BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260421T090513Z
LOCATION:Bldg. 6 - Room 104
DTSTART;TZID=Europe/Stockholm:20260701T143000
DTEND;TZID=Europe/Stockholm:20260701T150000
UID:submissions.pasc-conference.org_PASC26_sess163_msa196@linklings.com
SUMMARY:Single Precision Trajectories in 4D‑Var Data Assimilation
DESCRIPTION:Jorge Bandeiras, Patrick Gillies, Balthasar Reuter, and Ioan H
 adade (ECMWF)\n\nFour‑dimensional variational (4D‑Var) data assimilation i
 s underpinning the generation of high‑quality initial conditions required 
 for ECMWF’s weather forecasts and Destination Earth's (DestinE) Digital Tw
 ins. A major cost driver in this system is the nonlinear (NL) model trajec
 tory, whose forward integrations become increasingly expensive at the high
  spatial resolutions used in DestinE. Building on previous work, the curre
 nt IFS cycle CY50R1 introduced single‑precision arithmetic within the NL t
 rajectory to substantially reduce computational cost while preserving fore
 cast skill.\n \nAchieving this balance required a targeted precision strat
 egy. While single precision was adopted broadly in the forecast model to m
 aximize efficiency, specific components - particularly observation operato
 rs - were retained in double precision where necessary to avoid degrading 
 the representation of observational information. This hybrid approach ensu
 red numerical robustness in sensitive operators such as RTTOV, while still
  enabling significant speed‑ups in the overall system.\n \nBenchmark exper
 iments demonstrated a reduction of approximately 30% in the computational 
 cost of the NL forecast. The integration of single‑precision trajectories 
 into 4D‑Var marks an important step toward scalable, high‑resolution data 
 assimilation for next‑generation Earth system modelling.\n\nDomain: Climat
 e, Weather, and Earth Sciences\n\nSession Chairs: Balthasar Reuter (ECMWF)
  and Claudia Frauen (DKRZ)\n\n
END:VEVENT
END:VCALENDAR