VERIFIED FACTS — Post 48 “The Hemisphere That Caught Up” (Satellite Radiance Assimilation)

Verification agent pass, 2026-07-20. This table is CANONICAL — it overrides the four research dossiers where they conflict. Verdicts: CONFIRMED / DISPUTED / CORRECTED / COULD NOT VERIFY (the writer must OMIT every COULD-NOT-VERIFY row). Wayback column: URLs marked “saved 2026-07-20” received a fresh 302 snapshot request this session; the generic web.archive.org/web/<url> form resolves to the latest capture. Wikipedia AMSU and Anthes pages returned Wayback 429/520 (rate-limited) this session but have long-standing existing captures — the generic form still resolves.

Fact Verdict Primary URL Wayback URL Confidence Notes
SH 500 hPa forecast historically trailed NH by ~1 day of skill; the two anomaly-correlation curves converged by the mid-2000s CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/10.1256/003590002321042135 https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2001/12238-some-aspects-improvement-skill-numerical-weather-prediction.pdf High Simmons & Hollingsworth 2002; same frame paper as Post 47
S&H 2002 abstract verbatim: “The improvement amounts to about a one-day gain in predictability of mean-sea-level pressure and 500hPa height over the last decade in the northern hemisphere, with a similar gain over the last three years in the southern hemisphere.” CONFIRMED ECMWF Tech Memo 342 https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2001/12238-some-aspects-improvement-skill-numerical-weather-prediction.pdf High Quote read verbatim off TM342 PDF this session; safe to quote
S&H 2002 payoff line verbatim: “One-day forecast errors have been reduced so much in the southern hemisphere that medium-range forecasts for the region have become almost as skilful as those for the northern hemisphere.” CONFIRMED ECMWF Tech Memo 342 https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2001/12238-some-aspects-improvement-skill-numerical-weather-prediction.pdf High Verbatim; saved 2026-07-20
Simmons & Hollingsworth 2002, QJRMS 128:647–677, DOI 10.1256/003590002321042135 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/10.1256/003590002321042135 (Wiley record; full text HTTP 402 paywalled) High Citation correct; full text paywalled, use TM342 for quotes
VTPR = operational 8-channel IR sounder, flown on NOAA-2 through NOAA-5; routine soundings Nov 1972–Feb 1979 CONFIRMED https://www.ncei.noaa.gov/products/vertical-temperature-profile-radiometer https://web.archive.org/web/https://www.ncei.noaa.gov/products/vertical-temperature-profile-radiometer High NCEI VTPR page; “8-channel”, NOAA 2–5, Nov 1972–Feb 1979 confirmed. Per-channel split (7 CO2 + 1 window) NOT re-verified — present as “8-channel”
NOAA-2 (ITOS-D) launched 15 October 1972 CONFIRMED https://www.ncei.noaa.gov/products/vertical-temperature-profile-radiometer (NCEI, above) Medium Launch day widely documented; NCEI confirms 1972 operational start, exact day from general record
TIROS-N launched 13 October 1978 (11:23 UTC), sun-synchronous ~829–845 km; carried TOVS CONFIRMED https://en.wikipedia.org/wiki/TIROS-N https://web.archive.org/web/https://en.wikipedia.org/wiki/TIROS-N High Saved 2026-07-20
TOVS = HIRS/2 (20 ch: 19 IR + 1 vis) + MSU (4 ch, 50–58 GHz O2) + SSU (3 ch, 15 µm CO2 pressure-modulated) CONFIRMED https://www.class.noaa.gov/release/data_available/tovs_atovs/index.htm https://web.archive.org/web/https://www.class.noaa.gov/release/data_available/tovs_atovs/index.htm Medium Standard TOVS spec; channel counts consistent across sources, not each re-pulled to a single primary
Eyre 1989, “Inversion of cloudy satellite sounding radiances by nonlinear optimal estimation. I”, QJRMS 115:1001–1026 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49711548902 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49711548902 High Vol/pages verified; Part II is 115:1027–1037
1D-Var operational on TOVS at ECMWF June 1992; Eyre, Kelly, McNally, Andersson & Persson 1993, “Assimilation of TOVS radiance information through one-dimensional variational analysis”, QJRMS 119 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49711951411 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49711951411 High From Eyre 2007 30-yr review + Wiley record
Andersson, Pailleux, Thépaut, Eyre, McNally, Kelly & Courtier 1994, “Use of cloud-cleared radiances in three/four-dimensional variational data assimilation”, QJRMS 120 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49712051707 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49712051707 High DOI corrected to qj.49712051707 (research file cited no DOI)
ECMWF 3D-Var operational 30 January 1996 (direct TOVS cloud-cleared radiances), ran to 24 November 1997 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49712455002 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.49712455002 High Courtier et al. 1998 (3D-Var I); dates confirmed
ECMWF 4D-Var operational 25 November 1997 (first direct radiance assimilation in 4D-Var) CONFIRMED (Post 47 spine) — High Consistent with 3D-Var ending 24 Nov 1997
NOAA-15 (NOAA-K) launched 13 May 1998 — first spacecraft to fly AMSU (ATOVS era begins) CONFIRMED https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit https://web.archive.org/web/https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit High Wayback 520 this session (rate-limited); page has existing captures
AMSU-A: 15 channels between 23.8 and 89 GHz (O2-band temperature sounding) CONFIRMED https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit https://web.archive.org/web/https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit High  
AMSU-B: 5 channels between 89 and 183.3 GHz (moisture) CONFIRMED https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit https://web.archive.org/web/https://en.wikipedia.org/wiki/Advanced_microwave_sounding_unit High  
ATOVS (incl. AMSU-A) radiances operational at ECMWF May 1999 CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2008/9341-progress-achieved-assimilation-satellite-data-numerical-weather-prediction-over-last-30-years.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2008/9341-progress-achieved-assimilation-satellite-data-numerical-weather-prediction-over-last-30-years.pdf Medium McNally et al. 1999; month from Eyre review, saved 2026-07-20
AIRS on Aqua launched 4 May 2002; 2378 infrared channels, 3.7–15.4 µm CONFIRMED https://en.wikipedia.org/wiki/Atmospheric_infrared_sounder https://web.archive.org/web/https://en.wikipedia.org/wiki/Atmospheric_infrared_sounder High Saved 2026-07-20. ECMWF AIRS assimilation start 2003–04 (medium, minor)
IASI on MetOp-A launched 19 October 2006; 8461 spectral samples, 645–2760 cm⁻¹ (15.5–3.62 µm) CONFIRMED https://en.wikipedia.org/wiki/Infrared_atmospheric_sounding_interferometer https://web.archive.org/web/https://en.wikipedia.org/wiki/Infrared_atmospheric_sounding_interferometer High Saved 2026-07-20; “~8461” in the brief is exact
COSMIC/FORMOSAT-3 launched 15 April 2006 (Vandenberg, single Minotaur), six satellites; world’s first operational GPS radio-occultation mission (US–Taiwan) CONFIRMED https://www.eoportal.org/satellite-missions/formosat-3 https://web.archive.org/web/https://www.eoportal.org/satellite-missions/formosat-3 High Confirmed via eoPortal/NESDIS
GPS/MET (Microlab-1) demonstrated RO 1995; CHAMP RO from 2000 CONFIRMED https://www.cosmic.ucar.edu/what-we-do/gnss-radio-occultation https://web.archive.org/web/https://www.cosmic.ucar.edu/what-we-do/gnss-radio-occultation Medium Rocken et al. 1997 validation; consistent across sources
ATMS = 22-channel scanning microwave radiometer (AMSU-A + AMSU-B/MHS heritage) CONFIRMED https://en.wikipedia.org/wiki/Advanced_Technology_Microwave_Sounder https://web.archive.org/web/https://en.wikipedia.org/wiki/Advanced_Technology_Microwave_Sounder High Was a FLAG; now confirmed. Saved 2026-07-20
Suomi NPP launched 28 October 2011 (carried first ATMS) CONFIRMED https://en.wikipedia.org/wiki/Suomi_NPP https://web.archive.org/web/https://en.wikipedia.org/wiki/Suomi_NPP Medium Widely documented; Wayback 429 this session (rate-limited). Primary not re-pulled but date firmly established
Variational bias correction (VarBC) operational in ECMWF IFS September 2006 CONFIRMED https://nwp-saf.eumetsat.int/site/download/ECMWF-VarBC.pdf https://web.archive.org/web/https://nwp-saf.eumetsat.int/site/download/ECMWF-VarBC.pdf High Was a FLAG (month unconfirmed); ECMWF/NWP-SAF VarBC doc confirms “operational implementation in September 2006”
Dee 2004, “Variational bias correction of radiance data in the ECMWF system” (ECMWF workshop, 28 Jun–1 Jul 2004) CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2004/8930-variational-bias-correction-radiance-data-ecmwf-system.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2004/8930-variational-bias-correction-radiance-data-ecmwf-system.pdf High Saved 2026-07-20
Auligné, McNally & Dee 2007, “Adaptive bias correction for satellite data in a NWP system”, QJRMS 133:631–642 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.57 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.57 High Citation correct
Cardinali 2009, “Monitoring the observation impact on the short-range forecast”, QJRMS 135:239–250, DOI 10.1002/qj.366 CONFIRMED https://rmets.onlinelibrary.wiley.com/doi/10.1002/qj.366 https://web.archive.org/web/https://rmets.onlinelibrary.wiley.com/doi/10.1002/qj.366 High Original adjoint-FSO paper
FSOI ranking (data Jun–Oct 2011): “The largest contribution to decreasing the forecast error is provided by AMSU-A (∼25%), IASI, AIRS, AIREP (aircraft data) and GPS-RO observations account for 10% of the total impact, respectively. TEMP and SYNOP surface pressure observations contribute by 5% followed by AMVs and HIRS (∼4%), then by ASCAT and DRIBU (3%). All other observations contribute to less than 3%.” CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2013/16937-observation-impact-short-range-forecast.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2013/16937-observation-impact-short-range-forecast.pdf High Source p.—lines 397-399 reads “AIREP (aircraft data and GPS-RO observations account for 10%…” with an unclosed paren (typo in original); rendered here and in the post as “AIREP (aircraft data)” closing the paren editorially. AMSU-A = single most valuable system
Langland & Baker 2004, “Estimation of observation impact using the NRL atmospheric variational data assimilation adjoint system”, Tellus A 56:189–201 — introduced adjoint-based FSOI CONFIRMED https://www.researchgate.net/publication/312359703 (ResearchGate record) Medium Citation confirmed; full text not fetched. Do NOT attach a specific AMSU-A % to this paper
Joo, Eyre & Marriott 2013, “The Impact of MetOp and Other Satellite Data within the Met Office Global NWP System Using an Adjoint-Based Sensitivity Method”, MWR 141:3331–3342 CONFIRMED (AMS/MWR record) — Medium Citation confirmed; AMSU-A and IASI the leading satellite contributors (qualitative). Exact % paywalled — writer must not quote a number
Data-denial OSE (TM839, Bormann et al. 2019): SH extratropics — withholding microwave radiances degrades day-3 500 hPa error by 11%, significant to day 9; NH extratropics — conventional 10%, microwave 6% at day 3 CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2019/18859-global-observing-system-experiments-ecmwf-assimilation-system.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2019/18859-global-observing-system-experiments-ecmwf-assimilation-system.pdf High Fetched + pdftotext in research; saved 2026-07-20. Preferred hard numbers for the SH-collapse point
Denial of a single AMSU-A instrument costs ~0.5% in forecast scores; total satellite loss → “catastrophic degradation” in SH (TM711, English et al. 2013) CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2013/9301-impact-satellite-data.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2013/9301-impact-satellite-data.pdf High Saved 2026-07-20
~1/4 of the 30-yr forecast improvement is from better observations, ~3/4 from better assimilation/model (Dee & Uppala 2009 via TM711) CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2013/9301-impact-satellite-data.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2013/9301-impact-satellite-data.pdf Medium Important nuance: convergence is satellite-DRIVEN, not satellite-ONLY
Richard (Rick) Anthes born 9 March 1944, St. Louis, Missouri; President of UCAR 1988–2012; key figure in GPS/MET and COSMIC CONFIRMED https://en.wikipedia.org/wiki/Richard_A._Anthes https://web.archive.org/web/https://en.wikipedia.org/wiki/Richard_A._Anthes High Birthplace added. Wayback 520 this session (rate-limited); page has existing captures
RTTOV lineage: Eyre & Woolf 1988 → Eyre 1991 (ECMWF Tech Memo 176, first RTTOV) → Rayer 1995, Rizzi & Matricardi 1998, Saunders & Matricardi 1999 (TM282); now EUMETSAT NWP SAF, Roger Saunders lead CONFIRMED https://nwpsaf.eu/oldsite/deliverables/rtm/papers/tm282.pdf https://web.archive.org/web/https://nwpsaf.eu/oldsite/deliverables/rtm/papers/tm282.pdf Medium Lineage per NWP SAF/Eyre review
Saunders et al. 2018, “An update on the RTTOV fast radiative transfer model (currently at version 12)”, GMD 11:2717, DOI 10.5194/gmd-11-2717-2018 CONFIRMED https://gmd.copernicus.org/articles/11/2717/2018/ https://web.archive.org/web/https://gmd.copernicus.org/articles/11/2717/2018/ High Citation correct
Roger Saunders was ECMWF Satellite Section head 1995–1999 CONFIRMED (ECMWF, research) — Medium Career fact from research; not re-pulled to primary this session
Eyre 2007 ECMWF Seminar review “Progress achieved on assimilation of satellite data in NWP over the last 30 years” — the richest single narrative source CONFIRMED https://www.ecmwf.int/sites/default/files/elibrary/2008/9341-progress-achieved-assimilation-satellite-data-numerical-weather-prediction-over-last-30-years.pdf https://web.archive.org/web/https://www.ecmwf.int/sites/default/files/elibrary/2008/9341-progress-achieved-assimilation-satellite-data-numerical-weather-prediction-over-last-30-years.pdf High Full text read in research; saved 2026-07-20
NCEP milestones: SSI/3D-Var operational 1991; direct clear-sky IR+MW radiance assimilation in GFS added ~1995 (Derber & Wu 1998) DISPUTED https://www.emc.ncep.noaa.gov/emc/pages/numerical_forecast_systems/ncep_data_assimilation.php https://web.archive.org/web/https://www.emc.ncep.noaa.gov/emc/pages/numerical_forecast_systems/ncep_data_assimilation.php Medium Research files give both “June 1991” (SSI 3D-Var) and “Oct 1995”/”1998” (direct radiance) — these are TWO milestones, not a contradiction. Writer should distinguish: SSI/3D-Var operational 1991; direct radiance assimilation from mid-1990s. Exact direct-radiance date not pinned to primary
Lawrence “Larry” McMillin — exact birth and death dates COULD NOT VERIFY — — — OMIT. Two 2024 obituaries surfaced (Joliet IL Caterpillar assembler b.1946; Kinston NC 1935–2024) — NEITHER confirmed as the NOAA/NESDIS scientist. Do not print any birth/death year for McMillin. His career facts (NOAA/NESDIS, retrieval, split-window) are fine
McMillin split-window original year “1975” (McMillin 1975, JGR 80:5113–5117) as the primary origin COULD NOT VERIFY — — Low Primary 1975 paper not fetched; year comes from secondary literature. McMillin 1984 JGR (multiple-window SST, DOI 10.1029/JC089iC03p03655) IS confirmed. Writer: attribute split-window to “McMillin, mid-1970s” or cite the 1984 paper; do not assert a hard 1975 primary without hedge
Langland & Baker 2004 and Joo et al. 2013 exact AMSU-A percentages COULD NOT VERIFY — — — OMIT numbers. Both paywalled; only ECMWF/Cardinali ~25% is solid. State AMSU-A/IASI as “leading contributors” qualitatively for those two studies
“SH without satellites: +50% error short-range, +25–35% medium-range, ≈24 h of lost skill” COULD NOT VERIFY — — — OMIT. Web-summary figure only; not in any fetched primary. Use TM839 (11% at day 3, significant to day 9) + S&H “~1 day” instead to make the same point
Present-day exact SH-vs-NH day-N where 500 hPa AC curves meet (e.g. “both reach 80% AC at day X”) COULD NOT VERIFY — — — OMIT specific day numbers. Only the qualitative “curves now coincide” + S&H “~1 day / almost as skilful” are safe
1990s 5-day 500 hPa AC magnitudes (~0.6→0.75 NH, ~0.4→0.625 SH) COULD NOT VERIFY — — — OMIT as quoted figures; secondary/order-of-magnitude only, not read off a primary chart

Verification summary

  • Every load-bearing date, channel count, and citation in the four dossiers CONFIRMED against a primary/authoritative source.
  • Two former FLAGs RESOLVED to CONFIRMED: VarBC operational September 2006; ATMS = 22 channels.
  • The two anchor QUOTES (S&H 2002 abstract “one-day gain… last three years in the southern hemisphere”; and the “almost as skilful as… northern hemisphere” line) read VERBATIM off ECMWF TM342 — safe to quote.
  • The FSOI ranking read VERBATIM off the Cardinali 2013 ECMWF lecture notes (data Jun–Oct 2011) — AMSU-A ~25%, single most valuable system.
  • Wayback: fresh 302 snapshot requests succeeded this session for TM342, Cardinali 2013 notes, TM711, TM839, Eyre 2007 review, Dee 2004, Wikipedia IASI/AIRS/ATMS/TIROS-N. Wikipedia AMSU and Anthes and NCEI VTPR and Suomi-NPP returned 429/520 (Wayback rate-limiting) — those pages have existing long-standing captures, so the generic web.archive.org/web/<url> form still resolves.