Rainfall and its variation are not only important constituents for the global hydrological cycle, but also for all living organisms on our planet. As one of the most important types of rainfall variation, the diurnal cycle has been studied by many meteorologists with numerous valuable results having been produced (Yang and Smith, 2006). Over both land and ocean, the strongest diurnal cycles are found in regions of intense convective activity, and different precipitation types may peak at different times (Dai, 2001; Nesbitt and Zipser, 2003).

Over land, intense solar radiation heating increases lower-tropospheric temperature and therefore produces instability during the daytime, while radiation cooling decreases instability overnight, which is conducive to the observed and simulated late afternoon/early evening maximum rainfall (e.g., Yang and Slingo, 2001; Nesbitt and Zipser, 2003). But this idealized diurnal rainfall cycle is affected by local factors such as sea breezes and mountain winds, which provide irregular radiation forcing and change the rainfall accordingly (Yang and Slingo, 2001).

Most studies show a weaker diurnal cycle over the ocean compared to land because water bodies feature a smaller and slower response to radiation warming and cooling. Over the ocean, the maximum rainfall usually happens in the early morning, which can be explained by many processes, including direct radiation-convection (Randall et al., 1991), radiation-dynamic convection (Gray and Jacobson, 1977), changes in moisture content with radiation cooling and heating (Sui et al., 1997), and possibly pressure tides (Dai, 2001). However, the maximum rainfall sometimes happens over the ocean in the afternoon. Sui et al. (1997) showed that afternoon convective showers are more evident in the large-scale undisturbed periods when the diurnal SST cycle is strong. Chen and Houze (1997) found the afternoon rainfall maximum is a consequence of surface-cloud-radiation interactions. Continental influences could modulate the oceanic diurnal cycle through land breezes and gravity waves leading to a more pronounced diurnal cycle (Yang and Slingo, 2001).

Over China, many recent studies have found that the warm-season rainfall often generates over the highlands in the local mid-afternoon hours and then propagates eastward or southeastward to adjacent lowlands during the night or early morning over different regions (Wang et al., 2004; Yu et al., 2007a, b). The influence of the summer monsoon on the diurnal variations of warm-season precipitation over East China has also been noted in these studies. Besides, some published studies on the diurnal variations of precipitation over East Asia examined one or several subperiods of the warm season (Chen et al., 2009; Yuan et al., 2010; Bao et al., 2011; Xu and Zipser, 2011). Based on satellite or station rain gauge data, Hirose and Nakamura (2005) investigated the diurnal cycle of rainfall over East Asia in four seasons, and Li et al. (2008) studied the seasonal variation of the diurnal cycle of rainfall in southern contiguous China.

Recently, some scientists have analyzed the interannual variation of the diurnal cycle over America and Africa, finding that the diurnal cycle is different between wet and dry years and that the diurnal cycle of rain is corrected by the ENSO and MJO (Karen, 2004; Poveda et al., 2005).

Using nine years of Tropical Rainfall Measuring Mission (TRMM) 2A25 data, this paper presents a comparative analysis of the mean diurnal cycle and its seasonal and interannual variation for three kinds of rain (convective, stratiform, and total) over the Tibetan Plateau and its downstream regions.

Based on the vertical pattern of the profiles and on the horizontal variability of the echo, precipitation is classified into convective rain, stratiform rain and ‘ other’ by the TRMM Precipitation Radar (PR) (Kummerow et al., 2000). Different rain types have different cloud thermal and dynamic structures, as well as microphysical characteristics (Zipser and Lutz, 1994). The 2A25 near-surface rain product is produced directly from TRMM PR; it is calculated from the attenuation-corrected Z_e-profiles by using a power law: R = a(Z_e)^b. Z_e is the attenuation corrected reflectivity factor, R is the rain rate, and a and b are two parameters determined from the rain type and the terrain height for each ray. Despite the error caused by surface clutter and the limitations of the scattering algorithm, PR observations have advantages in detecting shallow rainfall and removing snow signals, especially over the Tibetan Plateau with steep terrain (Masafumi et al., 2008).

For better understanding the diurnal cycle of precipitation over the Tibetan Plateau and its downstream regions, three typical regions are selected in this study. The first is the Tibetan Plateau region (30-35° N, 80-105° E), whose altitude is mostly higher than 3000 m (average: 4530 m). The second is the Yangtze River basin region (30-35° N, 105-120° E), with an average altitude of 566 m. And the third is the East China Sea region (30-35° N, 120-140° E), 80% of which is ocean. All three regions are at the same latitude, meaning there is no influence of latitude on their diurnal cycles (Dai and Trenberth, 2004).

Instead of rainfall amount, the probability density function (PDF) of rainfall amount is used to study the diurnal cycle of rain in this paper, because it is found to more reasonable than the former when performing the comparative analysis. In order to clean out the diurnal deviation caused by data samples, in the beginning of this study, we deal with the data using a smoothing method (× mean hourly pixel number/pixel number at local time).

The mean diurnal cycle of total rain over the Tibetan Plateau, Yangtze River basin, and East China Sea is analyzed here, and we also analyze the diurnal cycle of convective rain and stratiform rain over these regions— except the Tibetan Plateau because of the uncertainty of classification of convective and stratiform rain in TRMM 2A25 over this region (Fu and Liu, 2007). After comparing the PDFs of rainfall amount (thick lines) and rain area (rain pixels observed by TRMM) (thin lines) over these three regions in Fig. 1, it is simple to see that the diurnal curves are very close; therefore, the description of diurnal rain area is omitted here, i.e., the diurnal rainfall is used to represent the diurnal cycle of precipitation in this paper.

As shown in Fig. 1a, the convective rain minimum (2.85%) happens at 1100 LST over the Yangtze River basin, while the largest PDF (5.45%, 1.9 times the minimum) is at 1500 LST. Besides, there is a secondary rain maximum at 0200-0600 LST. Different from the Yangtze River basin, the diurnal cycle of convective rain over the East China Sea appears to be the anti-phase of the continental diurnal cycle. The convective rain minimum (3.55%) happens at 1800 LST over the East China Sea, while the largest PDF (5.38%, 1.52 times the minimum) is at 0600 LST. Influenced by solar radiative heating (Yang and Slingo, 2001), convective rain happens more frequently in the second half of the day (1200-2400 LST) on land, and the maximum also occurs during this time period (around 1500 LST). But over ocean, convective rain is mainly controlled by the static radiation-convection mechanism and dynamic radiation-convection mechanism (Randall et al., 1991); thus, the diurnal cycle switches and there is more convective rain in the first half of the day (0000-1200 LST) than in the second half of the day, and the maximum is also in the first half of the day (around 0600 LST).

From the diurnal PDF of stratiform rain (Fig. 1b), we know that over the Yangtze River basin, the largest PDF (4.76%, 1.34 times the minimum, weaker than convective rain) of stratiform rain appears at 0400-1000 LST over the Yangtze River basin, with the minimum (3.54%) at 1800 LST. There is no apparent difference between diurnal stratiform rain and diurnal convective rain over the East China Sea, both peaking in the early morning (0600 LST; 5.21%, 1.52 times the minima, equal to convective rain) and obtain their minima (3.42%) in the afternoon.

Figure 1c shows the diurnal PDF of total rainfall amount. The diurnal variation rate of total rain over the Tibetan Plateau is the largest among the three regions. The PDF reaches a minimum (2.99%) at 0900 LST over the Tibetan Plateau, then increases as solar radiation heating increases, and reaches a maximum at 1700 LST (5.47%), which is 1.83 times larger than the minimum. The diurnal cycle of the Tibetan Plateau features a second peak at 0300 LST, three hours ahead of the maximum over the Yangtze River basin (Wang et al., 2004; Yu et al., 2007a, b). Over the East China Sea, the characteristics of diurnal total rain are very similar to diurnal convective rain and diurnal stratiform rain, because there is no apparent phase deviation between diurnal convective rain and diurnal stratiform rain. But over the Yangtze River basin, the phase deviation between diurnal convective rain and diurnal stratiform rain is obvious, which causes the diurnal cycle of total rain to have two peaks: one at 0600 LST caused by stratiform rain, and the other at 1500 LST caused by convective rain.

Figure 1

	Figure Option ViewDownloadNew Window
	Figure 1 The probability density function (PDFs) of rainfall amount and rain area of (a) convective rain, (b) stratiform rain, and (c) total rain over the Tibetan Plateau, Yangtze River Basin (YRB), and East China Sea (ECS).

In order to study the seasonal variation of diurnal precipitation over the Tibetan Plateau and its downstream regions, a comparative analysis of the diurnal rain cycles in March-May (MAM), June-August (JJA), September-November (SON), and December-February (DJF) over the Tibetan Plateau, Yangtze River basin, and East China Sea is performed.

For convective rain (Figs. 2a and 2b), the diurnal cycle over the Yangtze River basin (Fig. 2a) differs greatly among the four seasons: the maximum appears at 0300, 1500, 0100, and 0000 LST in MAM, JJA, SON, and DJF, respectively; all of them (except for JJA) are in the first half of the day; and the standard deviation reaches 6.95 h. Over the East China Sea (Fig. 2b), the diurnal cycle is similar in all four seasons: the maximum happens in the first half day, at 0600, 0500, 0500, and 0700 LST in MAM, JJA, SON, and DJF, respectively, and the standard deviation is 0.96 h— much smaller than for the Yangtze River basin.

Over the Yangtze River basin (Fig. 2c), the stratiform rain maximum occurs at 0600, 0600, 0200, and 1100 LST in MAM, JJA, SON, and DJF, respectively, and all in the first half of the day. The standard deviation of these peaks is 3.69 hours— much lower than that of convective rain (6.95 h), meaning the seasonal variation of diurnal stratiform rain over the Yangtze River basin is weaker than that of diurnal convective rain. Over the East China Sea (Fig. 2d), stratiform rain peaks at 0500, 0500, 0600, and 0700 LST in MAM, JJA, SON, and DJF, respectively, and the standard deviation is 0.96 h— equivalent to that of convective rain.

Over the Tibetan Plateau (Fig. 3a), the most violent period of diurnal total rain happens in MAM, not in JJA. This is because the rainy season has not arrived yet in MAM; the sky is clean, so the sensible heating is stronger than in JJA (Gao et al., 1984). The diurnal cycle of total rain is very similar in MAM, JJA, and SON; the first peak is at 1700-1800 LST, the second peak is at 0300 LST, and the minimum is at 0900 LST. In DJF, the diurnal variation of total rain is weakest; and, different from the other three seasons, it has three PDF peaks (0300, 1100, and 1700 LST).

Over the Yangtze River basin (Fig. 3b), the diurnal total rain is somewhere between the diurnal convective rain and diurnal stratiform rain in MAM, JJA, SON, and DJF, as the diurnal cycle of convective rain is quite different from diurnal stratiform rain. The seasonal variation of diurnal total rain over the East China Sea (Fig. 3c) is similar to diurnal convective rain and diurnal stratiform rain, because there is no obvious difference between convective rain and stratiform rain in each season.

Figure 2

	Figure Option ViewDownloadNew Window
	Figure 2 The PDFs of (a, b) convective rain and (c, d) stratiform rain in the four seasons (MAM, JJA, SON, and DJF) over the (a, c) Yangtze River basin and (b, d) East China Sea.

Figure 3

	Figure Option ViewDownloadNew Window
	Figure 3 The PDF of total rain in the four seasons (MAM, JJA, SON, and DJF) over the (a) Tibetan Plateau, (b) Yangtze River basin, and (c) East China Sea.

From Figs. 2 and 3 we know that the most violent seasonal variation of diurnal precipitation happens over the Yangtze River basin. This is because the Yangtze River basin is under the control of the Asian summer monsoon, Asian winter monsoon, westerlies, and some other general circulation factors, and is located between the Tibetan Plateau and East China Sea. The most complicated rain systems are situated over the middle to lower reaches of the Yangtze River, and the rainfall is differs greatly among the four seasons. The East China Sea is also affected by many general circulation systems, but the specific heat capacity of the ocean is very large, so the seasonal variation is more stable than over the Yangtze River basin. The Tibetan Plateau is so high that many general circulation factors cannot exert a substantial influence; rather, it is mainly affected by westerlies, so the rain systems are simpler than elsewhere and the seasonal variation of diurnal rain is weak.

The yearly change of diurnal precipitation over the Ti-betan Plateau and its downstream regions is analyzed in this section using TRMM 2A25 data from 1998 to 2006 (Fig. 4).

Over the Yangtze River basin (Figs. 4a-c), the diurnal cycle of total rain takes on a clear interannual feature. The peak of total rain happens at 1500 LST in the afternoon from 1998 to 2002, but changes to 0300 LST in the early morning from 2003 to 2006. The diurnal variation of convective and stratiform rain also shows different phases before and after 2002. The phase of diurnal convective precipitation jumps straight from 1500-1800 LST to 0200-0500 LST from 1998-2002 to 2003-2006, while stratiform precipitation shifts from 1200 LST in 1999 to 0300 LST in 2005 generally.

No obvious differences can be found among years in the diurnal cycle of total rain, convective rain, and stratiform rain over the East China Sea (Figs. 4d-f); the maximum always happens at 0500-0700 LST. Meanwhile, the interannual change of diurnal total rain is weak over the Tibetan Plateau (Fig. 4g) and peaks at 1500-1800 LST in each year.

Besides, in the La Niñ a year (June 1999 to May 2000), the amplitude of diurnal rain enhances over the Yangtze River basin and East China Sea, while the opposite changes can be found in the El Niñ o year (June 2002 to May 2003), consistent with results using rain gauge observations in the tropical Andes of Colombia (Poveda et al., 2005).

Figure 4

	Figure Option ViewDownloadNew Window
	Figure 4 LST-year cross sections of the PDFs of (a, d, g) total rain, (b, e) convective rain, and (c, f) stratiform rain over the (g) Tibetan Plateau, (a-c) Yangtze River basin, and (d-f) East China Sea. Shading represents PDFs (%) larger than 4.5; dashed rectangles mean El Niñ o or La Niñ a years.

Using nine years (1998-2006) of TRMM 2A25 data, based on the PDF of rainfall, this paper presents a comparative analysis of the average diurnal cycle and its seasonal and interannual variation for convective rain, stratiform rain, and total rain over the Tibetan Plateau and its downstream regions of the Yangtze River basin and East China Sea. The results can be summarized as follows:

The diurnal convective rain is stronger than diurnal stratiform rain over the Yangtze River basin, and the convective rain peaks in the afternoon while the stratiform rain maximum happens in the early morning. There is no apparent difference between diurnal stratiform rain and diurnal convective rain over the East China Sea; both peak in the early morning and reach a minimum in the afternoon. The diurnal total rain over the Tibetan Plateau is stronger than over its downstream regions, and peaks in late afternoon.

The seasonal variation of diurnal rain over the Yangtze River basin is much more violent than over the Tibetan Plateau and East China Sea, as the seasonal variation of rain systems over there is more severe and complex than over the other two regions. The seasonal variation of diurnal convective rain is more apparent than diurnal stratiform rain over the Yangtze River basin.

Yearly change of the diurnal cycle happens clearly over the Yangtze River Basin for total rain, convective rain, and stratiform rain. The maximum of total rain happens in the afternoon during 1998-2002 over the Yangtze River basin, while it peaks in the early morning during 2003-2006. No clear phase differences can be found among years in the diurnal cycle over the East China Sea and Tibetan Plateau. In the La Niñ a year, the amplitude of diurnal rain enhances over the Yangtze River basin and East China Sea, and weakens in the El Niñ o year. More data and long-term datasets are needed to verify the relationship between the diurnal cycle and ENSO.