Impact of cloud radiative heating on East Asian summer monsoon circulation

The impacts of cloud radiative heating on the East Asian Summer Monsoon (EASM) over southeastern China (105°–125°E, 20°–35°N) are addressed by using the Community Atmosphere Model version 5 (CAM5). Sensitivity experiments demonstrate that the radiative heating of clouds leads to a positive effect on the local EASM circulation over southeastern China. Without the radiative heating of clouds, the EASM circulation and precipitation would be much weaker than that in normal conditions. The longwave heating of clouds dominates the changes of EASM circulation. The positive effect of clouds on EASM circulation is explained by the thermodynamic energy equation, i.e. the different heating rate between cloud base and cloud top enhances the convective instability over southeastern China, which consequently enhances updraft. The strong updraft would further result in a southward meridional wind above the center of the updraft through Sverdrup vorticity balance.


Introduction
East Asian Summer monsoon (EASM) has prominent impacts on the local climate of East Asia (Wang and Li 2004). In each boreal summer, the southwesterly wind along the southern flank of the western Pacific subtropical high carries moist warm air from the Pacific and meets the cold continental air mass over Southeastern China. The East Asian subtropical front is thus established, which weakens the baroclinicity significantly and increases the frequency of deep convections Chan 2005, Wang et al 2008). Along the frontal zone, a significant portion of the convective precipitation extends from the Yangtze River valley eastward to Japan, due to the high occurrence frequency of deep convections (Ding and Chan 2005). Correspondingly, prevailing low clouds in winter are replaced by high clouds in summer over southeastern China, including deep cumulus and cirrus (e.g., Yu et al 2001, Luo et al 2009. The cloud is traditionally regarded as a result of atmospheric circulation changes associated with the EASM system. It is well recognized that the latent heats released by lifting condensations lead to positive effects in the development of EASM. However, the radiative effects of clouds are somewhat overlooked. Clouds are generally thought to warm the underlying atmosphere by their greenhouse effects, and cool the atmosphere underneath the cloud and heat the cloud aloft through their reflection and absorption of shortwave radiation. Such a vertical difference in heating rates between the cloud base and cloud top is strong enough to cause convective instability in anvils (Ackerman et al 1988, Mather et al 2007. We thus hypothesize that clouds can feedback on the EASM circulation through its different radiative heating effects between the underlying air and cloud top. This hypothesis is examined by performing numerical experiments with the Community Atmosphere Model version 5 (CAM5). We show evidence that, without accounting for the cloud radiative heating, EASM circulation and monsoon precipitation would (significantly) become weaker.

Data, model and methods
The datasets used in this study consist of the following seasonally mean fields: ( (2)The CPC Merged Analysis of Precipitation (CMAP) data from 1979 to 2011 is used to describe the climatological monsoon rain band (Xie and Arkin, 1996).
(3)CloudSat and CALIPSO (Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations) data from 2006 to 2011 are used to depict the vertical structure of cloud (Stephens et al 2002 6 ). CloudSat, which carries the Cloud Profiling Radar with its estimated minimum detectable signal of −30 dBZ, provides reflectivity (from Level 2 CloudSat Geometrical Profiling Product) and heating rates (from Level 2 CloudSat Fluxes and Heating Rates Product) that are used in this study.
Numerical experiments are performed by using CAM5 (Neale et al 2011). Stratiform clouds and cumulus in CAM5 are treated by separate cloud parameterizations. The shallow cumulus scheme is described by Park and Bretherton (2009) and deep cumulus is treated by the Zhang and McFarlane scheme McFarlane 1995, Neale et al 2008). Stratiform cloud macrophysics is described by Park et al (2014) and Gettelman et al (2010). The model applies the Rapid Radiative Transfer Model for general circulation models (RRTMG) to treat the radiative transfer (Mlawer et al 1997, Iacono et al 2000. In order to examine how the cloud radiative heating affects EASM over southeastern China, we perform three sensitive experiments (table 1), in which the all-sky radiative fluxes over (105°-125°E, 20°-35°N , see the box of figure 1(a)) are set equal to their clear-sky counterparts in shortwave (NS), longwave (NL) and net radiative (NR) calculations, respectively.
In addition, a control run without these changes is performed (CNTL). Results from three sensitive experiments are then compared to the control run. All experiments are run with 30 vertical levels and a horizontal resolution of 2°and forced by the same prescribed climatological sea surface temperatures (SSTs). All experiments are run 6 years; the outputs of the last 5 years are used in the analysis (see table 1). In order to better compare with satellite observations, the Cloud Feedback Model Intercomparison Project Observation Simulator Package (COSP) (Kay et al 2012) is configured in this study.

Results
The spatial patterns of the summer mean precipitation and winds (figure 1(a)) show the subtropical high with its southwesterly winds prevailing on Southeastern China, which carries abundant moisture from the South China Sea to Southeastern China at 850 hPa (figure 1(a)). The rain belt is consistent with the wind field, which extends along the Mei-Yu frontal area (∼30°N) from the Yangtze River valley to southern Japan.
Consistent with the EASM horizontal circulation, the vertical EASM circulation replaces the domination of Hadley cell over southeastern China (Ye andYang 1979, Zhou andLi 2002). A prominent feature of the vertical EASM circulation is that a strong updraft accompanied with its southward upper outflow prevails at southeastern China (figure 2(a)). Correspondingly, the seasonal mean rain belt is well synchronized with the updraft, which moves northward beyond 20°N (figure 2(a)). Following the updraft of the vertical EASM circulation, high clouds dominate Southeastern China (105°-125°E, 20°-35°N) during the summer. The contoured frequency by altitude diagrams (CFAD) of radar reflectivity for cloudy profiles over southeastern China derived from CloudSat/CALIPSO is shown in figure 3(a). As the minimum detectable signal of the Cloud Profiling Radar is −30 dBZ, only results with a reflectivity larger than −30 dBZ are shown. A strong, while narrow, radar reflectivity range extends to high altitudes (12 km). It indicates that the prevailing of anvils is mainly constituted by small ice crystals. The strong and unbroken frequency belt appears from 1 km to 14 km, representing the existence of convective clouds (Luo et al 2009). The high frequencies of radar reflectivity above freezing level (4 km) indicate large ice-phase cloud particles. While the frequency belt between −5 dBZ and 10 dBZ below 4 km are probably due to raindrops and liquid-phase cloud droplets. CAM5 generally reproduces the 3D structures of EASM circulation and cloud system ( figure 2(b)). It is notable that the simulated location of western Pacific subtropical high shifts northward relative to the observation ( figure 1(b)). Moreover, the simulated vertical circulation of EASM is broader and higher than the observation, and the rain belt is located at the North China Plain (∼40°N) instead of the Yangtze River valley (∼30°N) ( figure 2(b)). Consequently, the simulated radar reflectivity frequency belt vertically penetrates into a higher altitude and covers a wider range ( figure 3(b)), which mean more clouds (both liquid-and ice-phase) and higher cloud tops are produced by CAM5. Figures 1-4 show that clouds play a significant role in influencing the structure and magnitude of EASM by changing the vertical profiles of radiation, including circulation and precipitation. When the cloud longwave radiative effects are ignored (NL run), the weak cloud greenhouse effect causes a negative temperature anomaly of 0.8 K at the mid-level of the troposphere, which further enhances the stability and triggers an anomalous subsidence over southeastern China ( figure 2(c)). The strong anomalous subsidence thus results in several responses in the local EASM circulation and precipitation. First, the anomalous subsidence reduces the lifting condensations and cloud occurrence frequency, as the negative anomalous frequencies of cloud radar reflectivity indicates reduction in cumulus and anvils ( figure 3(c)). Second, the anomalous subsidence also produces an anomalous anticyclone located in southeastern China and cyclone appears in the western North Pacific. Their northeasterly winds dominate southeastern China, which cuts down the prevailing moisture transports from the South China Sea at 850 hPa, as shown in figure 1(c). This results in a negative anomalous precipitation in southeastern China and a positive anomalous precipitation in tropical oceans ( figure 1(c)).
The responses of circulation and precipitation to the shortwave cloud heating over southeastern China are weaker than those to the longwave cloud heating ( figure 1(d)). Ignoring shortwave cloud heating causes warmer conditions at low-and mid-levels that tends to enhance the local instability, however, the changes in temperature are weak ( figure 2(d)). Therefore, the cloud longwave radiative heating dominates the overall cloud radiative impacts on EASM circulations ( figure 1(e)).
Why does the local cloud radiative heating lead to such a huge impact on the EASM circulation over southeastern China? We use thermodynamic energy Ignoring horizontal advection, the thermodynamic energy equation can be simplified as (Ackerman et al 1988, Mather et al 2007: where Q represents the heating rates, w means the vertical velocity and H is the scale height. The Brunt-Vaisala frequency is defined by N , (1) shows that the radiative heating is balanced by the adiabatic cooling associated with the updraft. In other words, the absence of cloud radiative heating tends to produce an anomalous subsidence.
Over southeastern China, the cloud radiative heating is generally reproduced by CAM5 (figures 4(a) and (b)). However, CAM5 shows the averaged net radiative heating rate caused by clouds is approximately 0.55 K day −1 at 6 ∼ 12 km, which is larger than 0.47 K day −1 in observation (figures 4(a) and (b)). Moreover, the simulated location of the maximum radiative heating is higher compared with observations, which is consistent with the vertical profiles of cloud and EASM circulation (figures 2 and 3). When the cloud radiative effects are ignored (NR run), there is a radiative cooling anomaly of about −0.5 K day −1 compared with the control run (figure 4(c)) and most of this anomaly comes from the longwave cloud radiative effects (not shown), similar to what we found in figures 1-2. Based on equation (1), such a radiative deficit generates an anomalous subsidence rate of about 8 hPa day −1 with its peak appearing at around 6 ∼ 8 km, where the maximum radiative cooling occurs ( figure 4(c)). This contributes substantially to EASM circulation, since the updraft of EASM circulation is about −20 hPa day −1 over southeastern China in CAM5.
The changes in southward upper outflow of EASM can be explained by Sverdrup vorticity balances (Zhou et al 2009), that a meridional wind (v) tends to respond to the subsidence caused by cloud radiative cooling anomaly. The Sverdrup vorticity balance is (Rodwell and Hoskins 2001): is the Coriolis parameter, a 2 cos β = Ω φ is its meridional gradient and p ω ∂ ∂ represents vertical velocity gradient of omega. Thus, northward winds are generated above the maximum center of cloud radiative cooling anomaly, because of the positive vertical gradient of omega ( figure 4(c)). The vertical circulation forced by cloud radiative cooling anomaly has the opposite structure to EASM circulation (figure 4(c)).

Conclusion
Based on the sensitivity experiments of CAM5 model, we show evidence that the cloud radiative heating over southeastern China (105°-125°E, 20°-35°N) has significantly positive impacts on the local EASM circulation and precipitation. When cloud radiative heating is ignored over southeastern China, the model produces an anomalous subsidence and a negative temperature anomaly. Meanwhile, anomalous anticyclone and cyclone occur in southeastern China and the western North Pacific, respectively, which block the low level moisture transport from tropical oceans. These result in a weak EASM. This is because cloud heats the atmosphere at cloud base and cools the atmosphere at cloud top. Based on thermodynamic energy equation, absence of cloud radiative heating induces an anomalous radiative cooling of −0.5 K day −1 at the mid-level troposphere, which increases the atmospheric stability and would be balanced by a strong anomalous subsidence of about 8 hPa day −1 . The anomalous subsidence subsequently reduces the occurrence frequency of cloud and precipitation and results in a local anomalous anticyclone and a weak subtropical high. Moreover, it further reduces the meridional wind through Sverdrup vorticity balance at high levels of the troposphere. The anomalous vertical circulation caused by absence of cloud radiative heating is opposite to EASM vertical circulation. Therefore, cloud radiative heating leads to a positive effect on the EASM.
Our results imply that clouds and their radiative heating over southeastern China are also significant factors in the simulations of EASM. Bias in clouds may lead to significant bias in simulating EASM circulation and precipitation. Clarifying the importance of cloud radiative heating will not only provide a useful understanding of EASM, but also a possible way to reduce the bias in the simulation of EASM.