seaclid / cordex southeast asia

SEACLID / CORDEX Southeast AsiaGemma T. Narisma, Manila Observatory

On behalf of the SEA Regional Climate Initiative (SEARCI)

Coordinator: Fredolin Tangang (The National University of Malaysia)Indonesia: Edvin Aldrian, Dodo Gunawan (BMKG)Malaysia: Fredolin Tangang, Liew Juneng (UKM)Philippines: Gemma Narisma, Faye Cruz (Manila Observatory)Thailand: Jerasorn Santisirisomboon (Ramkhamhaeng University)

Patama Singhruck (Chulalongkorn University)Vietnam: Phan Van Tan, Thanh Ngo-Duc (NVU Hanoi University of Science)

YMC Workshop, 28-30 Jan 2015, CCRS Singapore

Southeast Asia Regional Climate Initiative (SEARCI), August 2012

• Scientists from Vietnam, Malaysia, the Philippines, Indonesia and Thailand in workshop hosted by the VNU Hanoi University of Science

• Motivation: To have a platform for regional collaboration on climate related research for SEA and build the capacity of the region in regional climate science

South Asia

East Asia

CORDEX domainsCORDEX domains

CORDEX Southeast Asia

Nov 2012

Jun 14,2013

Initial correspondence with WCRP (Michel Rixen) and CORDEX (Colin Jones)

Formal invitation by Ghassem Asrar (WCRP) for SEACLID to join the CORDEX network

• SEACLID activities streamlined to CORDEX• Possible contribution of additional simulations by CORDEX

affiliated centers over SEACLID domain (e.g. India, Australia, UK)• Capacity development and training for SEACLID in coordination

with CORDEX-Asia

May 2013 APN proposal approved for funding under the ARCP Programme for 3 years beginning October 2013

Pledged Commitments in SEACLID / CORDEX SEA

Country GCM Institution & Country developed the GCM

RCP RCM

Vietnam CNRM-CM5 Centre national de Recherches Meteorologiques, France RCP8.5, 4.5 RegCM4

Philippines HadGEM2 Hadley Centre, UK RCP8.5, 4.5 RegCM4

Thailand MPI-ESM-MR Max Planck Institute for Meteorology, Germany RCP8.5, 4.5 RegCM4

Thailand EC-Earth EC-Earth consortium RCP8.5, 4.5 RegCM4

Indonesia CSIRO MK3.6 CSIRO, Australia RCP8.5, 4.5 RegCM4

Malaysia CanESM2 Canadian Centre for Climate Modeling and Analysis, Canada RCP8.5, 4.5 RegCM4

Malaysia IPSL-CM5A-LR Institute Pierre-Simon Laplace, France RCP8.5, 4.5 RegCM4

Malaysia GFDL-ESM2M GFDL, USA RCP8.5, 4.5 RegCM4

Australia CNRM-CM5 Centre national de Recherches Meteorologiques, France RCP8.5 CCAM

Australia CCSM4 NCAR, USA RCP8.5 CCAM

Australia ACCESS1.3 CSIRO, Australia CCAM

Hong Kong SAR CCSM or CESM NCAR, USA WRF

United Kingdom HadGEM2-ES Hadley Centre, UKMO PRECIS

South Korea HadGEM2-AO Hadley Centre, UKMO WRF

RegCM4 Experiments Setup• Domain: 36 km, 81.14°E-143.86°E; 15.04°S ~ 39.84°N• PBL: Holtslag (1990)• Radiation: CCSM • Large scale moisture: SUBEX (Pal et al. 20)• Land-surface treatment: BATSe• Cumulus parameterization:

- Grell / Arakawa-Schubert (closure)- MIT Emanual- MIT (O) / Grell (L)- Grell (O) / MIT (L)- Grell / Fritch-Chappell (closure)- Kuo

• Ocean flux treatment: - BATSe- Zeng (iocnrough=1)- Zeng (iocnrough=2)

• Lateral boundary conditions: ERA Interim• Run length: 1989 – 2008 (20 years)

18 Simulations (100% completed)

DJF Grell AS MIT Eman Eman(l) Grell (o) Grell FC Grell (l) Eman (o) Kuo

OCEAN FLUX

Model enhances the cold bias

DJF Grell AS MIT Eman Eman(l) Grell (o) Grell FC Grell (l) Eman (o) Kuo

High correlation for north half of domainPoor correlation in south half of domain

APHRODITE data: – daily rainfall (V1003R1), 1951-2007– 0.25º– Monsoon Asia (60E-150E, 15S-55N)

http://www.chikyu.ac.jp/precip/products/index.html

Slide courtesy of Thanh Ngo-Duc, Vietnam

Annual Precipitation Biases (vs CRU)

Too wet !!!

Still too wet over the equatorial regions!

Slide courtesy of Juneng Liew, Malaysia

…cont.

Slide courtesy of Juneng Liew, Malaysia

Validation Issue

• Variations among observational products can be large.• Compare with all the available datasets.

Rainfall (Highest – Lowest)

Observation datasets - APHRODITE, CRU, GPCC, TRMM (1999 onward)

Slide courtesy of Juneng Liew, Malaysia

Seasonal Precipitation Spatial Comparison

• Correlation – ~0.5-0.7 • Inter-model variations higher during the winter season.

DJF JJA

Slide courtesy of Juneng Liew, Malaysia

Key questions:

• What are the key processes and mechanisms in the MC that are not captured by the regional climate model, resulting to poor model performance and inability to capture seasonal dynamics?

• What are the appropriate adjustments and modifications to cumulus parameterization schemes for the model to properly simulate rainfall and rainfall dynamics?

• How do variations in SSTs affect the simulated model climatology in the MC?

• Does the model adequately capture the diurnal cycle, Madden-Julian Oscillation (MJO), and the extreme rainfall events associated with MJO and ENSO events and are the associated physical dynamics simulated well?

Coordinator: Fredolin Tangang (The National University of Malaysia)Indonesia: Edvin Aldrian, Dodo Gunawan (BMKG)Malaysia: Fredolin Tangang, Liew Juneng (UKM)Philippines: Gemma Narisma, Faye Cruz (Manila Observatory)Thailand: Jerasorn Santisirisomboon (Ramkhamhaeng University)

Patama Singhruck (Chulalongkorn University)Vietnam: Phan Van Tan, Thanh Ngo-Duc (NVU Hanoi University of Science)

(http://www.ukm.edu.my/seaclid-cordex)

ERA Interim APHRODITE ERAIN-APHRO

D. P. Dee et al (2011)Yatagai et al. (2012) and Yasutomi et al. (2011)

Bias from the Boundary Condition

Biases outside the variability of APHRODITE

MIT

MIT(O)/Grell(L)

R1R1R2R2R3R3R4R4

R5R5 R6R6 R7R7R8R8

R9R9R10R10R11R11R12R12

R13R13R14R14

R15R15R16R16R17R17

R18R18R19R19R20R20

MIT(L)/Grell(O)

Grell/FC Kuo

PRECIPITATION Annual Cycles (vs 4 Obs.): Correlation Coefficient

Grell/FC

South

North

Slide courtesy of Juneng Liew, Malaysia

TEMPERATURE Correlation per region (aphrodite)

R1R1R2R2R3R3R4R4

R5R5 R6R6 R7R7R8R8

R9R9R10R10R11R11R12R12

R13R13R14R14

R15R15R16R16R17R17

R18R18R19R19R20R20

South

North

• WCRP sponsored exercise – a framework to downscale CMIP5 output.

• Providing a global coordination of Regional Climate Downscaling for improved regional climate change adaptation and impact assessment.

• Within AR5 timeline and beyond.

Model improves the cold bias of ERAIN

Cold bias esp during

cold months

Model warm bias

Cold bias: Jan-May

Warm bias: June-Dec

Warm biasCold bias

Summary Analysis: Seasonal Cycle, PDFs

CORRELATION

BIASMODEL SKILL

Where does MIT Emanuel perform best?