR’ Sigma0 meant to be higher by 1dB w.r.t. to RA-2. This will be in injected in the recommended Level 2 Neural Network algorithm upgrade. Specifications ready by January 2003.

MWR Recommendations (2/2): 

MWR Recommendations (2/2) Continuous monitoring of the 36 GHz channel gain drift and LO CH2 behavior is recommended, to assess possible stabilization. Execute some further testing, the redundant LO at 36.5 GHz could be used for an observation period. Nevertheless this is not recommended at this stage. Linear correction factors on the MWR Brightness Temperatures and RA-2 Sigma0 to be implemented in the Level 1B processing to correct for potential drifts

Orbit Validation Team Results: 

Orbit Validation Team Results

Orbit Validation Team Conclusion: 

Orbit Validation Team Conclusion Results based on tracking residuals and on orbit comparisons are very encouraging: the 3 cm goal of accuracy for the radial component is realistic Further evaluation through external tests is required: use of altimeter data OVT plan: use the first 3 consecutive 35-day cycles made available by the F-PAC Important recommendation: The ultimate goal being to get optimal error estimation both in terms of RMS and importantly in terms of geographical distribution, Then, the global DORIS tracking data acquired onboard are required for the on ground orbit processing

RA-2/MWR CCVTeam Recommendation 1/4: 

RA-2/MWR CCVTeam Recommendation 1/4 Supply 100% DORIS Doppler to SSALTO in less than 2 days as a highest priority. A gap in DORIS data deteriorates the Altimetry and this propagates for several orbits into the Altimetry product

RA-2/MWR CCVTeam Recommendation 2/4: 

RA-2/MWR CCVTeam Recommendation 2/4 Include the DORIS Navigator data in the fast delivery Products (FDGDR, FDMAR). Reported efforts done in this direction since the Cal Review but need to sustain the effort for this implementation

RA-2/MWR CCVTeam Recommendation 3/4: 

RA-2/MWR CCVTeam Recommendation 3/4 Publish individual echoes processing algorithm.

RA-2/MWR CCVTeam Recommendation 4/4: 

RA-2/MWR CCVTeam Recommendation 4/4 ERS-2 / ENVISAT tandem mission configuration after the three cycles data up-take, the CCVTeam can release the constraint on the 30 minutes separation Making same measurement 30 mn later is not useful any longer after the Cross-Calibration data is acquired. Recommendation: Move one satellite for a more optimal space-time sampling of the Earth Options discussed in terms of mission objectives Oceanography Marine Geodesy

Slide45: 

http://envisat.esa.int