(Created page with "==Abstract== The climate system models from Beijing Climate Center, BCC_CSM1.1 and BCC_CSM1.1-M, are used to carry out most of the CMIP5 experiments. This study gives a gener...")
 
m (Scipediacontent moved page Draft Content 747869469 to Xiao-Ge et al 2015a)
(No difference)

Revision as of 11:18, 15 May 2017

Abstract

The climate system models from Beijing Climate Center, BCC_CSM1.1 and BCC_CSM1.1-M, are used to carry out most of the CMIP5 experiments. This study gives a general introduction of these two models, and provides main information on the experiments including the experiment purpose, design, and the external forcings. The transient climate responses to the CO2  concentration increase at 1% per year are presented in the simulation of the two models. The BCC_CSM1.1-M result is closer to the CMIP5 multiple models ensemble. The two models perform well in simulating the historical evolution of the surface air temperature, globally and averaged for China. Both models overestimate the global warming and underestimate the warming over China in the 20th century. With higher horizontal resolution, the BCC_CSM1.1-M has a better capability in reproducing the annual evolution of surface air temperature over China.

Citation

Xin, X.-G., T.-W. Wu, and J. Zhang, 2013: Introduction of CMIP5 experiments carried out with the climate system models of Beijing Climate Center. Adv. Clim. Change Res.,4 (1), doi: 10.3724/SP.J.1248.2013.041.

Keywords

CMIP5 ; climate system model ; experiment ; BCC_CSM

1. Introduction and objective

A climate system model is an important tool to study the mechanisms for past climate change and to project future climate change. The Couple Model Intercomparison Project (CMIP) organized by Working Group on Couple Modeling (WGCM) serves as an important platform for the evaluation of climate models and the promotion of further development of climate models. Experimental data of the models participating in CMIP has been widely used in climate research. After four phases of CMIP, the fifth phase (CMIP5) started in September, 2008. A new set of climate model experiments is prescribed aiming to address outstanding questions that arose in the IPCC AR4 assessment. As in previous phases of CMIP, results from this new set of simulations will provide valuable scientific information for IPCC AR5 (scheduled to be published in 2014).

The CMIP5 simulations have being an important mission of climate modeling groups worldwide since the release of the CMIP5 experimental design [ Taylor et al., 2011 ]. Great efforts have been made by Beijing Climate Center (BCC) in preparing the forcing data, adding the required output variables, performing the experiments and standardizing the output data. The model outputs are available on the PCMDI website (http://pcmdi9.llnl.gov ).

In comparison with CMIP3, some new experiments were added in the CMIP5, including decadal predictions (hindcasts and projections), coupled carbon/climate model simulations, and several diagnostic experiments for understanding the long-term simulations. The number of the experiments in CMIP5 is much larger than in CMIP3. Some of the experiments are inter-connected with each other. The aim of this study is to introduce the experiments and provide main information on the simulations carried out at the BCC. We believe the introduction will provide useful information for researchers on how CMIP5 experiments are performed at the BCC and promote the wide application of the experimental dataset.

2. Models and forcing data

There are two versions of climate system models of BCC, BCC_CSM1.1 and BCC_CSM1.1-M participating in CMIP5. Both models are fully coupled global climate-carbon models including interactive vegetation and the global carbon cycle. The only difference between the two models is the atmospheric module. The atmospheric module in BCC_CSM1.1 is BCC_AGCM2.1, while it is BCC_AGCM2.2 in BCC_CSM1.1-M, with a horizontal resolution of about 2.8° (T42) and about 1° (T106), respectively. There are 26 levels in vertical direction in both AGCMs with the same dynamical and physical processes. The description and performance of BCC_AGCM2.1 can be found in Wu et al. [ Wu et al., 2008  and Wu et al., 2010 ] and Wu [2012] .

BCC_CSM1.1 and BCC_CSM1.1-M use the same ocean component (MOM4_L40), land component (BCC_AVIM1.0), and sea ice component (Sea ICE Simulator, SIS). BCC_AVIM1.0 includes biogeophysical, ecophysiological and soil carbon-nitrogen dynamical modules [ Ji et al., 2008 ]. The ocean model and sea ice model are both from the Geophysical Fluid Dynamics Laboratory (GFDL). The horizontal resolution in the ocean model is 1° longitude by 0.33° latitude with tripolar grid [ Griffies et al., 2005 ]. There are 40 levels in the vertical direction in the ocean component. The biogeochemistry module to simulate the ocean carbon cycle in MOM4_ L40 is based on the protocols from the Ocean Carbon Cycle Model Intercomparison Project-Phase 2 (OCMIP2,http://www.ipsl.jussieu.fr/OCMIP/phase2/ ). SIS has the same horizontal resolution as MOM4_L40, with three layers in the vertical direction, one snow cover layer and two equally sized sea ice layers [ Winton, 2000 ].

The external forcing in the experiments is comprised of greenhouse gases (GHGs), ozone, aerosols, volcanic cruptions, carbon emissions, and solar variability. The GHGs include CO2 , N2 O, CH4 , CFC11  and CFC12 . The aerosol properties consist of sulfate aerosols, sea salt, black carbon, organic carbon, and soil dust. The temporal resolution of the sulfate aerosols dataset is 10 years. Only direct effects of aerosols are considered in the models. The time interval is one year for GHGs, the solar constant and carbon emissions. These forcing data except the volcanic aerosols are all provided by CMIP5. The volcanic dataset is from Ammann et al. [2003] . This dataset uses the aerosol optical depths to describe the volcanic activity, which was used by the NCAR CCSM3 and CCSM4 in the historical simulation of CMIP [ Meehl et al., 2006  and Meehl et al., 2012 ]. The ozone data and volcanic aerosols data are both in monthly intervals.

3. CMIP5 experiments

The CMIP5 experiments carried out with the BCC models are classified into three categories: longterm climate simulations, climate simulations with coupled carbon/climate models, and decadal prediction experiments. The atmosphere-only experiments are not introduced here, though most of the experiments were carried out.

3.1. Long-term climate simulations

The long-term climate simulations include preindustrial experiments, historical simulation, future climate projections, climate attribution experiments, paleoclimate experiments and idealized CO2 experiments. Main information of each experiment is shown in Table 1 . The short name in the table is the standard abbreviation within the CMIP5 experiments . The model version T42 denotes the BCC_CSM1.1 and T106 denotes the BCC_CSM1.1-M.

Table 1. Long-term climate simulations in CMIP5
Short name Experiment name Model version Ensemble size Time period Forcing fields
piControl Pre-industrial control T42,T106 1 500 years N/A
historical Historical T42,T106 3 1850–2012 GHG,SD,Oz,Sl,Vl,SS,Ds,BC,OC
rcp85 RCP8.5 T42,T106 1 2006–2300 GHG,SD,Oz,Sl,SS,Ds,BC,OC
rcp45 RCP4.5 T42,T106 1 2006–2300 GHG,SD,Oz,Sl,SS,Ds,BC,OC
rcp26 RCP2.6 T42,T106 1 2006–2300 GHG,SD,Oz,Sl,SS,Ds,BC,OC
rcp60 RCP6 T42,T106 1 2006–2099 GHG,SD,Oz,Sl,SS,Ds,BC,OC
historicalGhg GHG-only T42 1 1850–2012 GHG
historicalNat Natural-only T42 1 1850–2012 Sl,Vl
midHolocene Mid-Holocene T42 1 100 years N/A
past1000 Last millennium T42 1 850–1850 GHG,Vl,Sl
1pctCO2 CO2 increase by 1% per year T42,T106 1 140 years CO2
abrupt4xCO2 Abrupt 4xCO2 T42,T106 1 150 years CO2

Forcing abbreviation: N/A, all forcing fixed; SD, direct effect of sulfate aerosol; SS, sea salt; BC, black carbon; OC, organic carbon; Ds, dust; Oz, ozone; Sl, solar irradiance; Vl, volcanic aerosol

In the pre-industrial (piControl) experiment, the GHGs, aerosols, ozone and solar irradiance are fixed at the year 1850. An output of 500 years is supplied after 100 years’ spin-up. This simulation serves as the baseline and provides initial conditions for the historical simulation, the paleoclimate experiments, and the idealized CO2  experiments. Results of this experiment are used to estimate unforced variability of the model and to diagnose the climate drift in an unforced system.

The historical simulation is equivalent to the 20th century simulation (20C3M) of the CMIP3. The experiment starts from the piControl run and integrates data from 1850 to 2012 with the external forcing changing with time. The external forcing includes GHGs, the solar constant, volcanic activity, ozone and aerosols. The forcing data for 1850–2005 is taken from observations. The forcing for 2006–2012 is based on the assumptions in the RCP8.5. The historical simulation has three ensemble members initialized from different points of the piControl run. The findings show that BCC_CSM1.1 has a better ability in reproducing the historical global temperature change in the 20th century than the earlier version BCC_CSM1.0 [ Xin et al., 2013 ].

The Representative Concentration Pathway projections include RCP8.5, RCP6.0, RCP4.5 and RCP2.6. The number following RCP represents the assumed radiative forcing of 8.5, 6.0, 4.5 and 2.6 W m− 2  by 2100, respectively. The detailed description of the RCPs can be found in van Vuuren et al. [2011a] . No volcanic forcing is included in the RCP simulations. The GHGs, solar constant, ozone and aerosols are all changing with time. The solar constant contains a stable cycle of 11 years. All RCPs except RCP6.0 have a 200 years extended simulation beyond 2100 (2100–2300). In the extended simulation, only GHGs and solar constant change with time. The ozone and aerosols are both fixed at the value of the year 2100 in the respective scenarios. Among the RCPs, RCP2.6 is the lowest scenario necessary to keep the increase of the global mean temperature below 2°C of the pre-industrial conditions [ Meinshausen et al., 2011  and van Vuuren et al., 2011b ]. Evaluations show that the maximum warming in RCP2.6 simulated by BCC_CSM1.1 is 2.0°C and 2.12°C by BCC_CSM1.1-M relative to the pre-industrial conditions [ Xin et al., 2012 ]. The higher warming in BCC_CSM1.1-M is consistent with its larger transient climate response to GHGs as shown in section 4.1 .

The climate attribution experiments include historicalGhg and historicalNat. The varying forcing of historicalGhg only includes GHGs. The historicalNat only considers the forcing of natural factors including solar irradiance and volcanic aerosols. The simulation time is the same as in the historical experiment (1850–2012). The GHGs for 2005–2012 adopt the forcing data from the RCP8.5. The two experiments are mainly used to determine whether the model response to GHGs and natural factors can be identified in the historical period.

The paleoclimate experiments done by BCC include midHolocene and past1000. The experiments are also included in the Paleoclimate Modelling Intercomparison Project Phase III (PMIP3). The mid-Holocene simulation is initiated from the piControl with the orbit parameter and GHGs fixed at the values during the mid-Holocene. The other forcings used are the same as in the piControl run. A one hundred years’ output is provided. In the past1000 simulation, the forcings changing with time include the solar constant, volcanic aerosols, GHGs and orbit parameter. Other forcings, such us ozone and anthropogenic aerosols, are the same values as in the piControl experiment. The details of the forcing data can be found in Schmidt et al. [2011] . The simulation starts from the spin-up results with the forcing fixed at the year 850 and stops in 1850.

The idealized CO2  experiments include 1pctCO2 and abrupt4xCO2. In the two experiments, only CO changes with time, while other forcings are fixed to the same values as in the piControl run. In 1pctCO2, an 140 years’ simulation was carried out with a 1% per year increase in CO2  concentration. In abrupt4xCO2, the CO2  concentration is set to 4 times of the preindustrial level at the first year with a spin-up for 150 years. Both 1pctCO2 and abrupt4xCO2 are initiated from the same point in the piControl run. The 1pctCO2 experiment is used to measure transient climate responses. The abrupt4xCO2 experiment is used to evaluate the equilibrium climate sensitivity of the model. The latter experiment can also be used to diagnose the fast response of the radiation process when the CO2  concentration changes.

The BCC_CSM1.1 carried out all experiments of the long-term climate simulations mentioned above. The BCC_CSM1.1-M carried out all experiments except for the paleoclimate experiments and the climate attribution experiments (historicalGhg and historical-Nat).

3.2. Carbon-cycle climate experiments

The difference between the carbon-cycle climate experiments and the long-term climate experiments is that the former use carbon emissions instead of CO2  concentration as the external forcing. In the carbon-cycle climate experiments, the CO2  concentration is simulated by the carbon-cycle process in the model. For the model itself, reproducing the evolution of global CO2  concentration is one of the main indicators to estimate whether the simulated carbon cycle is reasonable. The climate simulations with a carbon cycle include the pre-industrial experiment (esm-Control), the historical experiment (esmHistorical), the RCP8.5 projection (esmRcp85) and sensitivity experiments diagnosing the carbon-climate feedbacks (Table 2 ).

Table 2. Simulations with fully coupled carbon/climate models in CMIP5
Short name Experiment name Model version Ensemble size Time period Forcing fields
esmControl ESM pre-industrial control T42,T106 1 250 years N/A
esmHistorical ESM historical T42,T106 1 1850–2012 GHG,SD,Oz,Sl,Vl,SS,Ds,BC,OC
esmRcp85 ESM RCP8.5 T42,T106 1 2006–2099 GHG,SD,Oz,Sl,SS,Ds,BC,OC
esmFixClim1 ESM fixed climate 1 T42 1 140 years CO2
esmFixClim2 ESM fixed climate 2 T42 1 1850–2100 CO2
esmFdbk1 ESM feedback 1 T42 1 140 years CO2
esmFdbk2 ESM feedback 2 T42 1 1850–2100 CO2

The pre-industrial experiment (esmControl) is similar to piControl except that the CO2  concentration is not prescribed but simulated by the model. There is no external source of carbon emissions in the esm-Control run. An output of 250 years is provided after the spin-up of 100 years.

In the historical simulation with a carbon cycle (esmHistorical), the historical carbon emissions are used as the external forcing. Other forcings are the same as in the historical experiment. The integration time is from 1850 to 2012. The forcing data beyond 2005 comes from the RCP8.5 dataset.

The future projection experiment with a carbon cycle is abbreviated esmRcp85. The carbon emissions of the RCP8.5 dataset are used in this experiment. Other forcings are the same as in the RCP8.5 experiment. A simulation from 2006 to 2099 is carried out. This experiment can provide estimates of future climate change with carbon climate feedbacks impacting atmospheric CO2  and climate conditions.

The climate-carbon feedback diagnostic experiments include esmFixClim and esmFdbk, which are designed to estimate the strength of the carbon-climate feedback. In esmFixClim, the CO2 concentration in the radiative process is fixed as in the pi-Control run, but the CO2  in the carbon cycle changes with time. The change in CO2  can be the same as in 1pctCO2 (esmFixClim1) or as in the historical simulation from 1850 to 2005 and RCP4.5 from 2006 to 2100 (esmFixClim2). In esmFdbk, the CO2  concentration in the carbon cycle is fixed, but the CO2  concentration in the radiative module varies with time. The CO2  forcing can be the same as in 1pctCO2 (esmFdbk1) or as in the historical simulation and RCP4.5 (esmFdbk2). The integration time is 140 years for esmFixClim1 and esmFdbk1, and from 1850 to 2100 for esmFixClim2 and esmFdbk2, respectively.

The BCC_CSM1.1 carried out all carbon-cycle climate experiments, while the BCC_CSM1.1-M only performed esmControl, esmHistorical and esmRcp85.

3.3. Decadal prediction experiments

The decadal prediction experiments are newly added within CMIP5, as there is considerable interest in exploring to what degree future climate changes can be attributed to the initial climate state, in particular to the observed ocean state. The ocean initial conditions should be representative for the observed anomalies at the starting date. In BCC_CSM1.1 simulations, the nudging method is used to initialize the ocean state from 50°N to 50°S with a full field of observed ocean temperature. The observation dataset used is the monthly global ocean temperature reanalysis data from the Simple Ocean Data Assimilation (SODA). The time-interval for the nudging method is 1 d. The decadal prediction experiments carried out by the BCC_CSM1.1 are listed in Table 3 .

Table 3. Decadal prediction experiment in CMIP5
Short name Experiment name Model version Ensemble size Time period Forcing fields
decadalXXXX Initialized in year 1960,1965, 1970,1975,1980,1985,1990, 1995,2000,2005 T42 4 30 years GHG,SD,Oz,Sl,Vl,SS,Ds,BC,OC
decadalXXXXmv Addition years during 1960–2005 T42 3 10 years GHG,SD,Oz,Sl,Vl,SS,Ds,BC,OC
noVolcXXXX Volcano-free hindcasts T42 3 10 years GHG,SD,Oz,Sl,SS,Ds,BC,OC
volcIn2010 Prediction with 2010 hindcasts T42 3 10 years GHG,SD,Oz,Sl,Vl,SS,Ds,BC,OC

Note: XXXX represents the first year of the prediction

There are two groups of decadal prediction experiments prescribed by CMIP5. In group 1, the decadal prediction experiments are carried out with initial dates nudging towards the end of 1960, 1965, 1970, 1975, 1980, 1985, 1990, 1995, 2000 and 2005. In group 2, 30-year simulations should be finished with initial dates nudging to the end of 1960, 1980 and 2005. Three ensemble members are required and can optionally be increased to 10 in the simulations of these two groups. The simulations with the BCC_CSM1.1 are not exactly the same as the requirements. All prediction experiments (decadalXXXX) are integrated for 30 years with four ensemble members. The start date are the 1st September, 1st November, 1st December in the previous year before prediction and 1st January of the first prediction year. In addition, except for the ten prediction years (1960, 1965, …, 2005) of the two groups, decadal prediction experiments initialized in the rest years from 1960 to 2006 are also carried out. The integration time is 10 years with three members. The start dates are the 1st September, 1st November in the previous year before the prediction year and 1st January of the first prediction year. The external forcings of these decadal experiments are the same as in the historical simulation until 2005 and as in RCP4.5 beyond 2005.

The decadal experiments without volcanic activities (noVolcXXXX) are also carried out. The external forcing are the same with decadalXXXX but without volcanic activities. The initial dates are nudging toward the end of 1960, 1975, 1980, 1985 and 1990. There is an additional run with a Pinatubo-like volcanic eruption imposed in 2010 initialized near the end of 2005 (volcIn2010). Both noVolcXXXX and volcIn2010 have three ensemble members with 10-year simulation. These experiments enable an assessment of the impacts of volcanic eruptions on decadal predictions.

4. Model results

Results of the two models in the 1pctCO2 experiment and the historical simulation are evaluated in the following to serve as a basic reference for the models’ ability in projecting future climate conditions.

4.1. Transient climate response

The transient climate response (TCR) is an important metric to estimate the model sensitivity to the forcing of GHGs. The TCR is estimated with the 1pctCO2 experiment on the basis of an increase of CO2  by 1% per year. The value of TCR is defined as the globally averaged surface temperature increment at the time of CO2  doubling in 1pctCO2 experiment [ Cubasch et al., 2001 ]. As shown in Figure 1 , the simulated temperature grows linearly with increases in the CO2  concentration. The comparison of the two models shows that the temperature in BCC_CSM1.1-M is higher for the first 70 years, but afterwards lower than that in BCC_CSM1.1. The TCR is 1.85°C for BCC_CSM1.1 and 1.94°C for BCC_CSM1.1-M. In the IPCC AR4, the 10% and 90% confidence limits of multiple models are about 1 and 3°C, with a mean TCR of 1.8°C [ Meehl et al., 2012 ]. The mean TCR is 2.0°C for 16 climate models participating in CMIP5 [ Geoffroy et al., 2012 ]. Thus, sensitivities of both BCC models fall well within the multi-model range and the BCC_CSM1.1-M is closer to the multi-model mean (MME) of CMIP5.


Changes in global mean surface air temperature in 140 years relative to the ...


Figure 1.

Changes in global mean surface air temperature in 140 years relative to the first year based on the 1pctCO2 experiment with BCC_CSM1.1 and BCC_CSM1.1-M

4.2. Historical simulation of mean temperature in China and globally

Verified performances of climate models in reproducing past climate features are the basis of projections of future climate change. Here, the simulation of BCC models is compared with observations and the MME of the CMIP5 models. The MME has been believed to better reproduce the climate response to external forcing than individual model [ Sun and Ding, 2008  and Annan and Hargreaves, 2011 ]. The observed time series of global temperatures are derived from the HadCRUT3 dataset provided by the Climatic Research Unit [ Brohan et al., 2006 ]. China averagely mean time series of surface air temperature during 1906–2005 is from Tang and Ren [2005] . The Chinese-domain average for the model results is defined as the area-weighted average of three rectangular boxes: (28°–50°N, 80°–97.5°E), (22.5°–43°N, 97.5°–122.5°E), and (43°–54°N, 117.5°–130°E), as defined in Zhou and Yu [2006] .

Figure 2 shows the evolution of global mean surface air temperature from 1861 to 2005 simulated with the two BCC models and 18 other models participating in the CMIP5. All models can reasonably reproduce the warming trend in the 20th century. The model results show better agreement with the observation after the 1950s. Results of both BCC models are close to the MME and within the range of the multiple models results. The warming trends projected in the BCC models are higher than in the observation. None of the models can capture the observed warm peak in the 1940s. The warming amplitude during the early 21st century (2000–2005) relative to 1971–2000 mean is 0.45°C for the BCC_CSM1.1 and 0.62°C for the BCC_CSM1.1-M, which are higher than the observed values (0.33°C). The result of BCC_CSM1.1 is closer to the MME value (0.48°C). During 1861–2005, the inter-annual correlation of annual global mean temperature with the observation is 0.88 for the BCC_CSM1.1 and 0.83 for the BCC_CSM1.1-M. The correlation coefficient of the MME in CMIP5 (0.88) is a little higher than that of CMIP3 (0.87) as presented in Zhou and Yu [2006] .


Globally averaged surface air temperature anomalies from 1861 to 2005 relative ...


Figure 2.

Globally averaged surface air temperature anomalies from 1861 to 2005 relative to the 1971–2000 mean as simulated within CMIP5 (the number followed the model label is the correlation coefficient between each model and the observation during 1861–2005)

Simulations of regionally averaged temperature in China are presented in Figure 3 . In comparison with the global conditions, the model discrepancy is larger in the simulation of mean temperature in China. The BCC_CSM1.1 and BCC_CSM1.1-M results are also close to the MME. Results of the two models also show good resemblance to the observation except for the stronger inter-annual variability and the missing of the warm peak in the 1940s. During 1906–2005, the correlation with the observation is 0.50 for BCC_CSM1.1 and 0.55 for BCC_CSM1.1-M. So the high resolution model BCC_CSM1.1-M has better ability in reproducing the evolution of regional mean temperature in China. This is also evident in the models from Japan (MIROC4h, MIROC5, MIROC-ESM-CHEM). Among those three models, the high resolution model MIROC4h has the highest correlation coefficient of 0.60 with the observation. The MME has the highest performance when correlated with the observations, with a correlation coefficient of 0.68, which is larger than the MME result in the CMIP3 (0.55) shown in Zhou and Yu [2006] . This may be due to the fact that the resolutions of many models in the CMIP5 are higher than their previous versions participating in the CMIP3. In the early 21st century, the warming relative to 1971–2000 is 0.69°C for the observation and 0.63°C for the MME. Both of the BCC models underestimate the warming amplitude by about 0.15°C.


Same as in Figure 2, but for regionally averaged surface air temperature in ...


Figure 3.

Same as in Figure 2 , but for regionally averaged surface air temperature in China

5. Conclusions

In this study, the CMIP5 experiments carried out with the BCC climate models (BCC_CSM1.1 and BCC_CSM1.1-M) are described. Both models are coupled carbon/climate models. The atmospheric module of BCC_CSM1.1-M has a higher horizontal resolution. The main information of the experiments is described, including the experimental design, external forcings and main purpose.

The simulation results of the 1pctCO2 experiment and the historical simulation are examined in order to estimate the models’ sensitivity and their general performance in reproducing the historical climatic changes. In the 1pctCO2 experiment, the response of the temperature is stronger with the BCC_CSM1.1-M than with the BCC_CSM1.1 in the first 70 years, but vice versa afterwards. BCC_CSM1.1-M is closer to the CMIP5 multiple models mean results. In the historical run, the two BCC models perform well in simulating the historical evolution of global temperature. During 1861–2005, the temporal correlation coefficient of observed annual global mean temperature with the BCC_CSM1.1 is 0.88, identical to the MME result and higher than that of the BCC_CSM1.1-M. For the regional surface air temperature in China, the BCC_CSM1.1-M shows a better capability according to the inter-annual correlation coefficient of 0.55 during 1906–2005. The two models overestimate the global warming but underestimate the warming amplitude over China in the early 21st century.

Acknowledgements

This work was jointly supported by the National Basic Research Program of China (973 Program) under No. 2010CB951903, the National Science Foundation of China under Grant No. 41105054, 41205043, and the China Meteorological Administration under Grant No. GYHY201106022, GYHY201306048, CMAYBY2012-001.

Received: 8 November 2012

References

  1. Ammann et al., 2003 C.M. Ammann, G.A. Meehl, W.M. Washington, et al.; A monthly and latitudinally varying volcanic forcing dataset in simulations of 20th century climate; Geophys. Res. Lett., 30 (2003), p. 1657 10.1029/2003GL016875
  2. Annan and Hargreaves, 2011 J.D. Annan, J.C. Hargreaves; Understanding the CMIP3 multimodel ensemble; J. Climate, 24 (2011), pp. 4529–4538
  3. Brohan et al., 2006 P. Brohan, J.J. Kennedy, I. Harris, et al.; Uncertainty estimates in regional and global observed temperature changes: A new dataset from 1850; J. Geophys. Res., 111 (2006) 10.1029/2005JD006548, D12106
  4. Cubasch et al., 2001 U. Cubasch, G.A. Meehl, G.J. Boer, et al.; Projections of future climate change; J.T. Houghton (Ed.), et al. , The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press (2001), pp. 525–582
  5. Geoffroy et al., 2012 O. Geoffroy, D. Saint-Martin, A. Ribes; Quantifying the sources of spread in climate change experiments; Geophys. Res. Lett., 39 (2012) 10.1029/2012GL054172, L24703
  6. Griffies et al., 2005 S.M. Griffies, A. Gnanadesikan, K.W. Dixon, et al.; Formulation of an ocean model for global climate simulations; Ocean Sci., 1 (2005), pp. 45–79
  7. Ji et al., 2008 J.-J. Ji, M. Huang, K.-R. Li; Prediction of carbon exchange between China terrestrial ecosystem and atmosphere in 21st century; Science in China Series D: Earth Science, 51 (2008), pp. 885–898
  8. Meehl et al., 2006 G. Meehl, W.M. Washington, B.D. Santer, et al.; Climate change projections for the twenty-first century and climate change commitment in the CCSM3; J. Climate, 19 (2006), pp. 2597–2616
  9. Meehl et al., 2012 G.A. Meehl, W.M. Washington, J.M. Arblaster, et al.; Climate system response to external forcings and climate change projections in CCSM4; J. Climate, 25 (2012), pp. 3661–3683
  10. Meinshausen et al., 2011 M. Meinshausen, S.J. Smith, K. Calvin, et al.; The RCP greenhouse gas concentrations and their extensions from 1765 to 2300; Climatic Change, 109 (2011), pp. 213–241
  11. Schmidt et al., 2011 G.A. Schmidt, J.H. Jungclaus, C.M. Ammann, et al.; Climate forcing reconstructions for use in PMIP simulations of the last millennium (v1.0); Geosci. Model Dev., 4 (2011) doi:10.5194/gmd-4-33-2011.
  12. Sun and Ding, 2008 Y. Sun, Y.-H. Ding; An assessment on the performance of IPCC AR4 climate models in simulating interdecadal variations of the East Asian summer monsoon; Acta Meteorologica Sinica, 22 (2008), pp. 472–488
  13. Tang and Ren, 2005 G.-L. Tang, G.-Y. Ren; Reanalysis of surface air temperature change of the last 100 years of China; Climatic and Environment Research (in Chinese), 10 (2005), pp. 791–798
  14. Taylor et al., 2011 K.E. Taylor, R.J. Stouffer, G.A. Meehl; An overview of CMIP5 and the experiment design; Bull. Amer. Meteorol. Soc. (2011) 10.1175/BAMS-D-11-00094.1
  15. van Vuuren et al., 2011a D.P. van Vuuren, J. Edmonds, M. Kainuma, et al.; The representative concentration pathways: An overview; Climatic Change, 109 (2011), pp. 5–31
  16. van Vuuren et al., 2011b D.P. van Vuuren, E. Stehfest, M.G.J. den Elzen, et al.; RCP2.6: Exploring the possibility to keep global mean temperature increase below 2°C; Climatic Change, 109 (2011), pp. 95–116
  17. Winton, 2000 M. Winton; A reformulated three-layer sea ice model; J. Atmos. Oceanic Technol., 17 (2000), pp. 525–531
  18. Wu, 2012 T.-W. Wu; A mass-flux cumulus parameterization scheme for large-scale models: Description and test with observations; Clim. Dyn., 38 (2012), pp. 725–744
  19. Wu et al., 2008 T.-W. Wu, R.-C. Yu, F. Zhang; A modified dynamic framework for atmospheric spectral model and its application; J. Atmos. Sci., 65 (2008), pp. 2235–2253
  20. Wu et al., 2010 T.-W. Wu, R.-C. Yu, F. Zhang, et al.; The Beijing Climate Center for atmospheric general circulation model (BCC-AGCM2.0.1): Description and its performance for the present-day climate; Clim. Dyn., 34 (2010), pp. 123–147
  21. Xin et al., 2013 X.-G. Xin, Y.-J. Cheng, F. Wang, et al.; Asymmetry of surface climate change under RCP2.6 projections from the CMIP5 models; Adv. Atmos. Sci., 30 (3) (2013) doi: 10.1007/s00376-012-2151-3.
  22. Xin et al., 2012 Y.-J. Xin, Y.-J. Cheng, F. Wang, et al.; Asymmetry of surface climate change in RCP2.6 projected by multiple CMIP5 models; Adv. Atmos. Sci. (2012) doi:10.1007/s00376-012-2151-3.
  23. Zhou and Yu, 2006 T.-J. Zhou, R.-C. Yu; Twentieth-century surface air temperature over China and the globe simulated by coupled climate models; J. Climate, 19 (2006), pp. 5843–5858

Notes

. http://cmip-pcmdi.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf

Back to Top

Document information

Published on 15/05/17
Submitted on 15/05/17

Licence: Other

Document Score

0

Views 18
Recommendations 0

Share this document

claim authorship

Are you one of the authors of this document?