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:20190719T085744Z
LOCATION:HG D 1.1
DTSTART;TZID=Europe/Stockholm:20190613T174500
DTEND;TZID=Europe/Stockholm:20190613T181500
UID:submissions.pasc-conference.org_PASC19_sess118_msa108@linklings.com
SUMMARY:Physics-Informed Generative Learning to Emulate Unresolved Physics
  in Climate Models
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Emerg
 ing Application Domains, Climate and Weather, Physics\n\nPhysics-Informed 
 Generative Learning to Emulate Unresolved Physics in Climate Models\n\nKas
 hinath, Wu, Albert, Prabhat\n\nSimulating Earth's climate often involves s
 olving nonlinear coupled PDEs with multi-scale physics that cannot be full
 y resolved and requires parameterizations for sub-grid scale phenomena. Th
 erefore, reliable and accurate models to parameterize unresolved and under
 -resolved physics, such as atmospheric convection, remain an important req
 uirement for simulating Earth's climate. Recently, Machine Learning has pr
 oven to be successful in many data-driven tasks, including in mimicking di
 stributions of processes in complex systems using a flavor of deep neural 
 networks called generative adversarial networks (GANs). GANs have also bee
 n designed to generate solutions of PDEs governing complex systems without
  having to numerically solve these PDEs, by using examples from hi-fidelit
 y simulations or experimental data as training data. We present a physics-
 informed GAN that incorporates constraints of conservation laws and certai
 n statistical properties of the system obtained from the training data. By
  training both standard GANs and physics-informed GANs as emulators of Ray
 leigh-Benard convection, we show that the physics-informed GAN is more rob
 ust and better captures high-order statistics of turbulent atmospheric con
 vection. This work has great potential as an alternative to the explicit m
 odeling of closures for unresolved physics or parameterizations, which acc
 ount for a major source of uncertainty when simulating turbulent flows and
  Earth’s climate.
END:VEVENT
END:VCALENDAR