Performance evaluation and climate projections over Sub
Download
Report
Transcript Performance evaluation and climate projections over Sub
COSMO User Seminar 2012
Offenbach, 8 March 2012
Performance evaluation and climate projections
over Sub-Saharian Africa
with COSMO-CLM
Paola Mercogliano, CIRA and CMCC
Edoardo Bucchignani, CIRA and CMCC
Myriam Montesarchio, CMCC
Alessandra Zollo, CMCC
1. Outlook
• Introduction
•Short summary of activity shown at CLM Assembly 2011
•The CLUVA project
•The domains simulated
•List of simulations
•The West domain: validation and climate projections
•The Lower East domain: validation and climate proj.
• Conclusions
2. CLUVA – Climate change and urban vulnerability in Africa
Project Co-ordinator: AMRA, Center of Competence in the field of Analysis and
Monitoring of Environmental Risk, Italy
The project objective is to develop methods and knowledge to be applied to African cities, to
manage climate risks, to reduce vulnerabilities and to improve their coping capacity and
resilience towards climate changes. The project will explore the issues of climate change
vulnerability, resilience, risk management and adaptation in selected African cities with local
partners.
The aim is to set up methods and work out probabilistic scenarios of climate change
affected hazards having a resolution that fits for regional and urban systems (for the 5
selected cities) and related uncertainties. More detailed aims are:
•
To produce downscaled regional climate scenarios (IPCC scenarios: RCP4.5 and RCP8.5)
for selected African areas surrounding the African cities of interest, at high resolution
(about 8 km).
•
To produce very high resolution projection (about 1-2 km) for the climate of some African
cities using specific and accurate statistical techniques
3. Areas of interest for CLUVA 1950-2050
Spatial Resolution: 8 km
EAST Domains:
WEST Domain:
(18 W -15.17 E; 3.3 – 16.8 N)
465 x 190 grid points
U (34.4 – 42.9 E; 6.1N – 12.5N)
120 x 90 grid points
L (34.5 – 41.3 E; 11.8S – 2.1S)
95 x 135 grid points
St.Louis
(16.5 W, 16.03 N)
Ougadougou (1.55 W, 12.37 N)
Douala
(9.71 E, 4.045 N)
Addis Abeba (38.75 E, 9.02 N)
Dar es Salaam (39.27 E, 6.82 S)
4. Orography of the three areas
West
Upper
East
Lower
East
5. Details of the Numerical simulations
8 km resolution
•
•
•
•
•
•
Supercomputer used:
Cluster of 30 IBM P575 nodes (32 cores per node)
Driving data: CMCC-MED 80 km resolution
Model version: cosmo_090213_4.8_clm13
Time step: 40 sec.
Numerical scheme: Runge-Kutta 2-time level HE-VI integration
Validation: CRU data and observed datasets for the 5 cities.
6. List of simulations
•
•
Scenario A1B West domain 1970- 2065
Scenario RCP 4.5 Lower East domain 1950-2050
•
•
•
Scenario RCP 4.5 Upper East domain 1950-2050 (*)
Scenario RCP 4.5 West domain 1950-2050
Scenario RCP 8.5 Lower East domain 1950-2050
•
•
Scenario RCP 8.5 Upper East domain 1950-2050 (*)
Scenario RCP 8.5 West domain 1950-2050
(*): completed, but not yet post processed.
7. Mean temperature bias with CRU
1971-2000
DJF
Cold bias between -1
and -2 degrees.
In the north part, up to
-5
JJA
Hot bias between 2 and 3
degrees.
Higher values in the west
part.
(COSMO-CRU)
8. Mean precipitation (mm/month) bias with CRU
1971-2000
DJF
Underestimation of
about 25% in the
south coastal area.
Good agreement in the
other parts.
JJA
A strong wet bias is
registered in the south
coastal area.
Underestimation in other
parts.
9. Seasonal cycle of temperature
(COSMO vs Observations)
St. Louis
Max
Applied Bias correction
(Sperna et. Al 2010):
Tcorr T (Tobs T )
T
: 30-year daily average temperature
Mean
Maximum bias in April (20C)
10. Seasonal cycle of temperature
Douala
Max
Mean
(COSMO vs CRU)
Ouagadougou
11. T2m variation : future (2021-2050) vs past (1971-2000)
A1B
DJF
General increase of
temperature, up to 2.4o;
it is more evident in
winter.
JJA
12. T2m variation : future (2021-2050) vs past (1971-2000)
RCP4.5
DJF
Less evident increase of
temperature, especially
in summer.
In winter, significant
increase in the northern
part.
JJA
13. T2m variation: future (2021-2050) vs past (1971-2000)
RCP8.5
DJF
Larger
increase
of
temperature in summer
with respect the other
scenarios.
In winter, the increase is
evident only in the
northern part.
JJA
14. Precipitation variation: future (2021-2050) vs past (1971-2000)
A1B
DJF
There is a big
difference between
winter and summer.
In winter, there is a
slight decrease of
precipitation, while in
summer there is a
general increase with
some exceptions.
JJA
mm/month
15. Precipitation variation: future (2021-2050) vs past (1971-2000)
RCP4.5
There are differences
between winter and
summer.
In winter, there is a
general increase of
precipitation, while in
summer there is a
behavior similar to A1B
DJF
JJA
16. Precipitation variation: future (2021-2050) vs past (1971-2000)
RCP8.5
In winter, there is a
general increase of
precipitation, similar to
RCP4.5
DJF
In summer there is a
behavior similar to A1B
and RCP4.5
JJA
17. T2m trend (A1B vs RCP 4.5)
Ouagadougou
St.Louis
18. Precipitation trend (A1B vs RCP 4.5)
Ouagadougou
St.Louis
19. Mean temperature bias with CRU
(COSMO-CRU)
1971-2000
In winter, there is a
cold bias between -2
and -3 degrees.
In some parts, up to
-5
In summer , there
is a hot bias
between 1 and 2
degrees.
DJF
JJA
20. Seasonal cycle of temperature
(COSMO vs CRU)
Dar es Salaam
Max
In winter , there is
a cold bias
especially evident in
the maximum
values of daily
temperature.
Mean
21. T2m variation: future (2021-2050) vs past (1971-2000)
RCP4.5
In winter, two
different areas are
visible, but both
characterized by an
increase of
temperature.
In summer, a
general increase of
1.5o C is evident.
DJF
JJA
22. T2m variation: future (2021-2050) vs past (1971-2000)
RCP8.5
With this scenario,
the increase of
temperature is
more uniform and
evident with respect
to RCP4.5.
DJF
JJA
23. Precipitation variation: future (2021-2050) vs past (1971-2000)
RCP4.5
In winter, we
register reductions
in the central part
of the domain and
increases along the
coast.
In summer, a
general reduction is
evident.
DJF
JJA
24. Precipitation variation: future (2021-2050) vs past (1971-2000)
RCP8.5
In winter, similar
behavior as RCP4.5,
but less evident
increase along the
coast.
In summer, similar
behavior as RCP4.5.
DJF
JJA
25. Conclusions
•Numerical results related to the simulation of the climate of the west and the
east lower domain at high resolution have been shown.
•for the west domain: cold bias in winter, and hot bias in summer with
respect to CRU. Better agreement registered with observed data provided
by project partners. Quite good agreement of precipitations in winter, while
in summer there is a strong bias;
•for the east lower domain: the temperature is underestimated in winter
and overestimated in summer with respect to CRU; better agreement with
observed data, especially in summer;
•for the west domain: in the future, the temperature is projected to
increase, especially in winter with all the three scenarios; in winter the
precipitation is projected to slightly increase by RCP4.5 and 8.5; in summer
is projected to increase by all the scenarios.
•for the east lower domain: in the future, the temperature is projected to
increase by both the scenarios, especially in winter.
Thanks