## E 60e 90e 120e 150e 180 150w 120w 90w 60w

Figure 1. Precipitation climatology (1979—2004) in (a) the boreal summer (JJA) and (b) the boreal winter (DJF) from the Climate Precipitation Center (CPC) Merged Analysis of Precipitation (CMAP; Xie and Arkin, 1997).

Plateau. The Asian summer monsoon has several subregional systems, including the Indian summer monsoon, the western North Pacific summer monsoon, and the East Asian summer monsoon (Goswami et al., 1999; Lau et al., 2000; Wang and Fan, 1999; Wang and LinHo, 2002). Each subregional monsoon system has its own evolution. Furthermore, the Asian summer monsoon rain zone seems to extend even farther southwestward into Africa. Some studies indicate a connection between the African summer rainfall and the Indian summer rainfall (Liu and Yanai, 2001; Camberlin, 1997). Thus, the Asian summer monsoon system should be treated as one very broad system over the Eurasian-African continent and tropical oceans (Webster and Yang 1992; Yang et al., 1992). On one hand, the Asian summer monsoon as a whole is associated with land-ocean contrasts between the Eurasian-African continent and neighboring oceans, such as the Pacific Ocean and the Indian Ocean. On the other hand, local forcings, such as local sea surface temperature (SST), may affect the Asian summer monsoon and create different subregional summer monsoon systems.

The Asian summer monsoon rain zone marches with the season which is associated with solar heating. The Asian summer monsoon also shows interannual variation in the magnitude and position of the monsoon rainfall with and without El Niño/Southern Oscillation (ENSO) influences. Figure 2 shows precipitation anomalies in the ENSO growing (year 0) and decaying (year 1) years and the non-ENSO year (Chou et al., 2003; CTY hereafter). Under ENSO influences, a strong (weak) western North Pacific (WNP) summer monsoon occurs over the western North Pacific and the Philippine Sea in the El Nino (La Nina) growing year and the La Niña (El Niño) decaying year [Figs. 2(b) and 2(c)]. To its north, this strong WNP summer monsoon often accompanies a decrease of the northeast-southwest tilting summer monsoon rainfall. In other words, the northeast-southwest tilting Asian summer monsoon rainfall decreases (increases) in the El Niño (La Nina) growing year [Fig. 2(b)] and the La Niña (El Nino) decaying year [Fig. 2(c)]. This interannual variation of the Asian summer monsoon indicates a dipole pattern of the precipitation anomalies along the coast of East Asia, and may affect the position of the northern edge of the Asian summer monsoon rain zone. Without ENSO influences, the Asian summer monsoon also shows a similar interannual variability [Fig. 2(a)]. This implies that ENSO affects some processes that can also affect the inter-annual variation of the Asian summer monsoon. Besides the impacts of ENSO on the Asian summer monsoon strength and location, ENSO also affects the WNP summer monsoon onset via the remote and local SST anomalies (Wu and Wang, 2000).

To examine dynamical mechanisms limiting the poleward extent of summer monsoons, an atmospheric model with intermediate complexity is coupled with a simple land-surface model and a mixed-layer ocean with prescribed ocean heat transport (Neelin and Zeng, 2000; Zeng et al., 2000). A brief description of the model can be found in Sec. 2. Moisture and moist static energy budgets are used to discuss the summer monsoon mechanisms, and the derivations are shown in Sec. 3. We first use an idealized monsoon system to discuss those mechanisms in a general sense (Chou et al., 2001; CNS hereafter), and then compare the impacts of the various mechanisms on the Asian summer monsoon, a realistic summer monsoon system (Chou and Neelin, 2003; CN03 hereafter; Chou, 2003, CHOU hereafter). Conclusions follow.

2. Quasi-equilibrium Tropical

Circulation Model

To examine the ocean-atmosphere-land feedbacks on monsoons, a coupled ocean-atmosphere-land model of intermediate complexity

(a) Pa(JJA): strong-weak WNPSM non-ENSO years

(b) ENSO developing years

(c) ENSO decaying years

Figure 2. Composite difference of the CMAP summer rainfall anomalies between the strong and weak WNP summer monsoons for (a) the non-ENSO years, (b) the ENSO developing years and (c) the ENSO decaying years (from CTY). The contour interval is 1 mm day-1. Areas with the significance level at 5% by the two-sample t test are shaded. The thick dashed line shows the JJA precipitation climatology at 5 mm day-1.

Figure 2. Composite difference of the CMAP summer rainfall anomalies between the strong and weak WNP summer monsoons for (a) the non-ENSO years, (b) the ENSO developing years and (c) the ENSO decaying years (from CTY). The contour interval is 1 mm day-1. Areas with the significance level at 5% by the two-sample t test are shaded. The thick dashed line shows the JJA precipitation climatology at 5 mm day-1.

(Neelin and Zeng, 2000; Zeng et al., 2000) is used (CNS). Based on the analytical solutions derived from the Betts-Miller moist convective adjustment scheme (Betts and Miller, 1993), typical vertical structures of temperature, moisture and winds for deep convection are used as leading basis functions for a Galerkin expansion (Neelin and Yu, 1994; Yu and Neelin, 1994). The resulting primitive equation model makes use of constraints on the flow by quasi-equilibrium thermodynamic closures and is referred to as QTCM1 (quasi-equilibrium tropical circulation model with a single vertical structure of temperature and moisture for deep convection). Because the basis functions are based on vertical structures associated with convective regions, these regions are expected to be well-represented and similar to a general circulation model (GCM) with the Betts-Miller moist convective adjustment scheme. Far from convective regions, QTCM1 is a highly truncated Galerkin representation equivalent to a two-layer model. With its intermediate complexity, between a full GCM and simpler models, this model is easier to analyze, is numerically faster than a GCM, and has more physical processes than simpler models.

To accompany the representation of dynamics in this model, a cloud-radiation scheme (Chou and Neelin, 1996; Zeng et al., 2000), simplified from the full GCM radiation schemes (Harshvardhan et al., 1987; Fu and Liou, 1993), is included. The longwave radiation parame-trization is derived from Harshvardhan et al. (1987) with the help of a Green's function method to calculate upward and downward longwave radiative fluxes. The shortwave radiation scheme is a simple linear calculation of shortwave fluxes at the surface and atmospheric absorption based on Fu and Liou (1993). The deep and cirrocumulus/cirrostratus cloud fraction is estimated with an empirical parame-trization (Chou and Neelin, 1999). This physical parametrization package can give near GCM-like accuracy in the determination of radiative flux exchange at the surface under suitable circumstances. Caveats for the application here include lack of prognostic low and middle clouds. A simple atmospheric boundary layer that assumes a steady state and a vertically homogeneous mixed layer with fixed height (Stevens et al., 2002) is implemented.

Similar in structure to the physical parame-trization package, a land-surface model (Zeng et al., 2000) with intermediate complexity includes the essentials of complex land-surface models, such as the biophysical control on evapotranspiration and surface hydrology, but retains computational and diagnostic simplicity. This model is designed to capture land-surface effects for climate simulation at time scales longer than a day. It uses a single land-surface layer for both the energy and water budget. Since the heat capacity of land is small, the net surface heat flux is essentially close to zero at a time scale longer than a day. Surface heat flux is balanced by solar radiation, longwave radiation, evaporation and sensible heat flux. The prognostic equation for soil moisture induces precipitation, evaporation, surface runoff and ground runoff. Surface runoff increases as the soil moisture reaches a saturation value that depends on the surface type.

To study the interaction between the atmosphere and land, fixed SST is commonly assumed in climate models in order to simplify the simulation. To avoid any artificial effects induced by the contrast of the fixed SST boundary condition and the interactive land surface, an ocean surface layer with active thermodynamics is included in the simulations presented here. Instead of coupling a complicated ocean general circulation model, a slab mixed-layer ocean model with a fixed mixed-layer depth of 50 m is used. By specifying Q flux, which crudely simulates divergence of ocean transport (Hansen et al., 1988, 1997), SST can be determined by the energy balance between surface radiative flux, latent heat flux, sensible heat flux and Q flux. The Q flux can be obtained from observations or ocean model results (Doney et al., 1998; Keith, 1995; Miller et al., 1983; Russell et al., 1985). In general, the Q flux varies from ocean to ocean as well as from season to season. QTCM version 2.3 is used here, with the solar radiation scheme slightly modified.

3. Moist Static Energy Analysis

3.1. Moist static energy budget

The vertically integrated thermodynamic and moisture equations can be written as dt{T) + {v ■ VT) + {udps) = {Qc) + H + Snet + Rnet,

where v is horizontal velocity, w is pressure velocity, H is sensible heat flux, and E is evaporation. Temperature T and specific humidity q are in J kg-1, absorbing heat capacity at constant pressure (Cp) and latent heat per unit mass (L). The dry static energy s is defined as s = T + $, with $ the geopotential. The vertical integral (X ) through the whole troposphere pT is defined as 1/gfXdp. The longwave radiation Rnet and the shortwave radiation Snet are the net radiative energy into the atmospheric column. When vertically integrated, the con-vective heating Qc will cancel with the moisture sink Qq, since horizontal moisture transport by convective motion is negligible. Thus, precipitation can be estimated through P =

Because of the cancellation between (Qc) and (Qq ), adding the thermodynamic equation (1) and the moisture equation (2) can derive the moist static energy (MSE) equation:

The net heat flux into the atmospheric column Fnet, positive when heating the atmosphere, can be derived from

where subscripts s and t on the solar (S^ and ST) and longwave (R and R^) radiative terms denote the surface and model top, and R\ « 0 has been used. The net heat flux at the top of the atmosphere (TOA) is

and the net heat flux at the surface is

Positive Ftnet and Fsnet indicate downward heat fluxes. The net longwave and shortwave radiative fluxes in (1) can then be estimated through Snet = SS - Si - SS + Si and Rnet = -Rl - RS + Rl.

3.2. Horizontal advection associated with v^ and vx

In the experiments for examining the effect of horizontal transports, the terms v • VT and v • Vq are usually set to zero (CNS; Chou and Neelin, 2001; CN01 hereafter). However, suppressing these terms creates errors in the conservation of energy and moisture. To refine the experiments, we divide the horizontal winds into nondivergent part v^ and irrotational part vx. By definition,

When taking an area average over the domain with v^ • n « 0 through the entire boundary,

(a) JUL CLIMATOLOGY: OBS Fnet

(a) JUL CLIMATOLOGY: OBS Fnet

where n is the normal direction of the boundary and X is T or q. Thus, suppressing the advection terms associated with v^, e.g. v^ ■ Vq = 0, will not affect domain-integrated conservation in the experiments for examining the effect of the horizontal transport. Usually, v^ is much larger than vx (six times larger in CN03), so the effect of suppressing the terms v^ ■ VT and v^ ■ Vq is still substantial.

4. Mechanisms for the Poleward Extent of Summer Monsoons

Land-ocean heating contrast has been considered the heart of the summer monsoon circulation. Because of ocean heat storage and transport, the heating of the atmosphere initiated by solar radiation is much stronger over continental regions than over ocean regions. Figure 3 shows the net energy into the atmospheric column Fnet. Strong heating (positive Fnet) is found all over the summer continents, while weak heating or even cooling is over the ocean regions. Thus, strong land-ocean heating contrast is over the summer hemisphere. From the moisture budget (2) and MSE budget (3), the summer monsoon rainfall and the associated convection are affected not only by local processes via evaporation and Fnet, but also by the moisture and MSE transports via cross-continent flow and a feedback from the summer monsoon circulation. The transport

(a) Control run

## Post a comment