Corsica: a Cal/Val experiment to link offshore and coastal altimetry

Size: px

Start display at page:

Download "Corsica: a Cal/Val experiment to link offshore and coastal altimetry"

Felicia Boyd
6 years ago
Views:

1 orsica: a al/val experiment to link offshore and coastal altimetry P. Bonnefond (1), P. xertier (1),. Laurain (1),. uillot (2),. uinle (2),. Picot (2), P. Féménias (3) (1) /eoazur, ophia-ntipolis, France (2), oulouse, France (3) /, Frascati, taly

established in 1998 (equipped with 4 pressure

) upports continuous monitoring of Jason-2

altimeter readings connected to tide gauges

intensive P buoy and catamaran surveys along

2 orsica alibration ite orsica bsolute ltimeters alibration enetosa calibration site established in 1998 (equipped with 4 pressure tide gauges.) upports continuous monitoring of Jason-2 (and formerly /P and Jason-1) pen-ocean altimeter readings connected to tide gauges via detailed local geoid model Derived from intensive P buoy and catamaran surveys along ground track. xtension to jaccio (25) and apraia (24) pen-ocean verification location for P zodiac deployments. jaccio configuration upports continuous monitoring of L/LtiKa (and formerly -2, nvisat) Fiducial point near jaccio equipped with P/FL/ D. jaccio radar tide gauge (H)ew one since 29/9/16 (moved on 212/4/3) ome tracks of ryoat-2 and HY2- cross the geoids allowing absolute calibration 8th oastal ltimetry Workshop, ctober 214 onstance, ermany D P 2

ime of losest pproach () 4Hz altimetric H low pass filtered at 1Hz orsica bsolute ltimeters alibration P H (2): direct method (1): indirect method P H H D 2 ethods to compute H bias: - ndirect: need

to avoid any geoid slope and potential ocean dynamics effects 2 independent instruments to compute H bias: - From tide gauge: - () H from altimetry needs to be corrected for geoid - From P

3 ime of losest pproach () 4Hz altimetric H low pass filtered at 1Hz orsica bsolute ltimeters alibration P H (2): direct method (1): indirect method P H H D 2 ethods to compute H bias: - ndirect: need to correct from geoid slope and potential ocean dynamics effects between in situ and altimetric measurements - Direct: in situ instrument needs to be as close as possible from altimetric measurement to avoid any geoid slope and potential ocean dynamics effects 2 independent instruments to compute H bias: - From tide gauge: - () H from altimetry needs to be corrected for geoid - From P measurement (P aboard a zodiac located under the track, LV): - (1) Using geoid correction to average all the altimetric H (noted P-mean) - (2) omputation at the Point of losest pproach = no need to correct from geoid (noted P-P) 8th oastal ltimetry Workshop, ctober 214 onstance, ermany & U 3

4 5 orsica bsolute ltimeters alibration Bias (mm) ? ltika bias with tide gauge ltika bias with P (mean) ltika bias with P (P) ycle veraged H bias with D- products (6 common cycles) ndirect Direct P ide gauge P (mean) P (P) -127 mm (s=27 mm) -53 mm (s=3 mm) -6 mm (s=21 mm) ~-7 mm differences: instrumental bias? geoid or oceanic signal from tide gauge to offshore? ll together? 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 4

5 L/ltiKa H bias as a function of across-track distance jaccio pass #13: D, cycle 1 to 17-5 ltika bias with tide gauge 15 orsica bsolute ltimeters alibration ycle 1 H bias (mm) mm last maneuver to reach the nominal nvisat ground track during cycle 7 (28/1/213) 5-15 nominal nvisat ground track (~middle of the geoid) cross track distance to tide gauge (km) everal maneuvers were needed to reach the nominal ground track, it can be divided into 3 parts: 1- cycle 1 to 4: ground track located in the western part => contamination from anguinaires islands 2- cycle 5 to 7: ground track located in the eastern part => contamination from apu di muro 3- from cycle 8: ground track located in the center part => no a priori contamination except very close to the coast in the northern part mpact on the averaged H bias: 48 mm (H bias cycles 1-7 compared to cycles 8-17) Better stability since cycle 8: 2 mm rms (31 mm rms on the whole set) nominal nvisat ground track (~middle of the geoid) cross track distance to tide gauge (km) 3.2 km 3.6 km 8.2 km apu di uro 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 5 K P

6 nviat & L/ltiKa ltimeter alibration nvisat (D-: cycle 1 to 93) / L/ltiKa (D-: cycle 1 to 17) anguinaires islands 2 bias with tide gauge (2 km smoothed) ltika bias with tide gauge (2km smoothed) LV apu di uro 4.5 km orsica bsolute ltimeters alibration Bias (mm) mm Distance from coast (km) o make the comparison easier the ltika H bias has been shifted to the 2 H bias by the difference of their H biases (523 mm). 4.4 km apu di uro veraged ground track since cycle 8 (~5m eastward from nvisat nominal track) L K his plot shows the average H bias in function of the distance to the coast: ven if the land contamination is much smaller than for the nvisat (2) altimeter, it is estimated to be at the level of 2 mm in vicinity of the anguinaires islands and apu di uro : the theoretical ltika footprint radius is 4 km, so ltika should not been theoretically impacted However, even by selected data from cycle 8, the structures identified in the above figure remain. P 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 6

7 4 3 5 ld ide auge: ean = 2.3 mm / td = 7.3 mm ew ide auge: ean = mm / td = 12.3 mm enetosa (4 & 5): ean = +.4 mm / td = 8.8 mm orsica bsolute ltimeters alibration Differences P - (mm) Bias (mm) ltika bias with tide gauge ltika bias with P (mean) ltika bias with P (P) ycle 29/9/14: ide gauge replacement 212/4/2: ide gauge displacement and sensor replacement 212/7: Use of the P-zodiac 212/12/5: hange of P antenna for the reference (J, ) P / D ide gauge location, a clear instrumental bias has been identified from the 2 instruments ü -23 mm (after tide gauge replacement in september 29). ü -3 mm since the L/ltiKa launch (very stable, only 5 mm rms). his bias remains unsolved: ü J antenna change should not have impact (taken into consideration in the processing) ü omparisons with the same enetosa site do not exhibit any bias U 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 7

8 Bias (mm) D-: mean=-117mm, td=33mm D-: mean=-17mm, td=31mm D-: mean=-14mm, td=3mm since cycle 8 mean=-13mm, td=21mm mean=-86mm, td=21mm mean=-83mm, td=17mm /D-: in patch P2 since cycle 11 D-: whole set in patch P2 orrections used: - for ionosphere - Wet troposphere model - Product B orsica bsolute ltimeters alibration 5 D- - D- (mm) -5-1 orbit-range: mean=-13mm, td=22mm bias : mean=-11mm, td=27mm dry (mm) wet model (mm) ean = +.4 mm / td = 1.9 mm ean = -3.1 mm / td = 16.1 mm D he continuous time series of the tide gauge is very useful to study the stability of the H bias as well as to cross-compare the different products tandard deviation (31 mm) compararable to typical Jason-2 one (~35 mm) ean radial orbit differences between DD and (-13 mm): comparable to orbit errors analyzed over urope using short-arc orbit technique (-2 mm) Differences between D- and D- H bias are due to dry and wet tropo and linked to differences between predicted and computed WF model H bias difference of only 3 mm between D- and D- / D / D 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 8

9 bsolute H biases from tide gauge since cycle 8: D-: -13 ±7 mm (9 cycles) D-: -86 ±7 mm (9 cycles) D-: -83 ±6 mm (7 cycles) orsica bsolute ltimeters alibration omparison between tide gauges and P-zodiac (D-): ide gauge: -86 ±7 mm (indirect method) P (mean): -53 ±12 mm (semi-indirect method) P (P): -6 ±9 mm (direct method) 26 mm difference between tide gauge and P (P) methods/instruments ü 3 mm comes from instrumental differences tide gauge location): this remains unsolved ü ther effects: ocean dynamics? high resolution model is in development to estimate the impact but it should be small WH monitoring using P: ü ltimeter WH higher by ~7 cm adiometer monitoring using P: ü adiometer dryer by ~1mm (mainly land contamination) ain impact: ü o major impact on the H bias even during the leopatra storm (213/11/18) but radiometer is wetter by ~5 mm L U 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 9

ycle 5 orsica bsolute ltimeters alibration @ tide gauge @ calenv L

10 ycle 5 orsica bsolute ltimeters tide calenv L ~2cm 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 1

11 orsica bsolute ltimeters alibration P zodiac K L/ltiKa ground track shifted by ~2.5 km across-track toward east after the maneuvers performed end of July 213 F F => Use of real ground track prediction to plan the P-zodiac deployment 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 11

orsica bsolute ltimeters alibration Y L 1 2 2 igma blooms 15 igma (db) 1 5 1 2 3 4 5 6 7 8 9 1

12 orsica bsolute ltimeters alibration Y L igma blooms 15 igma (db) ycle 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 12 12

ainfall recorded @ jaccio airport orsica bsolute ltimeters alibration leopatra storm (213/11/18): - 2 mm/h @ overflight time (18h) - ome missing data but no clear impact on

13 ainfall jaccio airport orsica bsolute ltimeters alibration leopatra storm (213/11/18): - 2 overflight time (18h) - ome missing data but no clear impact on the H bias - ore impact on the radiometer - ~5 mm wetter than P and model - lope ~3 times stronger than model P 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 13

14 Wet tropospheric correction from ltika (mm) L/ltiKa leopatra storm Wet tropospheric correction from P (mm) Using P data from permanent receiver (J) and pressure from jaccio weather station, the wet tropospheric correction is computed and compared to radiometer (no P data for cycle 1): - ycle 8 clearly departs from the series: heavy rain during the leopatra storm - Without cycle 8, orrelation: 91% (slope =.85 / bias at origin = -4 mm) - Without cycle 8 radiometer exhibits a -1mm bias (dryer) compared to P; relatively strong standard deviation (~24 mm) compared to Jason-2 (14 mm) but the number of cycle is small Differences (mm) nviat & L/ltiKa: radiometer - WF jaccio: pass 13 W - WF (7 years) ltika - WF (15 cycles) Latitude orsica bsolute ltimeters alibration D V P 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 14

15 25 2 L/ltiKa orsica bsolute ltimeters alibration WH from atellite (cm) WH from P buoy (cm) P buoy measurements also provide the sea height variations due to waves. he standard deviation on the P buoy sea height residuals (σ shr ) is the root square sum of σ gps and σ wave where σ wave is the standard deviation of P buoy measurements due to waves and σ gps the internal error of P buoy measurements; the P buoy internal error was estimated by processing kinematically a quasi-static session and is at the level of 2.6 cm (σ gps ). σ shr 2 = σ gps 2 + σ wave2 ; so, σ wave = (σ shr 2 - σ gps2 ). WH (or H1/3) is then deduced from the formula below (tewart, 28): WH buoy = 4.σ wave WH monitoring using P (±5min at overflight time): Differences (P-altimeter): -7 cm WH bias (12 cm standard deviation) orrelation: 99% (slope = 1.11 / bias at origin = cm) With such a correlation, the P-buoy that was primarily dedicated to measure the absolute sea surface height bias appears to be also an interesting solution to validate WH from altimeters with enough precision. P B U Y F W H 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 15

orsica bsolute ltimeters alibration ü ~18% of valid (and good) data @ 2-3 km ü ~46% @ 3-6 km ü ~83% @ >6 km Percentage of 4 Hz valid data 1 9 8 7 6 5 4 3 2 1 Data

16 orsica bsolute ltimeters alibration ü ~18% of valid (and good) 2-3 km ü 3-6 km ü >6 km Percentage of 4 Hz valid data Data distribution in function of coastal distance jaccio pass # 13: D, cycle 1 to Distance from coast (km) 8th oastal ltimetry Workshop, ctober 214 onstance, ermany 16 D L H

Altimeter Range Corrections

Altimeter Range Corrections Schematic Summary Corrections Altimeters Range Corrections Altimeter range corrections can be grouped as follows: Atmospheric Refraction Corrections Sea-State Bias Corrections