Effects of Diurnal Cycle on a Simulated Asian Summer Monsoon

2 downloads 253 Views 4MB Size Report
Dec 15, 2012 - MYUNG-SEO KOO .... which can be divided into two subsystems, the Indian (or .... insula, Bay of Bengal, and the western coast of India.
8394

JOURNAL OF CLIMATE

VOLUME 25

Effects of Diurnal Cycle on a Simulated Asian Summer Monsoon SONG-YOU HONG Department of Atmospheric Sciences and Global Environment Laboratory, Yonsei University, Seoul, South Korea

MASAO KANAMITSU Scripps Institution of Oceanography, University of California, San Diego, San Diego, California

JUNG-EUN KIM Cooperative Institute for Research in the Atmosphere, Colorado State University, Fort Collins, and National Oceanic and Atmospheric Administration/Earth System Research Laboratory, Boulder, Colorado

MYUNG-SEO KOO Department of Atmospheric Sciences and Global Environment Laboratory, Yonsei University, Seoul, South Korea (Manuscript received 27 January 2012, in final form 31 May 2012) ABSTRACT This study investigates the effects of the diurnal cycle on monsoonal circulations over Asia in summer with a focus on precipitation. To this end, two sets of experiments are designed in a regional climate modeling framework forced by reanalysis data. The control experiment is a normal integration in which radiation is computed hourly, whereas the no-diurnal experiment is an experimental integration in which the daily averaged solar flux is computed once a day. Analysis of the results from the two experiments reveals that the diurnal cycle enhances the daily averaged sensible heat flux over land and the latent flux over oceans, which means that daytime net solar heating exceeds nighttime cooling in terms of the effects in surface climate and monsoonal circulations. Seasonal precipitation increased by about 3% over land and 11% over oceans. The surface hydroclimate over land is strongly influenced by the interaction between land and the atmosphere, and results in cooler surface temperatures except over the Tibetan Plateau. Over oceans, a robust increase in precipitation results from enhanced planetary boundary layer mixing. The diurnal cycle over the Tibetan Plateau region is found to decrease surface albedo by melting snow during the daytime, which contributes to the formation of the thermal low near the surface and the Tibetan high in the upper troposphere. The resultant monsoonal precipitation is modulated by an increase (decrease) in precipitation over northern (southern) India. This modulation results in the summer monsoon over East Asia being shifted northward.

1. Introduction The diurnal cycle at the surface and in atmosphere has been of great interest to atmospheric scientists in part due to uncertainty in the underlying dynamical and thermodynamical characteristics of the phenomenon (Dai and Trenberth 2004). In early studies, the characteristics of the diurnal cycle of precipitation have been identified by employing the analysis of gauge observation

Corresponding author address: Song-You Hong, Department of Atmospheric Sciences, Yonsei University, Seoul 120-749, South Korea. E-mail: [email protected] DOI: 10.1175/JCLI-D-12-00069.1 Ó 2012 American Meteorological Society

data. For example, Brier (1965) demonstrated that precipitation variations in the United States are related to solar and lunar tidal forces. Wallace (1975) classified the diurnal cycles of precipitation and thunderstorm frequency over the United States and summarized their general characteristics. McGarry and Reed (1978) investigated the phase and amplitude of diurnal variations in convective activity during the Global Atmospheric Research Program (GARP) Atlantic Tropical Experiment (GATE). These observational studies revealed the maximum of precipitation to occur in the late afternoon to early evening hours over land and at night or in the early morning over oceans. However, these general tendencies have been shown to be somewhat modulated by

15 DECEMBER 2012

HONG ET AL.

geographical (e.g., Dai 2001) and seasonal (e.g., Oki and Musiake 1994) characteristics. Diurnal oscillation studies have rapidly progressed with the development of improved methods for estimating precipitation at fine spatial resolutions, such as the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) (Huffman et al. 2007). Moreover, specific precipitation characteristics such as amplitude, phase, and frequency have been analyzed for convection systems over geographically different regions including over the tropics (Takayabu 2002), China (Zhou et al. 2008), East Asia (Koo et al. 2009), and South America (Su et al. 2008). The modeling community has also addressed the diurnal variation of precipitation. For example, Randall et al. (1991), Dai and Trenberth (2004), and Basu (2007) evaluated the diurnal cycle of precipitation in a general circulation model (GCM) system. These model evaluation studies can enhance our understanding of the important mechanisms that drive the diurnal cycle and validate various physical parameterizations. Zhou and Wang (2006) studied the geographic effect and offshore migration of diurnal cycle over the New Guinean region using a high-resolution regional climate model (RCM), which was further extended to a global cloud-resolving model by Sato et al. (2009). Wang et al. (2007) investigated the effect of lateral convective entrainment/detrainment rates on a simulated diurnal cycle of precipitation over the Maritime Continent and the surrounding oceans using a RCM. Koo and Hong (2010) investigated the diurnal variations of precipitation over East Asia, as simulated by two RCMs, and found that the cumulus parameterization modulates the simulated phase at maximum precipitation over land, whereas the amplitude is more highly controlled by the boundary layer processes. In addition to the evaluation studies of atmospheric models in reproducing the phase and amplitude of precipitation, it has been also recognized that the diurnal cycle of precipitation is an inevitable aspect of the climate system (Neale and Slingo 2003). Neale and Slingo conducted an experimental run without the islands of the Maritime Continent in a GCM and demonstrated that the Maritime Continent strongly influences the western Pacific in a simulated global climatology. They emphasized the need for a realistic representation of convective organization over regions that have complex land–sea terrains. Their results also implied that low-resolution GCMs poorly represent mesoscale circulations having a distinct diurnal cycle and therefore reflect degraded global large-scale circulations. Sato et al. (2008) later demonstrated that a resolution of less than 7 km is necessary to realistically simulate the phase of the precipitation diurnal cycle over the Tibetan Plateau (TP).

8395

Wapler et al. (2010) proved that a cloud-resolving model with grid spacing of a few kilometers is necessary to reproduce intense localized systems that are representative of diurnally forced continental convection over the Maritime Continent. These studies reinforce the fact that diurnal forcing strongly affects precipitation mechanisms over complex land–sea terrains, which in turn modulate large-scale circulations and embedded dynamics such as intraseasonal oscillations. The purpose of this study is to examine the role of the diurnal cycle on the summer monsoon circulations over Asia, focusing on precipitation. Despite increasing our understanding of the effects of diurnal characteristics on precipitating convection and associated mechanisms, previous studies do not clarify how the diurnal effect of solar forcing influences large-scale monsoonal circulations in observational or modeling studies. Although a physical link between precipitation and diurnal temperature range has been established through the analysis of station observation data from around the globe (e.g., Dai et al. 1999), the circulations that result from interactions between precipitation and temperature have not been elucidated. The Asian summer monsoon (ASM), which can be divided into two subsystems, the Indian (or South Asian) summer monsoon (ISM) and the East Asian summer monsoon (EASM), is the largest, most energetic, and most complex monsoon system in the world. The ASM system is basically caused by thermal contrasts between land and ocean (e.g., Tian and Yasunari 1992; Zhao et al. 2007) but the contribution of the diurnal cycle to summer monsoon precipitation has not been quantified. As a heat source in the warm season, the TP is the region of highest near-surface potential temperature. As a result, air moves inwards and upward along the flanks of the plateau and tends to yield a cyclonic circulation near the surface (Hoskins 1991). Therefore, discerning the effect of diurnal variation on the monsoonal system is scientifically interesting and feasible. To isolate the effect of diurnal variation on the monsoon circulations, we utilize a RCM system whose largescale forcing is prescribed by reanalysis data. Two sets of experiments are performed: specifically one employing the diurnal cycle and one excluding the diurnal variation of solar flux. The differences between the simulation results of the two runs are expected to elucidate the impact of the diurnal cycle on ASM circulations. The model, experimental setup, and selected case are described in section 2. The results of our control experiment are evaluated in section 3. Section 4 addresses the changes in precipitation and other large-scale features of the ASM system that are caused by the diurnal cycle. Finally, we summarize our findings and provide concluding remarks in section 5.

8396

JOURNAL OF CLIMATE

VOLUME 25

2. Model and experimental setup a. Model setup The National Centers for Environmental Prediction (NCEP) Regional Spectral Model (RSM) employs twodimensional cosine series to spectrally represent perturbations of pressure, divergence, temperature, and mixing ratio, and a two-dimensional sine series to represent the perturbation of vorticity. The linear computations of horizontal diffusion and semi-implicit adjustment are considered only for the perturbations so as to eliminate the error that would arise from reevaluating the linear forcing from the base fields. Details of the spectral representation and perturbation methods of the NCEP RSM are given in Juang and Kanamitsu (1994) and Juang et al. (1997). The physical package employed in the RSM is well documented in Yhang and Hong (2008). It includes longwave and shortwave radiative transfer by considering the interaction between clouds and radiation, the planetary boundary layer (PBL) process, deep and shallow convections, large-scale precipitation, gravity wave drag, a hydrology model, and vertical and horizontal diffusion. The model domain, centered at 27.58N, 1058E, includes the whole TP and covers both the EASM and ISM regions as well as the Indian and western North Pacific Oceans (Fig. 1a). The model grid consists of 151 (west–east) by 112 (north–south) grid lines with a horizontal separation of approximately 60 km on a Mercator projection and a vertical spacing of 28 sigma layers. This is the typical resolution used by the regional climate modeling community. Initial conditions and large-scale forcing are obtained from the 6-hourly NCEP–Department of Energy (DOE) Atmospheric Model Intercomparison Project (AMIP) II reanalysis (R-2) data (Kanamitsu et al. 2002). Note that the large-scale data from R-2 are needed over the domain of interest since the RSM computes some of dynamical processes on a perturbation, which is the deviation of the RSM forecasts from the derived lowerresolution base fields. This base field from the R-2 will be referred to as large-scale forcing in this study. In addition, the spectral nudging technique of Kanamitsu et al. (2010) was applied to the simulated large-scale fields to suppress a possible synoptic-scale drift. Note that this spectral nudging is relatively weak since the nudging is applied to the vorticity component only for wave lengths greater than 1000 km. Observed sea surface temperature (SST) is updated daily from the optimum interpolation SST (OISST) weekly dataset (Reynolds and Smith 1994). We chose the integration period from April to July but, because the major monsoonal rain period in the Indo-Pacific and East Asia

FIG. 1. (a) Model domain with orography, contours at 500-m intervals and shadings greater than 3000 m, and (b) TMPA precipitation (mm month21) during the period May–July (MJJ) 2004. The inner three dotted boxes in (a) are the Tibetan (solid), Indian (dashed), and East Asian (dash-dotted) regions used for skill score calculations in Table 1.

comprises only the latter three months, we discarded the month of April and verified only May–July (MJJ).

b. Case description A typical summer in East Asia is characterized by long-lasting rainy days during June and July accompanied by a quasi-stationary monsoon front that is referred to as mei-yu in China, changma in South Korea, and baiu in Japan, and a hot spell from late July to mid-August. Over South Asia the rainy and hot weather extends into September (Ding and Chan 2005). This study evaluates the year 2004, which was comprehensively reviewed by Levinson et al. (2005). This year had the summer that

15 DECEMBER 2012

HONG ET AL.

showed typical monsoon behavior in East Asia with a slightly dry condition and only limited regional-scale impacts associated with the El Nin˜o–Southern Oscillation (ENSO) event. This case is further supported by the extensive discussion given by Song et al. (2010) on the observational characteristics and basic evaluation of RSM results. Song et al. conducted a series of numerical experiments using the RSM with various TP heights ranging from a flat surface to 140% of the actual height and investigated the thermal and dynamical effects of the TP on monsoon circulations. In the 2004 MJJ, in addition to the persistent rainband associated with the intertropical convergence zone (ITCZ) near 158N over the western North Pacific, two monsoonal precipitation bands are distinct, as shown in the TMPA dataset (Huffman et al. 2007) (Fig. 1b). One is the ISM, which extends northward from the tropical monsoon areas near the equator to the Indochina peninsula, Bay of Bengal, and the western coast of India. The other one is the precipitation band associated with the EASM, which includes rainfall over southeastern China, the mei-yu over the Yangtze River Valley, the changma in Korea, and the baiu in Japan. The beginning of the local rainy season in these regions is well associated with strengthening of the spring southwesterly wind, tropospheric upward motion, and the convergence of low-level water vapor over southeastern China (Zhao et al. 2007).

8397

two experiments. However, in cloudy conditions, daily averaged fluxes at the surface can differ through cloud– radiation interactions since cloud information for radiation properties is updated when the radiation module is activated, which is every 1 h and 24 h for the CTL and NDI runs, respectively. In the model, there are two types of cloudiness for the inclusion of cloud effects on radiation. Convective cloud amount, due to the convective parameterization scheme, is the averaged quantity for preceding period, whereas the amount of stratiform clouds, due to the large-scale motion, is a function of instantaneous value of relative humidity when the radiation module is activated [see Shimpo et al. (2008) for further details]. Thus, the calculated cloud effect on radiation in the CTL run is equivalent to the average of the solar radiation calculated 24 times at the end of each hour, using the modeled cloud amount every hour. One may argue that the calculated cloudiness in the NDI run has a strong dependency upon local time when the largescale information is updated since the stratiform clouds use the instantaneous value of relative humidity. To avoid the aforementioned issue on cloud–radiation interactions and a possible dependency of simulated climatology on the large-scale data when updated, four sets of experiments with initial times of 0600, 1200, and 1800 UTC 31 March and 0000 UTC 1 April 2004 are designed. Validation of the model setup is given in the following section.

c. Experimental setup The conventional regional climate experiment forced by the analyzed data updates the large-scale forcing at the lateral boundaries as frequently as possible. For example, Song et al. (2010) used the R-2 data to provide large-scale forcing every 6 h. The shortwave and long wave radiation fluxes are computed every hour. The control (CTL) run with the diurnal cycle follows the conventional approach, but the analyzed data from the R-2 are updated every 24 h to remove the diurnal variation of large-scale forcing at the lateral boundaries. The nodiurnal (NDI) run is the same as the CTL run but without the diurnal variation of solar forcing in the radiation algorithm. To exclude the diurnal variation of solar forcing, the radiation fluxes are computed every 24 h when the R-2 data are updated, and the resulting daily-mean solar flux is prescribed during the integration for a given date. That is, the NDI simulation maintains a constant amount of solar heating throughout the day. The resulting solar flux is a function of solar declaration angle and latitudes, but not the zenith angle. In clear-sky conditions, the daily-averaged amount of solar radiative fluxes at the top of the atmosphere (TOA) and at the surface for a given date is the same in the

3. Validation of model setup and control (CTL) simulation We first validate the configuration of the model setup for the diurnal cycle experiment prior to evaluating its effects on the simulated Asian monsoon. The CTL experiment is shown to accurately reproduce precipitation and large-scale climatology, before the effects of diurnal variation on monsoon behavior are investigated. Largescale features from the CTL experiment are then evaluated against the R-2 data, and precipitation is evaluated against the TMPA observation data. Figure 2a compares the diurnal variation of downward solar fluxes at ground level from the CTL and NDI runs. The CTL flux shows a sine curve during daytime and zeros at night, as expected, whereas the averaged flux from the NDI run is constant. Even though the same solar forcing is used in both the CTL and NDI experiments, the averaged solar flux from the CTL run is greater than that from the NDI run (Table 2), as discussed in detail in the next section. The simulation results from both CTL and NDI experiments are strongly affected by the local time of the large-scale update since the resulting solar flux is directly

8398

JOURNAL OF CLIMATE

VOLUME 25

FIG. 2. Diurnal variation of (a) downward shortwave flux (W m22) at surface for the 0000 UTC run averaged over land for MJJ 2004 and (b) surface temperature (8C) from the four experiments with the four different times for the large-scale update. The lines with open circles designate the diurnal cycle from the CTL experiments, and the corresponding straight lines designate the constant solar heating from the NDI experiments.

related to the local solar zenith angle at the time the radiation is computed. Accordingly, the surface temperature ranges from highest to lowest at the times of 1200, 0600, 1800, and 0000 UTC, in that order, due to the solar hour at which large-scale forcing is updated (Fig. 2b). This indicates that simulation results depend on the external large-scale forcing. Also, the cloud–radiation interaction is influenced since the calculated clouds for the radiation module depend on the solar hour at which the radiation scheme is called. However, for each of the large-scale inputs, the surface temperature results of the CTL and NDI runs maintain the same relative relationship, with those of the CTL experiment being warmer during the daytime and cooler during the nighttime than those of the NDI experiment. This consistent relationship between four experiments with different solar hours for updating the external forcing and modeled clouds reflects the justification of the model setup designed in this study, which provides confidence in investigating differences in precipitation and related meteorological features between the CTL and NDI runs. The assumption underlying the experimental setup is that the average results from the four CTL experiments, which imposed large-scale forcing at different local times, should very closely reproduce the observed climatology. The CTL experiment, which combines the results of the four external forcings, reproduces the seasonal mean precipitation fairly well (cf. Figs. 1b and 3). The RSM is able to capture the major precipitation over the Indian Ocean tropical monsoon near the equator, the western North Pacific near the Philippines, the ISM (western coast of India, northeastern India, and Bay of Bengal), and the East Asian subtropical monsoon. However, the local maximum over the Philippines is underestimated.

The monsoon band over East Asia is also displaced northward relative to observation. In terms of the amount, the model overestimates precipitation by about 11% over land and slightly underestimates oceanic convection by about 1%. The large-scale features associated with the ASM circulation are also well reproduced (Fig. 4). At 850 hPa (Figs. 4a,b) the westerly Somalia jet, which is important for warm and moist air transport toward India and South Asia, is simulated fairly well compared to the R-2 data. The slight dryness shown by the CTL simulation near southern India, the Indochina peninsula, and the South China Sea contributes to the dry bias in simulated precipitation (see Fig. 3). General patterns at the 500-hPa

FIG. 3. The monthly averaged precipitation (mm month21) during the period of MJJ 2004 from the CTL run.

15 DECEMBER 2012

HONG ET AL.

8399

FIG. 4. (a),(b) Relative humidity (%, shading) and wind vectors (m s21, arrows) at 850 hPa and (c),(d) temperature (8C, dotted) and geopotential height (gpm, solid) at 500 hPa from the (left) R-2 and (right) CTL.

geopotential height (GPH) and temperature fields of the CTL results are largely consistent with those of the R-2 data (Figs. 4c,d). For example, the model reproduces the two planetary-scale troughs in the midlatitudes and the northwestern Pacific high, as well as the thermal low over the TP. Figure 5 depicts the vertical profiles of differences in temperature and humidity between the CTL simulation and the R-2 data. Over land, the domain-averaged temperature from the CTL run has a maximum error of about 20.5 K near the 400-hPa level and a warm bias of 0.4 K near the surface (Fig. 5a). The unstable structure of the CTL simulation reflects the overestimated precipitation. The temperature bias of the CTL experiment is small over the oceans where it is less pronounced than over land, except near the surface where there is a cold bias. A distinct moisture deficit in the lower levels confirms that precipitation activities are exaggerated over land (Fig. 5b). Over oceans moistening near the surface may be due to the cold bias, causing an increase in relative humidity.

The different large-scale updates, plotted individually, are similar in pattern to that shown by the sum of the four simulations. However, the overall increase or decrease in the temperature/GPH varies in response to the solar hour of the corresponding external forcing that is updated daily. For instance, the 0000 UTC simulation shows a colder temperature at the 500-hPa level than for the other times, whereas the 1200 UTC run shows a warmer temperature than that of the sum of the four simulations, as shown in Fig. 4d. The plots in Figs. 4b and 4d closely reflect those of a conventional experimental setup in which the large-scale data are updated every 6 h (not shown). Further analyses of surface variables such as precipitation and horizontal surface fluxes also showed a consistent sensitivity across all sets of the external forcing experiment, confirming the appropriateness of the model setup employed herein. Thus, only the sum of the four simulation results will be discussed in the following section. Also worth considering is the intraseasonal variation of diurnal effects on the formation of the seasonal climate.

8400

JOURNAL OF CLIMATE

VOLUME 25

FIG. 5. Vertical profiles for (a) temperature (8C) and (b) relative humidity (%) averaged over land and oceans for CTL minus R-2.

Figure 6 shows the daily evolution of precipitation during the three-month integration period. One can tell that the RSM is capable of reproducing the observed seasonal evolution of daily rainfall. Three peaks in precipitation over land that occurred around mid-May, midJune, and July are reasonably well simulated by both CTL and NDI experiments. The periods of break over the oceans in late May and early July are also well reproduced. A close inspection of the trend reveals that

the NDI experiment reduces the amount of precipitation over both land and oceans as compared to the CTL simulation. The differences in the two simulations do not exhibit a specific intraseasonal evolution, which consequently leads to reliability in changes in the seasonal mean due to the diurnal cycle. In terms of the diurnal variation in precipitation (not shown), the simulated trend over land from the CTL run showed the afternoon peak about 2 h ahead of that seen in the TMPA

FIG. 6. Time series of daily precipitation (mm day21) averaged over (a) land and (b) oceans.

15 DECEMBER 2012

HONG ET AL.

8401

FIG. 7. Changes in (a) precipitation (mm month21), (b) MMJ-averaged surface temperature (8C), and (c) sensible and (d) latent heat fluxes (W m22) due to the diurnal cycle (CTL minus NDI).

precipitation analysis observation, which is a deficiency in the RSM simulation (Koo and Hong 2010). Similarly, over the oceans the CTL run showed a morning peak and evening minimum about 2 and 8 h early, respectively, relative to those seen in the TMPA observation. Consequently, in the following section we will examine the seasonal characteristics of diurnal effects that are accumulated for three months.

4. Effects of the diurnal cycle (CTL minus NDI) The effects of the diurnal cycle on surface properties and upper-level features will be separately analyzed over land and oceans due to fundamental differences between surface and atmospheric interactions over these environments. The local effect of diurnal forcing on the change in precipitation is assessed by changes in surface fluxes, planetary boundary layer, and precipitation processes. An assessment of the feedback between changes in precipitation and associated monsoonal circulations is

also attempted. Finally, a dynamic influence of the Tibetan high on the monsoon climate is discussed.

a. Overall impacts Figure 7 shows the differences in surface variables that are due to the diurnal cycle. Results for precipitation, cloudiness, surface temperature, and PBL height are tabulated in Table 1, and the surface energy results are shown in Table 2. Figure 7a shows the amount of precipitation to increase considerably under the influence of the diurnal cycle except over the southeastern Asian region including the Indochina peninsula, the Philippines, and adjacent oceans. From Table 1, it is clear that the increase in precipitation is more pronounced over the oceans than over land. The amount of monsoonal precipitation is enhanced over India, although the band of precipitation is shifted northward over East Asia. The diurnal cycle drives a general cooling of the surface temperature except for over the TP and Mongolia (Fig. 7b). The warming seen over the TP is localized and prominent

8402

JOURNAL OF CLIMATE

TABLE 1. Changes in precipitation (mm day21), and cloudiness (%) for high (H), middle (M), and low (L) clouds, surface temperature (Tsfc; 8C), and PBL height (PBL; m) due to the diurnal cycle (CTL minus NDI), and their percentages (parentheses). Cloudiness

Whole Land Ocean Tibet Indo-Pacific East Asia

Precipitation

H

M

L

Tsfc

PBL

0.32 (7.62) 0.12 (3.09) 0.51 (10.9) 0.71 (27.2) 0.60 (12.6) 0.05 (1.1)

22.1 (28.1) 21.4 (26.7) 22.6 (29.2) 1.4 (8.7) 21.1 (24.4) 22.1 (27.9)

20.8 (23.3) 0.2 (0.7) 21.6 (26.7) 4.6 (18.5) 0.7 (2.6) 21.5 (26.6)

20.5 (22.7) 20.2 (21.3) 20.8 (23.8) 20.1 (23.4) 0.2 (0.8) 20.6 (23.0)

20.6 (20.2) (21.2) (20.4) 20.0 (20.0) 0.48 (0.18) 20.3 (20.1) 20.7 (20.2)

131.9 (27.2) 66.7 (9.5) 191.0 (65.3) 186.7 (33.8) 145.6 (27.1) 105.2 (26.1)

with a magnitude greater than 3 K. A distinct increase in sensible heat flux is seen over land, especially at high latitudes (Fig. 7c), whereas the enhanced latent flux is more pronounced over the oceans (Fig. 7d). Interestingly, the areas of enhanced latent flux and increased precipitation are largely coincident over the Indian Ocean, but the overlap is not seen elsewhere. The diurnal cycles of differences in surface fluxes and precipitation are contoured in Fig. 8. Specifically, the three-month averages of a field for a fixed local time, incremented at 1-h intervals, are shown. In Fig. 8a, we can see that the CTL run demonstrates a typical diurnal evolution of precipitation over land with an afternoon peak and suppressed nighttime activity, whereas the corresponding simulation from the NDI run shows a nearly constant precipitation intensity. A detailed examination of precipitation from the NDI run revealed

VOLUME 25

a diurnal variation due to natural variability, but its magnitude was very small, as seen in the figure. It appears that the amount of increased precipitation during the daytime overwhelms the nighttime deficit, such that an overall increase is seen in seasonal precipitation, as depicted in Fig. 7a. The sensible and latent heat fluxes from the CTL run exhibit a typical diurnal cycle in response to the solar forcing (Figs. 8b,c), whereas the averaged flux from the CTL run is greater than that of the NDI run (see Table 2). When the diurnal cycle is taken into account, precipitation over the oceans is intense during both daytime and nighttime periods (Fig. 8d). The enhanced nighttime precipitation can be explained by the increase in relative humidity that is due to cooler temperatures over the oceans. Specifically, the cooler nighttime temperatures yield a low-level structure that is conducive to the initiation of precipitating convection, as was found by Dai (2001). Accordingly, the minimum precipitation seen in the CTL run, relative to the NDI run, can be attributed to the reduced relative humidity in the evening. Because the SST is fixed within each day, the variation of sensible heat flux is not as distinct over the oceans as over land (Fig. 8e), but the addition of latent heat flux is noticeable over both land and the oceans when diurnal forcing is applied (Fig. 8f). The variation of latent flux is seen to follow the sensible heat flux with a maximum in the early evening and a minimum in the early morning. Wind speed and the PBL height also maintained a consistent relative variation. This confirms that boundary layer evolution is affected by solar forcing, as further discussed below.

b. A physical link for the enhanced precipitation The vertical profiles of the differences in temperature and humidity are shown in Fig. 9. The diurnal cycle causes

TABLE 2. Changes in downward solar flux at the surface (DSW), upward solar flux at the surface (USW), downward longwave flux at the surface (DLW), upward longwave flux at the surface (ULW), net radiative flux at the surface (Rnet), ground heat flux (GH), sensible heat flux (SH), latent heat flux (LH), albedo, and water equivalent of accumulated snow depth (Snow; kg m22) due to the diurnal cycle (CTL minus NDI), and their percentages (parentheses).

Whole Land Ocean Tibet Indo-Pacific East Asia

DSW

USW

DLW

ULW

Rnet

GH

SH

LH

Albedo

Snow

12.7 (4.7) 9.7 (3.4) 15.4 (6.0) 7.2 (2.1) 10.1 (3.8) 13.9 (4.9)

0.1 (0.2) 22.4 (22.8) 2.3 (15.7) 212.8 (29.2) 21.6 (22.9) 1.0 (2.26)

20.5 (20.1) 20.2 (20.1) 20.7 (20.2) 6.2 (2.6) 1.0 (0.3) 21.2 (20.3)

22.9 (20.6) 26.1 (21.4) 20.0 (20.0) 4.2 (1.2) 21.6 (20.3) 23.6 (20.8)

15.0 (10.3) 17.9 (17.5) 12.4 (6.7) 22.3 (29.5) 14.3 (11.0) 15.3 (9.0)

0.1 (1.9) 0.2 (1.9) 0.0 (202.9) 1.1 (8.1) 0.2 (7.5) 0.2 (4.6)

6.9 (28.6) 16.5 (40.4) 21.8 (219.5) 18.8 (55.9) 7.0 (26.3) 4.4 (22.5)

6.7 (8.7) 2.8 (5.2) 10.2 (10.4) 3.5 (14.3) 9.4 (11.2) 8.4 (12.0)

20.3 (22.0) 20.7 (22.4) 20.0 (20.1) 23.4 (28.8) 20.7 (23.8) 20.1 (20.9)

20.5 (240.4) 20.9 (276.9) 20.1 (26.4) 24.7 (279.0) 21.0 (279.0) 20.1 (240.7)

15 DECEMBER 2012

HONG ET AL.

8403

FIG. 8. Diurnal variation of (a),(d) precipitation (mm h21), (b),(e) sensible heat flux (W m22), and (c),(f) latent heat flux (W m22) over (left) land and (right) oceans for CTL minus NDI.

warming in the lower troposphere over land, with a maximum of about 0.3 K at 850 hPa, and a slight cooling near the surface (Fig. 9a). This warming can be explained by the increase in sensible heat flux from the surface that is driven by the diurnal effect and that causes enhanced mixing of the boundary layer. The PBL height is increased by 9.5% (Table 1). The enhanced turbulent mixing also results in a reduction in moisture near the surface (Fig. 9b). The diurnal cycle also results in positive heat fluxes that transport heat and moisture upward and

result in the increase in moist static energy near the surface. This addition of heat and moisture in the lower troposphere helps to enhance convective initiation and thereby increase the amount of precipitation that reaches the ground. The diurnal effect also enhances PBL mixing over the oceans, as evidenced by increases in temperature near the surface and cooling at 900 hPa (Fig. 9a). There is a prominent increase in PBL height (Table 1), which is normally less than 1 km over the oceans. Because it

8404

JOURNAL OF CLIMATE

VOLUME 25

FIG. 9. As in Fig. 5 but for CTL minus NDI.

causes humidity to decrease near the surface and increase above, the diurnal cycle generally enhances PBL mixing (Fig. 9b). This enhanced mixing transports moisture from the ocean into the atmosphere, but does not significantly affect the sensible heat flux because SST is fixed. The accumulative moisture near the surface during daytime results in cloud formation when the temperature cools during nighttime. Thus, the enhanced PBL mixing due to the diurnal cycle results in the increase in oceanic precipitation shown in Fig. 7a. Although precipitation is increased, cloudiness is decreased under the influence of the diurnal cycle (Table 1), and higher elevation clouds are reduced more than lower elevation clouds. This indicates that the precipitation in the NDI run is produced in association with more clouds than is precipitation in the CTL run. In the absence of the diurnal cycle, clouds are constantly present, of a moderate intensity, and result in a lesser amount of precipitation than that yielded by the deeply precipitating, sporadically developed clouds associated with the diurnal cycle. This indicates that, as in nature, the amount of detrained water vapor varies by convection type and can be smaller than that of clouds with a constant precipitation rate that develop in the absence of the diurnal cycle. The increase in downward shortwave flux that reaches the surface confirms the overall effect of the diurnal cycle. That is, the increase in surface precipitation that accompanies surface cooling is balanced with the reduction in cloudiness associated with the increase in downward shortwave radiation that enhances surface flux.

It is interesting to note that the diurnal cycle results in a cooler land surface, relative to that seen in the NDI run, even though the net radiation is greater (Tables 1 and 2). Figure 2 shows that, in the CTL run, the nighttime cooling overcomes the daytime heating to affect an average surface temperature that is lower than the constant temperature of the NDI run. Moreover, there are persistent turbulent motions in the NDI run due to constant shortwave radiation, which in the CTL run is turned off during nighttime. Nighttime in the CTL run provides continuous cooling until sunrise, which overwhelms daytime sinusoidal solar heating, whereas the NDI simulation maintains a constant amount of solar heating throughout the day.

c. Impacts on dynamic circulations In the previous subsection, we examined a physical link between the enhanced precipitation over land and oceans. However, precipitation is distinctly reduced over some areas centered over the Indochina peninsula, South China, and Japan, as seen in Fig. 7a. We attempt to interpret this reduction as a dynamical feedback between the direct, local response and modulated large-scale circulations. A comparison of Figs. 7a and 10 shows that areas of enhanced precipitation over the oceans largely coincide with convergence areas. Enhanced daytime PBL mixing due to the diurnal cycle is the primary driver of the increased precipitation shown in Fig. 8a, but there could be subsequent feedback caused by changes in the largescale circulation that increase moisture in other locations.

15 DECEMBER 2012

HONG ET AL.

FIG. 10. Differences in the MJJ 850-hPa wind (arrows) and vertically integrated moisture convergence (shading).

Northerly flows into the equatorial western Pacific, where precipitation is prominently increased, and southerly flows in East Asia induce a divergent area over South China, the Philippines, and the adjacent oceans where precipitation is decreased. The monsoonal precipitation band in East Asia is shifted northward. The enhanced northward transport of moisture in East Asia can be associated with the dynamical effect of the TP that is caused by the diurnal cycle, as seen in Fig. 11. Figure 11a shows that the thermal effects of diurnal forcing result in prevailing surface heating over the TP. As discussed in Yanai and Wu (2006),

8405

surface heating over the TP produces cyclonic (positive) vorticity in the shallow lower layer and anticyclonic (negative) vorticity in the deep upper layers aloft. These anomalous features appear in Fig. 11a, albeit with a small magnitude. To compensate the negative vorticity anomaly in the upper troposphere above the TP, alternative positive vorticity anomalies centered at 278 and 398N are generated. Over the northern slope in particular, it appears in a deeper layer than near the Himalayas due to increased surface heating. The change in upper-level wind speed is less than 1 m s21, but diverging flows and an increase in the geopotential heights are still visible (Fig. 11b), indicating that the magnitude of the TP high is slightly influenced by the diurnal cycle. Thus, we can say that the enhanced thermal forcing over the TP associated with the thermal high in the upper troposphere is caused by the diurnal cycle. This enhanced thermal forcing results in a dynamical influence on monsoon circulations over India and East Asia. Song et al. (2010) also examined the effects of the TP on the Asian monsoon by changing the height of the TP and verified that the TP enhances precipitation in northern India both thermodynamically and dynamically. They also confirmed that the TP presence shifts the East Asian summer monsoon northward, which in turn decreases precipitation over South China. These features are reproduced here by the diurnal cycle. Thus, it is concluded that the enhanced thermal forcing of the TP that is due to the diurnal cycle further modulates the distribution of precipitation by changing dynamical circulations.

FIG. 11. Difference of (a) temperature (8C, shading) and absolute vorticity (106 s21, contours) for Tibet averaged over 778–1028E and (b) surface temperature (8C, shading) and 200-hPa wind vectors (m s21, arrows) and geopotential height (thick solid lines).

8406

JOURNAL OF CLIMATE

VOLUME 25

FIG. 12. Differences in transient eddies (1024 s21, shading) and advection (m s21, arrows) for (a) temperature (8C) and (b) moisture (g kg21) averaged for MJJ.

Finally, the warming over the TP, which is not seen over other land areas, must be explained (Fig. 7a). The warming can be attributed to the changes in albedo and snow (Table 2). The amount of snow over the TP is decreased by 79% when the diurnal cycle is considered, as the daytime surface temperature over that region exceeds the freezing level and the snow melts in the CTL run. The melting effect cools the surface, but the enhanced solar heating due to the decrease in albedo results in a net increase in surface temperature. Meanwhile, snow keeps accumulated in the NDI run since a surface temperature is consistently below freezing level over the TP and it causes the increase in surface albedo. It can be deduced that transient eddies that result from the diurnal cycle also contribute to the warming seen over the TP (Fig. 12). Eddies are computed by averaging the daily mean of the diurnal deviation of the advection quantity for 3 months. Eddies of large magnitude appear over complex terrain such as that of the TP. Diurnal flux forcing is also significant over Korea, Japan, and eastern Siberia where mountain slopes are great. The convergence of heat and moisture over the TP is pronounced, which can also contribute to increases in surface temperature and precipitation.

5. Concluding remarks We have examined the effects of the diurnal cycle on the monsoonal circulations over Asia in summer using a RCM with and without the diurnal cycle. The analysis of the results from the two experiments reveals that the diurnal cycle enhances sensible heat flux over land and latent heat flux over the oceans. The corresponding increase in downward solar flux at the surface with the

reduction in cloudiness, especially of high clouds, is found to result from the diurnal cycle. Seasonal precipitation is increased by about 3% over land and 11% over the oceans, and the surface hydroclimate over land is strongly influenced by the interaction between land and the atmosphere, which results in cooler surface temperatures except over the TP. Over the oceans the diurnal cycle elicits a distinct increase in precipitation by enhancing planetary boundary layer mixing. The diurnal cycle is also found to contribute to the formation of the Tibetan high in the upper troposphere, which consequently influences the East Asian monsoon and the Indian monsoon climate. This study does not address the changes in the planetaryscale system that are attributable to the diurnal cycle, due to not using a GCM approach; therefore the study results may be called into question. However, the GCM approach has its own set of limitations, such as the decrease in predictability associated with integration time and the coarse resolution at which GCMs conducted. The RCM approach used in our study has advantages such as high resolution, the ability to maintain realistic boundary conditions, and large-scale forcing, which make it possible to isolate regional feedback from the model solution. Therefore, although precise quantitative measurement of the changes in the monsoonal circulation, due to the diurnal cycle, may not be possible because of the lack of interaction with large-scale forcing, the principal role of the diurnal cycle in the ASM can be firmly deduced. Still, despite the shortcomings of the GCM approach, we feel that the same kind of experiments should be applied to the global model setup to understand the resulting climatology within a fully interactive largescale system.

15 DECEMBER 2012

HONG ET AL.

Another limitation of our study is that it examines only the single year 2004, although it is a near-normal year in terms of ASM activity. To develop robust and quantitative measures of the impacts of the diurnal cycle on the TP, further investigations of a long-term period using a RCM are required. Also, we indicate that the 60-km resolution is not high enough to resolve the diurnal cycle of precipitation correctly. Our results can be further robust for the model setup at higher resolutions with the improved physics package. Nevertheless, this study is valuable because it is the first attempt at using a RCM to investigate the diurnal cycle effect. Acknowledgments. This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (20120000158), and by the Korea Meteorological Administration Research and Development Program under Grant CATER 2012-3084.

REFERENCES Basu, B. K., 2007: Diurnal variation in precipitation over India during the summer monsoon season: Observed and model predicted. Mon. Wea. Rev., 135, 2155–2167. Brier, G. W., 1965: Diurnal and semidiurnal atmospheric tides in relation to precipitation variations. Mon. Wea. Rev., 93, 93–100. Dai, A., 2001: Global precipitation and thunderstorm frequencies. Part II: Diurnal variations. J. Climate, 14, 1112–1128. ——, and K. E. Trenberth, 2004: The diurnal cycle and its depiction in the Community Climate System Model. J. Climate, 17, 930– 951. ——, F. Giorgi, and K. E. Trenberth, 1999: Observed and modelsimulated diurnal cycles of precipitation over the contiguous United States. J. Geophys. Res., 104 (D6), 6377–6402. Ding, Y. H., and J. C. Chan, 2005: The East Asian summer monsoon: An overview. Meteor. Atmos. Phys., 89, 117–142. Hoskins, B. J., 1991: Towards a PV-u view of the general circulation. Tellus, 43A, 27–35. Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 38–55. Juang, H.-M. H., and M. Kanamitsu, 1994: The NMC nested regional spectral model. Mon. Wea. Rev., 122, 3–26. ——, S.-Y. Hong, and M. Kanamitsu, 1997: The NCEP regional spectral model: An update. Bull. Amer. Meteor. Soc., 78, 2125– 2143. Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. M. Fiorino, and G. L. Potter, 2002: NCEP-DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643. ——, K. Yoshimura, Y.-B. Yhang, and S.-Y. Hong, 2010: Errors of interannual variability and trend in dynamical downscaling of reanalysis. J. Geophys. Res., 115, D17115, doi:10.1029/ 2009JD013511.

8407

Koo, M.-S., and S.-Y. Hong, 2010: Diurnal variations of simulated precipitation over East Asia in two regional climate models. J. Geophys. Res., 115, D05105, doi:10.1029/ 2009JD012574. ——, ——, and J. Kim, 2009: An evaluation of the Tropical Rainfall Measuring Mission (TRMM) Multi-Satellite Precipitation Analysis (TMPA) data over South Korea. Asia-Pac. J. Atmos. Sci., 45, 265–282. Levinson, D. H., and Coauthors, 2005: State of the climate in 2004. Bull. Amer. Meteor. Soc., 86, S1–S86. McGarry, M. M., and R. J. Reed, 1978: Diurnal variations in convective activity and precipitation during phases II and III of GATE. Mon. Wea. Rev., 106, 101–113. Neale, R., and J. Slingo, 2003: The maritime continent and its role in the global climate: A GCM study. J. Climate, 16, 834–848. Oki, T., and K. Musiake, 1994: Seasonal change of the diurnal cycle of precipitation over Japan and Malaysia. J. Appl. Meteor., 33, 1445–1463. Randall, D. A., Harshvardhan, and D. A. Dazlich, 1991: Diurnal variability of the hydrological cycle in a general circulation model. J. Atmos. Sci., 48, 40–62. Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate, 7, 929–948. Sato, T., T. Yoshikane, M. Satoh, H. Miura, and H. Fujinami, 2008: Resolution dependency of the diurnal cycle of convective clouds over the Tibetan Plateau in a mesoscale model. J. Meteor. Soc. Japan, 86A, 17–31. ——, H. Miura, M. Saatoh, Y. N. Takayabu, and Y. Wang, 2009: Diurnal cycle of precipitation in the tropics simulated in a global cloud-resolving model. J. Climate, 22, 4809– 4826. Shimpo, A., M. Kanamitsu, S. F. Iacobellis, and S.-Y. Hong, 2008: Comparison of four cloud schemes in simulating the seasonal mean field forced by the observed sea surface temperature. Mon. Wea. Rev., 136, 2557–2575. Song, J.-H., H.-S. Kang, Y.-H. Byun, and S.-Y. Hong, 2010: Effects of the Tibetan Plateau on the Asian summer monsoon: a numerical case study using a regional climate model. Int. J. Climatol., 30, 743–759. Su, F., Y. Hong, and D. P. Lettenmaier, 2008: Evaluation of TRMM multisatellite precipitation analysis (TMPA) and its utility in hydrologic prediction in the La Plata basin. J. Hydrometeor., 9, 622–640. Takayabu, Y. N., 2002: Spectral representation of rain profiles and diurnal variations observed with TRMM PR over the equatorial area. Geophys. Res. Lett., 29, 1584, doi:10.1029/ 2001GL014113. Tian, S.-F., and T. Yasunari, 1992: Time and space structure of interannual variations in summer rainfall over China. J. Meteor. Soc. Japan, 70, 585–596. Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103, 406–419. Wang, Y., L. Zhou, and K. Hamilton, 2007: Effect of convective entrainment/detrainment on the simulation of the tropical precipitation diurnal cycle. Mon. Wea. Rev., 135, 567–585. Wapler, K., T. P. Lane, P. T. May, C. Jakob, M. J. Manton, and S. T. Siems, 2010: Cloud-system-resolving model simulations of tropical cloud systems observed during the Tropical Warm Pool-International Cloud Experiment. Mon. Wea. Rev., 138, 55–73.

8408

JOURNAL OF CLIMATE

Yanai, M., and G. X. Wu, 2006: Effects of the Tibetan Plateau. The Asian Monsoon, B. Wang, Ed., Springer, 513–549. Yhang, Y. B., and S.-Y. Hong, 2008: Improved physical processes in a regional climate model and their impact on the simulated summer monsoon circulations over East Asia. J. Climate, 21, 963–979. Zhao, P., R. Zhang, J. Liu, X. Zhou, and J. He, 2007: Onset of southwesterly wind over eastern China and associated atmospheric circulation and rainfall. Climate Dyn., 28, 797–811.

VOLUME 25

Zhou, L., and Y. Wang, 2006: Tropical Rainfall Measuring Mission observation and regional model study of precipitation diurnal cycle in the New Guinean region. J. Geophys. Res., 111, D17104, doi:10.1029/2006JD007243. Zhou, T., R. Yu, H. Chen, A. Dai, and Y. Pan, 2008: Summer precipitation frequency, intensity, and diurnal cycle over China: A comparison of satellite data with rain gauge observations. J. Climate, 21, 3997–4010.