-
Extreme Precipitation by High Resolution RegCM3Over East
AsiaJing ZHENG , Zhenghui Xie Institute of Atmospheric Physics
(IAP),CAS, ChinaXunqiang Bi the Abdus Salam International Centre
for Theretical Physics (ICTP), Italy
-
OutlineIntroductionModel & Exp.Results Discussion
-
OutlineIntroductionModel & Exp.Results Discussion
-
Giorgi & Bi, 2000:
The internal model variability exhibits a pronounced summer
maximum. It significantly influences the day-to-day model solution,
especially for summer precipitation.Also affected by the model
internal variability was the frequency of occurrence of heavy daily
precipitation events.
A regional climate model simulation is characterized by an
intrinsic level of internal variability which can be excited by any
type of perturbation and is regulated by synoptic conditions,
season, model domain, region of application, and specific
simulation period.
-
Giorgi & Bi, 2000:
The internal model variability exhibits a pronounced summer
maximum. It significantly influences the day-to-day model solution,
especially for summer precipitation.Also affected by the model
internal variability was the frequency of occurrence of heavy daily
precipitation events.
A regional climate model simulation is characterized by an
intrinsic level of internal variability which can be excited by any
type of perturbation and is regulated by synoptic conditions,
season, model domain, region of application, and specific
simulation period.
-
Resolution?Model? Its internal model variability?the extreme
precipitation,Yangzi and Huaihe Valley summer, 1998 compared to
60km resolution & observational data.the results indicated that
high resolution can give more details about the region of the
extreme precipitation. Additionally, the maximum of the extreme
precipitation were different by the two simulations of different
resolutions.Difficulties in simulating precipitation, esp. extreme
prec.20km resolution of RegCM3
-
OutlineIntroductionModel & Exp.Results Discussion
-
Summary of RegCM3 CoreDynamics:MM5 Hydrostatic (Grell et al
1994)Radiation:CCM3 (Kiehl 1996)Large-Scale Clouds &
Precipitaion:SUBEX (Pal et al 2000)Cumulus convection:Anthes-Kuo
(1977)Grell (1993)Emanuel (1991)Boundary Layer:Holtslag (1990)
Tracers/Aerosols/dust:Qian et al (2001); Solmon et al (2005);
Zakey et al. (2006)Land Surface:BATS (Dickinson et al 1993)SUB-BATS
(Giorgi et al 2003)Ocean FluxesBATS (Dickinson et al 1993)Zeng et
al (1998)ComputationsParallel Code (Bi, Gao, Yeh)Multiple
PlatformsMore User-Friendly Code(Pal et al 2006; Since Giorgi et al
1993ab)ICTP
-
Experiment Design Region East ChinaPeriod 1998. Jun. ~ Aug.
(June: Spinup time)Resolution 2.5o60km20km (80*100, (32N,112E))ICBC
Experiment ERA40.vs.NNRP2Cumulus Scheme Experiment (ERA40)Domain
Size experiment (NNRP2, icup=4)Buffer zone experiment (ERA40,
icup=4)
-
60km20km
-
ICBC -- ERA40 .vs. NNRP2ERA40NNRP2?
-
OutlineIntroductionModel & Exp.ResultsDiscussion
-
Time Series for PrecipitationNNRP2 vs ERA40 (icup=4) 60km
20kmERA40(icup=1,2,4)60km20km4 observational stations: Wuhan,
Hefei, Nanjing, Shanghai
-
ERA[i1]ERA[i2]ERA[i4]NNRP[i4]EnsembleObs.
-
ERA[i1]ERA[i2]ERA[i4]NNRP[i4]EnsembleObs.
-
ERA[i1]ERA[i2]ERA[i4]NNRP[i4]EnsembleObs.
-
ERA[i1]ERA[i2]ERA[i4]NNRP[i4]EnsembleObs.
-
Domain Size Experiment (NNRP2,i4)LargeSmallObs
-
Buffer Zone Experiment (ERA40,i4)BZ12BZ18 Obs
-
GPCPRegCM3CMAPCRUMet. stationsSpatial pattern
-
Spatial pattern ERA40_i4NNRP2_i4NNRP2_SERA40_BZtpr .vs. prcv
July1998
-
Spatial pattern ERA40_i4NNRP2_i4NNRP2_SERA40_BZtpr .vs. prcv
Aug1998
-
OutlineIntroductionModel & Exp.Results Discussion
-
Preliminary conclusion & Discussionsimulations were
different by different resolutions/ domain sizes/ cumulus schemes/
forcing fields.Resolution: higher resolution can give more details
about the region of the extreme precipitation. Domain: it depends,
but smaller seems much higher peaks.Cumulus schemes: MIT-Emanuel
seems better here.Forcing fields: ERA40 southward rain band [vs.
NNRP2 ]Difficulties in simulating precipitation, esp. extreme
prec.Resolution & Precipitation?Cumulus scheme? Domain size?
Forcing fields?Internal model variability?Land Component of the
Model? More accurate physical process?A CASE STUDY ----
-
Thanks for Your Attention!
Small Domain size seems much higher peaks than large
domain.()ERA40NNRP