BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A Highly Efficient Machine Learning-Based Ozone Parameterization f
 or Climate Sensitivity Simulations - Yiling Ma\, Karlsruhe Institute of Te
 chnology (KIT)\, Germany
DTSTART:20251104T110000Z
DTEND:20251104T120000Z
UID:TALK237430@talks.cam.ac.uk
CONTACT:Yao Ge
DESCRIPTION:Biography: Yiling Ma is a PhD student at Institute of Meteorol
 ogy and Climate Research (IMK)\, Karlsruhe Institute of Technology (KIT) s
 ince 2023\, mainly working on developing hybrid approach that integrate ma
 chine learning with physical climate models to improve ozone modeling\, pa
 rticularly in the context of a changing climate. She holds a BSc in Atmosp
 heric Science and a MSc in Climate Dynamics (2016-2023). Her research inte
 rests involve climate change\, machine learning application in climate sci
 ence\, atmospheric chemistry modelling\, ocean-atmosphere interaction. \n
 \nAbstract: Atmospheric ozone is a crucial absorber of solar radiation and
  an important greenhouse gas. However\, most climate models participating 
 in the Coupled Model Intercomparison Project (CMIP) still lack an interact
 ive representation of ozone due to the high computational costs of atmosph
 eric chemistry schemes. In this talk\, I will present a machine learning p
 arameterization (mloz) to interactively model daily ozone variability and 
 trends across the troposphere and stratosphere in standard climate sensiti
 vity simulations\, including two-way interactions of ozone with the Quasi-
 Biennial Oscillation. We demonstrate its high fidelity on decadal timescal
 es and its flexible use online across two different climate models -- the 
 UK Earth System Model (UKESM) and the German ICOsahedral Nonhydrostatic (I
 CON) model. With atmospheric temperature profile information as the only i
 nput\, mloz produces stable ozone predictions around 31 times faster than 
 the chemistry scheme in UKESM\, contributing less than 4% of the respectiv
 e total climate model runtimes. In particular\, we also demonstrate its tr
 ansferability to different climate models without chemistry schemes by tra
 nsferring the parameterization from UKESM to ICON. This highlights mloz’
 s potential for widespread adoption in CMIP-level climate models that lack
  interactive chemistry for future climate change assessments\, particularl
 y when focusing on climate sensitivity simulations\, where ozone trends an
 d variability are known to significantly modulate atmospheric feedback pro
 cesses.
LOCATION:Chemistry Dept\, Unilever Lecture Theatre and Teams
END:VEVENT
END:VCALENDAR