QESM: The QUEST Earth System Model

1

QESM The QUEST Earth System Model _at_Reading
Manoj Joshi, Jonathan Gregory, Allan Spessa,
Annette Osprey _at_CEH Wallingford Eleanor Blyth,
Doug Clark _at_UEA Erik Buitenhuis, Corinne Le
Quéré _at_Leeds Ken Carslaw, Graham
Mann _at_Cambridge Luke Abraham, Olaf Morgenstern,
John Pyle _at_Lancaster Oliver Wild and many
others...
2

• QESM the QUEST Earth System Model
• An unprecedented attempt to simulate the
  direction and strength of a wide range of
  biogeochemical processes and feedbacks affecting
  the functioning of the Earth system under
  contemporary and likely future climate scenarios.
• Coordinated and managed by the NCAS-Climate and
  the Walker Institute. Its development builds upon
  the efforts of several exisiting modelling
  consortia currently engaged in the development of
  the climatic and biogeochemical models
• http//www.quest-esm.ac.uk/

3
Climate models have historically been comprised
of the following components atmospheric
dynamics/physics ocean dynamics/physics sea-ice la
nd surface processes

Atmospheric chemistry aerosols mostly offline
as climatologies
Atmospheric physics dynamics

Ocean biology
Land processes
Ocean and sea-ice
4
Earth system models are comprised of the
following components atmospheric
dynamics/physics ocean dynamics/physics sea-ice oc
ean biology atmospheric chemistry land surface
processes dynamic terrestrial vegetation

Atmospheric chemistry aerosols
Atmospheric physics dynamics

Ocean biology
Ocean biology
Dynamic vegetation
Land processes
Ocean and sea-ice
5
QUESM will be comprised of the following
components HadGEM3-A NEMO (Nucleus for European
Modelling of the Ocean) CICE (Community Ice
CodE) QPFT UKCA JULES (Joint UK Land Environment
Simulator) ED (Ecosystem Dynamics)

6

• Planned and ongoing work on QESM
• The QESM is being developed in parallel with the
  Met Office Hadley Centres HadGEM3 model (an
  updated version of HadGEM1/1a)
• There is a close relationship in timetable and
  components
• HadGEM3-A (the atmosphere component of HadGEM3)
  will be run at N96 or above QESM atmosphere is
  planned to run at N48
• HadGEM3-AO will (probably) have an ocean
  resolution of 1, while QESM will run at 2
• The coarser resolution enables greater
  flexibility, such as running with 60 atmospheric
  layers in the vertical, and running it on systems
  such as the QUEST cluster

7

• QESM progress so far
• Three major changes have been tested in the N48
  atmospheric model since the freezing of
  HadGEM2-OA in early 2007
• a parameterisation of surface-atmosphere exchange
  at coastal points (this is still under
  development)
• implementation of adaptive detrainment on shallow
  convection
• changes to the convective cloud/precipitation
  scheme when the cloud depth is small (lt3 km)
  (also under development)

8

• QESM progress so far
• A parameterisation of surface-atmosphere exchange
  at coastal points (this is still under
  development)
• This is an attempt to solve the problem of land
  drag dominating the momentum budget at coastal
  points, lowering windspeed, and lowering moisture
  exchange
• The surface-lowest model layer windshear is
  tweaked upwards over the ocean fraction, and
  reduced over the land fraction
• It's a bit ad-hoc, but it does dramatically
  improve moisture in the BL and rainfall over the
  Maritime continent (probably also by broadly
  representing sea breezes)
• Its presence also fixes the AMOC drift in the N48
  model, which is now 17-18 Sv in the latest
  integration

9

• QESM progress so far
• A parameterisation of surface-atmosphere exchange
  at coastal points (this is still under
  development)

10

• QESM progress so far
• A parameterisation of surface-atmosphere exchange
  at coastal points (this is still under
  development)

Reduced these tendencies
11

• QESM progress so far
• implementation of adaptive detrainment on shallow
  convection
• changes to the convective cloud/precipitation
  scheme when the cloud depth is small (lt3 km)
• both of these improve the rainfall field,
  especially the monsoon

12

• QESM progress so far
• implementation of adaptive detrainment on shallow
  convection, changes to the convective
  cloud/precipitation scheme when the cloud depth
  is small atmos-only results

13

• QESM progress so far
• In coupled mode, there are still problems, but
  the N48 AO model with the latest fixes displays
  performance that is comparable (ish) with
  HiGEM/HadGEM, e.g. JJA rainfall

14

• QESM progress so far
• e.g. NINO3 timeseries

15

• QESM progress so far
• e.g. ENSO periodicity

16

• QESM progress so far and future plans
• Progress with the N48 model
• viable! (remember this wasn't the case earlier
  with the N48 model's "P minus E" problems)
• N48 HadGEM2-A is 4.5x as quick as N96 HadGEM2-A
• There are still ongoing issues similar to
  HadGEM/HiGEM
• Future work
• We're soon moving from HadGEM2 to HadGEM3
• a different ocean means different coupled
  problems, so there's not much point in more
  tweaking

17

• QESM progress so far and future plans
• Progress with the N48 model
• viable! (remember this wasn't the case earlier
  with the N48 model's "P minus E" problems)
• N48 HadGEM2-A is 4.5x as quick as N96 HadGEM2-A
• There are still ongoing issues similar to
  HadGEM/HiGEM
• Future work
• We're soon moving from HadGEM2 to HadGEM3
• a different ocean means different coupled
  problems, so there's not much point in more
  tweaking ...for the moment
• Keeping in contact with the Met O's HadGEM3
  development

18

• QESM progress so far and implications for HiGEM
• Perhaps the HadGEM3 implementation of the coastal
  mod can be tested in HiGEM- might have an effect
  on the Maritime Continent and the SW
  Pacific/Andes region (c.f HadGEM1a/WAN)
• Adaptive detrainment/entrainment has had
  beneficial effects on everything from the N48
  model to the global NWP model the river outflow
  scheme presently in HadGEM2 has had beneficial
  effects
• The convective cloud changes would probably
  improve HiGEM performance
• As for other coupled biases- sorry, not really my
  job anymore
• but that doesn't mean I haven't got one or two
  ideas

19

Thank you for your attention
20

21