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:20230831T095747Z
LOCATION:Sertig
DTSTART;TZID=Europe/Stockholm:20230628T140000
DTEND;TZID=Europe/Stockholm:20230628T143000
UID:submissions.pasc-conference.org_PASC23_sess109_msa264@linklings.com
SUMMARY:Building Data-Driven Surrogate Models at Appropriate Granularity U
 sing High-Resolution Climate Simulation and Coupling Library
DESCRIPTION:Minisymposium\n\nHisashi Yashiro (National Institute for Envir
 onmental Studies); Takashi Arakawa (CliMTech Inc., The University of Tokyo
 ); and Shinji Sumimoto and Kengo Nakajima (The University of Tokyo)\n\nWhe
 n building a climate model emulator using machine learning, we are concern
 ed about the generalization performance of the surrogate model and the ext
 rapolation of the simulation results. It is very problematic to evaluate a
 n emulator that implicitly includes too many physical processes and has be
 en trained to reproduce only the training data that can exist now. Nor wil
 l scientists be able to use the results of such emulators to expand scient
 ific knowledge. We extended the coupling library that combines components 
 of climate models, such as atmospheric and oceanic models, and developed a
  system for constructing scheme-level surrogate models. With the functiona
 lity provided by this coupling library named h3-Open-UTIL/MP, we extract t
 he inputs and outputs of a particular physical scheme under simulation and
  transform them to different spatiotemporal resolutions as needed. That da
 ta is fed on-the-fly to a Python-based machine learning suite. By having t
 he surrogate model behave like a faithful "accelerator" to the original ph
 ysical model, we can perform long-term simulations with reduced anxiety ab
 out unexpected behavior. Also, by constructing a parameterization model of
  the conventional (low-resolution) climate model based on km-scale simulat
 ion, improvements in reproducibility and reduction of computational worklo
 ad are expected.\n\nDomain: Climate, Weather and Earth Sciences\n\nSession
  Chair: Tobias Weigel (German Climate Computing Centre, DKRZ)
END:VEVENT
END:VCALENDAR