Transcript present ability and future plans - wcrp strategic framework 2005-2015

Regional/local climate projections:
present ability and future plans
Richard Jones: WCRP workshop on regional climate,
Lille, 14-16 June 2010
© Crown copyright Met Office
Research funded by
Outline – version 1
• Regional climate projections and predictions:
present ability
• Detailed (spatially and temporally) climate
projections: present ability
• Regional and detailed climate projections and
predictions: future plans
• Expected outcomes of these plans and the
resulting implications
© Crown copyright Met Office
Outline – version 2
• What we can say and why
• What we cannot say and why
• What we are planning next
• What will this (not) deliver
© Crown copyright Met Office
Regional climate projections and
predictions: present ability
Large-area “average” indications of ranges of plausible (and
in some cases “likely”, i.e. predicted) climate changes for
all regions via e.g. CMIP3, other (ensemble) global models
• IPCC AR4 states that there will likely (or very likely) be :
• seasonal temperature increases in all regions
• sea-level rises in all regions
• seasonal precipitation changes (either increased or
decreases) in many regions
© Crown copyright Met Office
Observed, simulated and projected
temperature ranges with human
and natural forcings
Significant “predicted” regional temperatures rises, “likely” ranges
given
IPCC
© Crownby
copyright
Met Officebased on model responses to observed forcing etc.
Regional sea-level rise - China
Global to regional sea level rise
ratio in 11 CMIP3 models
© Crown copyright Met Office
Temperature/precipitation changes:
Asia, A1B emissions, 2090s, CMIP3
Significant highlatitude precip.
increase via
moister
atmosphere;
West Asian dries
via warmingdriven lower
relative humidity
and surface
drying in spring
© Crown copyright Met Office
Fig
Regional climate projections –
what we can say and why
The following broad overview of likely sub-continental changes
(i.e. a range of values in which the observed change is
expected to lie), i.e.:
• seasonal temperature increases in all regions – we
understand the processes and models respond correctly to
observed forcings
• sea-level rises in all regions – likely range for global average
from observations and process understanding with regional
wider due to model-dependent variability in patterns
• seasonal precipitation in many regions – confident where we
understand the processes and when the dominant ones are
driven by warming
© Crown copyright Met Office
Detailed climate projections:
present ability and limiting factors
Detailed temperature changes (including extremes) can
often be inferred from the large-scale changes
Similarly, some detailed precipitation changes can be
inferred but sampling issues often limit the information
Various factors, including climate variability and processes
that are poorly understood or not represented, act to limit
the information that can be obtained
© Crown copyright Met Office
Changes in Caribbean islands in
global and regional models
Pattern of change
over the seasurface is very
different using
different GCMs
Temperature
changes over
the islands are
larger, more
consistent and
we are confident
of the processes
involved
© Crown copyright Met Office
ECHAM4- A2
HADCM3 – A2
ECHAM4 – B2
HADCM3 – B2
The influence of internal variability
Run 1, winter
Changes in average intensity of top
5% of wet days, SRES A2, 2080s
forecast with 3-member ensemble
Run 2, winter
Run 1, summer
Run 2, summer
• Many areas of significant change
in each ensemble member
• Key point is not where change is
significant but reliable
Run 3, winter
Run 3, summer
• Changes similar on large-scales
• Locally considerable differences
between 3 realisations
© Crown copyright Met Office
-80 -40 -20 -10 -5 5 10 20 40 80
Change (%)
Use signal to noise analysis for
robust changes Mean, winter Mean, summer
• Can discern significant changes
over much of Europe in winter
and parts of Europe in summer
Top 5%, summer
Top 5%, winter
• Changes quantified with
reasonable accuracy where
|SNR|>, i.e. only in limited areas
• Robust change limited even in
mean with 3-member ensemble
and is less for extremes
Top 1%, winter
Top 1%, summer
Kennett, E. J., D. P. Rowell, R. G. Jones, and E. Buonomo,
2008: Robustness of future changes in local precipitation
extremes. J. Climate, doi: 10.1175/2008JCLI2082.1.
© Crown copyright Met Office
-10 -5 -2 -1
0
1
2
5
10
Climate variability and information
available from a climate projection
Single 30 year integrations can be insufficient to infer
detailed changes in precipitation
With 3-member ensemble sampling top 5%, changes
at grid box level discernible over much of Europe in
winter, but less than half of Europe in summer
Changes quantified only in some northern European
locations
More than 3x30yr ensemble integrations needed for
significant local changes over large parts of
Mediterranean in winter and central and eastern
Europe in summer
© Crown copyright Met Office
Future precipitation: a regional
example – the Indian monsoon
Changes over
India are
mostly large
and positive,
supported by
physical
insight
© Crown copyright Met Office
Top 5% precipitation change drivers
– 3 member ensemble experiments
Percentage change in extreme precipitation for each mechanism
Warming
Winter
Summer
© Crown copyright Met Office
Large-scale
Soil moisture
Total change
Some information on changes in
top 5% daily precipitation intensity
Increases across much of Europe in winter are
dominated by increasing atmospheric moisture in a
warmer climate
The prediction of increased extreme precipitation for
Europe as a whole in winter is reliable
At the regional scale (~1000 km or less) an enhanced
increase or a decrease due to changing circulation
patterns is possible.
In summer there are competing processes which tend to
respectively enhance and reduce precipitation
© Crown copyright Met Office
Factors influencing ranges of local
precipitation projections
Contributions to the ranges in
projected winter precipitation
changes for 2070-99 relative
to 1961-90 at selected 25km
grid squares in an ensemble
global/regional projection.
Internal variability, carbon
cycle, downscaling, account
for ~50%, i.e. not accounting
for these reduces range of
change around projected or
likely value by factor of 2
© Crown copyright Met Office
Detailed climate projections –
what we cannot say and why
Regional patterns of changes, even in the mean, for a given
decade or thirty-year period under transient climate
change are affected by sampling (only small areas of
change in precipitation quantified with a 3x30y sample) –
and the situation is worse for extremes
Spatially and temporally detailed changes are often unclear
because they results from:
(a) processes we are uncertain about, e.g. large-scale
circulation changes,
(b) a combination of processes which we cannot quantify or do
not represent
© Crown copyright Met Office
What we are planning next
CMIP5 will deliver, for centennial projections, an
increase in the processes simulated and more
models with higher resolution
CMIP5 will deliver ensemble decadal projections
initialised from the observed state of the system
including 30-year projections from 2005
CORDEX is providing a framework for, and will deliver
for Africa, coordinated downscaling of the CMIP5
centennial projections using the rcp4.5 and rcp8.5
concentration scenarios
© Crown copyright Met Office
CMIP5 Experiments
© Crown copyright Met Office
What this will (not) deliver
A broad picture of sub-continental changes likely to lie in given
ranges (similar to those in AR4) from the CMIP5 GCMs
This information on changes will include more spatial detail and
a greater range of processes and thus provide a more
realistic range of possible outcomes around the likely value
The CMIP5 30-year initialised ensembles should provide
reasonable sampling of patterns of near-term projected
climate change where downscaling would not add information
CORDEX will further enhance the realism of ranges around
likely centennial climate changes for Africa
Little additional information will be provided on other aspects of
detailed climate change
© Crown copyright Met Office
Conclusions
To provide information on detailed patterns of even mean climate
change for periods under transient climate change requires
sample-sizes that can only be derived from ensembles
Even with ensemble sampling, uncertain aspects of climate
change such as possible circulation change or the magnitude
of competing processes (even if we have confidence in their
sign of change) will limit what we can say with confidence on
detailed climate change
AR5 is likely to improve on the picture of area-averaged regional
climate change providing some increases in spatial detail and
more realistic range of possible outcomes around the likely
value (from including more processes) and, for the near term,
ensemble 30-year initialised projections – but little will be
provided on other aspects of detailed climate change
© Crown copyright Met Office