Transcript High-resolution studies in NERC

High Resolution Climate Modelling
in NERC (and the Met Office)
Len Shaffrey, University of Reading
Thanks to: Pier Luigi Vidale, Jane Strachan,
Dave Stevens, Ian Stevens, Marie-Estelle
Demory, Julia Slingo, Malcolm Roberts, Jen
Catto, Kevin Hodges, Adrian New
High Resolution Climate Modelling

HiGEM is based on the Met Office Hadley Centre coupled ocean-atmosphere
climate model, HadGEM1
HiGEM was jointly developed
by the National Centre for
Atmospheric Science, the
University of East Anglia, the
National Oceanographic
Centre and the Met Office
HiGEM has an atmospheric
resolution of 1.25ox0.83o
longitude by latitude (90km),
and 38 levels in the vertical
The ocean resolution is 1/3ox1/3o, which allows the model to begin resolve
ocean eddies (an eddy-permitting resolution). The ocean model has 40 levels in
the vertical
Centennial length runs have been integrated
A 60km atmosphere-only model (NUGAM) has also been developed.
Does increasing the resolution improve the representation of regional climate
and weather?





ENSO in HiGEM
DJF El Nino SST
Composites
Observations
HiGEM
HadGEM
Shaffrey et al.
2009, J. Clim.
Tropical Instability Waves
Instantaneous SSTs (lines) and surface windstress divergence (colours) from HadGEM1 (left)
and from HiGEM (right) over the Tropical Pacific (Shaffrey et al. 2009, J. Clim.)
The eddy heat transport convergence from Tropical Instability Waves warm the Tropical
Pacific cold tongue, improving the mean state (Roberts et al. 2009, J. Clim.).
Extratropical cyclones in HiGEM
• Global climate models generate their own weather.
This makes it difficult to evaluate weather systems in
climate models
• An emerging methodology is to evaluate the structure
of extratropical cyclones using a compositing technique
• Results will be shown from the HiGEM high-resolution
global coupled climate model and the ERA-40 ECMWF
global reanalysis. These have very similar horizontal
resolutions (approx. 90km)
•Catto, J. et al. 2010, Journal of Climate
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
First work out the
tracks of the strongest
storms using the
850hPa vorticity and
Kevin Hodges feature
tracking program
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
Then work out where
the storms reach
maximum strength (in
this case maximum
850hPa vorticity)
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
We then select the
region around the
these points..
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
We then select the
region around the
these points…and note
which way they are
moving
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
We then select the
region around the
these points…and note
which way they are
moving
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
Since the storms are all
propagating in different
directions we rotate the
structures …
Creating composite
structures for extreme
storms in HiGEM and
ERA-40
Since the storms are all
propagating in different
directions we rotate the
structures …and then
average.
+
+
=
Storm Structure
A
A
Propagation
ERA-40
B
HiGEM
B
Composites of system relative 925hPa wind speed (ms-1) from the 100 most
intense NH wintertime storms
Storm Structure
ERA-40
HiGEM
Composites of system relative winds and potential temperatures from the 100
most intense NH wintertime storms
Storm Structure
CCB
CCB
ERA-40
CCB = Cold Conveyor Belt
HiGEM
WCB = Warm Conveyor Belt
Composites of system relative winds and potential temperatures from the 100
most intense NH wintertime storms
Location of the simulated tropical storms
Observed
Track densities (storm transits/month) Northern Hemisphere
High resolution
Low resolution
Modelled
Courtesy of P-L. Vidale and Jane Strachan
Intensity of simulated tropical storms
Wind Speed near the surface
Frequency
Resolution
V.Low
Low
Mid
High
(300km)
(150km)
(90km)
(60km)
Although high-resolution climate models are beginning to capture the
geographical location of tropical storms, at these resolutions, we can’t
capture intensity.
Courtesy of P-L. Vidale and Jane Strachan
Summary
•Increasing resolution reduces some biases, and improves the
representation of ENSO, blocking, NH stationary wave patterns and
stratocumulus decks (Shaffrey et al. 2009, J. Clim.)
•Higher resolution climate models are starting to resolve structures in
weather systems, for example intense extratropical cyclones (Catto
et al., 2010, J. Clim.). Can also simulate locations and number of
Tropical Cyclones (although not intensity).
•Increasing resolution is not a panacea. Major biases remain, e.g.
both HadGEM1 and HiGEM poorly represent the distribution of
Indian summer monsoon rainfall.
Future Directions
•The HiGEM and NUGAM models are being used extensively in a range of
projects, VOCALS, Willis RN, Met Office PACE, Queensland CCE, etc...
•Experimental decadal predictions using HiGEM are being produced for
CMIP5. The results of CMIP5 will be used to inform the next IPCC
assessment report (AR5)
•NERC is involved in the development of the next generation Met Office
high resolution coupled climate model, HadGEM3-H (Atmos: 60km; Ocean:
NEMO ORCA0.25 grid)
• Currently seeking funding from the NERC Storm Risk Mitigation
programme to study the impact of climate change on extratropical cyclones
• Consistent tracking diagnostics across the CMIP5 models
• Using the ECMWF Athena results, investigate the impact of climate
change on intense extratropical cyclones, and their upscale effects,
in very high-resolution global atmospheric models
A test: The 97/98 El Nino
A test of the system,
forecasting the
1997/1998 ENSO
from Dec 1996.
Timeseries of
Nino3.4 SST
anomalies from
observations (black),
the assimilation run
(thick red) and three
hindcasts (thin red).