利用虚拟机的灵活性和快速部署能力,设计并实现了任务部署与调度系统(task deployment and dispatch system,TDDS).TDDS能够根据用户的需求,为用户的计算任务提供可以进行资源配置的集群计算环境,满足了用户对不同操作系统、不同应用程序和不同计算资源的需求.TDDS还使用了负载均衡策略,以提高物理集群资源的利用率.提出了两种虚拟机部署策略,用以加快虚拟集群部署的速度.TDDS尽量控制虚拟机系统镜像的大小和访问频率,以提高部署的效率.实验表明,TDDS系统能够快速灵活地部署用户所需要的计算环境和计算资源,负载均衡的调度策略也切实提高了物理集群处理任务的能力,提高了集群的使用率.
The historical simulation of phase five of the Coupled Model Intercomparison Project （CMIP5） ex- periments performed by the Beijing Climate Center cli- mate system model （BCC_CSM1.1） is evaluated regard- ing the time evolutions of the global and China mean sur- face air temperature （SAT） and surface climate change over China in recent decades. BCC CSM1.1 has better capability at reproducing the time evolutions of the global and China mean SAT than BCC_CSM1.0. By the year 2005, the BCC_CSM1.1 model simulates a warming am- plitude of approximately I℃ in China over the 1961- 1990 mean, which is consistent with observation. The distributions of the warming trend over China in the four seasons during 1958-2004 are basically reproduced by BCC CSM1.1, with the warmest occurring in winter. Al- though the cooling signal of Southwest China in spring is partly reproduced by BCC_CSM1.1, the cooling trend over central eastern China in summer is omitted by the model. For the precipitation change, BCC_CSM1.1 has good performance in spring, with drought in Southeast China. After removing the linear trend, the interannual correlation map between the model and the observation shows that the model has better capability at reproducing the summer SAT over China and spring precipitation over Southeast China.
The multi-model ensemble （MME） of 20 models from the Coupled Model Intercomparison Project Phase Five （CMIP5） was used to analyze surface climate change in the 21st century under the representative con- centration pathway RCP2.6, to reflect emission mitigation efforts. The maximum increase of surface air temperature （SAT） is 1.86℃ relative to the pre-industrial level, achieving the target to limit the global warming to 2℃. Associated with the ＂increase-peak-decline＂ greenhouse gases （GHGs） concentration path- way of RCP2.6, the global mean SAT of MME shows opposite trends during two time periods： warming during 2006-55 and cooling during 2056-2100. Our results indicate that spatial distribution of the linear trend of SAT during the warming period exhibited asymmetrical features compared to that during the cool- ing period. The warming during 2006-55 is distributed globally, while the cooling during 2056-2100 mainly occurred in the NH, the South Indian Ocean, and the tropical South Atlantic Ocean. Different dominant roles of heat flux in the two time periods partly explain the asymmetry. During the warming period, the latent heat flux and shortwave radiation both play major roles in heating the surface air. During the cooling period, the increase of net longwave radiation partly explains the cooling in the tropics and subtropics, which is associated with the decrease of total cloud amount. The decrease of the shortwave radiation accounts for the prominent cooling in the high latitudes of the NH. The surface sensible heat flux, latent heat flux, and shortwave radiation collectively contribute to the especial warming phenomenon in the high-latitude of the SH during the cooling period.
