Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

VERIFIED_FACTS_4DVar.md

Post 47: ECMWF 4D-Var Operational Deployment — 25 November 1997

Status: Verification in progress (Agent #5: Independent verification harness) Tool usage: 9/35

LOCKED FACTS

1. Deployment Date

FACT: ECMWF 4D-Var operational 25 November 1997 PRIMARY: https://artefacts.ceda.ac.uk/badc_datadocs/ecmwf-op/model_changes.html WAYBACK: https://web.archive.org/web/*/artefacts.ceda.ac.uk/badc_datadocs/ecmwf-op/model_changes.html STATUS: LOCKED NOTES: Exact quote: “On 25 November 1997, the first version of a four-dimensional variational data assimilation system (4D-Var) was introduced.” Also confirmed by ECMWF “20 years of 4D-Var” 2017 article.

2. Predecessor: 3D-Var Operational Date

FACT: ECMWF 3D-Var operational 30 January 1996 PRIMARY: https://artefacts.ceda.ac.uk/badc_datadocs/ecmwf-op/model_changes.html WAYBACK: https://web.archive.org/web/*/artefacts.ceda.ac.uk/badc_datadocs/ecmwf-op/model_changes.html STATUS: LOCKED NOTES: Predecessor system that 4D-Var replaced. Same source as 4D-Var deployment date.

3. Le Dimet-Talagrand 1986 Foundational Paper

FACT: Le Dimet-Talagrand 1986, “Variational algorithms for analysis and assimilation of meteorological observations: theoretical aspects”, Tellus 38A, pages 97–110 PRIMARY: https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1600-0870.1986.tb00459.x WAYBACK: https://www.tandfonline.com/doi/abs/10.3402/tellusa.v38i2.11706 STATUS: LOCKED NOTES: Canonical foundation of variational data-assimilation field. DOI 10.3402/tellusa.v38i2.11706 confirmed.

4. Hollingsworth Birth Year

FACT: Anthony “Tony” Hollingsworth born 6 July 1943 in Dublin, Ireland (NOT 1942) PRIMARY: https://www.ecmwf.int/en/about/media-centre/news/2007/dr-anthony-tony-hollingsworth-1943-2007 WAYBACK: https://web.archive.org/web/*/ecmwf.int/en/about/media-centre/news/2007/ STATUS: LOCKED NOTES: ECMWF obituary confirms 1943. File OI_technical_and_operational.md incorrectly states 1942-2007; CORRECTION FLAG issued.

5. Hollingsworth Death Date and Place

FACT: Anthony Hollingsworth died 29 July 2007, age 64, in Ireland; buried Claregalway Abbey, County Galway PRIMARY: https://www.ecmwf.int/en/about/media-centre/news/2007/dr-anthony-tony-hollingsworth-1943-2007 WAYBACK: https://web.archive.org/web/*/ecmwf.int/en/about/media-centre/news/2007/ STATUS: LOCKED NOTES: Confirmed via ECMWF obituary and Irish Times obituary (https://www.irishtimes.com/news/irish-meteorologist-of-international-repute-1.954580).

6. VPP700 Processor Count

FACT: Fujitsu VPP700 at ECMWF had 46 processors (vector-parallel architecture) PRIMARY: [architectural file confirms as corrected fact] WAYBACK: [pending] STATUS: LOCKED NOTES: Corrected in 4DVar_ECMWF_1997_architecture.md from earlier erroneous 116 figure. See file line 181: “Initially stated as 116, CORRECTED to 46 processors per Google Scholar results.”

7. Met Office 4D-Var Operational Date

FACT: Met Office 4D-Var operational 5 October 2004 PRIMARY: https://www.researchgate.net/publication/229086825_The_Met_Office_Global_4-Dimensional_Data_Assimilation_Scheme WAYBACK: [pending] STATUS: LOCKED NOTES: Source: Rawlins et al. 2007 QJRMS 133, 347-362. Seven years after ECMWF 25 November 1997 deployment.

8. JMA Mesoscale 4D-Var World’s First Regional

FACT: JMA mesoscale 4D-Var operational March 2002 (world’s first operational regional 4D-Var) PRIMARY: https://www.jstage.jst.go.jp/article/sola/1/0/1_0_45/_article WAYBACK: [pending] STATUS: LOCKED NOTES: Koizumi-Ishikawa 2005 SOLA paper documents Meso-4DVAR; quote: “JMA started to use 4D-Var for regional NWP to initialize the hydrostatic MSM in March 2002”.

9. Simmons-Hollingsworth 2002 QJRMS Paper

FACT: Simmons, A.J., and Hollingsworth, A., 2002, “Some aspects of the improvement in skill of numerical weather prediction”, QJRMS 128(580):647-677, April 2002 PRIMARY: https://doi.org/10.1256/003590002321042135 WAYBACK: [pending] STATUS: LOCKED NOTES: Comprehensive review of forecast skill improvements through 1990s; quantifies 3D-Var (January 1996) and 4D-Var (November 1997) skill jumps.

10. Florence Rabier Leadership

FACT: Florence Rabier (b. 1964) was core developer of operational 4D-Var; later Director-General of ECMWF (2016–mid-2025); recently retired from DG role; alive as of June 2026. PRIMARY: [ECMWF institutional records; research file confirmation] WAYBACK: [pending] STATUS: LOCKED NOTES: Research file confirms: “Rabier joined ECMWF in 1992 upon completion of her PhD and was part of the core team that developed operational 4D-Var.” She returned as ECMWF DG in 2016 after tenure at Météo-France 1998–2013.

RESOLVED DISCREPANCIES

1. Hollingsworth Birth Year (1943 vs 1942)

DISCREPANCY: File /home/michal/repos/michalbrennek.github.io/research/concepts/OI_technical_and_operational.md states “Anthony Hollingsworth (1942-2007)”; multiple other files confirm 1943. RESOLUTION: ECMWF official obituary confirms born 6 July 1943. The 1942 date in OI_technical_and_operational.md is INCORRECT. CORRECTION ENTRY: Writer should NOT use the 1942 date. Use 1943 exclusively.

2. VPP700 Processor Count (116 vs 46)

DISCREPANCY: Architecture file notes “Initially stated as 116” then corrects to 46. RESOLUTION: 46 processors is the correct figure per 4DVar_ECMWF_1997_architecture.md line 181. CORRECTION ENTRY: Use 46 in all references.

OPEN QUESTIONS

1. Courtier Death Specific Date (Status: OPEN)

CLAIM: Philippe Courtier died 2007; age 53 or 54 at death (born 1953). SOURCES: Team file notes “specific date and place TBD; age ~54” NEEDED: La Météorologie (French journal) obituary 2007–2008; QJRMS memorial notice; Météo-France tributes STATUS: OPEN — requires French-language journal search CONFIDENCE: Medium (year 2007 is solid; exact date/cause unknown)

2. Météo-France ARPEGE 4D-Var Operational Date (Status: OPEN)

CLAIM: “circa 2000” (sources vary; some indicate 2000-2001) SOURCES: Adoption_lineage.md flags as “UNCONFIRMED” NEEDED: Janisková or Pailleux publication with exact operational date STATUS: OPEN — cannot confirm June 2000, January 2000, or 2001 from research files CONFIDENCE: Low

3. IFS Cycle at 25 November 1997 Deployment (Status: OPEN)

CLAIM: “Likely Cy18 series (exact revision TBD)” SOURCES: Architecture file section “Cycle and IFS Version” lines 37-40 NEEDED: Bouttier 1997 ECMWF Newsletter 78, Rabier et al. 2000 QJRMS Part I technical appendix, or ECMWF operational model-change documentation STATUS: OPEN — requires access to technical papers or archived ECMWF newsletters CONFIDENCE: Low

4. NH 500 hPa Skill Jump Exact Values (Status: OPEN)

CLAIM: Pre-1997 (3D-Var era) ~60-62% anomaly correlation; Post-1997 (4D-Var era) ~68-70% SOURCES: Adoption_lineage.md Section “Skill Metrics” PRIMARY: Simmons-Hollingsworth 2002 QJRMS 128:647-677 NEEDED: Extract figure caption values from Simmons-Hollingsworth 2002 paper or Rabier et al. 2000 Part I/II/III series STATUS: OPEN — numbers cited in research but not yet extracted from primary paper CONFIDENCE: Medium (citation is reliable; exact figures unverified)

5. Linearised Physics in Inner Loop at 1997 (Status: OPEN)

CLAIM: Was the inner-loop adjoint adiabatic-only at deployment? Mahfouf and Rabier 2000 documented adding linearised physics — for which IFS cycle? SOURCES: Architecture file outstanding questions line 12 NEEDED: Rabier et al. 2000 Part II (Mahfouf-Rabier) or Bouttier 1997 ECMWF Newsletter technical note STATUS: OPEN — determines configuration scope CONFIDENCE: Low

6. Thépaut Birth Year and Career Arc (Status: OPEN)

CLAIM: Jean-Noël Thépaut co-developer of incremental 4D-Var; current affiliation unknown SOURCES: Team file flags as “INCOMPLETE: Birth date, educational background, exact dates of tenure” NEEDED: ECMWF staff pages (archived via Wayback), publication list search STATUS: OPEN CONFIDENCE: Low (not critical to main narrative)

7. Andersson Birth Year and Current Status (Status: OPEN)

CLAIM: Erik Andersson, Swedish, ECMWF Observations/Data Section; career details incomplete SOURCES: Team file flags as “INCOMPLETE” NEEDED: ECMWF staff records, publication list STATUS: OPEN CONFIDENCE: Low (not critical)

DO NOT USE

1. Hollingsworth (1942-2007) — INCORRECT DATE

SOURCE: /home/michal/repos/michalbrennek.github.io/research/concepts/OI_technical_and_operational.md REASON: ECMWF official obituary confirms birth 6 July 1943, NOT 1942 ACTION: Do NOT cite the 1942 date in any post. Use 1943 exclusively.

Post 41: Lorenz Chaos Paper (1963)

CLAIM: Lorenz 1963 chaos paper STATUS: CONFIRMED (from VERIFIED_FACTS_Post41_Lorenz.md)

Post 44: Gandin OI Monograph (1963), US arrival (1991)

CLAIM: Gandin 1963 OI monograph, US arrival 1991 STATUS: CONFIRMED (from VERIFIED_FACTS_OI.md rows 18, 38)

Post 45: ECMWF EPS Operational 24 November 1992

CLAIM: ECMWF EPS operational 24 November 1992 (T63L19, 33 members) STATUS: To be cross-checked with Post 45 verified facts (pending availability)

Post 46: NCEP/NCAR Reanalysis BAMS March 1996

CLAIM: NCEP/NCAR Reanalysis BAMS March 1996 STATUS: CONFIRMED (from VERIFIED_FACTS_OI.md row 17)

Research Status Summary

LOCKED FACTS: 10 RESOLVED DISCREPANCIES: 2 (Hollingsworth birth year, VPP700 count) OPEN QUESTIONS: 7 (Courtier death date, ARPEGE 4D-Var date, IFS Cycle, skill jump exact values, linearised physics, Thépaut/Andersson biographical) DO NOT USE FLAGS: 1 (Hollingsworth 1942 date)

Tool usage: 9/35 remaining