Introduction

Having steadily improved their general performance in simulation of global climate, Coupled Global Climate Models (CGCMs) are now the primary tools to study the climatic processes, natural variability and response to external forcings. However, due to their significant complexity and the climate change community's requirement for long-term integration, the atmospheric components of the most CGCMs usually have their horizontal resolutions limited from 400 km to 125 km, so they are unable to provide reliable climatic information at regional scale, which depends on the treatment of processes at the unresolved scales for CGCMs. In view of the demand to provide integrated assessments of climate change impacts on regional scale, the new challenges are to estimate accurate timing (occurrence) and intensity of regional climate change, analyze its environmental and socio-economic consequences, and provide integrated assessments to society and the environment. Therefore, so called downscaling methods have been developed to generate detailed regional information.

There are two widely used downscaling approaches: the dynamical and empirical downscaling. Dynamical downscaling refers to the methods that derive regional climate information by using high resolution numerical climate models, such as high-resolution Atmosphere-only Global Climate Model (AGCMs) and nested regional climate model (RCM) which use outputs from CGCM or observations as boundary and initial condition.

Currently many research facilities have AGCMs with resolution of 100km and finer; a resolution of 50 km will likely be the normal for AGCMs in the near future (Bengtsson, 1996; May and Roeckner, 2001; Deque and Gibelin, 2002; Govindaswamy, 2003) (IPCC 2004). Besides uniform high-resolution AGCM, variable-resolution (including stretched-grid) AGCMs are viewed as attractive downscaling approach due to their abilities to increase resolution over interested regions while maintaining a full interaction in other regions of the globe (VRGCM; e.g., Deque and Piedelievre, 1995; Krinner et al., 1997; Fox-Rabinovitz et al., 2001;

McGregor et al., 2002; Gibelin and Deque, 2003). The results show that VRGCMs are able to capture finer scale details over the focus regions, and at the same time retain the their global skill (IPCC 2004).

Another dynamic downscaling is to generate realistic regional climate information through regional climate models (RCMs), which are driven by either GCM or observation, which exert correct large circulation through providing boundary and initial conditions. RCMs were first successfully applied in regional climate simulation in late 1980 (Dickinson et al., 1989; Giorgi and Bates, 1989), and coupled with the driving forcings in one-way mode, i.e., there's no feedback from the RCM simulation to the driving GCMs. Not until recently two-way technique has been developed to nest RCM with GCMs (Lorenz and Jacob, 2005), which will allow feedback from the RCM onto the GCM. In addition, more climate components, such as regional ocean and sea ice, hydrology, aerosol, atmospheric chemistry and dynamical land surface process, have been included in RCMs (Gao and Yu, 1998; Xue et al., 2000; Maslanik et al., 2000; D.scher et al., 2002; Rinke et al., 2003; Debernard et al., 2003; Schrum et al., 2003; Meier et al., 2004; Rummukainen et al., 2004). Due to its ability to capture the mesoscale nonlinear effects under perturbed forcing conditions and provide coherent information between multiple climate variables, dynamical downscaling is widely used to reproduce varying climates around the world.

Empirical downscaling refers to the method which predicts future climate change by first establishing a statistic relationship between large scale atmospheric state and regional variables which are derived from historical data, then applying the relationship to model simulation results for future regional climate projection. The methods are computationally inexpensive though they have the drawback that they require long time series of reliable, homogeneous station data and assume that the derived statistical relationships will remain unaltered under perturbed climate.

RCMs have been applied widely in research of current and future climate over United States (Giorgi et al., 1994, 1998; Gutowski et al., 2004), Central America (Giorgi and Mearns, 1996; Anderson et et al., 2003;), West America (Dickinson, 1989; Giorgi, 1990b, 1991; Giorgi et al., 1993c; Anderson et al., 2002), Europe (Giorgi et al., 1990; Giorgi and Marinucci, 1991; Marinucci, 1992; Christensen et et al, 1997; Machenhauer et al, 1996, 1998; Noguer et al., 1998; Pal et al., 2004; Raisanen, 2004), Australia (Walsh and McGregor, 1995), Africa (Sun et al., 1999), Central Asia (Small et al., 1999), North Pole (Rinke et al., 1999; Wei et al., 2002) and East Asia-West Pacific area (Giorgi et al., 1999) etc. RCMs have shown promising performance in reproducing the regional details in surface climate forced by topography, lakes, the costal line, land use distribution and vegetation coverage, etc.(Christensen et al., 1998; Machenhauer et al., 1998).

The East Asian summer monsoon is characterized by marked variability at seasonal, interannual and interdecadal time scales (Fu and Zheng 1998a). The timing of monsoon onset and the irregular pace of its seasonal, northward progression can influence water availability for agriculture and urban consumption remarkably (Fu et al., 2003; Tao and Chen 1987; Wu and Zhang 1998). Inter-annual variability, such as those linked with the ENSO cycle, affects the frequency of droughts, floods and other weather extremes that occur during the summer monsoon period (Fu and Teng 1993; Ju and Slingo 1995; Fasullo and Webster 2002). On decadal to century time scales, the rapidly growing economy and population of East Asia present anthropogenic influences that may also alter monsoon behavior (Fu and Zheng 1998a; Quan et al. 2003). However coarse resolution climate models generally fail to give satisfactory simulations of the East Asian monsoon (Lau and Yang 1996; Yu et al. 2000).

A series of researchers have used RCMs for simulating the regional climate of East Asia (e.g. Lee, 1992, 2000; Liu et al., 1994, 1996; Hirakuchi et al., 1995; Ji et al., 1997; Qian et al., 1999; Leung et al., 1999; Giorgi et al., 1999; Kato et al., 1999, 2001; Hong, 1999; Gao et al., 2001, 2003; Zhao Zongci et al., 1997, 1999; Fu Congbin et al., 1998; Qiang Yongfu, 1999; Liu Liping, 1999, 2000; Wang Shuyu, 1999; Chen Ming, 2000; Shi Xueli, 2001; Li Qiaoping et al., 2004; Zhang Yinjuan et al, 2005). Many of these studies have shown that RCMs can simulate the spatial detail of monsoon climate better than GCMs (e.g.Liu et al. 1994, 1996; Fu et al. 1998b; Lee and Suh 2000).

Efforts have been made to investigate the role the resolution and topography in East Asian precipitation simulation (Gao et al., 200). The results show that model resolution can significantly influence the simulated East Asian large scale precipitation patterns, and the effect of resolution is more important during the mid to late monsoon months when the precipitation is dominated by smaller scale convection. Other simulations have been carried out to study the impacts of man-made modification of surface character and land cover/land use on regional precipitation. Wu et al. (2006) examined such as the construction of Three Gorges Dam (TGD) on regional climate using the high-resolution Pennsylvania State University-National Center for Atmospheric Research (PSU-NCAR) fifth generation Mesoscale Model (MM5). The numerical simulation indicates that the land use change associated with the TGD construction has increased the precipitation in the region between Daba and Qinling mountains, and reduced the precipitation in the vicinity of the TGD after the TGD water level abruptly rising from 66 to 135 m in June 2003. The results are agreed with inde pendent analysis of satellite data sets, i.e., rainfall rate from TRMM (Tropical Rainfall Measuring Mission) and land surface temperature from MODIS (Moderate Resolution Imaging Spectroradiometer).

However, multi-year simulations must be used to identify models' uncertainties in simulating climatology and climatic variations, and systematically evaluate the RCMs' application for Asia. A Regional Climate Model Inter-comparison Project (RMIP) for Asia therefore has been established to study the performance of an ensemble of regional climate models (RCMs) when simulating East Asian climate. Up to ten research groups from the Asia-Pacific area have been participating in the project. One of the objectives of RMIP is to further improve RCM simulations of East Asian climate by evaluating their strengths and weaknesses in a common framework (Fu et al. 2003).

Considering the relatively high degree of uncertainty in regional climate change information of East Asia derived from GCMs, which results from the scenarios construction such as future emission variations and GCMs'modeling of the climate responses to a given scenario, RMIP also try to compare the regional climate scenarios of a set of RCMs and provide multi-RCM ensemble of regional climate change of Asia.

This chapter presents the latest progresses of regional climate modeling of China and East Asia, including mainly (1) the development of a regional integrated environment model system (RIEMS); (2) Simulation of climate of China and East Asia through inter-comparison of a set of regional climate models; and (3) RCM's response to increasing atmospheric greenhouse gases and aerosols and to changing land use/land cover in Asia.

0 0

Post a comment