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:20260731T133316Z
LOCATION:Bldg. 6 - 002
DTSTART;TZID=Europe/Stockholm:20260701T140000
DTEND;TZID=Europe/Stockholm:20260701T160000
UID:submissions.pasc-conference.org_PASC26_sess122@linklings.com
SUMMARY:MS5A - Data-Driven Regional Weather Modeling: Towards Trustworthy 
 Convection-Resolving Forecasts
DESCRIPTION:Access the recording\n\nOrganizer(s): Oliver Fuhrer (MeteoSwis
 s, ETH Zurich), and Laure Raynaud (Météo-France)\n\nData-driven weather pr
 ediction has advanced rapidly in recent years, with machine-learning-based
  models now complementing traditional numerical weather prediction. While 
 global data-driven models have demonstrated impressive skill, extending th
 ese approaches to regional, convection-resolving forecasting introduces ne
 w scientific and computational challenges. At kilometer and sub-kilometer 
 scales, models must represent complex physical processes, integrate high-f
 requency observations, and provide trustworthy uncertainty estimates, part
 icularly for extremes. This minisymposium focuses on recent progress and o
 pen challenges in data-driven regional weather modeling. Topics include ge
 nerative diffusion models for convective-scale downscaling, graph-based ne
 ural network architectures for high-resolution domains, training strategie
 s that improve generalization across scales and regions, and operational p
 erspectives from national meteorological services. The session emphasizes 
 scientific trustworthiness, evaluation, and physical consistency, and disc
 usses how these requirements interact with high-performance computing work
 flows and model design. By bringing together experts from academia and ope
 rational forecasting, the minisymposium provides a forum to assess the sta
 te of the art and explore pathways toward reliable, convection-resolving d
 ata-driven forecasts, with relevance to a wide range of computational scie
 nce domains.\n\nComputational and Verification Challenges in Data-Driven A
 tmospheric Downscaling\n\nUsing generative machine learning for performing
  atmospheric downscaling (super-resolution for meteorological data) is of 
 growing interest, as the methods used for data-driven downscaling are comp
 utationally inexpensive compared with statistical downscaling methods or w
 ith data-driven models for for...\n\n\nMary McGlohon and Petar Stamenkovic
  (MeteoSwiss, ETH Zurich); David Leutwyler, Xavier Lapillonne, and Oliver 
 Fuhrer (MeteoSwiss); Fabian Bösch, Lukas Drescher, and Henrique Mendonça (
 ETH Zurich / CSCS); Sebastian Schemm (University of Cambridge); and Siddha
 rtha Mishra (ETH Zurich)\n---------------------\nFrom Numerical to Data-Dr
 iven Regional Forecasting: Challenges from a Scientific and Operational Pe
 rspective\n\nAI has opened a new path for atmospheric modeling, with gains
  in both quality and computational efficiency. At Météo-France, like at ot
 her national weather services, the topic of AI for weather prediction has 
 developed rapidly and is being explored from various angles, with applicat
 ions to both glob...\n\n\nLaure Raynaud (Météo-France)\n------------------
 ---\nTowards Operational Data-Driven Regional Forecasting at Convection-Re
 solving Scales\n\nTranslating global data-driven weather models to regiona
 l, convection-resolving prediction remains a key scientific challenge. In 
 Alpine environments, orographic precipitation, convective initiation, and 
 valley flows demand kilometer-scale resolution and hourly to sub-hourly ou
 tput to provide physic...\n\n\nCarlos Osuna, Claire Merker, Alberto Pennin
 o, Andreas Pauling, Daniele Nerini, Francesco Zanetta, Hugues de Laroussil
 he, Jonas Bhend, Katrin Ehlert, and Mary McGlohon (MeteoSwiss); Michele Ca
 ttaneo (Swiss data science center); and Oliver Fuhrer and Radi Radev (Mete
 oSwiss)\n---------------------\nMulti-Domain: Improving Generalization Acr
 oss Scales and Regions\n\nThe domain of weather forecasting is currently u
 ndergoing a significant transformation driven by advances in machine learn
 ing. Following these developments, high-resolution regional models have em
 erged. Among these regional models is the stretched-grid model (SGM), a gl
 obal model with an increased s...\n\n\nSophie Buurman (KNMI), Aram Farhad 
 Shafiq Salihi and Even Marius Nordhagen (Norwegian Meteorological Institut
 e), Mario Santa Cruz (ECMWF), Michiel van Ginderachteren (RMI Belgium), an
 d Thomas Nils Nipen (Norwegian Meteorological Institute)\n\nDomain: Climat
 e, Weather, and Earth Sciences, Physics, Computational Methods and Applied
  Mathematics\n\nSession Chair: Oliver Fuhrer (MeteoSwiss, ETH Zurich)
END:VEVENT
END:VCALENDAR