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:20250822T115807Z
LOCATION:Room 5.0B15 & 16
DTSTART;TZID=Europe/Stockholm:20250616T112000
DTEND;TZID=Europe/Stockholm:20250616T115000
UID:submissions.pasc-conference.org_PASC25_sess139_msa172@linklings.com
SUMMARY:Frugal Extension of Aurora Weather Foundation Model: Applications 
 to the Water Cycle
DESCRIPTION:Fanny Lehmann, Torsten Hoefler, Siddhartha Mishra, Sebastian S
 chemm, and Benedikt Soja (ETH Zurich)\n\nFoundation models have achieved r
 emarkable accuracy in short- to medium-range weather forecasts, primarily 
 focusing on atmospheric variables. However, predicting new physical variab
 les typically requires training or fine-tuning the model with additional d
 atasets, incurring significant costs. We show that new variables can be le
 arned directly from the latent space of the Aurora foundation model. Our f
 rugal extension involves training lightweight decoders using a small datas
 et, specifically a subset of ERA5 and MSWEP. These decoders accurately pre
 dict surface variables related to the water cycle, establishing the first 
 baseline for many of these variables. For precipitation, our decoder achie
 ves results comparable to those in the literature. This work presents an a
 ffordable method to extend foundation models beyond atmospheric prediction
 s. It also suggests that Aurora captures an internal representation of the
  Earth system, contributing to a better definition and understanding of fo
 undation models.\n\nDomain: Climate, Weather, and Earth Sciences, Computat
 ional Methods and Applied Mathematics\n\nSession Chairs: Christian Lessig 
 (ECMWF, Otto-von-Guericke-Universitat Magdeburg) and Ilaria Luise (CERN)\n
 \n
END:VEVENT
END:VCALENDAR
