Satellite Altimetry Status Products Part 2 Do GODAE

Satellite Altimetry Status & Products Part 2 Do GODAE models need upcoming altimetry missions ? G. Dibarboure G. Larnicol J. Lambin 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 1 -

Introduction • ! Starting points ( ): recent history shows that we should not take altimetry for granted – Operations of old missions difficult to carry on (GFO, Jason-1, ENVISAT) – Decision process for new missions long and complex (Jason-3 not fully approved) – Specific OSE/OSSE/impact studies were performed by CLS ( ) upon requests from CNES & ESA to provide material in a time span compatible with the decision process √ • ∞ Old missions decommissioned, future ones not approved (risk : no data whatsoever) What is at stake for the long term ( ) ? – – Integration of “subpar” missions in DUACS not certain (risk : no SLA product whatsoever) – The role of GODAE as a prescriber for the altimetry constellation • ? What should be the status of routine GODAE indicators when a satellite is added/lost ? Questions for GODAE members ( ) considering “operational metrics” – – – Can we use OSE/OSSE outputs to refresh evidence that GODAE models need altimetry ? Can routine metrics show evidence that GODAE specifically needs mission X or Y ? Does this still holds if they are subpar (degraded performance, aging, drifting orbit) ? Should OSSEs performed on the observing system itself remain a quick and cheap complement to model-based recommendations ? 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 2 -

Recent altimetric events (1/2) • Jason tandem now fully operational – Excellent results and consistency between Jasons (July 08) – Used by DUACS in August 2008 (anomaly on Jason-1) – But the Jason tandem started only in early March 09 (no green light from project, conservative Cal/Val approach) – With more operational evidence, could we have gained up to 6 months of tandem – General statements (3 - 4 sats) and tandem results were considered well-published Mostly ignored in 2008 ! 1 m 2. 1 m Map of Absolute Dynamic Topography Animation : Jason-2 vs Tandem ∞ • WARNING : Nasa funding for Jason-1 after 2010 !– ? – Recently jeopardized (do you need a tandem anyway? ) Can we use GODAE outputs to illustrate the need for more tandem data and the specific need for extended Jason -1 operations ? (deadline : July) 1 m 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 3 - 2. 1 m

Recent altimetric events (2/2) • ! • GFO : stopped after 11 years of service after a critical onbard failure Twice, Navy (& NOAA) asked for evidence that operations were still worth funding despite her aging problems and limited coverage √ • Inputs based on operational DUACS metrics provided with highlights of GFO contribution Operations successfully defended twice ? • ∞ • Can routine metrics from models provide more evidence if a similar case arise ? Can we deploy them fast enough to be compatible with the decision process ? Should minimalistic outputs from DUACS remain an official complement to model studies as a contribution to a dedicated GODAE task force ? Formal mapping error of the multi-satellite objective analysis in % of the signal variance (black line), and missing GFO data on ocean (grey histogram) and quality levels (colors) -Blue level : 3 missions nominal -Green level : 2 missions 100%, GFO 50% -Orange level : 1 mission unavailable -Crimson level : 1 mission unavailable + anomaly on a second mission 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 4 -

What about altimeter oldies ? • Jason-1 status (8 yo > extended lifespan) : • Intruments are performing well (tandem > JA 1/TP) • Most redundant equipment safeties burnt already • The next major failure might be the last one • ENVISAT status (7 yo > extended lifespan) : • Degraded quality (S-Band lost) • Option considered by ESA : drifting phase by mid-2010 Any use for altimetry users ? • Geodetic phase of ERS-1 in DUACS since 2008 • Extended phase of ENVISAT can be used by DUACS after 2010 (but with additional errors) • Multi-satellite mapping OSE : a drifting+degraded ENVISAT is still a noteworthy sampling addition • Are drifting data relevant for GODAE models ? Despite the additional errors ? ! % of additional variance when geodetic data from ERS-1 are added to T/P Map of SLA (1994/07/20) for TP (left) and TP+ERS 1 geodetic (right) √ ? 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 5 -

Upcoming additions to DUACS • Alti. Ka : likely available in DUACS in 2011 probably in line with ENVISAT quality with minor concerns about influence of rain • Cryo. Sat : • Launched in December 09, data possibly available during the commissionning phase (if requirement is adamant) ! • Ocean L 2 product still being discussed but not secured (Cryo. Sat-specific requirements from GODAE might help) • Limitation : no dual frequency, no radiometer, drifting orbit √ • ? • ∞ • OSSE DUACS Same results as for drifting ENVISAT (sampling wins over error budget) Would GODAE models also benefit from improved processing and minimized errors? (ongoing project SLOOP) More generally, would drifting altimetry (geodetic+ocean) be acceptable for models despite the additional errors ? 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 6 - Mapping-based impact study for Cryo. Sat and Alti. Ka : mapping reconstruction error normalized by the mapping error of Jason-2 alone

Other opportunities • Sentinel 3 tandem ? √– !–– ∞– • Sampling from one S 3 is good (mapping OSSE) S 3 -B = fully redundant sampling (for altimetry, OC-oriented) Alternative orbit options can better exploit the S 3 A/S 3 B tandem We will need more evidence to back-up the need for a S 3 altimetry tandem HY-2 ? (Chinese mission with CNES contribution) – Availability in Near Real Time not confirmed – Actual quality level still unknown – New ground track + geodetic phase after 2 years – Should we run a dedicated OSSE to back-up the need for HY-2 data ? ? • A second Geosat follow-on ? – Jason-class quality ? Open data policy ? – Ground track and orbit-related sampling values well-known 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 7 - Mapping-based OSSE for Sentinel-3 A: mapping reconstruction error in the Gulf Stream

Three altimeters in operations ? ∞ • OSE/OSSE figures of merit can be used in a combinatorial probabilistic model : • To quantify the ability of a given satellite to strengthen the constellation • • √ • • To see if mission X or Y would be a better option To identify false improvements (redundant sampling) or critical periods Probability model to get • 3 fully operational missions (100% of the time) • On different ground tracks (Jason-1 / Jason-2, Alti. Ka / ENVISAT) • Cal. Val phase = redundant sampling (failure prob cancelled by Ja-2) • Interleaved phase = probabilities stack • At nominal quality level (e. g : mapping OSSE says Cryo. Sat = 50%) Typical figures • Nominal satellite lifespan = 75% chance to have the ground track covered • Probabilistic death at 2. 2 * nominal lifespan • Risk of failure at launch (or early life anomalies) not taken into account 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 8 -

Do GODAE models need a reference mission ? ! • √ • • ? • • ∞ • Jason-3 is not fully approved (do we need it anyway ? ) Primary use : high-precision and reference ground track (ongoing OSE-like work on MSL applications) Can be used to minimize geographically correlated errors on other missions (TP era ? ) In 2013+, other missions might achieve a good POD level (acceptable large scale errors ? TBC) DUACS maps do need a large scale reference What about GODAE models ? Would mesoscale-oriented missions be enough or do models need good accuracy on large scale as well ? What new GODAE metrics (OSSE/routine indicators) could help illustrate long-term need for a reference mission ? Regional MSL trend differences between Jason and GFO (-10/+10 mm/year) Global MSL trend for Jason and ENVISAT : after 2004, Jason was used to detect and minimize errors on ENVISAT 2 nd GODAE Observing System Evaluation Workshop - June 2009 -

Conclusion (1/2) ! • Altimetry constellation status √ • CLS performed many OSE or OSSE-like studies √ • Advantages of the DUACS approach : analyses limited to the observing system – – – New Jason tandem operational and ENVISAT still active OK status today Constellation remains extremely fragile (no spare, Jason-1 operations stopped in 2011? ) Funding new missions and operations on old satellites is difficult General considerations (3 – 4 satellites) are not enough (we need to ask for mission X or Y) Even what might seem obvious requirements (Jason & S 3 tandems, Jason-3) need new or « refreshed » scientific evidence Altimetry should not be taken for granted – Generally short and mission/application oriented useful when we either lack time or money – The outputs are now used to build DUACS’ operational Key Performance Indicators – Useful as a cheap complement to detailed model-based outputs – Easy to set up and customize (comparing 20+ variants is possible helpful for early designs) – More sensitive to subtle differences (changes in orbit, payload, processing) 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 10 -

Conclusion (2/2) ? • ∞ • This work does not replace in-depth impact studies based on models – GODAE model requirements would be an order of magnitude stronger if provided with solid evidence based on recent OSE/OSSE, metrics, and indicators (agency/mission specific) – If routine model metrics can supplement OSSE studies, it is important to exploit them to illustrate major events on the altimetric constellation – DUACS impact studies can incorporate new metrics derived from model-based outputs Integrating subpar missions and Real Time : high or low priority for DUACS (sampling vs error) ? To be a strong altimetry prescriber, GODAE needs a specific task force – Able to run quick impact studies to assess the gain/loss associated to a major change in the constellation In a time span compatible with the agencies’ calendar needs – Able to make mid-term requirements about upcoming opportunities with new and solid evidence – To have more weight, impact studies performed without models need to remain in line with model requirements, and they need to be one component of this GODAE task force 2 nd GODAE Observing System Evaluation Workshop - June 2009 - 11 -