ostst, hobart, march 2007 - 1- future altimeter systems : is mesoscale observability guaranteed for...

OSTST, Hobart, March 2007- 1-

Future altimeter systems : is mesoscale observability guaranteed for operational oceanography?

Future altimeter systems : is mesoscale observability guaranteed

for operational oceanography?

P.Escudier, G.Dibarboure, J.Dorandeu, CLS, Space Oceanography Division



Overview

• Until recently, the exceptional longevity of T/P and GFO added to the performance of Jason and ENVISAT were hiding the risk of future data gaps.

– The current missions are still performing well but exceeding designed lifetime

• With the current launch schedule (Jason2, CryoSat, AltiKa, Sentinel…) can we guarantee a sufficient mesoscale observability for new and developing applications and for operational oceanography ?

• Use a probabilistic model to better assess the odds of having an observing system (multiple altimeters) accurate enough :

1

GEOSATT/PERS

T/PERSGFO

JASONT/Ptandem

ENVISATGFO

JASONENVISAT

GFO

Oct1992

T/PT/PERS

2 3 4 3Jan

2000Oct

2002Sep2005

1 2Dec1993

Mar1995

Nov1986

Dec1988

?

JASON-2Cryosat2?Sentinel?GMES?AltiKa?

20082009

1234

Insufficient for offline meso-scale applications

Minimum for off-line applications

Robust for offline applications (barely for real time)

Robust sampling for real time applications

Color Legend

• Minimum needed to monitor mesoscale structures :

– 2 satellites in delayed time (offline),

– 3 satellites or more in near real time (Pascual et al)

Question asked in Venice meeting (2006)



EKE (cm²/s²)

J1 J1+E2

J1+E2+TP J1+E2+TP+G2

EKE on [october 2002 – june 2003]

+ 40% J1

+ 15% J1+E2

50 350

EKE in the Mediterranean sea depending on the satellite configuration



Approach and model used

• Lifespan probability model of a satellite driven by platform and payload design. – Probability function approximated with a simple model fitted on lifespan

computations from AAS

00.2

0.4

0.6

0.8

1

0 1 2 3 4 5 6 7 8 9 10 11 12

Launch (100% success rate assumed)

Lifespan (75% survival chance)

Mission should not outlive 2.2 * nominal lifespan

• Probability function for Each mission based– launch date, – expected lifespan as announced– specific cases taken into account

• Cryosat considered as 50% operational on ocean : payload limitation, non-repetitive ground track

• Lifetime limitation – ie propellant limit for ENVISAT

• Missions grouped by ground track :– redundant measurements are not

stacked, but likelihood to have the ground track covered increases

• Individual probability function combined to compute combinatorial probabilities Probability function used for an altimeter with a 5 year

lifespan



Model validation• The probability model is validated using historical datasets.

– actual coverage on ocean taken into account : platform/payload events, …

• Agreement between model prediction and observed reality fairly good

• Most significant differences :

– GFO post-launch anomalies

– unexpected longevity of T/P

• In both cases, the event is against the base probability assumptions on the satellite : 100% success rate at launch, and 2.2 * lifespan max

01

23

45

1992 1994 1996 1998 2000 2002 2004 2006

Unexpected post-launch anomalies on GFO

T/P and ERS still in operations (50% chance)

Exceptional longevity of T/P

(and GFO)

01

23

45

1992 1994 1996 1998 2000 2002 2004 2006

Unexpected post-launch anomalies on GFO

T/P and ERS still in operations (50% chance)

Exceptional longevity of T/P

(and GFO)

Model Reality

ERS1

ERS2

T/P

GFO

Jason1

ENVISAT

Data losses

Number of Altimeters in operation



Will altimetry provide the input needed by near real time applications ?

• Probability to have the minimum observing system needed to monitor mesoscale : – 2 satellites for delayed time applications (red), – 3 satellites for near real time applications (blue)

• Critical situations– Probability less than 10% in 2008

for near real time applications,Probability less than 25% until CryoSat.

– Odds fall down to 30% after the scheduled end of ENVISAT in 2010, and until both Sentinel-3 are operational

• Results are arguably optimistic:– altimeters assumed to be

operational just after launch – 100% launch success rate

• HY-2 to be added• Jason-3 to be confirmed• Launch dates to be tuned

according to latest infos 00.2

0.4

0.6

0.8

1

2007 2008 2009 2010 2011 2012 2013 2014

Jason-2 (old T/P track, not stacking

with Jason-1)

CryoSat

Jason-1 tandem phase (new ground track covered)

AltiKa

End of ENVISAT

Sentinel-3x 2

00.2

0.4

0.6

0.8

1

2007 2008 2009 2010 2011 2012 2013 2014

Jason-2 (old T/P track, not stacking

with Jason-1)

CryoSat

Jason-1 tandem phase (new ground track covered)

AltiKa

End of ENVISAT

Sentinel-3x 2

Jason -3

Probability to get 2 satellites availableProbability to get 3 satellites available



Jason tandem phase• The probabilistic model was used to explore alternate scenarios and to assess

the impact of various decisions on the minimum observability needed for Near Real Time applications and operational oceanography

• As an illustration, shifting Jason-1 on a different orbit (tandem phase, similar to the Jason/TP phase in 2003/2005) would improve the NRT odds by 20%

• The cross-calibration of both Jasons is mandatory, but if the tandem phase does not start six months after launch, the odds of getting the minimum NRT data reach 0%

00.2

0.4

0.6

0.8

1

2008 2009 2010 2011 2012

Tandem phase after 6 months ? 0% chance of minimum NRT coverage

Tandem phase after 6 months(up to 20% chance increaseto get the minimum NRTobservation)

00.2

0.4

0.6

0.8

1

2008 2009 2010 2011 2012

Tandem phase after 6 months ? 0% chance of minimum NRT coverage

Tandem phase after 6 months(up to 20% chance increaseto get the minimum NRTobservation)

Probability to have the minimum NRT observability without a Jason-1 tandem mission (green), or with a Jason-1 tandem phase 6 months (red) or 1 year (blue)

after the launch of Jason-2



Exploiting CryoSat data

• Similarly, an additional 20% chances can be gained if CryoSat data are fully exploited with a complete coverage on ocean (ice + ocean as primary mission), and better geophysical corrections/references :

– improved MSS to balance out the non-repetitive ground track,

– Improved model for ionosphere correction

– Improved model for wet troposphere correction

00.2

0.4

0.6

0.8

1

2009 2010 2011 2012 2013 2014

20% chance increase to get the minimum observation in NRT with perfect ocean coverage and optimized processingProbability to have the minimum NRT

observability with opportunistic ocean data from an ice-oriented CryoSat

(blue) and with perfect ocean coverage and improved geophysical corrections

(green)



Summary and Conclusions

• Minimum number of satellites in operations needed to monitor mesoscale structures :– 2 satellites in delayed time (offline studies)

– 3 satellites or more in near real time

• Simple probability model used to assess the odds of having an observing system (multiple altimeters) accurate enough to provide the input measurements needed for offline and Near Real Time applications

• Validation with historical data : good consistency and global predictions but unexpected boolean events (T/P longevity, launch failure…) Improvements possible on model

• Offline mesoscale observability is likely (P>60% after Jason-2)

• Odds of meeting the NRT requirements are very poor until Jason-3 + both Sentinel-3 (P<25% until 2010, and 40% until 2011). It is ironically when operational oceanography and NRT applications start to be fruitful that NRT altimeter data may be lacking.

• Odds can be improved with a Jason tandem phase or a better use of CryoSat on ocean

• One more reliable altimeter launched before 2013 would almost double the NRT odds

ostst, hobart, march 2007 - 1- future altimeter systems : is mesoscale observability guaranteed for...

Documents