Abstract

Knowing the relationship between local convective precipitation and boundary layer processes is critical for forecasting rainstorms. For the South China Sea area, such a forecast is particularly important during the monsoon season. During such a season, the authors examined the boundary layer features at three sites as part of the South China Sea Monsoon Experiment—Boundary Layer Height (SCSMEX-BLH) experiment. The sites are spread from inland to over sea along a 43.4-km line. Here the authors analyze SCSMEX-BLH data from an intensive observing period that includes a convectively suppressed (inactive) period, a period influenced by a tropical storm, and a convectively active monsoon period. Some preliminary findings include the following: 1) The absorption of shortwave radiation over the sea is the primary driver of the land–sea temperature difference. The difference produces a diurnal variation below 400 m, with a warmer surface layer over the coast at night. 2) In the inactive and storm periods, the sensible heat flux is larger than that in the active period, whereas in the active period, the heat flux (primarily latent heat flux) over sea is significant. Also in the active period, the depth of the mixed layer inland is smaller than that in other periods, but the depth on the coast is always higher than that in other periods. 3) In the active period at night, as a monsoon vapor surge advects horizontally over the warm sea surface, a large latent heat flux driven by strong winds aids the growth of marine cumulus, which eventually develop into inland cumulonimbus bringing inland rainfall.

1. Introduction

The Asian–Australian monsoon is the largest monsoon system in the world, with a major portion of it being the East Asian summer monsoon (EASM) (Tao and Chen 1987; Wang 2006). A possible precursor to EASM development is the South China Sea summer monsoon (SCSSM) (Tao and Chen 1987; Lau and Yang 1997). For example, the occurrence of droughts and floods in southeast China has a close relationship with the activities of the EASM and the onset of the SCSSM (Ye and Huang 1991; Zhao 1999; Ding et al. 2005).

The original South China Sea Monsoon Experiment (SCSMEX) was a large-scale, multination, atmospheric, and oceanic field experiment (1996–2001) that aimed to better understand the onset, maintenance, and variability of the summer monsoon over the South China Sea (SCS) (Lau et al. 2000; Ding et al. 2004). A particular goal was to better understand the large-scale interactions of SCSSM—for example, the onset of the SCSSM, including possibly strong influences from extratropical systems as well as from convection over the Indian Ocean and the Bay of Bengal (Lau et al. 2000). The study found that a strong monsoon over the SCS usually leads to less precipitation over the middle and lower reaches of the Yangtze River basin but more precipitation in north China. A weak monsoon has the reverse precipitation trend (Ding et al. 2004).

Boundary layer processes and the sea surface temperature are crucial aspects of monsoon precipitation. The mean boundary layer structures over the central Arabian Sea and the north central Bay of Bengal had been examined in the Indo-Soviet Monsoon Experiment of 1977 (MONSOON 77) and the International Monsoon Experiments (MONSOON 79) (Holt and Raman 1986, 1987). Afterward, the Monsoon Trough Boundary Layer Experiment (MONTBLEX) focused on atmospheric boundary layer processes in the monsoon-trough area of northern India (Goel and Srivastava 1990). Later, Liu and Ding (2000) found that the depth of the mixed layer decreased, and even disappeared, both in the northern and southern SCS after the monsoon onset during SCSMEX. Yan et al. (2007) found a higher variation in the heat fluxes over the sea after the monsoon onset compared to that before. Mo and Juang (2003) analyzed the general features of monsoon rainfall and the diurnal cycle along the western slopes of the Sierra Madre Occidental during the North American monsoon. They found that a warm sea surface in the Gulf of California increased the rainfall (and vice versa). Later, Roxy (2014) pointed out that the precipitation will continually increase with increasing sea surface temperature, though a time lag exists between the two, and suggested that having realistic ocean–atmosphere coupling is crucial even for short-term monsoon weather forecasts.

How the atmospheric, land, and ocean systems communicate in the monsoon system affects the local precipitation characteristics (Webster et al. 1998). Use of an atmospheric general circulation model coupled with an ocean model cannot yet predict summer monsoon rainfall, a problem arising because the lower boundary conditions are poorly known (Wang et al. 2005). Therefore, we need more observations in the SCS area to better understand the surface–air interactions.

Accurate forecast of a rainstorm in a monsoon season still remains difficult. This difficulty may arise from the complexity of tropical convection (Betts 1974), the ever-changing relationships between stability and monsoon convection in the break and active monsoon events, as well as influences from tropical storms (McBride and Frank 1999). Webster et al. (1998) point out that the spatial distribution of convection in a monsoon season is far more complicated and suggest more observations to better understand the convection process. Furthermore, although local convective precipitation likely has a close relationship with the local air–land–sea boundary layer process (Huang 1986; Xue 1999; Higgins et al. 2006), few experiments have analyzed the characteristics of the boundary layer at the coastal area during the typhoon rainy season in southern China.

Here we analyze the atmospheric boundary layer height during the South China Sea monsoon period, as part of SCSMEX-BLH, in summer 2012. The SCSMEX-BLH study focuses on the local convective precipitation in the coastal area of the southern China coast, with vertically aligned measurements at three observational sites (inland, coastal, and over sea). We aim to better understand the relationship between the local convective precipitation and the boundary layer process during an inactive monsoon period, a tropical-storm period, and an active monsoon period. We find that the heat flux (primarily latent heat flux) over the sea plays an important role in the local precipitation in the active period.

2. Background

Regions of southern China have either a tropical- or a subtropical-maritime monsoon climate. Southern China, with an annual rainfall of 1000–2800 mm, has the most rainfall, the highest number of rainstorms, and the longest rain season in China. Thus, this region endures severe floods (Huang 1986; Shi 2003). Because SCS supplies water and supports the regional rainfall for southeast China, most of the floods are closely related to the SCSSM, which is primary climatic system in summer (Zhu et al. 1986; Tao and Chen 1987; Ding 1994; Simmonds et al. 1999; Wu et al. 2003). The flood season has two rainfall peaks. The first peak, the first rainy season (also called the “preflood” season), occurs from April to June. The second peak, the typhoon rainy season, occurs from July to September. Rainfall in the latter season mainly comes from tropical synoptic systems (Huang 1986; Xue 1999). According to the distribution of monthly rainfall amounts over southern China (1971–2012), the first rainy season contributes 42% and the typhoon season 33% to the annual total precipitation (Y. Luo 2015, personal communication). Assuming the SCSSM occurs from April to October (Tu and Huang 1944; Huang 1986), then the SCSSM can contribute over 75% to the annual total precipitation.

Synoptic systems always influence local precipitation. In the Gulf of California, moisture surges closely relate to tropical cyclones in the eastern Pacific basin (Higgins and Shi 2005). In the Indo-Gangetic plains, convective precipitation closely relates to the south Asian monsoon trough (Choudhury and Krishnan 2011). And in the typhoon rainy season on the southern China coast, the local precipitation characteristics relate closely to major weather systems such as tropical storms and the monsoon trough (Tao and Chen 1987; Huang and Tang 1996; Qian and Tang 2010; Chang et al. 2012). When two such weather systems appear one after another, severe torrential floods or debris flows occur on the coast (Xue 1999).

Apart from the influence of the synoptic systems, the local precipitation characteristics in the southern China coast relate closely to topographic features and coastal features. Features such as the windward slope and the convergence region of a mountain produce more rainfall. Furthermore, a low-level jet plays an important part in the rainy season—for example, by causing the rainfall to increase during the late night and morning (Huang 1986; Xue 1999; Chen et al. 2013). Although topographic features and the low-level jet clearly influence rainfall in the region, our focus here is on the local surface–air exchange and interactions. These processes are important during the typhoon rainy season.

An earlier rainstorm experiment in 1977–82 was run to understand the climatic and synoptic characteristics of rainstorms in the first rainy season (Huang 1986). However, few experiments have focused on the typhoon rainy season after mid-June. The experiment reported here aims to fill this gap.

3. Observational site and data

a. Observational sites

Instruments are mounted at three sites: The coastal site A (within 50 m of the sea), herein “coast site,” is the Marine Meteorological Science Experiment Base (MMSEB; Huang and Chan 2011) at Bohe, Maoming (21.45°N, 111.32°E, 7.0 m MSL), which is operated by the Institute of Tropical Marine and Meteorology (ITMM). The sea site B, herein “sea site” or “over sea,” is the Integrated Observation Platform for Marine Meteorology (IOPMM, 21.44°N, 111.39°E, operated by MMSEB). The inland site C, herein “inland site,” is the Dianbai National Climate Observatory (DNCO, 21.55°N, 110.99°E, 31.6 m MSL). The three sites lie along a straight line that is roughly along the prevailing southeasterly flow during the active period of the monsoon (Ding and Reiter 1983; Tao and Chen 1987) (Figs. 1 and 2). The distance from the inland to the coast site is about 35.2 km, whereas that from the coast to over sea is about 8.2 km. Hence, the whole cross section is about 43.4 km. Furthermore, the distance from the inland to the nearest coastline is about 11.2 km and from over sea to the nearest coastal line is about 6.5 km (Fig. 1b).

Fig. 1.

Observational site locations. (a) The coastal site is MMSEB (21.45°N, 111.32°E), the sea site is IOPMM (21.44°N, 111.39°E), and the inland site is DNCO (21.55°N, 110.99°E). (b) Distances from different locations and from coastal lines (from Google Earth).

Fig. 1.

Observational site locations. (a) The coastal site is MMSEB (21.45°N, 111.32°E), the sea site is IOPMM (21.44°N, 111.39°E), and the inland site is DNCO (21.55°N, 110.99°E). (b) Distances from different locations and from coastal lines (from Google Earth).

Fig. 2.

The three observational sites. (a) A bird’s-eye view of the coast site (MMSEB): a1 = observational center; a2 = office and data center; a3 = living quarters; a4 = Zhizai Island, a 100-m iron tower inside, not used in this experiment; and a5 = Dazhuzhou Island, and the IOPMM is behind it (main photo courtesy of Weikang Mao on 17 May 2012). (b) Oversea site (IOPMM), operated by ITMM. The upper-right corner of the image is the mounting position of the Gill Windmaster Pro and Li-cor 7500A instruments at the 23.4-m level in summer 2012 (main photo courtesy of Weikang Mao on 27 Jun 2014). (c) The inland site (DNCO) (main photo courtesy of Huijun Huang on 24 May 2012).

Fig. 2.

The three observational sites. (a) A bird’s-eye view of the coast site (MMSEB): a1 = observational center; a2 = office and data center; a3 = living quarters; a4 = Zhizai Island, a 100-m iron tower inside, not used in this experiment; and a5 = Dazhuzhou Island, and the IOPMM is behind it (main photo courtesy of Weikang Mao on 17 May 2012). (b) Oversea site (IOPMM), operated by ITMM. The upper-right corner of the image is the mounting position of the Gill Windmaster Pro and Li-cor 7500A instruments at the 23.4-m level in summer 2012 (main photo courtesy of Weikang Mao on 27 Jun 2014). (c) The inland site (DNCO) (main photo courtesy of Huijun Huang on 24 May 2012).

b. Observational data

Operations covered a period including one SCS summer monsoon, from May 2012 to July 2013. Table 1 is a summary of the observation sites, instruments, and data. The operationwide dataset includes a normal period from May 2012 to July 2013 and two 5-day intensive observing periods (IOPs), the first starting at 1400 LST 28 June 2012 and the second at 0800 LST 29 August 2012. These data include the wind, temperature, and humidity of the atmospheric boundary layer. Together with the air–land and air–sea surface data, the complete dataset reflects the characteristics of the atmospheric boundary layer from the inactive to the active monsoon period. The observation system includes two wind profilers (inland, coast), two iron towers with gradient- and flux-observation systems (inland, sea) (Fig. 2), and two GPS sonde observation systems (inland, coast). The iron tower inland is 70 m high and contains instruments to measure the gradients of wind, temperature, and humidity at three levels (10, 20, and 40 m) and instruments to measure fluxes on 10-m levels. Other instruments include an automatic weather system, radiometers, and instruments to measure ground temperature and humidity, skin temperature, and soil heat flux (Fig. 2c, Table 1). All the inland instruments can observe the air–land interaction to represent an inland site in this study.

Table 1.

The observation sites, instruments, and data.

The observation sites, instruments, and data.
The observation sites, instruments, and data.

The sea site lies above ~15 m of water, and the platform forms a regular triangle about 7 m on a side and about 11 m above sea level. The triangular platform has one vertex pointing north and the two others southwest and southeast. These southern corners each have a 10-m-high observational tower. The main observational tower is about 25 m high and has seven measurement-level heights. Measured from sea level, these heights are 13.4, 16.4, 20.0, 23.4, 27.3, 31.3, and 35.1 m (Fig. 2b). We collected data on wind direction, wind speed, atmospheric temperature, and humidity (at the 13.4-, 16.4-, 20.0-, 27.3-, and 35.1-m levels). In addition, we used a net radiometer at about 12 m above sea level, infrared CO2/H2O analyzers and ultrasonic anemometers at the 23.4- and 31.3-m levels, and a thermistor unit at 2, 4, and 6 m below the sea surface (Table 1). We calibrated the gradient of temperature and humidity at five levels and the sea temperature using a suite of thermometers and ventilated psychrometers. The datalogger at this site was synchronized to those at the other two sites prior to each observational period.

During the IOPs, we launched GPS Vaisala RS-92 sondes every 3 h simultaneously from the inland and coastal sites. The first IOP ran from 1400 LST 28 June to 1400 LST 3 July and the second from 0800 LST 29 August to 0800 LST 3 September. In total, 164 GPS sondes were released at 82 times (i.e., two at a time) during the two events. The sonde measurement accuracy for temperature is 0.2 K, for humidity 2%, for pressure 0.5 hPa, for wind direction 2°, and for wind speed 0.15 m s−1. The sampling rate is 0.5 Hz and the vertical resolution is 4–10 m. To achieve uniform data, we interpolate the GPS sonde data in the vertical with a 10-m interval and with a 3-h time interval. We use linear interpolated data to fill in two missed times (0200 LST 2 July and 0200 LST 3 July) at the coast. The interpolation use fore-and-aft time data and soundings from two sites to analyze their boundary layer structures. Calculation of the equivalent potential temperature follows Bolton (1980). All the meteorological variables obtained from the GPS sonde data are smoothed using a three-point smoothing method in the vertical direction. We also use the logbook entries and photos taken by the crew at coast and inland, which were taken at 0200, 0800, 1100, 1400, 1700, 2000, and 2300 LST.

We use Gill Windmaster Pro ultrasonic anemometers, with EddyPro 4.0 software (http://www.licor.com/env/products/eddy_covariance/software.html) to calculate the turbulence flux over time intervals of 30 min. The software uses the Foken et al. (2004) method to run quality control tests of the fluxes. The overall quality flags follow Spoleto agreement (Second CarboEurope QA/QC Workshop for Eddy-Covariance Measurements, Spoleto, Italy, January 2004) for CarboEurope-IP (http://www.bayceer.uni-bayreuth.de/qaqc/en/forschung/21826/QC_Spoleto.php), as shown in Table 2. We also use global reanalysis data on a 1° × 1° grid from the National Centers for Environmental Prediction (NCEP), obtained from the research data archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory. (http://rda.ucar.edu/datasets/ds083.2).

Table 2.

Overall quality flags of the EddyPro 4.0 software. RNcov is the parameter to evaluate the steady state of turbulence, and ITCσ is the parameter to evaluate the integral turbulence characteristic. More information is available in Foken et al. (2004).

Overall quality flags of the EddyPro 4.0 software. RNcov is the parameter to evaluate the steady state of turbulence, and ITCσ is the parameter to evaluate the integral turbulence characteristic. More information is available in Foken et al. (2004).
Overall quality flags of the EddyPro 4.0 software. RNcov is the parameter to evaluate the steady state of turbulence, and ITCσ is the parameter to evaluate the integral turbulence characteristic. More information is available in Foken et al. (2004).

c. Definition of boundary layer height scales

We define the mixed-layer depth as that of the nearly uniform layer, which we determine by the static stability method described in Stull (1988, his Fig. 5.17). We use a subjective determination of this depth, similar to that done by Johnson et al. (2001). The definition of virtual potential temperature used here is

 
formula

where is potential temperature, is the specific humidity, and is the liquid-water specific humidity. We set to zero for the calculation, as it is poorly known owing to a lack of measurements. Thereafter, we use the value (from GPS sonde) at all altitudes and subtract the mixed-layer value at each time to get the zero contour as the mixing depth indicator.

4. Synoptic background

In this study, we focus on the first IOP, which contains both convectively active and inactive monsoon conditions. The observational sites are affected by the active South China Sea summer monsoon before 1400 LST 28 June. Then, owing to the influence of the peripheral wind field of Tropical Storm Doksuri (2012), an inactive monsoon period follows (Fig. 3a). Later, the influence of the storm becomes dominant (Fig. 3b), with the storm making landfall at 0330 LST 30 June in the Nanshui town of Zhuhai (about 210 km east of the coast). After the storm, the South China Sea summer monsoon affects the observational sites beginning at 1400 LST 30 June, with a nearly uniform southeast wind direction and higher wind speed (Fig. 3c). The monsoon trough is at about 16°N. From 1400 LST 30 June to 1400 LST 1 July, the observational sites record a doubling in wind speed, being continuously affected by onshore monsoonal flow (Fig. 3d).

Fig. 3.

Wind vectors at 850 hPa and geopotential height at 500 hPa: (a) 1400 LST 28 Jun, (b) 2000 LST 29 Jun, (c) 1400 LST 30 Jun, and (d) 0200 LST 1 Jul. (From the NCEP Global Analysis on 1° × 1° grid spacing datasets.)

Fig. 3.

Wind vectors at 850 hPa and geopotential height at 500 hPa: (a) 1400 LST 28 Jun, (b) 2000 LST 29 Jun, (c) 1400 LST 30 Jun, and (d) 0200 LST 1 Jul. (From the NCEP Global Analysis on 1° × 1° grid spacing datasets.)

Two evident rainfall processes occur sequentially in the IOP1. The first one comes from the tropical storm and the second from the monsoon trough. Before either occurs, vapor advects horizontally toward the observation area, indicating an apparent vapor surge (Fig. 4), with a surge coming from the tropical storm being much stronger than that from the monsoon trough (Figs. 4a,b). The tropical storm leads to a squall line at the peripheral area, hitting the inland site at 2300 LST 29 June with a rainfall rate of 36.67 mm h−1 (Fig. 5a, Fig. 8e).

Fig. 4.

Average horizontal advection of vapor (10−4 g kg−1 s−1) in 1000–850 hPa and wind vectors at 1000 hPa: (a) 2000 LST 29 Jun and (b) 0200 LST 1 Jul. (From the NCEP Global Analysis on 1° × 1° grid spacing datasets.)

Fig. 4.

Average horizontal advection of vapor (10−4 g kg−1 s−1) in 1000–850 hPa and wind vectors at 1000 hPa: (a) 2000 LST 29 Jun and (b) 0200 LST 1 Jul. (From the NCEP Global Analysis on 1° × 1° grid spacing datasets.)

Fig. 5.

The composite radar reflectivity image of two precipitation cases: (a) for tropical-storm period at 2200 LST 29 Jun and (b) for active monsoon period at 0700 LST 1 Jul. (The black arrow shows the direction of cloud motion. The color bar shows the range of radar reflectivity. The image comes from Guangdong Province Observatory.)

Fig. 5.

The composite radar reflectivity image of two precipitation cases: (a) for tropical-storm period at 2200 LST 29 Jun and (b) for active monsoon period at 0700 LST 1 Jul. (The black arrow shows the direction of cloud motion. The color bar shows the range of radar reflectivity. The image comes from Guangdong Province Observatory.)

The rainfall caused by the monsoon trough can be distinguished from that caused by the tropical storm. During the active monsoon period, cumuli always occur over SCS and move inland across the coastal line, resulting in heavy convective precipitation from coast to inland sites. First, scattered cumuli form over the sea and move toward the coast. Then, some cumuli quickly develop and build into a larger cumulonimbus cloud while approaching the coast. Finally, these cumulonimbus reach the coast, bringing sudden rainfall (Fig. 5b). This case results in heavy rainfall (28.95 mm in one hour, 47.29 mm over the 3-h process) at the inland site (Fig. 8e).

We define the inactive and active monsoon period using a 55-yr average of wind speed of the SCSSM over the SCS region. At 850 hPa, the average speed of the prevailing wind is 5 m s−1 with standard deviation of about 1 m s−1. So for threshold values, we use 4 m s−1 for the inactive period and 6 m s−1 for the active period (Wu and Liang 2002; Gu et al. 2007). In this IOP1, the observational sites undergo three distinct synoptic periods: 1) an inactive monsoon period (herein, inactive period, defined as having a prevailing monsoon wind speed always below 4 m s−1 at 850 hPa), 2) a tropical-storm period (herein, storm period, defined as having a wind direction at 850 hPa that is largely due to tropical-storm processes), and 3), an active monsoon period (herein, active period, defined as having a prevailing monsoon wind speed always over 6 m s−1 at 850 hPa).

The wind direction (from the GPS sonde) varies similarly at inland and coast sites, and each period has distinct wind characteristics in the boundary layer. The inactive period has a south-southeast wind of speed below 8 m s−1 in the boundary layer. The storm period has a variable wind direction, whereas the active period has a uniform southeast wind, mostly exceeding 8 m s−1 (Fig. 6, Figs. 8a,b). The resulting inactive period is from 1400 LST 28 to 0200 LST 29 June, the storm period is from 0200 LST 29 to 1400 LST 30 June, and the active period is from 1400 LST 30 June to 1400 LST 03 July (Fig. 6). Although we consider the inactive period as the typical period overall, in this case (where it is caused by a typhoon), it is only 12 h long (i.e., not a full diurnal cycle). Thus, the observation sites are still affected by the previous active period, even at 0800 LST 28 June. Hence, because of the limited sampling of the inactive conditions, analyses from this period have larger uncertainties.

Fig. 6.

Wind velocity in the atmospheric boundary layer through the three periods: (a) inland and (b) coastal. The times marked with black squares have interpolated data. As marked across the top, the inactive period is from 1400 LST 28 to 0200 LST 29 Jun, the storm period is from 0200 LST 29 to 1400 LST 30 Jun, and the active period is from 1400 LST 30 Jun to 1400 LST 3 Jul.

Fig. 6.

Wind velocity in the atmospheric boundary layer through the three periods: (a) inland and (b) coastal. The times marked with black squares have interpolated data. As marked across the top, the inactive period is from 1400 LST 28 to 0200 LST 29 Jun, the storm period is from 0200 LST 29 to 1400 LST 30 Jun, and the active period is from 1400 LST 30 Jun to 1400 LST 3 Jul.

We now examine the synoptic backgrounds of the three periods. To do so, we determine profiles for the average equivalent potential temperature , temperature, and mixing ratio from the inactive period (8 profiles), the storm period (22 profiles), and the active period (52 profiles). These profiles are from both the coast and inland sites. Consider first their averaged and differential profiles of . Of the three periods, the storm period has the largest layer below the 743-hPa level, whereas the inactive period has the smallest layer in 743–352 hPa (Fig. 7a). Viewing the complete profiles, the storm period’s is nearly everywhere larger than that in the inactive period (Figs. 7a,b). However, the values of the active period are not always larger than of the inactive period. Specifically, the value of the active period is lower than of the inactive period in the following three layers: surface to 936, 853–747, and 352–162 hPa (Fig. 7b). The values in the storm period exceed those of the active period, which indicates a moister and warmer boundary layer in the storm period (Fig. 7b).

Fig. 7.

Average and differential equivalent potential temperature , differential temperature T, and mixing ratio r. (a) Average profiles of inactive period (8 profiles), storm period (22 profiles), and active period (52 profiles). (b)–(d) Differential , T, and r values from the storm period minus that of the inactive period and similarly for active period minus that of inactive period.

Fig. 7.

Average and differential equivalent potential temperature , differential temperature T, and mixing ratio r. (a) Average profiles of inactive period (8 profiles), storm period (22 profiles), and active period (52 profiles). (b)–(d) Differential , T, and r values from the storm period minus that of the inactive period and similarly for active period minus that of inactive period.

The temperature T and mixing-ratio r differences for these periods support the above picture. However, the temperature difference shows that the active period has a cooler boundary layer than that in the inactive period (Fig. 7c). Also, compared to that in the inactive period, the active period has a lower humidity near the surface (below 942 hPa) but a higher humidity in the upper boundary layer (above 942 hPa) (Fig. 7d). The largest difference among the three periods is the increase in vapor above the 942-hPa level in the storm and active periods, as compared to the inactive period.

5. Characteristics of atmospheric boundary layer

a. Meteorological variables

Here we examine the GPS sonde and rainfall data at the inland and coast. We compare the evolution of the main meteorological variables of the atmospheric boundary layer (from the sonde) and rainfall (from the automated meteorological station) between inland and coast (Figs. 89). The most evident phenomenon is a low-level jet at both sites that occurs in the boundary layer from 1800 LST 30 June to 1400 LST 1 July. The occurrence is just after the region becomes influenced by the tropical-storm circulation (Figs. 8a,b). The wind speed inland generally exceeds that at coast at heights of 200–1000 m, but, probably because of the greater surface friction, is less than that at coast below 200 m (Fig. 9a).

Fig. 8.

Evolution of the main meteorological variables of the atmospheric boundary layer (left) inland and (right) on the coast: (a),(b) wind speed (m s−1); (c),(d) temperature (K); (e),(f) surface precipitation (mm); (g),(h) relative humidity (%); (i),(j) mixing ratio (g kg−1); and (k),(l) equivalent potential temperature (K). (m),(n) Local deviation of , defined as the value at a given position minus the time-averaged value for the same position. The value (K) was averaged over the entire duration shown. White rectangles represent missed data and the black boxes on the abscissa indicate interpolated data. The period indicator on the abscissa is as in that for Fig. 6.

Fig. 8.

Evolution of the main meteorological variables of the atmospheric boundary layer (left) inland and (right) on the coast: (a),(b) wind speed (m s−1); (c),(d) temperature (K); (e),(f) surface precipitation (mm); (g),(h) relative humidity (%); (i),(j) mixing ratio (g kg−1); and (k),(l) equivalent potential temperature (K). (m),(n) Local deviation of , defined as the value at a given position minus the time-averaged value for the same position. The value (K) was averaged over the entire duration shown. White rectangles represent missed data and the black boxes on the abscissa indicate interpolated data. The period indicator on the abscissa is as in that for Fig. 6.

Fig. 9.

Difference of meteorological variables between inland and coast sites: (a) wind speed (m s−1), (b) temperature (K), (c) relative humidity (%), (d) mixing ratio (g kg−1), and (e) equivalent potential temperature (K). White rectangles represent missed data. The period indicator on the abscissa is as in that for Fig. 6.

Fig. 9.

Difference of meteorological variables between inland and coast sites: (a) wind speed (m s−1), (b) temperature (K), (c) relative humidity (%), (d) mixing ratio (g kg−1), and (e) equivalent potential temperature (K). White rectangles represent missed data. The period indicator on the abscissa is as in that for Fig. 6.

For inland and coast sites, the temperature is highest during the storm period in the boundary layer, but is generally lower during the active period than that in the previous inactive period (Figs. 8c,d). This lower temperature might be due to the larger wind speed and the evaporation of rain (Figs. 8e,f). Generally, the night temperature at coast exceeds that at inland and is less than at inland at daytime, especially below 400 m (Fig. 9b). The rainfall mainly occurs inland, not at the coast (Figs. 8e,f), because the cumulus always forms over the sea and then moves inland with the wind to develop into cumulonimbus. The maximum rainfall occurs during the storm period, when is largest, which represents the most unstable period (Fig. 5b, Figs. 8k,l).

Above 800 m, the relative humidities (RHs) are larger in the storm and active periods than in the inactive period (Figs. 8g,h), indicating more moisture. The increasing moisture is accompanied by sparse clouds (defined as RH > 98%) in the storm and active periods in the boundary layer. During the day, the relative humidity is larger at coast than inland, but at night the situation reverses, especially below 400 m (Fig. 9c). Compared to that at the coast, the larger temperature diurnal cycle at the inland site occurs because it has the lower thermal capacity. In the daytime, the surface temperature inland exceeds that at the coast, resulting in its lower relative humidity. But at night, the surface temperature inland decreases quickly, going below that at the coast, leading to a greater increase in relative humidity inland than at coast. The increasing depth of the moisture layer can be seen in the mixing-ratio pattern, which shows the greatest depth in the storm period and least depth in the inactive period (Figs. 8i,j). The increasing moisture-layer depth during the active period was also found by Lau et al. (2000), who also pointed out a deepened moist layer in the lower troposphere, signaling the onset of the monsoon. The difference of mixing ratio between coast and inland is more complicated. A diurnal variation also exists but does not show as clear a pattern as the relative humidity (Fig. 9d).

Because it has the largest values of temperature and humidity, the storm period also has the largest values (Figs. 8k,l). Below 600 m, the values in the inactive period exceed those in the active period, which is consistent with Fig. 7b. Nevertheless, the values have a different pattern between coast and inland sites. A more subtle variation can be seen from the local deviation of , defined as the value at a given position minus the time-averaged value for the same position (Figs. 8m,n). For example, two vertical cooling zones below 800 m occur inland, but not at the coast. The first vertical cooling zone occurs from about 0700 to 1400 LST 1 July, just after receiving 47.29 mm of rain over 3 h. Evaporation within the precipitation downdraft apparently lowers the temperature, resulting in a vertical cooling zone (Figs. 8c,m). A stronger vertical cooling zone occurs from about 2100 LST 1 July to 0100 LST 2 July because of longwave radiative cooling over land at night (Fig. 8m). This cooling agrees with Betts and Albrecht (1987), who showed that radiative cooling could lower values. In contrast, at the coast, no evident vertical cooling zones exist below 800 m in the active period because of the lower precipitation rate and the lower longwave radiative cooling over the coast at night (Fig. 8n). The difference varies diurnally below 400 m, with the coast having a larger-value layer at night and a lower-value layer in daytime, compared to inland (Fig. 9e). The larger -value layer at coast at night indicates a warmer surface layer over the coast that will suppress the radiative cooling effect and thus promote thermal mixing at night.

b. Characteristics of the air–land and air–sea interfaces

To better understand the air–surface interfaces, we now examine the surface radiometer data inland and over sea. The downward shortwave radiation (DSR) inland is similar to that over sea, except for an evident deviation on 3 July (LST). The reduction of DSR over the sea site on 3 July (LST) is due to an increase in cloudiness (Fig. 10a). The largest difference between the sites comes from the upward shortwave radiation (USR). Inland, the USR (average value 78.5 W m−2) greatly exceeds that over sea (average value 15.7 W m−2) during daytime (Table 3). This difference in USR between the inland and oversea sites is primarily due to the higher albedo inland. Similar characteristics appear in the longwave radiation. During daytime, the upward longwave radiation (ULR) is larger inland (average value 480.4 W m−2) than over sea (average value 469.7 W m−2) (Fig. 10b, Table 3). However, at night, the ULR over sea (average value 469.3 W m−2) is larger than that inland (average value 454.2 W m−2) as a result of the air temperature at night being higher over sea (Fig. 9b, Fig. 10b, Table 3). Averaging over 24 h, the ULRs are nearly the same at both sites (Table 3). Meanwhile, the downward longwave radiation (DLR) has a similar diurnal variation and average value for inland and over sea (Fig. 10b, Table 3). Therefore, the USR accounts for the difference of the averaged net radiation between inland and over sea. In other words, the absorption of shortwave radiation over sea is the primary cause of the land–sea difference during this period (Table 3).

Fig. 10.

Radiation inland and over sea. (a) Shortwave radiation. DSR_land = downward shortwave radiation inland; USR_land = upward shortwave radiation inland; DSR_sea = downward shortwave radiation over sea; and USR_sea = upward shortwave radiation over sea. (b) Longwave radiation. The notation is similar to that for the shortwave radiation. Averaging interval is 10 min. The period indicator on the abscissa is as in that for Fig. 6.

Fig. 10.

Radiation inland and over sea. (a) Shortwave radiation. DSR_land = downward shortwave radiation inland; USR_land = upward shortwave radiation inland; DSR_sea = downward shortwave radiation over sea; and USR_sea = upward shortwave radiation over sea. (b) Longwave radiation. The notation is similar to that for the shortwave radiation. Averaging interval is 10 min. The period indicator on the abscissa is as in that for Fig. 6.

Table 3.

Average radiation values inland and over sea, Rn is the net radiation (W m−2, from 1400 LST 28 Jun to 1400 LST 3 Jul).

Average radiation values inland and over sea, Rn is the net radiation (W m−2, from 1400 LST 28 Jun to 1400 LST 3 Jul).
Average radiation values inland and over sea, Rn is the net radiation (W m−2, from 1400 LST 28 Jun to 1400 LST 3 Jul).

The momentum flux inland always exceeds that over sea (Table 4). This flux at the inland site has less variation during the inactive and storm periods but has a larger variation during the active period, especially when the active period starts (Fig. 11a). However, the momentum flux over sea has different features, showing two peaks. One peak occurs during the storm period and the other at the start of the active period (Fig. 11a). Also, the average momentum flux over sea increases when the wind speed increases in the storm and active periods (Table 4, Fig. 8b).

Table 4.

Average momentum flux (Tau), sensible heat flux (Hs), and latent heat flux (LE) values inland and over sea (from 1400 LST 28 Jun to 1400 LST 3 Jul).

Average momentum flux (Tau), sensible heat flux (Hs), and latent heat flux (LE) values inland and over sea (from 1400 LST 28 Jun to 1400 LST 3 Jul).
Average momentum flux (Tau), sensible heat flux (Hs), and latent heat flux (LE) values inland and over sea (from 1400 LST 28 Jun to 1400 LST 3 Jul).
Fig. 11.

Fluxes at (left) inland and (right) over sea. Land measurements are 10 m above the surface and sea measurements are 23.4 m above the surface: (a) tau = momentum flux (kg m−1 s−2), (b) Hs = sensible heat flux (W m−2), and (c) LE = latent heat flux (W m−2). Averaging interval for the fluxes is 30 min. Missing data are either out of range or doubtful. Dotted colors mark the quality flags of the fluxes, as depicted in Table 2. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

Fig. 11.

Fluxes at (left) inland and (right) over sea. Land measurements are 10 m above the surface and sea measurements are 23.4 m above the surface: (a) tau = momentum flux (kg m−1 s−2), (b) Hs = sensible heat flux (W m−2), and (c) LE = latent heat flux (W m−2). Averaging interval for the fluxes is 30 min. Missing data are either out of range or doubtful. Dotted colors mark the quality flags of the fluxes, as depicted in Table 2. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

Consider now the sensible (Hs) and latent heat (LE) fluxes inland. Both fluxes have large diurnal variations, with respective averages of 10.75 and 77.57 W m−2. The sensible heat flux is less than the latent heat flux and is larger in the inactive period (averaging 21.49 W m−2) and the storm period (averaging 22.63 W m−2) than in the active period (averaging 3.08 W m−2). The lower average in the active period is due to a negative (downward) sensible heat flux at night (Fig. 11b). Similar to the sensible heat flux, the inland latent heat flux is largest in the storm period. But unlike the sensible heat flux, the latent heat flux is larger in the active period than in the inactive period. This behavior is consistent with the mixing ratio near the surface being smaller in the active than in the inactive period (Table 4, Fig. 7d). The time series of mixing ratio in Fig. 12a for the inland site supports this argument. The mixing ratio in the active period is, on average, 0.85 g kg−1 less than that in the inactive period (Fig. 12a). The drier near-surface air in the active period would result in a larger near-surface mixing-ratio gradient and, thus, larger latent heat fluxes during the active period.

Fig. 12.

Time series of mixing ratio and vertical velocity at inland and sea sites. (a) Mixing ratio (g kg−1); land measurements are 10 m above the surface and sea measurements are 13.4 m above the surface. (b) Vertical velocity (m s−1); land measurements are 10 m above the surface and sea measurements are 23.4 m above the surface. Suspect data are not shown, producing the gap. Averaging interval for the data is 30 min. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

Fig. 12.

Time series of mixing ratio and vertical velocity at inland and sea sites. (a) Mixing ratio (g kg−1); land measurements are 10 m above the surface and sea measurements are 13.4 m above the surface. (b) Vertical velocity (m s−1); land measurements are 10 m above the surface and sea measurements are 23.4 m above the surface. Suspect data are not shown, producing the gap. Averaging interval for the data is 30 min. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

The heat flux over sea has different characteristics. Unlike the case inland, the sensible and latent heat fluxes have no apparent diurnal variation. But in going from the inactive to active period, the sensible heat flux shows a decreasing trend, whereas the latent heat flux has an increasing trend. Furthermore, the largest latent heat flux value over sea is in the active period (Table 4, Fig. 11c). By comparing the heat flux between inland and over sea, we find that the heat flux inland is important in the inactive and storm periods.

However, the heat flux, primarily the latent heat flux, over sea is particularly large in the active period (Table 4). The large value might be due to a large decrease of mixing ratio over the sea, which indicates that the vapor surge near the sea surface was drier during the active period than that in the inactive period. After the inactive period, the mixing ratio decreases by an average of 1.49 g kg−1 in the active period (Fig. 12a). Moreover, the behavior of the average vertical velocity is quite different between inland and sea. (The vertical velocity is measured at different heights at the two sites, but we found that the vertical velocities at the higher level over the sea have the same trend as that measured at roughly the same height as that over the inland site.) Moreover, the two sites have different trends between the inactive and active periods. Inland has more descending airflow in the active period and less in the inactive period (decreasing on average by 0.05 m s−1). Over sea has more ascending airflow in the active period than in the inactive period (increasing on average by 0.07 m s−1) (Fig. 12b). That the sea site has more ascending airflow while the inland site has more descending airflow indicates that more convergence occurs over the sea, which may promote cumulus formation and more divergence inland.

c. Depth of the mixed layer

The depth of the mixed layer is determined from the profile, which comes from the GPS sonde. Sondes are released at the inland and coast sites; however, because the coastal site is within 50 m of the sea, and because the prevailing wind comes from the sea, we consider the mixed layer at the coast the same as that over sea.

In the atmospheric boundary layer, has its largest value in the storm period before the precipitation both inland and at coast. The precipitation downdraft leads to lower values in the rest of the storm period and the following active period. The values near the surface in the active period are also less than those in the inactive period. The values inland also have greater diurnal variation than at coast (Figs. 13a,b). These features determine the different ways the mixed layer develops at the two sites. The depth of the mixed layer at inland peaks in the daytime and always disappears at night, whereas that at the coast has less diurnal amplitude and sometimes has a larger value at night (Figs. 13c,d). Also, the depth of the mixed layer inland decreases in the later active period, less than that in the inactive period (Fig. 13c). In contrast, the coast’s mixed-layer depth always has much larger values in the active period than in the inactive and storm periods, especially at night (Fig. 13d).

Fig. 13.

Evolution of the virtual potential temperature and the depth of the mixed layer (left) inland and (right) at coast: (a),(b) virtual potential temperature (K); and (c),(d) depth of mixed layer (m), the green zero line indicates the mixed-layer depth. (e) Average diurnal variation of the mixed-layer depth in the active period. White rectangles represent missed data and the black box indicates interpolated data. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

Fig. 13.

Evolution of the virtual potential temperature and the depth of the mixed layer (left) inland and (right) at coast: (a),(b) virtual potential temperature (K); and (c),(d) depth of mixed layer (m), the green zero line indicates the mixed-layer depth. (e) Average diurnal variation of the mixed-layer depth in the active period. White rectangles represent missed data and the black box indicates interpolated data. On the abscissa, the lightest gray is the inactive period, middling gray is the stormy period, and black is the active period.

Consider the diurnal variation of the mixed-layer depth in the active period. Inland, there are two peaks: one at 1100 LST and another at 1700 LST. In contrast, the coast has a continuous high value from 1400 to 0500 LST the next day, indicating thermal mixing at night (Fig. 13e). Such thermal mixing starts from the absorption of shortwave radiation in the day and continues at night as a result of the relatively warm SST over the sea.

6. Precipitation mechanism in the active monsoon period

Consider the rainfall from 0310 to 0640 LST 1 July. This case is typical of a rainfall influenced by the monsoon trough in IOP1. This local case of convective precipitation, in the active period, also had a strong wind and a close relationship with the boundary layer process (Figs. 8a,b,e,f). Now we consider the mechanism for this precipitation.

First, at night in the active period, the sea, with its larger heat capacity, remains warmer at night than the land (Fig. 14). In particular, the soil temperature at a 5-cm depth averages 0.94°C lower than the water temperature at about 2 m below the sea surface from 2350 LST 30 June to 1200 LST 1 July. Meanwhile, the air temperature at 13.4 m over the sea averages 1.2°C warmer than that at 10 m over land before the rain (i.e., from 1930 LST 30 June to 0300 LST 1 July). This leads to a high mixed layer over sea at night (Fig. 13d, Figs. 14 and 15). Then, most of the precipitation occurs from 0310 to 0640 LST 1 July inland, and the precipitation downdrafts produce strong cooling over land (Figs. 8c and 14).

Fig. 14.

Temperatures near air–land and air–sea interfaces. T_soil_5cm = soil temperature, 5 cm below land surface; SST = sea surface temperature, about 2 m below sea surface; T_land_10m = air temperature, 10 m over land surface; and T_sea_13.4m = air temperature, 13.4 m over sea surface. The little black rectangle just above the abscissa indicates the main precipitation time from 0310 to 0640 LST 1 Jul. On the abscissa, the middling gray is the stormy period and black is the active period.

Fig. 14.

Temperatures near air–land and air–sea interfaces. T_soil_5cm = soil temperature, 5 cm below land surface; SST = sea surface temperature, about 2 m below sea surface; T_land_10m = air temperature, 10 m over land surface; and T_sea_13.4m = air temperature, 13.4 m over sea surface. The little black rectangle just above the abscissa indicates the main precipitation time from 0310 to 0640 LST 1 Jul. On the abscissa, the middling gray is the stormy period and black is the active period.

Fig. 15.

Sketch of the local precipitation at night in the active period. Hs = sensible heat flux; LE = latent heat flux; w = vertical velocity; ML = mixed layer (dotted line); Cu hum = cumulus humilis; Cu con = cumulus congestus; Cb = cumulonimbus; Wd = wind direction; = virtual potential temperature (K) (solid line); and r = mixing ratio (g kg−1) (dashed line). Boundary layer profiles from GPS sonde at 0200 LST 1 Jul are shown at the sides, with the left being inland and the right being the coast.

Fig. 15.

Sketch of the local precipitation at night in the active period. Hs = sensible heat flux; LE = latent heat flux; w = vertical velocity; ML = mixed layer (dotted line); Cu hum = cumulus humilis; Cu con = cumulus congestus; Cb = cumulonimbus; Wd = wind direction; = virtual potential temperature (K) (solid line); and r = mixing ratio (g kg−1) (dashed line). Boundary layer profiles from GPS sonde at 0200 LST 1 Jul are shown at the sides, with the left being inland and the right being the coast.

Second, when a monsoon vapor surge strongly advects horizontally over the warm sea surface at night, the drier vapor surge near the sea surface produces a large latent heat flux over the sea (Figs. 11 and 12a). The large latent heat flux, together with the increasing upward vertical velocity, helps produce scattered marine cumulus humilis (Fig. 15). By using a large-eddy simulation and concepts from bulk theory, Nuijens and Stevens (2012) pointed out that the strong winds with shallow marine cumulus convection will increase the latent heat flux and decrease the sensible heat flux, moistening the air and lowering the cloud base. Their mechanism agrees with the observations here. The cumuli humiles develop over the sea surface with strong thermal mixing and a dominant latent heat flux from the sea surface, and then gradually become cumulus congestus. In another cloud-resolving, large-eddy simulation, Xu et al. (2010) showed that the warm sea surface will increase the surface latent heat flux, which is the primary mechanism for increasing the cloud geometric thickness, liquid-water content, liquid-water path, and cloud optical depth. Therefore, the cumulus congestus keep growing over the warm sea surface, developing into cumulonimbus, resulting in local precipitation over the relatively cool inland surface (Fig. 15).

7. Discussion and conclusions

We ran the SCSMEX-BLH experiment on the southern China coast to better understand the local monsoon precipitation. Two types of rainfall events occurred during IOP1: one during the storm period and the other during the active period. To better understand this rainfall, we examined its relationship to the boundary layer processes, obtaining the following main results:

  1. The different thermal capacity of the land and sea results in a clear diurnal variation of below 400 m, with the coast having a larger -value layer at night and a smaller-value layer at daytime than inland. This indicates a warmer surface layer over the coastal area at night especially in the active period.

  2. The absorption of shortwave radiation over sea is the primary cause of the land–sea temperature difference during this period. The sensible heat flux in the inactive and storm periods is larger than that in the active period. However, the heat flux (mostly the latent heat flux) over sea plays an important role in the active period. Inland, the depth of the mixed layer decreases in the later active period to below that in the inactive period. In contrast, the depth of the coastal mixed layer is always higher in the active period than in the inactive and storm periods, especially at night.

  3. A warm sea surface promotes the formation and development of marine cumulus. When a monsoon vapor surge strongly advects horizontally over the warm sea surface at night, the wind increases the dominant latent heat flux over the sea. This latent heat flux, with the help of the upward vertical velocity, aids the growth of marine cumulus. The marine cumuli later develop into cumulonimbus, producing inland rainfall.

Thus, we argued that the strong advection of vapor from the monsoon, passing over the warm sea surface, leads to local precipitation inland. This precipitation mechanism does not rely upon topographical features (Ogura and Yoshizaki 1988; Becker and Berbery 2008; Kerns et al. 2010; Tu et al. 2014) and is distinct from the island precipitation over the ocean such as that in Hawaii (Chen and Nash 1994; Esteban and Chen 2008) and in Sumatra (Mori et al. 2004; Wu et al. 2009) but, nevertheless, may also occur over tropical oceans. Although we found that the SCS has a key influence on local inland rainfall, this study covered only 5 days. Further study with additional cases would help determine the generality of these relationships. Also helpful would be a high-resolution simulation that includes the local characteristics of the land–sea and air–sea interaction. Nevertheless, the results may help clarify the relation between the coastal boundary layer processes and inland precipitation in the South China Sea.

Acknowledgments

We thank two anonymous reviewers whose constructive suggestions and comments greatly improved this presentation. Special thanks to the crew of the Marine Meteorological Science Experiment Base at Bohe for their help in conducting the field program and providing the data. This study was supported jointly by the National Key Basic Research Program of China (973 Program; Grants 2014CB953903 and 2011CB403501), the National Natural Science Foundation of China (Grants 41175013 and 41475061), the Guangdong Science and Technology Plan Project (2012A061400012), and the early warning and forecasting technology for marine meteorology of the Weather Bureau of Guangdong Province.

REFERENCES

REFERENCES
Becker
,
E. J.
, and
E. H.
Berbery
,
2008
:
The diurnal cycle of precipitation over the North American monsoon region during the NAME 2004 field campaign
.
J. Climate
,
21
,
771
787
, doi:.
Betts
,
A. K.
,
1974
:
Thermodynamic classification of tropical convective soundings
.
Mon. Wea. Rev.
,
102
,
760
764
, doi:.
Betts
,
A. K.
, and
B. A.
Albrecht
,
1987
:
Conserved variable analysis of the convective boundary layer thermodynamic structure over the tropical oceans
.
J. Atmos. Sci.
,
44
,
83
99
, doi:.
Bolton
,
D.
,
1980
:
The computation of equivalent potential temperature
.
Mon. Wea. Rev.
,
108
,
1046
1053
, doi:.
Chang
,
C.-P.
,
Y.
Lei
,
C.-H.
Sui
,
X.
Lin
, and
F.
Ren
,
2012
:
Tropical cyclone and extreme rainfall trends in East Asian summer monsoon since mid-20th century
.
Geophys. Res. Lett.
,
39
,
L18702
, doi:.
Chen
,
G.
,
W.
Sha
,
M.
Sawada
, and
T.
Iwasaki
,
2013
:
Influence of summer monsoon diurnal cycle on moisture transport and precipitation over eastern China
.
J. Geophys. Res. Atmos.
,
118
,
3163
3177
, doi:.
Chen
,
Y.-L.
, and
A. J.
Nash
,
1994
:
Diurnal variation of surface airflow and rainfall frequencies on the island of Hawaii
.
Mon. Wea. Rev.
,
122
,
34
56
, doi:.
Choudhury
,
A. D.
, and
R.
Krishnan
,
2011
:
Dynamical response of the South Asian monsoon trough to latent heating from stratiform and convective precipitation
.
J. Atmos. Sci.
,
68
,
1347
1363
, doi:.
Ding
,
Y. H.
,
1994
:
Monsoons over China
. Kluwer Academic Publisher, 419 pp.
Ding
,
Y. H.
, and
E. R.
Reiter
,
1983
:
Large-scale hemispheric teleconnections with the frequency of tropical cyclone formation over Northwest Pacific and North Atlantic Oceans
.
Arch. Meteor. Geophys. Bioclimatol.
,
32A
,
311
337
, doi:.
Ding
,
Y. H.
,
C. Y.
Li
, and
Y. J.
Liu
,
2004
:
Overview of the South China Sea Monsoon Experiment
.
Adv. Atmos. Sci.
,
21
,
343
360
, doi:.
Ding
,
Y. H.
, and Coauthors
,
2005
: Relationship between variations of the Asian monsoon and the anomalous climate in China. Evolution of Climate and Environment in China (Vol. 1, in Chinese), Science Press,
398
454
.
Esteban
,
M. A.
, and
Y.-L.
Chen
,
2008
:
The impact of trade wind strength on precipitation over the windward side of the island of Hawaii
.
Mon. Wea. Rev.
,
136
,
913
928
, doi:.
Foken
,
T.
,
M.
Gӧckede
,
M.
Mauder
,
L.
Mahrt
,
B.
Amiro
, and
W.
Munger
,
2004
: Post-field data quality control. Handbook of Micrometeorology: A Guide for Surface Flux Measurement and Analysis, X. Lee, W. Massman, and B. Law, Eds., Kluwer Academic Publishers, 181–208.
Goel
,
M.
, and
H. N.
Srivastava
,
1990
:
Monsoon Trough Boundary Layer Experiment (MONTBLEX)
.
Bull. Amer. Meteor. Soc.
,
71
,
1594
1600
, doi:.
Gu
,
D. J.
,
Z. P.
Ji
,
D. X.
Wang
,
J. Y.
Liang
, and
Y. C.
Fang
,
2007
:
The relationship between SCS summer monsoon intensity and oceanic thermodynamic variables at different time scale
.
J. Trop. Meteor.
,
13
,
85
88
.
Higgins
,
R. W.
, and
W.
Shi
,
2005
:
Relationships between Gulf of California moisture surges and tropical cyclones in the eastern Pacific basin
.
J. Climate
,
18
,
4601
4620
, doi:.
Higgins
,
R. W.
, and Coauthors
,
2006
:
The NAME 2004 field campaign and modeling strategy
.
Bull. Amer. Meteor. Soc.
,
87
,
79
94
, doi:.
Holt
,
T.
, and
S.
Raman
,
1986
:
Observations of the mean and turbulence structure of the marine boundary layer over the Bay of Bengal during MONEX 79
.
Mon. Wea. Rev.
,
114
,
2176
2190
, doi:.
Holt
,
T.
, and
S.
Raman
,
1987
:
A study of mean boundary-layer structures over the Arabian Sea and the Bay of Bengal during active and break monsoon periods
.
Bound.-Layer Meteor.
,
38
,
73
94
, doi:.
Huang
,
J.
, and
P.
Chan
,
2011
:
Progress of marine meteorological observation experiment at Maoming of South China
.
J. Trop. Meteor.
,
17
,
418
429
.
Huang
,
S. S.
,
1986
: Rainstorm of the Pre-Flood Season in Southern China (in Chinese). Guangdong Science and Technology Press, 244 pp.
Huang
,
S. S.
, and
M. M.
Tang
,
1996
:
The early summer flood periods of southern China and the summer monsoon circulation of East Asia
.
J. Trop. Meteor.
,
2
,
12
25
.
Johnson
,
R. H.
,
P. E.
Ciesielski
, and
J. A.
Cotturone
,
2001
:
Multiscale variability of the atmospheric mixed layer over the western Pacific warm pool
.
J. Atmos. Sci.
,
58
,
2729
2750
, doi:.
Kerns
,
J.
,
B.
Wesley
,
Y.-L.
Chen
, and
M.-Y.
Chang
,
2010
:
The diurnal cycle of winds, rain, and clouds over Taiwan during the mei-yu, summer, and autumn rainfall regimes
.
Mon. Wea. Rev.
,
138
,
497
516
, doi:.
Lau
,
K.-M.
, and
S.
Yang
,
1997
:
Climatology and interannual variability of the southeast Asian summer monsoon
.
Adv. Atmos. Sci.
,
14
,
141
162
, doi:.
Lau
,
K.-M.
, and Coauthors
,
2000
:
A report of the field operations and early results of the South China Sea Monsoon Experiment (SCSMEX)
.
Bull. Amer. Meteor. Soc.
,
81
,
1261
1270
, doi:.
Liu
,
Y. J.
, and
Y. H.
Ding
,
2000
:
Evolution of the atmospheric stratification and mixed layer before and after monsoon onset over the South China Sea (in Chinese)
.
Climate Environ. Res.
,
5
,
459
468
.
McBride
,
J. L.
, and
W. M.
Frank
,
1999
:
Relationships between stability and monsoon convection
.
J. Atmos. Sci.
,
56
,
24
36
, doi:.
Mo
,
K. C.
, and
H. M. H.
Juang
,
2003
:
Influence of sea surface temperature anomalies in the Gulf of California on North American monsoon rainfall
.
J. Geophys. Res.
,
108
,
4112
, doi:.
Mori
,
S.
, and Coauthors
,
2004
:
Diurnal land–sea rainfall peak migration over Sumatera Island, Indonesian Maritime Continent, observed by TRMM satellite and intensive rawinsonde soundings
.
Mon. Wea. Rev.
,
132
,
2021
2039
, doi:.
Nuijens
,
L.
, and
B.
Stevens
,
2012
:
The influence of wind speed on shallow marine cumulus convection
.
J. Atmos. Sci.
,
69
,
168
184
, doi:.
Ogura
,
Y.
, and
M.
Yoshizaki
,
1988
:
Numerical study of orographic-convective precipitation over the eastern Arabian Sea and the Ghat Mountains during the summer monsoon
.
J. Atmos. Sci.
,
45
,
2097
2122
, doi:.
Qian
,
W. H.
, and
S. Q.
Tang
,
2010
:
Identifying global monsoon troughs and global atmospheric centers of action on a pentad scale
.
Atmos. Oceanic Sci. Lett.
,
3
,
1
6
, doi:.
Roxy
,
M.
,
2014
:
Sensitivity of precipitation to sea surface temperature over the tropical summer monsoon region—And its quantification
.
Climate Dyn.
,
43
,
1159
1169
, doi:.
Shi
,
P. J.
,
2003
:
Atlas of Natural Disasters System of China
. Science Press, 218 pp.
Simmonds
,
I.
,
D.
Bi
, and
P.
Hope
,
1999
:
Atmospheric water vapor flux and its association with rainfall over China in summer
.
J. Climate
,
12
,
1353
1367
, doi:.
Stull
,
R. B.
,
1988
: An Introduction to Boundary Layer Meteorology. Kluwer Academic Publishers, 666 pp.
Tao
,
S. Y.
, and
L. X.
Chen
,
1987
: A review of recent research on the East Asian summer monsoon in China. Monsoon Meteorology, C.-P. Chang and T. N. Krishnamurti, Eds., Oxford University Press, 60–92.
Tu
,
C.-C.
,
Y.-L.
Chen
,
C.-S.
Chen
,
P.-L.
Lin
, and
P.-H.
Lin
,
2014
:
A comparison of two heavy rainfall events during the Terrain-Influenced Monsoon Rainfall Experiment (TiMREX) 2008
.
Mon. Wea. Rev.
,
142
,
2436
2463
, doi:.
Tu
,
Z. W.
, and
S. S.
Huang
,
1944
:
The advance and retreat of the summer monsoon
(in Chinese).
Acta Meteor. Sin.
,
18
,
82
92
.
Wang
,
B.
,
2006
:
The Asian Monsoon.
Springer/Praxis, 787 pp.
Wang
,
B.
,
Q.
Ding
,
X.
Fu
,
I.-S.
Kang
,
K.
Jin
,
J.
Shukla
, and
F.
Doblas-Reyes
,
2005
:
Fundamental challenge in simulation and prediction of summer monsoon rainfall
.
Geophys. Res. Lett.
,
32
,
L15711
, doi:.
Webster
,
P. J.
,
V. O.
Magaña
,
T. N.
Palmer
,
J.
Shukla
,
R. A.
Tomas
,
M.
Yanai
, and
T.
Yasunari
,
1998
:
Monsoons: Processes, predictability, and the prospects for prediction
.
J. Geophys. Res.
,
103
,
14 451
14 510
, doi:.
Wu
,
P. M.
,
M.
Hara
,
J.-I.
Hamada
,
M. D.
Yamanaka
, and
F.
Kimura
,
2009
:
Why a large amount of rain falls over the sea in the vicinity of western Sumatra Island during nighttime
.
J. Appl. Meteor. Climatol.
,
48
,
1345
1361
, doi:.
Wu
,
S. S.
, and
J. Y.
Liang
,
2002
:
Intensity index of South China Sea monsoon and its variation characteristics
.
J. Trop. Meteor.
,
8
,
1
9
.
Wu
,
S. S.
,
J. Y.
Lian
, and
C. H.
Li
,
2003
:
Relationship between the intensity of South China Sea summer monsoon and the precipitation in raining seasons in China (in Chinese)
.
J. Trop. Meteor.
,
19
(
Suppl.
),
25
36
.
Xu
,
K.
,
A.
Cheng
, and
M.
Zhang
,
2010
:
Cloud-resolving simulation of low-cloud feedback to an increase in sea surface temperature
.
J. Atmos. Sci.
,
67
,
730
748
, doi:.
Xue
,
J. S.
,
1999
: Research on the Summer Heavy Rainstorm in Southern China in 1994 (in Chinese). China Meteorological Press, 185 pp.
Yan
,
J. Y.
,
J. M.
Liu
,
G. R.
Jiang
,
H.
Wan
,
Y. J.
Liu
, and
H. D.
Yao
,
2007
:
Advances in the study of air sea flux exchange over the South China Sea (in Chinese)
.
Adv. Earth Sci.
,
22
,
685
697
.
Ye
,
D. Z.
, and
R. H.
Huang
,
1991
:
Progress achievement and problems of study on cause and prediction of droughts and floods in Yellow River and the Yangtze River Basins (in Chinese)
.
Earth Sci. Adv.
,
6
,
24
29
.
Zhao
,
Z. G.
,
1999
: Summer Droughts and Floods in China and Environmental Fields (in Chinese). China Meteorological Press, 297 pp.
Zhu
,
Q. G.
,
J. H.
He
, and
P. X.
Wang
,
1986
:
A study of circulation differences between East-Asian and Indian summer monsoons with their interaction
.
Adv. Atmos. Sci.
,
4
,
466
477
.