Imperial College London
Alternate

Professor Joanna D. Haigh

Faculty of Natural Sciences, Department of Physics

Distinguished Research Fellow
 
 
 
//

Contact

 

+44 (0)20 7594 7770j.haigh Website

 
 
//

Assistant

 

Mr Luke Kratzmann +44 (0)20 7594 7770

 
//

Location

 

Blackett LaboratorySouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@inproceedings{Nowack:2019:10.5065/y82j-f154,
author = {Nowack, P and Ong, QYE and Braesicke, P and Haigh, J and Abraham, L and Pyle, J and Voulgarakis, A},
doi = {10.5065/y82j-f154},
pages = {263--268},
publisher = {UCAR},
title = {Machine learning parameterizations for ozone: climate model transferability},
url = {http://dx.doi.org/10.5065/y82j-f154},
year = {2019}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - Many climate modeling studies have demon-strated  the  importance  of  two-way  interactions  betweenozone and atmospheric dynamics. However, atmosphericchemistry models needed for calculating changes in ozoneare  computationally  expensive.  Nowack  et  al.  [1]  high-lighted  the  potential  of  machine  learning-based  ozoneparameterizations in constant climate forcing simulations,with  ozone  being  predicted  as  a  function  of  the  atmo-spheric  temperature  state.  Here  we  investigate  the  roleof additional time-lagged temperature information underpreindustrial  forcing  conditions.  In  particular,  we  testif  the  use  of  Long  Short-Term  Memory  (LSTM)  neuralnetworks can significantly improve the predictive skill ofthe parameterization. We then introduce a novel workflowto  transfer  the  regression  model  to  the  new  UK  EarthSystem Model  (UKESM). For this, we  show for the  firsttime  how  machine  learning  parameterizations  could  betransferred  between  climate  models,  a  pivotal  step  tomaking  any  such  parameterization  widely  applicable  inclimate  science.  Our  results  imply  that  ozone  parame-terizations  could  have  much-extended  scope  as  they  arenot bound to individual climate models but, once trained,could be used in a number of different models. We hope tostimulate  similar  transferability  tests  regarding  machinelearning  parameterizations  developed  for  other  Earthsystem model components such as ocean eddy modeling,convection,  clouds,  or  carbon  cycle  schemes.
AU  - Nowack,P
AU  - Ong,QYE
AU  - Braesicke,P
AU  - Haigh,J
AU  - Abraham,L
AU  - Pyle,J
AU  - Voulgarakis,A
DO  - 10.5065/y82j-f154
EP  - 268
PB  - UCAR
PY  - 2019///
SP  - 263
TI  - Machine learning parameterizations for ozone: climate model transferability
UR  - http://dx.doi.org/10.5065/y82j-f154
UR  - http://hdl.handle.net/10044/1/75531
ER  -
Download