Mechanisms for Diurnal Variability of Global Tropical ... - AMS Journals

5190

JOURNAL OF CLIMATE

VOLUME 19

Mechanisms for Diurnal Variability of Global Tropical Rainfall Observed from TRMM SONG YANG School of Computational Sciences, George Mason University, Fairfax, Virginia

ERIC A. SMITH NASA Goddard Space Flight Center, Greenbelt, Maryland (Manuscript received 8 April 2005, in final form 11 January 2006) ABSTRACT The behavior and various controls of diurnal variability in tropical–subtropical rainfall are investigated using Tropical Rainfall Measuring Mission (TRMM) precipitation measurements retrieved from the three level-2 TRMM standard profile algorithms for the 1998 annual cycle. Results show that diurnal variability characteristics of precipitation are consistent for all three algorithms, providing assurance that TRMM retrievals are producing consistent estimates of rainfall variability. As anticipated, most ocean areas exhibit more rainfall at night, while over most land areas, rainfall peaks during daytime; however, important exceptions are noted. The dominant feature of the oceanic diurnal cycle is a rainfall maximum in late-evening–early-morning (LE–EM) hours, while over land the dominant maximum occurs in the mid- to late afternoon (MLA). In conjunction with these maxima are pronounced seasonal variations of the diurnal amplitudes. Amplitude analysis shows that the diurnal pattern and its seasonal evolution are closely related to the rainfall accumulation pattern and its seasonal evolution. In addition, the horizontal distribution of diurnal variability indicates that for oceanic rainfall, there is a secondary MLA maximum coexisting with the LE–EM maximum at latitudes dominated by large-scale convergence and deep convection. Analogously, there is a preponderancy for an LE–EM maximum over land coexisting with the stronger MLA maximum, although it is not evident that this secondary continental feature is closely associated with the large-scale circulation. Neither of the secondary maxima exhibit phase behavior that can be considered semidiurnal in nature. Diurnal rainfall variability over the ocean associated with large-scale convection is clearly an integral component of the general circulation. Phase analysis reveals differences in regional and seasonal features of the diurnal cycle, indicating that underlying forcing mechanisms differ from place to place. This is underscored by the appearance of secondary ocean maxima in the presence of large-scale convection, along with other important features. Among these, there are clear-cut differences between the diurnal variability of seasonal rainfall over the mid-Pacific and Indian Ocean Basins. The mid-Pacific exhibits double maxima in spring and winter but only LE–EM maxima in summer and autumn, while the Indian Ocean exhibits double maxima in spring and summer and only an LE–EM maximum in autumn and winter. There are also evident daytime maxima within the major large-scale marine stratocumulus regions off the west coasts of continents. The study concludes with a discussion concerning how the observational evidence either supports or repudiates possible forcing mechanisms that have been suggested to explain diurnal rainfall variability.

1. Introduction Diurnal variation of precipitation on planet Earth was first reported in the early twentieth century by Hann (1901, 338–346) and since that time has spawned an immense literature. In general, rainfall maxima in

Corresponding author address: Dr. Eric A. Smith, NASA Goddard Space Flight Center, Mail Code 613.1, Greenbelt, MD 20771. E-mail: [email protected]

© 2006 American Meteorological Society

JCLI3883

the late-evening–early-morning hours (LE–EM) have been reported for open-ocean environments (e.g., Kraus 1963; Andersson 1970; Gray and Jacobson 1977; Jordon 1980; Albright et al. 1985; Randall et al. 1991; Imaoka and Spencer 2000), while mid- to late afternoon (MLA) maxima have often been reported for land (e.g., Ray 1928; Cook 1939; Kousky 1980; Hamilton 1981; Garreaud and Wallace 1997). However, such generalities do not describe all the intricacies inherent to rainfall’s diurnal cycle, and both past observational rainfall

15 OCTOBER 2006

YANG AND SMITH

datasets and cloud parameterizations in physical models have often lacked the completeness and reliability to ensure dependable analyses [see, e.g., Liu and Moncrieff (1998) and Lin et al. (2000) for different perspectives on this topic]. The main scientific objective of this study is to improve the current understanding of the earth’s diurnal rainfall variability by analyzing the newest and most reliable global-scale satellite data now being acquired from the Tropical Rainfall Measuring Mission [TRMM; see Simpson et al. (1996) and Kummerow et al. (2000) for overviews of the mission and salient rainfall results, respectively]. We seek to explain how large-scale and local forcing mechanisms at work over both land and ocean, but in distinct manifestations, produce coherent diurnal variations, to the extent that the TRMM observations can resolve the underlying variations spatially and temporally. However, before doing so, it is worthwhile to consider a number of past results. Table 1 summarizes the salient points concerning the various mechanisms described in the published literature as possible causes for diurnal variability in precipitation, mechanisms that will be referred to throughout the remainder of the paper. The appendix provides expanded explanations for these mechanisms. Diurnal variability of precipitation has long been reported to exhibit distinct spatial variations. For example, Albright et al. (1985) found a morning rainfall maximum in the intertropical convergence zone (ITCZ) of the central and eastern tropical Pacific Ocean and an afternoon rainfall maximum in the South Pacific convergence zone (SPCZ). Gray and Jacobson (1977), in studying the diurnal variability of rainfall over tropical oceans, suggested a coupled radiative– dynamics mechanism to explain the LE–EM maximum within deep cumulus convection. However, they also found that the 1974 Global Atmospheric Research Program (GARP) Atlantic Tropical Experiment (GATE) radar measurements in the eastern Atlantic exhibited the heaviest rainfall during late-morning to earlyevening periods, in conjunction with an LE–EM minimum, which was later corroborated by Jordon (1980). This means that the eastern Atlantic’s heavy rainfall maximum occurs ⬃6 h later than for most other ocean regions. If such an offset is real, this region’s diurnal control is likely due to some other distinct mechanism. Gray and Jacobson (1977) suggested that it could be associated with the relatively stronger degree of lowlevel vertical wind shear in the eastern Atlantic Tropics and the consequent frequent appearance of squall lines and other types of line convection. They also noted that it could be due to the affinity for that part of the ocean Tropics to maintain diurnally driven weather systems

5191

originating over continental West Africa and propagating out over the open ocean as incipient easterly waves. Other mechanisms have been investigated that would transfer diurnal variability from local to distant sources through the wavelike propagation of upper-tropospheric convergence–divergence patterns (e.g., Silva Dias et al. 1987) or land to ocean gravity wave dispersion (e.g., Yang and Slingo 2001). As noted by Zhao and Weare (1994) in an intraseasonal context, with application of the shallow water equations on an equatorial beta plane in which a wave–conditional instability of the second kind (CISK) model is used to represent cloud effects, the amplitude of the diurnal variability of cloudiness is found to determine the strength of the intraseasonal oscillations; note that Chen and Takahashi (1995) suggest that this cause–effect relationship is actually reversed. On a climate time scale, the strength of the low-level jet appears to determine the nature of diurnal precipitation variability within the central United States (see Hu 2003). From these and other studies, it becomes evident that the diurnal variability of precipitation is a response to processes operating at a number of time scales, almost always associated with spatial complexities. Moreover, any diurnal variation in precipitation directly translates to diurnal variation in atmospheric latent heating, which is often rainfall’s main feedback link to the atmosphere, and follows up diurnally modulated circulations and cloudiness features pertaining to rainfall. Diurnal cycles are manifested in many meteorological variables, as numerous studies have reported. These include 1) surface pressure and tropospheric geopotential height (Taylor 1929; Rosenthal and Baum 1956; Thiele 1966; Chapman and Lindzen 1970; Hollingsworth 1971; Haurwitz and Cowley 1973; Hamilton 1984; Lieberman 1991; Hays et al. 1994; Lieberman and Leovy 1995; Dai and Wang 1999; Petenko and Argentini 2002); 2) land surface temperatures and moisture along with the concomitant fluxes of sensible and latent heat (Pierce 1934; Fritz 1963; Wallace and Hobbs 1977; Shi and Chen 1984; Chen and Wang 1994; Markowski and Stensrud 1998; Ignatov and Gutman 1999); 3) sea surface temperature (Rosenthal 1960; Deschamps and Frouin 1984; Stramma et al. 1986; Webster et al. 1996); 4) upper air temperatures (Ballard 1933; Wallace and Patton 1970); 5) horizontal winds and boundary layer circulations (Roll 1965; Bonner 1968; Wallace and Hartranft 1969; Wallace and Todd 1974; McNider and Pielke 1981; Brill and Albrecht 1982; Douglas and Li 1996; Douglas et al. 1998; Deser and Smith 1998; Rife et al. 2002; Zhang and Zheng 2004); 6) large-scale horizontal moisture fluxes (Benton and Estoque 1954; Ras-

5192

JOURNAL OF CLIMATE

VOLUME 19

TABLE 1. A summary of the 12 foremost published mechanisms for explaining diurnal precipitation variability. Columns provide the acronym and succinct description (where -P denotes propagating mode), recommended references, characteristic spatial scale, characteristic regime insofar as land or ocean along with characteristic time period of maximum precipitation occurrence insofar as day or night, and a brief causal explanation of the mechanism itself. Authors make no claim that this is an exhaustive list of mechanisms.

Acronym name and description

Recommended reference(s)

Spatial scale

1

BLWO

Blackadar (1957) Wallace (1975)

Meso-␣

2

LSCH-P

Silva Dias et al. (1987)

Continental

3a LSH–SD Byers and Braham Meso-␥-␤-␣ [e.g., air mass (1948) thunderstorm Byers and (AMTS)] Rodebush (1948) Braham (1952) Gentry and Moore (1954) Ogura and Takahashi (1971)

3b LSH–DH (e.g., LSBC)

Schmidt (1947) Pearce (1955) Fisher (1961) Estoque (1961) Pielke (1974)

Meso-␣ to synoptic

Land (L) or ocean (O), day (D) or night (N)

Concise causal explanation

L (D or N)

Solar-forced convective boundary layer cycle enhances local convergence, mechanical convection, Richardson number instability, and strength of near-surface winds, ultimately producing a subsequent precipitation maximum. L (D and N) Solar heating diurnally regulates continental heating, thus modulating the large-scale upper-tropospheric divergence in which a positive divergence anomaly produces a precipitation maximum. This is a remote control mechanism when the divergence region itself propagates downstream, producing continuously phase-shifted (delayed) precipitation maxima generally associated with decaying amplitude. L (D) Static destabilization occurs as a result of both surface sensible and latent heating in which dry and moist static stability variations associated with diurnal thermal and moisture oscillations lead to the growth of dry and moist convective overturnings. Thunderstorms generally arise from conditional instability in which deep convective overturnings will not take place until the dry convective overturning process penetrates the PBL. Once initiated, deep moist convective overturnings reach as high as the tropopause and are typically associated with afternoon rain showers. Within environments of weak wind shear, showers take the form of AMTSs. With more significant shear, the associated momentum transport by showers leads to meso-␥, meso-␤, and meso-␣ organizations of convective overturnings over time (e.g., supercells) where larger-scale organizations prolong the period of afternoon rain into evening hours. L (D) Differential solar heating across land–sea boundaries (or inland surface thermal boundaries) produces daytime solenoidal circulations with surface flow from cool side to warm side and with warm side experiencing low-level convergence, convective lifting, and the formation of precipitation along a narrow mesoscale front, then a reversal of circulation at night in which the daytime warmside region experiences subsidence and precipitation suppression while the nighttime warm-side region may undergo cloud line formation accompanied by drizzle or light rain.

musson 1968); 7) large-scale vertical motion and mass divergence (Nitta and Esbensen 1974; McBride and Gray 1980; Reiter and Tang 1984; Nicolini et al. 1993; Sui et al. 1997; Krishnamurti and Kishtawal 2000); 8) cloud cover (Lavoie 1963; Short and Wallace 1980; Warren et al. 1986; Meisner and Arkin 1987; Morcrette 1991; Menzel et al. 1992; Bergman and Salby 1997; Wylie and Woolf 2002); 9) cloud size (Browner et al. 1977; Thiao and Turpeinen 1992; Machado et al. 1993);

10) mesoscale moisture convergence (Cooper et al. 1982); 11) thunderstorm frequency (Skaggs 1969; Wallace 1975; Easterling and Robinson 1985); 12) satellite infrared (IR) fluxes (Reed and Jaffe 1981; Murakami 1983; Fu et al. 1990; Janowiak et al. 1994; Faysash and Smith 2000); 13) outgoing longwave radiation (Duvel and Kandel 1985; Slingo et al. 1987; Liebmann and Gruber 1988; Hartmann et al. 1991; Ellingson and Ba 2003; Smith and Rutan 2003);14) daily temperature

15 OCTOBER 2006

5193

YANG AND SMITH TABLE 1. (Continued)

Acronym name and description 4

MTFPS-P

Recommended reference(s)

Spatial scale

Bonner (1963) Tripoli (1986) Tripoli and Cotton (1989a,b)

Meso-␣ to synoptic

Land (L) or ocean (O), day (D) or night (N) L (D and N)

Concise causal explanation Over north–south-aligned mountain environments with plains to the east (west) in westerly (easterly) regimes, and mean geostrophic cross-ridge winds, daytime solenoidal upslope flow into the leeside convergence zone situated along the nocturnal dryline position sets off the stage-by-stage process of thunderstorm development and propagation. As convection develops along the leeside convergence zone and dryline (producing westward tilted afternoon storms along mountain slopes), the upslope flow steadily breaks through the deepening PBL inversion on the upslope side of the dryline, while convective overturning erodes the PBL inversion downslope of the dryline position. This enables midtropospheric wind to steer and advect the convergence line and dryline eastward. Convergence into the dryline is reinforced by the upslope flow of moist air trapped below the PBL inversion downslope of the dryline (provided by surface evaporation and transpiration as daytime progresses) and the downward flux of upper-level well-mixed dry air and westerly momentum into the leeside convergence zone. The down flux of upper-level air also supports the downdraft development by mixing dry air into the convection zone bolstering evaporative cooling. As evening commences, focused deep convection in narrow squall line–type band structures has continued to erode the PBL inversion to the east, while the midtropospheric westerly steering current carries convection out onto the Great Plains. As evening progresses, the nocturnal PBL forms with a top well below the daytime PBL top, causing the new dryline to form back to the west along the lee side of the mountain. Nighttime PBL formation enables the nocturnal low-level southerly jet to form as moist air between daytime and nighttime PBL tops is released from friction and the mesoscale east–west pressure gradient. Convection spreads out as a blob structure instead of line structure, and inertial oscillation is set off within the trapped moist layer by the release from friction above the nocturnal PBL top, ultimately leading to supergeostrophic flow from the south providing convective blob abundant moisture at a rapid rate, supporting the deep convection that can couple with the upper troposphere. Low-level jet moisture pump and deep convection eventually lead to the formation of evening MCS over the Great Plains, a process involving inertial oscillation dynamics, coupling of convection with the upper troposphere, late-evening precipitation burst, formation of mobile quasigeostrophic shortwave and jet streak, and migration of MCS eastward via quasigeostrophic dynamics. Eastward motion is also supported by gravity wave dynamics that produce surface density currents arising from evaporatively cooled downdrafts. Cooled downdrafts and cold pools trigger new cell growth to the east, with the MCS precipitation maximum occurring over the central United States within the Mississippi–Ohio valley by early morning.

5194

JOURNAL OF CLIMATE

VOLUME 19

TABLE 1. (Continued)

Acronym name and description

Recommended reference(s)

Spatial scale

Land (L) or ocean (O), day (D) or night (N)

5

SRCL

Dai (2001)

Synoptic

L (N)

6

AGWT

Tripoli (1992)

Meso-␤ to meso-␣

O (N)

7

CGWM-P

Mapes et al. (2003a,b) Warner et al. (2003)

Meso-␣ to synoptic

O (D and N)

8

DRC

Ruprecht and Gray (1976a,b) Gray and Jacobson (1977)

Meso-␣ to synoptic

O (N)

9

OSH

Sui et al. (1997)

Planetary

O (D)

10

SRC

Synoptic to planetary

O (N)

11

SRCM

Kraus (1963) Ramage (1971) Randall et al. (1991) Tao et al. (1996) Sui et al. (1997)

Synoptic

O (N)

12

SDTM-P

Brier (1965) Brier and Simpson (1969) Lindzen (1978)

Planetary

L or O (D and N)

range (Saltzman and Pollack 1977; Karl et al. 1984; Gallo et al. 1996; Dai et al. 1999; Durre and Wallace 2001a,b); 15) lightning frequency (Israël 1952; Maier et al. 1984; Lopez and Holle 1986; King and Balling 1994; Watson et al. 1994; Camp et al. 1998; Boccippio et al. 2002); and even 16) diurnal meteorological contrasts between the inner planets, that is, differences between Earth’s diurnal meteorological behavior and the very large amplitude diurnal behavior of planet Mars (e.g., Sutton et al. 1978) and the very small amplitude diurnal behavior of planet Venus (e.g., Lewis 1971). All of these studies testify, in some fashion, to the meteorological influence of diurnally regulated solar or diabatic heating on either surface or atmospheric processes. Results from the published literature indicate that there are a number of possible mechanisms that could explain rainfall diurnal variations. The most popular explanations for the afternoon continental rainfall maximum are various manifestations of the land sur-

Concise causal explanation Enhanced nighttime cloud-top radiative cooling thermally destabilizes the upper atmosphere, thus increasing nighttime condensation, leading to a late-night maximum in drizzle and light precipitation analogous to the delayed SRC mechanism over ocean. Enhanced nighttime radiative cooling of a developing tropical cyclone’s anvil, which destabilizes the cloud top, thus trapping vertically propagating gravity wave energy, which amplifies the cyclone intensity and precipitation. Solar thermal forcing over a coastal land area produces a precipitation maximum, followed by delayed offshore precipitation maxima induced by gravity wave propagation. Oscillating horizontal convergent circulation established into a sustained cloud region from a surrounding cloud-free region through day–night modulation of clear-air subsidence by enhanced daytime water vapor absorption of solar radiation produces a late-night precipitation maximum. Afternoon SST heating induces an ocean precipitation maximum analogous to the LSH–SD mechanism. Enhanced nighttime cloud-top radiative cooling thermally destabilizes the upper cloud, thus increasing convection and precipitation. Enhanced nighttime radiative cooling (including SRC influence) for disturbed convective conditions increases humidities, thus amplifying condensation and precipitation. Lunar–solar gravitational radiative atmospheric tides induce the semidiurnal oscillation in surface pressure with consequent column mass convergence and precipitation maxima.

face heating (LSH) mechanism (see Pielke 2002), while the so-called “static radiation–convection” (SRC) and “dynamic radiation–convection” (DRC) mechanisms are customarily used to explain the early-morning oceanic rainfall maximum (e.g., Ramage 1971; Gray and Jacobson 1977).

2. Advent of comprehensive satellite datasets Given the historical scarcity of in situ rainfall measurements over ocean, satellite rainfall estimates have been used to provide more uniform and continuous data for investigating oceanic rainfall variability. Over the past two decades, a number of precipitation retrieval algorithms have been developed for various satellite instruments, the predominant techniques being applied to geosynchronous IR (GEO-IR) measurements and passive microwave (PMW) radiometer measurements. However, until recently, significant uncer-

15 OCTOBER 2006

YANG AND SMITH

tainties in the retrieval products of satellite remote sensing techniques have undermined their applicability to rainfall variability analysis. Before the launch of TRMM on 27 November 1997 (Thanksgiving Day in the United States), the most reliable estimates of instantaneous precipitation had been generated from special sensor microwave/imager (SSM/I) PMW measurements (e.g., Wilheit et al. 1994; Ebert and Manton 1998; Smith et al. 1998; Yang and Smith 1999). SSM/I radiometer measurements are obtained from Defense Military Satellite Program (DMSP) platforms—satellites notionally flown in pairs within separate (phase shifted) sun-synchronous orbit planes. Although quantifying diurnal variability from one or two sun-synchronous orbiters is a marginal exercise, by extended averaging, Sharma et al. (1991) and Chang et al. (1995) found clear-cut differences in annual mean ocean rainfall based on morning and evening passes of SSM/I measurements. The latter study found morning pass rainfall to be about 20% greater than evening pass rainfall, with the most significant differences occurring along the ITCZ. Our study is motivated by the new measurements from the TRMM satellite, whose non-sun-synchronous orbit was explicitly selected to provide diurnal sampling capability. TRMM’s orbit plane is inclined ⬃35° with respect to Earth’s equatorial plane. It was inserted into an orbit altitude of ⬃350 km after its launch in November 1997, then lifted to an altitude of ⬃400 km in August 2001 to preserve fuel for extended life. Studies of the diurnal variability of rainfall prior to the TRMM era, whether they were based on in situ (rain gauge, radar) or satellite (GEO-IR, sunsynchronous PMW) datasets, were always restricted by some type of spatial or temporal sampling limitation, on top of having inherent problems in retrievals from nonmicrowave data sources. Therefore, a comprehensive understanding of diurnal precipitation processes at a global scale has been lacking. With diurnal sampled measurements intrinsic to TRMM’s non-sun-synchronous orbit, and the high-quality, uniform, and consistent precipitation products over the ⫾35° latitude zone being produced by the three standard level-2 rain-rate algorithms applied to TRMM Microwave Imager (TMI) and TRMM precipitation radar (PR) measurements (see Okamoto 1988; Nakamura et al. 1990; Kummerow et al. 1998 for descriptions of the TRMM instruments), an improved understanding of global rainfall diurnal variability has gradually begun to emerge. Hong et al. (2005) recently demonstrated the improved representation of diurnal rainfall variability from TRMM measurements. Also, ongoing intercomparison analyses being conducted at the National Aero-

5195

nautics and Space Administration (NASA) Goddard Space Flight Center (GSFC) by this paper’s authors involving results from the three standard level-2 algorithms (i.e., the TMI-only, PR-only, and PR/TMI combined algorithms) now indicate agreement within 10% on a global–annual average scale. Using retrieval datasets from the three standard level-2 rain-rate retrieval algorithms for the 1998 annual cycle, this investigation initially demonstrates that the diurnal properties of the three types of retrievals are consistent (this lends credibility to diurnal analyses of the nonhomogeneous space–time properties of diurnal rainfall variability). Thereafter, systematic analyses of diurnal rainfall processes over land and ocean are carried out, with a focus on explaining the various control mechanisms at work producing the variations. In the following, section 3 describes the analysis methodology while section 4 describes the various TRMM datasets. Section 5 then presents analysis results at the large scale for three distinct time frameworks (day–night, 3 hourly, and monthly), with section 6 presenting analysis results at the seasonal time scale. Section 7 provides a final discussion and main conclusions.

3. Methodology The TRMM satellite is equipped to measure rainfall within the 35° north–south parallels (Simpson et al. 1996). There are approximately 16 orbits per day with ⬃92 min between orbits. Since the PR’s narrow-swath scan (⬃220 km) does not provide measurements much beyond the ⫾35° latitudes, only PR and TMI rain retrievals within the ⫾35° latitude limits are used. Due to the nature of the TRMM orbit, which samples a given spot approximately 26 times per month at the equator (increasing slightly outside the equatorial zone but significantly near the ⫾35° orbit reversal points), some type of space–time averaging must be performed to obtain representative diurnal quantities [see Bell and Reid (1993), Soman et al. (1995), and Salby and Callaghan (1997) for further studies concerning sampling requirements for the unambiguous analysis of diurnal variability]. For this study, the elemental space–time averaging operator extends over a 5° ⫻ 5° latitude–longitude grid box across a 3-h diurnal time interval. Thereafter, depending on the particular analysis procedure, additional areal averaging (partial zonal, partial meridional, or partial global) may be applied in the spatial framework, whereas either monthly or seasonal (3 month) averaging is applied in the temporal framework, with the further option of extending the elemental averaging

5196

JOURNAL OF CLIMATE

time interval by multiples of 3 h (e.g., 12-h half days or 24-h full days). Sensitivity tests suggest that diurnal statistics are largely, although not completely, stabilized at the monthly–diurnal (3 h) time scale over the 5° grid mesh [see Kishtawal and Krishnamurti (2001), Negri et al. (2002), and Sorooshian et al. (2002) for other studies addressing space–time averaging procedures used to examine diurnal signals in TRMM measurements]. It should be recognized that when considering monthly daytime–nighttime or daily averaging (12- or 24-h elemental averaging time interval), seasonal–diurnal averaging, or any type of areal averaging (note that all of these averaging procedures are used in the subsequent analyses), the residual uncertainties due to undersampling found with 1-month averaging on the 5° ⫻ 5° scale are effectively mitigated (to be discussed further in section 6). At various points during the analysis over the 5° grid mesh and for either monthly or seasonal temporal averaging, 3-hourly diurnal rainfall time series for the 1-day period (8 time steps) are expressed in terms of amplitudes and phases of the first Fourier harmonic of the 24-h period (i.e., the diurnal harmonic). The consequent grids of amplitude and phase factors then provide continuous global-scale fields of approximate peak rainfall intensities and associated clock times. This technique objectively decomposes the most salient space–time properties of diurnal rainfall variability within a phase resolution of 3 h [note that higherresolution phenomena, such as the 1-h phase shifts of diurnal peaks identified by Oki and Musiake (1994) in their hourly rain dataset, are not detectable with our dataset].

4. Description of datasets Precipitation retrievals from TRMM’s standard algorithms provide the most accurate rainfall from spaceborne remote sensing, as acknowledged in various blended precipitation algorithm studies, which use TRMM retrievals as the primary satellite calibration standard (e.g., Turk et al. 1999; Huffman et al. 2001), and other types of studies depending on inherent accuracy and precision in the retrievals (e.g., Simpson et al. 1998; Hou et al. 2000; Short and Nakamura 2000; Tao et al. 2006). The three standard TRMM rainfall algorithms whose retrievals are used in this study are called 2a12 (TMI-only), 2a25 (PR-only), and 2b31 (PR/TMI combined), all of which are referred to as standard level-2 (instantaneous) profile algorithms. The 2a12 algorithm is TRMM’s main level-2 TMIonly profile scheme, which uses cloud-model generated microphysics as solution guidance for microphysical

VOLUME 19

profiling based on the Bayesian technique (Kummerow et al. 1996, 2000; Olson et al. 2006; Yang et al. 2006). The 2a25 rain-rate algorithm is the main level-2 PRonly profile scheme, using a retrieval technique akin to the Hitschfeld and Bordan (1954) deterministic inversion solution, followed by the application of the surface reference technique (SRT) to account for residualattenuation correction error (Iguchi and Meneghini 1994; Iguchi et al. 2000; Meneghini et al. 2000; Meneghini and Kozu 1990). The 2b31 algorithm is TRMM’s main level-2 PR/TMI combined profile scheme based on an elaborate Bayesian modification to the 2a25 retrieval solution, augmented with an additional estimate of total path attenuation derived from TMI’s 10.7-, 19-, and 37-GHz channels (Haddad et al. 1997; Smith et al. 1997). The 2a12 algorithm is applied over the wide-swath domain (⬃780 km) observed by the conically scanning TMI radiometer while the 2a25 and 2b31 algorithms are applied over the narrow-swath domain (⬃220 km) observed by the crosstrack-scanning PR radar. The rainfall retrievals used in the study are the pixelresolution surface rain rates created by the above three standard level-2 TRMM algorithms. The TRMM algorithms have evolved through three versions since the launch of the TRMM satellite. The day-1 launch versions are referred to as V4, the first postlaunch refined versions as V5, and the newly processed second postlaunch versions as V6. There are some significant differences in rain-rate magnitudes between these versions; however, the rainfall patterns and their spatial– temporal variations are very similar (Ikai and Nakamura 2003; Dinku and Anagnostou 2005; Olson et al. 2006; Yang et al. 2006). At the time this study was completed, 1 test month of data had been reprocessed with the V6 algorithm (i.e., February 1998). Results from that month have been used to prepare the diagrams presented in Figs. 1 and 2 (to be discussed in section 5). The remaining results were produced from the first postlaunch version V5 (i.e., Table 2 and Figs. 3–14). Sensitivity tests have been conducted to ensure that all results obtained from the V5 version are only marginally changed with respect to V6 refinements. Hirose and Nakamura (2005) demonstrated that TRMM PR sampling produces biases at a 1-h temporal scale. As noted, in order to minimize the impact of this sampling bias but to allow the study of monthly modulated variations of the diurnal rainfall cycle, fullresolution rain rates are assigned to 5° ⫻ 5° grid meshes for each TRMM orbit. Then, all pixels within each grid box are accumulated and averaged over each month for the time intervals of 3 h, 12-h daytime and nighttime, and 24-h daily. The mean rain rate for each grid box is

15 OCTOBER 2006

YANG AND SMITH

5197

FIG. 1. The distributions of (a) daytime (0600–1800 MST) and (b) nighttime (1800–0600 MST) monthly rainfall accumulation over the global Tropics for February 1998 produced by the version 6 2b31, 2a25, and 2a12 standard TRMM profile algorithms (top, middle, and bottom, respectively). The color bar denotes mm accumulation.

given by the arithmetic average, that is, the total rainrate accumulation divided by the total pixels assigned to the box. This procedure is also applied to seasons whenever seasonal–diurnal rainfall variations are under

consideration (i.e., the diagrams of Figs. 9–14). We have found that the monthly–seasonal averages at a 3-h time step over the 5° ⫻ 5° grid mesh represent an optimal means to spatially and temporally decompose 1 yr

FIG. 2. The ratio of daytime to nighttime monthly rainfall accumulations for the TRMM algorithm 2a25 (as given in Figs. 1a,b). The grayscale denotes a unitless ratio.

5198

JOURNAL OF CLIMATE

TABLE 2. Correlations between diurnal rainfall variability amplitude and phase distribution quantities from the TRMM algorithms 2a12, 2a25, and 2b31 for January and July 1998. January 1998

2a12–2a25 2a12–2b31 2a25–2b31

July 1998

Amplitude

Phase

Amplitude

Phase

0.85 0.84 0.98

0.74 0.71 0.93

0.86 0.87 0.98

0.67 0.67 0.93

of TRMM data to investigate diurnal rainfall features. Nonetheless, these spatial and temporal resolutions could be increased by compositing multiple years of data.

5. Large-scale diurnal variability Coherent precipitation variations occur over a range of time scales from semidiurnal to decadal. Ultimately, all of these variations are important in the context of weather, climate, and hydrology, and thus it is important that all of these modes of variability are understood, measured, and hopefully at some point in the future, reproduced in prognostic environmental models. Because diurnal variability is such a pronounced component of the atmosphere, it has the possibility of resonating with variational modes at other time scales, and thus has to be given careful attention.

a. Day–night variations At the diurnal time scale, day–night variation is the prominent variability mode over most regions of the earth. Figures 1a,b illustrate maps of day and night rainfall accumulations for February 1998 from the V6 versions of the three level-2 TRMM standard algorithms 2a12 (TMI only), 2a25 (PR only), and 2b31 (PR/TMI combined). At the planetary scale, the day and night rainfall distribution patterns for all three algorithms are similar, which bodes well for deriving conclusions concerning monthly-scale diurnal variability from any individual algorithm. Inspection also reveals that the macroscale rainfall distribution patterns from day to night for any given algorithm are similar, with the most significant regional-scale day–night variations concentrated within the climatologically heavy rainfall regions, particularly the ITCZ, the SPCZ, the major rain forest regions, the Australian monsoon, and those portions of the baroclinic storm observed by the TRMM satellite. Qualitatively, it is not apparent from Figs. 1a,b whether any particular algorithm tends to run higher or lower than the others, therefore this issue is addressed quantitatively at a later point. It is also not immediately

VOLUME 19

apparent whether any given algorithm exhibits greater or lesser day–night changes, although each algorithm tends to exhibit its own unique regional features in this regard. Perhaps the most notable region where all three algorithms exhibit fairly dramatic day–night changes is within the V-like intersection of the ITCZ and SPCZ (referred to hereafter as the “V-sector”). Regardless of detailed day–night differences amongst the algorithms, the near-equivalent distribution patterns indicate that both narrow-swath (2a25, 2b31) and wide-swath (2a12) algorithms at the monthly scale are in close agreement vis-à-vis rain coverage for the two 12-h diurnal periods. As noted, the rainfall distribution patterns for the day and night periods are similar at the tropical macroscale although not equivalent. On the other hand, day–night diurnal amplitudes are considerably different for specific regions. Figure 2 depicts the spatial distribution of the ratio of daytime to nighttime rainfall accumulations (from Figs. 1a,b) for algorithm 2a25 [the ratio patterns for 2a12 and 2b31 (not shown) are virtually identical]. The main features evident in this diagram are that (i) major portions of the continents indicate values greater than 1.0, signifying more daytime than nighttime rainfall over those regions; (ii) major portions of the oceans indicate values less than 1.0, signifying more nighttime than daytime rainfall; (iii) some portions of the oceans exhibit coherent daytime maxima (ratios ⬎ 1.0), particularly eastern basins off the west coasts of continents and where extensive marine stratocumulus cloud fields persist, as well as very isolated ocean features, such as the mid Pacific Ocean, the southwest Indian Ocean, the central Atlantic Ocean, the eastern Caribbean Sea, the North Pacific, and the southeast Pacific; and (iv) some regions over various continents exhibit coherent nighttime maxima (ratios ⬍ 1.0), mainly portions of subtropical Africa, Australia, and South America, as well as eastern tropical South America. Over oceans, regions where ratios exceed 1.0 (daytime maxima) generally exhibit weak rainfall (with the exception of the persistent marine stratocumulus areas), and conversely for continents, regions where ratios fall below 1.0 (nighttime maxima) generally are subtropical. There are also a few small areas where ratios exceed 1.0, which are the result of retrieval noise, that is, when small-value nighttime denominators are accompanied by insignificant differences in day–night values, but with the noise-contaminated denominators slightly smaller than the correspondingly small and noisy daytime numerators (i.e., regions of very little rainfall accumulation). Overall, however, there is a greater area where nighttime rainfall exceeds daytime rainfall over

15 OCTOBER 2006

YANG AND SMITH

oceans in statistically significant terms, with the opposite behavior over continents. Insofar as the marine stratocumulus (MSC) areas, the results indicate that the associated rainfall, which is mostly drizzle and warm rain, peaks during daylight hours. This is counterintuitive to some who argue from a modeling perspective that the MSC drizzle maximum should occur at night along with a cloud cover maximum, peaking well after the period during daytime when solar-induced convective erosion of the stratocumulus cloud decks has been generated by cloud layer warming and thermodynamic decoupling. Notably, the diurnal properties of ocean MSC drizzle have never been thoroughly observed or modeled [see Bougeault (1985), Turton and Nicholls (1987), Duynkerke (1989), Blaskovic et al. (1991), Duynkerke and Hignett (1993), Rogers et al. (1995), Rozendaal et al. (1995), Smith and Kao (1996), and Garreaud and Muñoz (2004) for studies of the afternoon cloud minimum]. This is the result of the measurement of drizzle being so elusive, which in turn has upheld a consistent theory involving the diurnal behavior of cloudiness and rainfall as a coupled process. Regardless, there is observational evidence for a daytime maximum in precipitation even given the largely substantiated minimum in daytime cloud cover. For example, synthesizing the results found by Miller et al. (1998) and Ciesielski et al. (2001) from Atlantic Stratocumulus Transition Experiment (ASTEX) data, it appears that the daytime MSC boundary layer exhibits significantly greater water vapor mixing ratios and an afternoon maximum in precipitating drizzle arising from vigorous cumulus cells overshooting the trade inversion base. Therefore, and with some caution because TRMM algorithms exhibit their greatest uncertainties at the small-valued end of the rain-rate spectrum, these new results suggest that MSC stratocumulus rainfall coherently peaks during daytime throughout all ocean basins at the relatively large scales associated with these climatically important clouds. To better quantify algorithm-to-algorithm differences, and concomitantly their consistency in representing bulk diurnal rainfall behavior, the mean monthly zonal daytime and nighttime rainfall accumulations over ocean and land are examined for the entire 1998 calendar year (figures not shown). Over ocean for January 1998, algorithm 2a25 generates the lowest totals while 2a12 generates the highest. Over land for the same month, algorithm 2a25 again generates the lowest totals while 2b31 generates the highest (these same minima–maxima are found for the February 1998 V6 results). However, regardless of these magnitude differences, the zonal patterns are structurally similar. For

5199

January, the peak of heavy rainfall over ocean is confined within the 0°–5°S equatorial band but with large values on either side progressively decreasing into the subtropics. Over land, the heavy rainfall peak is spread out over a wider zone, between 0°–25°S. Over ocean, the day-to-night zonal average ratios are generally less than 1.0 for the heavy rainfall zones. Over land, the patterns of the day-to-night zonal average ratios are more complicated. Although most ratios exceed 1.0 in the heavy rainfall regions, there is no distinct heavy rainfall zone between 20°–10°S in which ratios are significantly less than 1.0. This is akin to ocean behavior and denotes a second mode of variability over the continents mimicking the primary ocean mode, although it is presumed that the cause of the nighttime continental mode may be different than that of the nighttime oceanic mode. Overall, it is found that there is greater intermittency in the primary diurnal regulation process over land in conjunction with its major MLA maximum than there is over ocean in conjunction with its major LE–EM maximum. These analyses have been conducted for the other months of 1998 with each set of monthly results indicating the same basic diurnal phenomena, distinguished only by the seasonally evolving spatial distributions of the heavy rainfall regions.

b. Mean monthly–diurnal cycle at 3-h time scale Figures 3a,b illustrate the 3-hourly evolution of the spatial distribution of accumulated rainfall for January 1998 based on algorithm 2a12. The large-scale spatial patterns from period to period have obvious similarities, but overall, they describe a dynamically changing situation largely produced by diurnal variations in rainfall intensity in the heavy rainfall regions and associated variations in the locations of the relative maxima. For example, over the Indian and Pacific Oceans, there are major peaks in the early-morning period between 0300 and 0600 mean solar time (MST; i.e., the LE–EM mode). Also, there is a secondary peak over the Pacific in the 1500–1800 MST period—the MLA mode. The tropical eastern Atlantic exhibits a late-morning maximum between 0900 and1200 MST, a feature described by Gray and Jacobson (1977) as a delayed morning oceanic rainfall maximum. Further inspection reveals additional major differences in rainfall peaks over the eastern and western Indian Ocean tied to a diurnally intermittent east–west dipole; there are similar processes taking place over the western and central Pacific Ocean. Over central South America, the major peak takes place between 1800 and 2100 MST, occurring right after the nominal MLA mode, with a significant secondary peak occurring in the 0600–0900 MST pe-

5200

JOURNAL OF CLIMATE

VOLUME 19

FIG. 3. The distributions of 3-hourly monthly rainfall accumulation (mm) over the (a) first four and (b) last four 3-h periods of the diurnal cycle for January 1998 produced by the 2a12 algorithm.

riod, corresponding to the LE–EM mode. Over southern Africa, two diurnal peaks are also evident in the 1500–1800 and 0300–0600 MST periods—modes corresponding to the two found over central South America (these same peaks are found in the 2a25 and 2b31 results). To more clearly emphasize the roles of oceans and continents in organizing the global-scale pattern of diurnal rainfall variability, Fig. 4 presents the diurnal– longitude cross section for January 1998, again for algorithm 2a12, averaged over the boreal winter’s heavy rainfall zone of 0°–15°S. This diagram reveals an eastto-west sequence of early-morning peaks, almost all over ocean, and a less clearly defined sequence of afternoon peaks, with each major continental area in the Southern Hemisphere (Africa, the Maritime Continent, and South America) exhibiting relative maxima along with other afternoon relative maxima situated over various sectors of the Indian and Pacific Oceans. There

is a suggestion of eastward transiting wave behavior in the western Pacific, although it is a subtle feature. To elucidate the primary modes of monthly diurnal variability but resolved over a spatial grid, Fourier analysis is applied to the January 1998 results for all three level-2 algorithms to produce a diurnal-amplitude

FIG. 4. Diurnal longitude cross section of rainfall accumulation (mm) over the 15°S–0° zone for January 1998 produced by the 2a12 TRMM algorithm.

15 OCTOBER 2006

YANG AND SMITH

5201

all systematic differences between the algorithms, and (ii) the fact that 2a12 wide-swath coverage is 3 times that of 2a25 and 2b31 narrow-swath coverage, noting that 2a25 and 2b31 are in greater agreement with each other than either is to 2a12]. Another key feature is that the spatial pattern of the strongest amplitudes largely mimics that of the heavy rainfall pattern. Regardless of whether this arises from statistical principles or from physical necessity, it is important at this stage to recognize that strong diurnal variability couples with heavy rainfall. The map of phase factors associated with algorithm 2a12’s amplitudes is illustrated in Fig. 5b. Phase is plotted in units of diurnal time, that is, the hour associated with the rainfall maximum of the first Fourier harmonic. The diagram also uses shading to denote the regions exhibiting the largest amplitude factors, that is, regions where the phase information is perhaps most important insofar as diurnal rainfall activity (as with amplitudes, the spatial patterns of phase factor for the other two algorithms are similar; see Table 2). An evident feature of the phase distribution is that the timing of the daily rainfall maximum is a heterogeneous variable. However, most ocean regions indicate some type of coherent morning maximum, with a notable exception in the tropical central Pacific just west of the date line where a coherent afternoon maximum is located, consistent with what Augustine (1984) found 20 yr ago. At the same time, the major continents of Africa, Australia, and South America all exhibit coherent afternoon maxima for significant portions of the strong amplitude regions and morning maxima for the weak amplitude regions. This mixture of early-morning and midafternoon maxima over both oceanic and continental domains in which land areas exhibit the greatest heterogeneity must be considered a fundamental property of rainfall diurnal variability.

FIG. 5. (a) The distributions of amplitude (mm) in diurnal rainfall variability for January 1998 produced by the 2b31, 2a25, and 2a12 TRMM algorithms. (b) Same as in (a), but for phase (MST at a maximum of the first diurnal Fourier harmonic) and just for the 2a12 algorithm. The shaded areas indicate the regions of the largest amplitudes.

c. Annual evolution of mean monthly diurnal variation

phase framework. On a 5° ⫻ 5° latitude–longitude grid mesh, the first Fourier harmonic is extracted for each grid cell along with its respective amplitude and phase factors. Mapping the amplitude factors, as shown in Fig. 5a, enables an examination of the varying intensity of the primary diurnal harmonic. One of the more salient features of this diagram, as quantified in Table 2 with correlation analysis, is that all three algorithms, although not in complete agreement in terms of amplitude intensity, are in considerable agreement insofar as the spatial distribution of the largest amplitudes [the amplitude differences arise from two sources: (i) over-

Figures 6a,b illustrate diurnal–monthly cross sections of rainfall over oceanic and continental regions from algorithms 2a12, 2a25, and 2b31 extending throughout the 1998 annual cycle. Over oceans, the distribution patterns for the three algorithms are similar, in which a predominant morning rainfall peak during the 0300– 0600 MST period (the LE–EM mode) occurs every month with its greatest amplitudes in January, May, August, and December. In addition, a secondary MLA peak occurs during the winter and spring months. Over continents, a consistent diurnal cycle is also evident for all three algorithms. A MLA peak persists each month with maximum amplitudes occurring during the autumn

5202

JOURNAL OF CLIMATE

FIG. 6. Diurnal–monthly cross sections of 3-hourly (a) oceanand (b) land-based rainfall accumulation (mm) for the entire TRMM coverage zone (35°N–35°S) over the 1998 annual cycle produced by the 2b31, 2a25, and 2a12 TRMM algorithms.

VOLUME 19

and winter months. As with oceans, there is a secondary LE–EM peak most evident during the months of January and September; an early study of diurnal rainfall variability by Belden (1936) found that at St. Joseph, Missouri, peak rainfall occurred during nighttime. Therefore, primary and secondary diurnal peaks occur over both oceans and continents, with their diurnal phasing reversed. On a monthly time scale, morning rainfall maxima over oceans and afternoon maxima over continents are persistent. The secondary maxima are generally a regular afternoon feature for oceans while a more intermittent morning feature for continents and tend to exhibit more seasonality in their occurrence than the primary maxima whose amplitudes vary from month to month but never disappear. The ocean result is substantiated by the recent study of Serra and McPhaden (2004), who reported from an analysis of moored buoy precipitation data from tropical Atlantic and Pacific Ocean areas that selected locations indicate both early-morning and afternoon maxima in rainfall accumulation. In addition, Mohr (2004) found from a TRMM-TMI data analysis over the Sahara desert that the bimodal precipitation behavior tends to occur north of 10°N, particularly in dry years. The most important feature of these month-to-month changes in the diurnal rainfall cycle is the seasonal variation of diurnal intensity, particularly in the largescale convergence zones. To emphasize this point, an analysis of the annual evolution of diurnal rainfall variation has been carried out over the heavy tropical precipitation regions embedded within the equatorial trough, including the Asian monsoon precipitation system, and the persistent rainfall portion of the V sector. Figure 7a illustrates time–longitude cross sections of diurnal cycle amplitude and phase averaged over the 10°N–10°S deep tropical latitudes produced by all three level-2 TRMM algorithms. To facilitate interpretation, the corresponding cross sections of accumulated rainfall produced by each of the three algorithms is provided in Fig. 7b. Each of the amplitude diagrams (lefthand side of Fig. 7a) consistently indicate propagation of the diurnal peak amplitude away from the international date line in the vicinity of the warm pool to the west, with weaker propagation from the same origin to the east. This is in response to the transition from the 1997–98 El Niño to the 1998–99 La Niña in which the strong active ENSO convection region over the warm pool diminishes once the large-scale zonal circulation cells begin to migrate to their so-called “neutral” configurations. However, we note that the propagation of the diurnal rainfall peak might also be associated with the seasonal characteristics of the diurnal rainfall cycle. In addition, the significant spatial variations of the di-

15 OCTOBER 2006

YANG AND SMITH

FIG. 7. (a) Month–lon cross sections of amplitude (mm; lefthand side) and phase (MST at a maximum of the first diurnal Fourier harmonic; right-hand side) of diurnal rainfall variability within the 10°S–10°N zone over the 1998 annual cycle, produced by three level-2 TRMM algorithms. The labels indicate the African continent (AF), Indian Ocean (IO), west central Pacific Ocean (W-CP), eastern Pacific Ocean (EP), South American continent (SA), and Atlantic Ocean (AO). The date line is shown with thick blue arrows and propagation paths are shown with thin red arrows. (b) The rainfall accumulation (mm) associated with (a).

5203

urnal rainfall cycle over the oceans and its coastal regions, especially around the Maritime Continent, could complicate an explanation of rainfall’s diurnal propagation. These uncertainties should be assessed in the future by analyzing multiple years of TRMM data. The evolution of the positioning of the peak diurnal amplitudes during 1998 corresponds almost exactly with the evolution of the location of total rainfall within the equatorial trough that same year. As evident in Fig. 7b, there is a migration in the maximum rainfall pattern away from the warm pool to both west and east, that is, the expected transformation process from an active to an inactive ENSO climate (see Bell et al. 1999; Trenberth et al. 2002). Moreover, these results emphasize that the annual cycle of rainfall diurnal variability is closely associated with the positions of heavy convective rainfall systems developing within the convergence regions of the deep Tropics, regions that undergo interannual shifts due to coupled ocean–atmosphere climate-scale perturbations. Another way to describe this variability is to simply note that the ENSO process includes within its multifaceted governing mechanisms a strong diurnal component associated with enhanced warm pool rainfall. Notably, the amplitude distribution patterns in Fig. 7a for each of the three algorithms are consistent. There is a persistent rainfall peak over central-west South America throughout the year with maximum amplitudes in September and October, accompanied by strong amplitudes in the central-east Pacific from February to May. Over the tropical Indian Ocean, the diurnal amplitude peaks in May and diminishes in September, with evidence of a monsoon surge-break cycle. There is also a large amplitude region in the west Pacific that exhibits the greatest values throughout the boreal autumn–winter–early spring period when the ITCZ is mainly active within the deep Tropics. Once boreal summer commences and the ITCZ both propagates northward and becomes more intermittent, the diurnal amplitudes diminish. There is also a shift in the highest amplitudes from the western end of the tropical west Pacific during spring toward the warm pool in early winter, associated with SPCZ convection and the excitation of enhanced convection in the wintertime Vsector. Over the tropical African rain forest, the largest amplitude peaks occur in September in conjunction with the rainy season. Over the equatorial Atlantic, weak diurnal amplitudes occur in December, possibly associated with convective waves. The corresponding phase distributions are found on the right-hand side of Fig. 7a. As with amplitudes, the three patterns of phase distribution are consistent. For the high-amplitude South American and African re-

5204

JOURNAL OF CLIMATE

gions, associated phase values are ⬃1200–1500 MST, indicating that the most intense rainfall activity occurs from early to midafternoon (or just prior to the MLA mode). Over the tropical Indian Ocean from early summer to midautumn [i.e., during the active southwest monsoon (May–October)], there are rainfall peaks in both early and late morning (0600–0900 MST/LE–EM mode and 0900–1200 MST), but mostly during the early afternoon (1200–1500 MST). The latter timing is consistent with a delayed early-morning ocean mode. There is an LE–EM peak in the west Pacific during the early part of the year, transitioning to the central Pacific in the latter part of the year. Conversely, and as discussed earlier in conjunction with Augustine (1984), early-afternoon peak amplitudes are found in the central Pacific during the early part of the year, with that phase feature shifting to the west Pacific as the year progresses. As with the Indian Ocean, this establishes seasonality in how the morning and afternoon diurnal rainfall maxima materialize throughout the deep tropical western Pacific. Similar phenomena are found in the eastern Pacific where morning maxima are most prominent, but there are peak diurnal amplitudes occurring in the early afternoon toward the eastern side of the region throughout July–September. In the Atlantic Ocean, the diurnal maxima can occur anytime from early morning to late morning depending on the time of year. Overall, the diurnal amplitude-phase analysis of the deep Tropics indicates that the fundamental diurnal harmonic of rainfall can exhibit highly localized geographic properties, but properties in which the most significant features are directly associated with the occurrence of heavy rainfall. Contrasting differences in diurnal variability between the Indian and Pacific Oceans are evident from the above analysis. Additional insight into these differences is given by the analysis of the evolution of diurnal variability along the meridional axis. Figure 8a illustrates month–latitude cross sections of the diurnal amplitude of rainfall variability produced by the 2a12 algorithm, averaged over the 50°–110°E (upper panel) and 150°E– 150°W (lower panel) longitude sectors. It is evident that within the Indian Ocean Basin (upper panel of Fig. 8a), the maximum amplitude is located at ⬃7.5°S in January, with a persistent relative peak amplitude at that same latitude throughout the year. This corresponds to persistent intraseasonal tropical convective waves that occur over the equatorial Indian Ocean (e.g., see Jones et al. 2004). However, the key feature in this diagram is that the peak amplitude of diurnal rainfall variability migrates from its ⬃7.5°S latitude position in January to an ⬃15°N latitude position in May, propagating farther northward to ⬃18°N in mid-June (where the amplitude

VOLUME 19

FIG. 8. Same as in Fig. 7, but for month–lat for (a) 50°–110°E and (b) 150°E–150°W sectors produced by the 2a12 TRMM algorithm.

achieves its maximum value for the year), farther northward to ⬃30°N in August, and finally retreating back to an equatorial position in December. This process describes the initiation, progression, and withdrawal of the southwest Indian monsoon but cast in a diurnal rainfall variation framework. In contrast to Indian Ocean variability, diurnal rainfall variability in the central Pacific Ocean (lower panel of Fig. 8a) is characterized by a less extended south-tonorth progression over the course of the year. Some-

15 OCTOBER 2006

YANG AND SMITH

what similar to the Indian Ocean, the maximum amplitude in January is ⬃10°S, but over the mid-Pacific, northward propagation extends only to ⬃10°N in July, retreating back to ⬃7°N in December. Moreover, and unlike the Indian Ocean, a relative maximum reemerges in the vicinity of 10°S in late summer and persists throughout the remainder of the year (August– December), an apparent bifurcation in the maximum amplitude progression. In fact, this process is the manifestation of diurnal rainfall variability concurrent with the annual migration of the ITCZ. Although it is not entirely consistent with the conventional understanding that the annual climatological ITCZ migration extends from 7°S to 10°N in the course of an annual cycle (Waliser and Gautier 1993), it is physically consistent with ITCZ migration dynamics. What actually takes place in terms of precipitation variability is first an increase in rainfall during the excitation of the V sector in January at ⬃10°S, that is, the amplified merger of the ITCZ and SPCZ that has been described in detail by Vincent (1994). This gives rise to the boreal winter maximum in the Southern Hemisphere seen in the lower panel of Fig. 8a. As boreal spring and summer commence, the ITCZ shifts north in response to the northerly migration of the trade winds needed to satisfy global mass-momentum conservation constraints. There is a concomitant shift in rainfall, and as summer wanes, the ITCZ begins its reverse migration southward. At the same time, the SPCZ begins to reactivate during austral summer, which produces the appearance of bifurcation in propagation activity. There is an additional high-amplitude feature appearing in the 30°–35°N zone at the end of 1998 during late summer–early winter (boreal). This represents storm-track surges into the subtropics associated with tropical storm activity (see Chang et al. 2002). The phase distributions associated with the amplitude diagrams given in Fig. 8a are shown in Fig. 8b. Over the Indian Ocean (upper panel), the dominant phase peak is from late morning to early afternoon for spring and summer, while early to midmorning for autumn and winter. This describes a shift from a delayed early-morning peak to an LE–EM peak, generally consistent with what has been described in the Fig. 7 result. Over the Pacific Ocean (lower panel), there is a nearly consistent LE–EM phase peak over the 20°S–20°N latitude zone except in the Southern Hemisphere during January–April. The early-year Northern Hemisphere subtropics exhibit a coherent MLA maximum where rainfall is weak, and possibly a counterpart in the Southern Hemisphere at nearly the same time. Clearly, the Pacific Ocean is the dominant basin of the two insofar as maintaining an LE–EM morning maximum in

5205

rainfall within the heavier rainfall region. In essence, the Indian Ocean exhibits an LE–EM maximum during its cold season, switching to an early-afternoon maximum during its warm season. This localized seasonal heterogeneity in amplitude and phase is intrinsic to diurnal rainfall variability, beyond the mean features produced by the primary LE–EM oceanic and MLA continental modes.

d. Possibility of semidiurnal tidal forcing One interpretation of the double maxima for both oceans and continents is that they are forced by the atmosphere’s semidiurnal tide induced by the S2 solar atmospheric heating mode (see Brier 1965 for discussion). However, in both cases, the phasing does not correspond to the semidiurnal mode, and thus all that should be said from our observational analysis is that if there is a semidiurnal harmonic embedded in the rainfall, then it is either being phase shifted by data averaging or dominated by other physical processes that produce the dual maxima [see Landin and Bosart (1985) for further discussion of this issue]. This interpretation is not inconsistent with that of Carbone et al. (2002), who investigated a very pronounced summertime eastward propagation of rainbands between the Rocky Mountains and Appalachian Mountains using very high space–time resolution radar data [2 km (15 min)⫺1]. Their study found that semidiurnal controls appeared to occur between the two mountain ranges distanced from the strong diurnal forcing produced by the mountains themselves, but that other physical processes controlling the diurnal cycle prevented the semidiurnal mode from dominating precipitation except along a single meridian situated between the two mountain ranges.

6. Seasonal-scale analysis of diurnal variability The results presented in the previous section define various coherent modes of diurnal rainfall variability at the monthly time scale, the most salient modes being the primary LE–EM and MLA modes over oceans and continents, respectively, and their associated MLA and LE–EM secondary modes. Attention is also given to the delayed early-morning mode over ocean and other fluctuations in timing around the primary modes, especially for continental rainfall. Intuitively, the primary oceanic and secondary continental modes would be produced by similar or equivalent physics [supported in part by the Tao et al. (1996), Sui et al. (1997, 1998), and Dai (2001) studies], as would the primary continental and secondary oceanic modes [as Sui et al. (1997, 1998)

5206

JOURNAL OF CLIMATE

VOLUME 19

FIG. 9. The tabulation of rain events over the TRMM satellite coverage domain during July 1998. The matrix of diurnal clocks, each divided into eight 3-h segments and each representing one 5° ⫻ 5° grid box, graphically illustrates rain sampling (symbolic clock key shown in the bottom right corner). Six gray shades denote how many rain events are observed by TMI radiometer in each grid box during each diurnal time segment in the course of a month. The grayscale in the top right corner shows values to the left that denote the number of rain events in terms of a range of values over the gray-shaded category (including zero category given as white) and values to the right that denote category-specific bin population percentages with respect to the total bin population (8 ⫻ 14 ⫻ 36 ⫽ 4032).

have suggested]. Nonetheless, these associations have not been objectively proven through the monthly analyses, and as stressed in the introduction, there are various competing theories as to the origin of diurnal variability over oceans and continents. For example, the studies of Silva Dias et al. (1987), Yang and Slingo (2001), Mapes et al. (2003a,b), Warner et al. (2003), and Rickenbach (2004) draw attention to how different surface and atmospheric mechanisms may manifest themselves in transferring diurnal signals from interior continents to coastal regions, and on into oceanic regions, or in the case of the modeling study by Sato and Kimura (2003), how diurnally controlled wind and water vapor concentration interact to propagate precipitation from the lee of a mountain out over the flat open plain. In this regard, it is important to recognize that the results presented in section 5, particularly the diagrams in Figs. 2 and 5b, provide concrete evi-

dence that the modes of diurnal variability are spatially heterogeneous on a monthly time scale over both oceanic or continental regions. Presumably, much of the observed diurnal mode heterogeneity is real. However, as noted in section 3, part is erroneous because of TRMM undersampling when considering an individual monthly grid box at a spatial scale of 5° ⫻ 5°. To quantify rain sampling for a month at this scale, Fig. 9 illustrates graphically over the TRMM coverage domain how many rain events are observed within the array of 5° ⫻ 5° grid boxes over the set of eight 3-h diurnal time segments in the course of the month of July 1998 as seen by the TMI radiometer. For the purpose of this analysis, a rain event is counted for a given 5° box within a given segment if rainfall is detected by TMI during the course of a TRMM overpass. In the diagram, a matrix of diurnal clocks, each divided into eight segments and each representing one

15 OCTOBER 2006

YANG AND SMITH

5° ⫻ 5° grid box, indicates by use of six gray shades how many times a rain event is observed in the course of the month (a symbolic clock key is seen in the lower right corner of the figure). The grayscale is located to the right of the map panels in which numeric values to the left of the bar denote rain event counting intervals for each of the six gray-shade categories (including a zero category), while values to the right of the bar denote category-specific bin population percentages with respect to the total bin population (indicated below the bar, i.e., 4032). It is apparent from Fig. 9 that much of the TRMM coverage zone exhibits robust sampling around the diurnal clock, that is, rain event counting within individual diurnal segments generally runs between 5 and 23. There is less robust rain event sampling within the climatological dry regions, that is, within (i) dry season central South America and southern Africa; (ii) the Saharan, Arabian Peninsula; Rajasthan, India; and central Australian deserts; (iii) the eastern sectors of subtropical high regions off North American, South American, and African west coasts where marine stratocumulus clouds prevail; and (iv) the western Arabian Sea. Over the comparatively wetter regions, there are also cases of individual diurnal segments for which rain event sampling is low (below 5) and even cases of zero. We emphasize that in the prior analyses discussed in section 5, only Figs. 3a,b illustrating the diurnally changing rainfall accumulation pattern for January 1998 and Figs. 5a,b illustrating the amplitude-phase patterns for the same month contain results at the 5° ⫻ 5° scale, noting that the discussion associated with these figures avoids exact quantitative interpretation. It is because of this intermittent undersampling at the 5° scale that caution must be given in interpreting results at this scale, that is, results that have not undergone further spatial or temporal averaging as has been done for the results shown in Figs. 1a,b; 2; 4; 6a,b; 7a,b; and 8a,b. Finally, one of the advantages of using the first harmonic Fourier analysis is that it overcomes some of the effects of undersampling because the scheme acts as a smoothing operator, thus imposing a good degree of spatial integrity on its rendition of diurnal variability. This is borne out in the following seasonal analysis where it is evident that the January 1998 amplitudephase distributions are similar to the winter season distributions (see in Figs. 14a and 15a).

a. Global interpretation at the seasonal scale Given the possible uncertainties with the monthly analyses conducted over the 5° grid mesh, we examine diurnal variability at the seasonal scale (3-month averaging framework), which serves to mitigate rain event

5207

FIG. 10. The diurnal variation of seasonal rainfall accumulation (winter, spring, summer, and autumn) averaged separately for land and ocean over the TRMM observation zone (35°S–35°N) for the 1998 annual cycle using the 2a12 standard v5 TRMM profile algorithm

undersampling and enables focusing on the most robust diurnal variability modes. Figure 10 involves the use of the boreal winter, spring, summer, and autumn seasonal classifications to illustrate how diurnal variability materializes when considering the entire TRMM observation zone. This diagram presents the diurnal variability of rainfall, averaged over ocean and land separately from 35°S to 35°N, for the four seasons of 1998. These results are obtained from TRMM algorithm 2a12 with the analysis using the following standard seasonal definitions: (i) winter consists of the months December– February (DJF); (ii) spring consists of March–May (MAM); (iii) summer consists of June–August (JJA); and (iv) autumn consists of September–November (SON). It is apparent in Fig. 10 that the primary and secondary diurnal modes identified in the monthly analyses are very pronounced in the seasonal analysis. For ocean, the primary LE–EM mode is the most evident and strongest during winter. The secondary MLA ocean mode is also the most pronounced during winter, but also readily apparent in the other three seasons. Over land, the primary MLA mode is strong for all four seasons, particularly for autumn. In the case of the secondary continental mode, the LE–EM peaks are the most apparent for the two transition seasons of spring and autumn, and although weaker, still detectable for winter and summer, with the winter season’s relative maximum tending toward early morning (EM) as opposed to late evening (LE), where it is positioned for the other three seasons. One might argue that the controls on diurnal variability are more complex over the continents because of the greater modulation of the amplitudes in the primary and secondary maxima, but the contrasts in amplitude between the primary winter

5208

JOURNAL OF CLIMATE

FIG. 11. Season–lon cross sections for four seasons of rainfall accumulation (mm) for the 10°S–10°N latitude zone over the 1998 annual cycle produced by the 2a12 algorithm.

ocean mode and its counterparts in the other three seasons tend to negate that argument. In any case, it is evident that there are seasonal fluctuations in the amplitudes of the diurnal maxima for a given mode but only minor variations in phases.

b. Regional interpretation at the seasonal scale Figure 11 is used to show how much global-scale coherency in the four prominent modes is preserved at a regional scale. This diagram presents the seasonal– diurnal longitude cross sections for the deep Tropics (10°S–10°N) over the annual cycle using algorithm 2a12. Starting in the Indian Ocean, (i) LE–EM peaks clearly occur in this basin, most pronounced in summer during the peak monsoon period but sustained for the rest of the year, (ii) relatively weak LE–EM peaks and rarer early-evening peaks occur over the Indonesian and far-west Pacific regions, (iii) seasonally intermittent but weak LE–EM peaks occur over the central-

VOLUME 19

west Pacific, (iv) strong LE–EM and MLA maxima occur in the central-east Pacific but only during winter and spring, (v) a strong LE–EM maximum occurs in the east Pacific during spring in conjunction with the somewhat rare early to midevening maximum, with the other three seasons tending toward diurnal quiescence, and (vi) seasonally fluctuating (weak to strong) LE–EM and MLA peaks occur over South America. These diurnal maxima always occur during spring, except for the central-west Pacific where its springtime diurnal maximum has begun propagating to the east in response to the development of the 1998 El Niño (see Bell et al. 1999). The Atlantic Ocean exhibits its own unique diurnal behavior. Here, weak LE–EM maxima occur during spring and summer, as well as the previously discussed delayed morning maxima that appear in summer and autumn. During wintertime, however, the Atlantic is diurnally quiet. Over Africa, it is evident that for the winter season there are both prominent LE–EM and MLA rainfall peaks, but they also occur during the other three seasons, noting that the afternoon maximum is sometimes delayed to early or midevening. As discussed in the previous section, most of the accentuated regional diurnal behavior is coupled with the intensity of the associated rainfall. The largest LE–EM peak in the Indian Ocean occurs in the summer when the monsoon produces annual maxima over the central and northern basins. Rainfall over portions of the Pacific Ocean is intermittently weak in the tropical zone (e.g., the central-west and central-east domains) and often over the subtropical zones. In such cases, the related afternoon peaks virtually always disappear. Throughout the spring–autumn period (either boreal or austral) when solar insolation is strong, the MLA maxima over land areas become strong and readily evident (a pronounced early-evening peak also occurs over the Indian Ocean during summer—a feature associated with the heavy summertime monsoon rains). LE–EM peaks almost always persist along the equatorial trough in the Pacific Ocean where ITCZ and SPCZ rainfall are the key features; however, it is only in confined regions highlighted by the central-west tropical Pacific [noted by Augustine (1984) in cloudiness data retrieved from geosatellite measurements] where the occasional prominent and sustained MLA maxima appear. Figures 12a,b present a similar analysis to that shown in Fig. 11, but in this case calculated in a seasonal– diurnal latitude cross sectional framework for rainfall averaged over the 50°–110°E and 150°E–150°W sectors (similar to the sectors defined in Figs. 8a,b; these results are also derived from the 2a12 algorithm). Over the

15 OCTOBER 2006

YANG AND SMITH

5209

FIG. 12. (a) Same as in Fig. 11, but for season–lat for the 50°–110°E sector. (b) Same as in (a), but for the 150°E–150°W sector.

Indian Ocean sector (Fig. 12a), a broad and strong LE– EM maximum is found throughout the four seasons, along with an indication of a delayed morning peak in spring and summer. There is also a broad but weaker LE–EM maximum in the Southern Hemisphere subtropics occurring south of 20°S during winter and spring, prior to the onset of the southwest monsoon. Moreover (as noted above), there is a relatively strong Northern Hemisphere subtropical MLA maximum in summer, when the heavy rainfall region extends to 35°N because of the penetration of the southwest–East Asian monsoon, but only a weak feature in spring. Clearly, as Indian Ocean rainfall propagates from 7° to 35° latitude during the progression from winter to summer, it carries with it a pronounced LE–EM diurnal mode. The Indian Ocean sector primary modes are also evident during autumn, except for the diurnal behavior disappearing from higher latitudes as the summer mon-

soon withdraws (the relatively greater rainfall accumulations in the morning and afternoon beyond 25°N are related to midlatitude storm systems). An identical analysis to what has been presented in Fig. 12a, except for a smaller region focusing on the strongest monsoon activity (not shown), was also conducted to verify the above features. These results reproduce the main features of Fig. 12a, except for indicating even larger LE– EM amplitudes associated with the intense monsoon rainfall. Over the central Pacific Ocean sector (Fig. 12b) during winter, the heaviest rainfall zone is between 25°S and 10°N, concentrated around 7°S and coincident with two distinct LE–EM and MLA peaks. Because of the northward movement of the ITCZ from its winter position in the Southern Hemisphere to its summer position in the Northern Hemisphere, the band of heavy rainfall in the central Pacific at approximately 7°S during winter evolves into two heavy rainfall bands in sum-

5210

JOURNAL OF CLIMATE

mer and autumn (around 7°N associated with the summer ITCZ and around 7°S associated with the SPCZ). Another rainfall area north of 20°N indicates a single MLA peak around 35°N, suggesting interactions with midlatitude systems. Similar features are seen in the Tropics during spring, except that the MLA peak is delayed until early evening. During summer and autumn, there are weak LE–EM maxima near 7°S, 7°N, and 20°N. It should be noted that the breadths of the summertime northern and southern rainband maxima are different. This difference denotes variants in the activation and timing of the principle mechanism(s) producing the LE–EM maxima in the ITCZ and SPCZ rain systems, but does not represent any basic departure from the behavior of the coherent and primary LE–EM oceanic mode. A further notable feature seen in both Figs. 12a,b is that the LE–EM rainfall peaks occur a few hours earlier in the heavy rainfall areas than in the surrounding areas. For example, the peak at the center of the wintertime mid-Pacific rainfall maximum is in the interval from 0300–0600 MST (LE) while the peaking in the surrounding areas takes place in the interval 0600–0900 MST (EM; i.e., a longer delay). Similar features are evident in the monthly analysis given in Fig. 4. These results, perhaps prone to alternate interpretations, are consistent with the ideas of Gray and Jacobson (1977), who maintained that the day–night oceanic cycle of radiative forcing and in-cloud convergence between deep convection and its surroundings (i.e., the DRC interaction) is responsible for underlying LE–EM rainfall peaks within the mid-Pacific, presumably in which the deepest convection would be the quickest to initiate the interaction. The regional results extracted from Figs. 11 and 12a,b are summarized in Fig. 13 to help differentiate diurnal rainfall variability over the tropical Indian Ocean (15°S–15°N, 50°–110°E) and the tropical mid-Pacific Ocean (15°S–15°N, 150°E–150°W). Over the Indian Ocean, the primary LE–EM maximum is persistent, appearing within its notional time frame during summer, autumn, and winter, but delayed to late morning (0900– 1200 MST) during spring. The secondary MLA oceanic peak appears unequivocally during spring and summer, but is not evident in autumn and winter. For the midPacific Ocean, the primary LE–EM peak persists for all four seasons while the secondary MLA peak is evident only for the spring and winter seasons.

c. Amplitude-phase interpretation at the seasonal scale Similar to what was done to produce Figs. 5a,b, amplitude-phase analysis is applied to the seasonal rainfall

VOLUME 19

FIG. 13. Diurnal–seasonal cross sections of rainfall accumulation (mm) for a region in the Indian Ocean (15°S–15°N/50°– 110°E; top) and a region in the Pacific Ocean (15°S–15°N/150°E– 150°W; bottom) over the 1998 annual cycle produced by the 2a12 TRMM algorithm.

averages so that spatial distributions of diurnal rainfall properties can be inspected systematically. Figures 14a,b present the amplitude distributions for winter and summer. The amplitude distributions for winter obtained from the 2a12, 2a25, and 2b31 algorithms are similar with respect to both general and specific features (Fig. 14a). Strong diurnal activity is observed within the major ITCZ and SPCZ oceanic convergence zones and over the South American and African continents as well as their surrounding areas. Maxima are located over Indonesia, the central-east Pacific Ocean, the eastern Indian Ocean, South America, and western equatorial Africa. Another active area is located within the extratropical Pacific where most rainfall emanates from midlatitude baroclinic storm systems and their interactions with tropical waves (Chang et al. 2002). During summer (Fig. 14b), peak rainfall amplitudes propagate northward with maxima located in the northern Bay of Bengal, the southeastern China Sea, the Japanese archipelago, and the eastern coastal Pacific, as well as North Africa and the eastern United States. The strong diurnal activity over the northern Bay of Bengal and South China Sea is related to the southwest East Asian summer monsoon. The strong maximum over Japan is related to the baiyu monsoon trough. A comparison of Fig. 5a with Figs. 14a,b indicates that the seasonal amplitude distributions from the three algorithms are more consistent than found with

15 OCTOBER 2006

YANG AND SMITH

FIG. 14. The distributions of amplitude (mm) of rainfall variability for the 1998 (a) winter and (b) summer produced by the 2b31, 2a25, and 2a12 TRMM algorithms. The shaded areas indicate amplitudes of 15, 30, and 45 mm, respectively, and the regions of the largest amplitudes for (a) and (b).

5211

monthly rainfall, whose spatial correlations were high to begin with (see Table 2). This simply denotes that the additional sampling gained by extending the averaging period by a factor of three helps reduce statistical noise. Furthermore, the results from Figs. 14a,b confirm that strong diurnal rainfall variability is most closely related to heavy rainfall systems—systems that undergo major seasonal variations in their locations but only minor variations in their diurnal characteristics. Figures 15a,b illustrate the associated phase distributions of the rainfall diurnal cycle for the winter and summer seasons. The related amplitudes of 15, 30, and 45 mm are indicated with shading, recalling that the phase distributions indicate times of peak rainfall for the first diurnal harmonic contributing to the total rainfall diurnal cycle (note that for both figures, the phase distribution patterns from the three algorithms are in qualitative agreement). The heterogeneity of phase is evident in these diagrams, but as noted with the Fig. 2 result, most ocean areas indicate distinct LE–EM maxima and most continental areas indicate distinct MLA maxima. Clearly, in winter, the high-amplitude ocean regions are dominated by the LE–EM mode, particularly over the ITCZ and SPCZ convergence zones where diurnal amplitudes are at their peak values. Nevertheless, afternoon phase peaks are found over the tropical central-west Pacific and east Pacific in the more moderate amplitude regions, early-afternoon phase peaks are apparent over the midlatitude storm track regions in both hemispheres, and an evening phase peak occurs over the South China Sea. Conversely, the MLA mode dominates the high-amplitude continental regions of Africa and South America. Different phase patterns materialize for the summer (Fig. 15b). An LE–EM maximum extends over most of the Pacific and Atlantic Oceans, while an early-afternoon maximum dominates the north Indian Ocean. In addition, LE–EM maxima are apparent over the Bay of Bengal and the inland Indian monsoon region, while an early-afternoon phase peak persists within the strong amplitude region over the South China Sea. An MLA phase peak is found over Africa as well as the Americas. Finally, there are early-afternoon maxima over the coastal areas of China [perhaps from the coastal gravity wave migration (CGWM) mechanism], coincident with late-morning and early-afternoon phase peaks within the mainland. These diurnal features of rainfall over the South China Sea and its adjacent regions are consistent with those from a cloudiness analysis by Asai et al. (1998). This denotes that diurnal rainfall variability over southern China, whose central-southern eastern zone is frequented by summertime mesoscale convec-

5212

JOURNAL OF CLIMATE

VOLUME 19

tive systems propagating off the Tibetan Plateau, is mostly heterogeneous during its convection season.

d. Closing discussion on seasonal-scale analyses Synthesis of the results presented in this section support the argument that diurnal rainfall variations are significant in both tropical and subtropical land and ocean areas, especially over heavy oceanic rainfall regions such as within the ITCZ, SPCZ, and Asian monsoon and the continental rain forest zone. Moreover, these results emphasize the dominance and perseverance of the two primary modes and the presence of the two secondary modes. Although the exact timings of rainfall maxima vary from place to place and from season to season and with a few notable exceptions [such as the delayed early-morning primary mode of the eastern tropical Atlantic and Indian Oceans during spring and summer and the delayed MLA mode to early or midevening occurring intermittently over Africa and South America (delayed primary) and selected portions of the Pacific Ocean (delayed secondary)], the coherent primary and secondary modes that have been described explain most diurnal variability that takes place in a time-averaged framework, especially at the seasonal scale. Therefore, the two-each primary and secondary modes define the mainstream behavior of diurnal rainfall variability.

7. Discussion and conclusions

FIG. 15. Same as in Fig. 14, but for phase (MST at a maximum of the first diurnal Fourier harmonic). The shaded areas indicate the regions of the largest amplitudes.

Rainfall retrievals from the three level-2 TRMM standard algorithms have been analyzed in a diurnal framework for developing a better understanding of the nature of diurnal rainfall variability, the space–time distribution of the various diurnal modes of variability, and the mechanisms producing the various modes. One year of TMI and PR data from 1998 has been used in conjunction with TRMM level-2 algorithms 2a12, 2a25, and 2b31 to conduct the study. Testing for repeatability among multiple algorithms is a means to help confirm the reliability of each of the individual algorithms, and given the consistent results produced by the three algorithms, it is concluded that the analyses presented in this investigation are sound. It is evident that the detailed diurnal properties of rainfall variability are more consistent between 2a25 and 2b31 (the narrow-swath PR-only and PR/TMI combined algorithms) than either is with 2a12 (the wide-swath TMI-only algorithm), which is expected since 2a12 provides 3 times the duty cycle (sampling coverage) that the two narrow-swath algorithms do. However, for monthly composited analyses, even on a 3-hourly basis, these differences are

15 OCTOBER 2006

YANG AND SMITH

minor insofar as the salient properties of diurnal rainfall variability. On a seasonal basis, the individual algorithm results are in even closer agreement, demonstrating that the less than one-third sampling capacity of the narrow-swath PR (relative to the wide-swath TMI) does not produce significant undersampling insofar as quantifying the pattern of diurnal rainfall variability and its evolution. In general, the investigation has shown that diurnal rainfall variability is a heterogeneous variable over both continents and ocean, but that within these two distinct heterogeneous patterns are both large- and small-scale coherent patterns, and most importantly, dominant features in amplitude and phase properties. First, there are more areas over continents with greater rainfall during daytime than nighttime by a wide margin and conversely, more areas over oceans with greater rainfall during nighttime than daytime by an even wider margin. Moreover, most continental daytime variability exhibits the MLA (1500–1800 MST) maximum while most oceanic nighttime variability exhibits the LE–EM maximum (0300–0900 MST). There are obvious departures from this behavior, but most phase departures from these two primary modes can be described as “opposed secondary modes,” that is, a secondary LE–EM maximum for continents and a secondary MLA maximum for oceans. In essence, these are the overriding properties of diurnal rainfall variability in the Tropics and subtropics. Our results appear to depart from a claim recently reported by Nesbitt and Zipser (2003), who examined the diurnal nature of the frequency of occurrence of (mesoscale convective system) MCS-type precipitation features indexed from TRMM TMI and PR measurements. One of their main conclusions, drawn from their method of analysis, was that diurnal rainfall variability over the ocean is weak. We do not find this to be the case when examining rainfall itself, which Nesbitt and Zipser (2003) did not consider. It should be noted that this may not be a contradiction in results when the two methodologies are examined in detail, although it might certainly be perceived as one. Within this framework of primary and secondary modes are several notable regional phenomena and several more instances where there are clear-cut regional departures from either the primary or secondary mode behavior. Of great interest is that diurnal rainfall variability is most prominent over heavy rainfall areas, particularly large-scale convergence zones (highlighted by the ITCZ, SPCZ, and most of their V-sector intersection), monsoon regions, and storm tracks. For example, it is in the V-sector that could be described as a superconvection zone where the largest diurnal precipi-

5213

tation amplitudes are found. Along this same line, the annual migrations of the ITCZ and SPCZ convergence zones produce coherent signals in diurnal rainfall. Thus, in general, the amplitude distribution of the diurnal rainfall cycle mimics that of the distribution of rainfall intensity, the salient question being, is this a required relationship emanating from Gaussian statistics or does intense rainfall carry within its hydrodynamic– thermodynamic makeup distinct mechanisms [such as SRC, SRC with enhanced moistening (SRCM), and DRC interactions] for producing diurnal variations that are not present in less intense rainfall systems? Another example of this phenomenon is how diurnal rainfall variability over the Indian Ocean follows the onset, northward propagation, and withdrawal of the summertime southwest East Asian monsoon. Also, the pattern of diurnal rainfall variability over the central Pacific Ocean responds to the development and withdrawal of the 1997–98 El Niño event. All such examples favor an argument that large-scale precipitation systems containing heavy convection excite either a static or dynamic mode of diurnal rainfall variability or both. Moreover, the results in Figs. 12a,b illustrating the gradual delay of the phase peaks away from the convection centers in the Indian and Pacific Ocean sectors provide support to the argument that the DRC interaction mechanism prevails, although this assertion must be treated cautiously. Clearly, there are areas of heavy rainfall over both continents and oceans that do not exhibit a primary mode, for example, the east Atlantic that exhibits a delayed early-morning mode or the central tropical Pacific that exhibits a distinct MLA mode. By the same token, there are various regional-scale areas of heavy rainfall that do not exhibit meaningful diurnal variability such as the northern part of South Africa and the central-east Pacific equatorial region. Moreover, there is at least one prominent rainfall case that exhibits weak rainfall but also contains highly coherent diurnal rainfall variations. This is the case of subtropical marine stratocumulus regions off the west coasts of North America, South America, central-north Africa, and Australia, which all exhibit daytime maxima. This process could be due to the secondary mode ocean surface heating (OSH) mechanism touted by Sui et al. (1997, 1998). However, we suggest it is more likely related to how solar-driven turbulence and daytime convection embedded in the MSC layers activate precipitation microphysics while at the same time producing cloud breakup and erosion. We emphasize that the diurnal behavior of precipitation in large-scale MSC fields deserves further study because part of the associated precipitation is drizzle below the detection limit of either

5214

JOURNAL OF CLIMATE

the PR or the TMI and because these widespread cloud fields are so climatically important since the atmosphere would steadily cool radiatively if the cloud cover were to steadily increase in magnitude. The pattern of the LSH–differential heating (DH) mechanism due to land–ocean transect differential heating appears over the west Australian and South African coasts during their summer seasons in sufficient enough strength to confirm the durability of that mechanism at the monthly scale. Moreover, without an alternative explanation, the relative frequency of secondary mode, coherent LE–EM diurnal peaks over continental areas argues in favor of the SRC over land (SRCL) mechanism. By the same token, no outstanding evidence of the large-scale continental heating (LSCH) mechanism appears over the rain forests of South America or Africa at the monthly or seasonal time scales, although this does not mean that a strong and large-scale convection zone cannot emit a propagating and diurnally coherent divergence pattern (accompanied by rainfall) away from the convection center in the course of a few days to a week. Nonetheless, these results suggest that the LSCH diurnal mechanism is not a first-order diurnal precipitation phenomenon. The same can be said of the semidiurnal tidal mode (SDTM) mechanism in that no sustained processes mimicking this semidiurnal signal are evident in the analyses. On the other hand, there is evidence of the CGWM or Mobile Terrain-Forced Precipitating System (MTFPS) mechanisms. Diurnal processes akin to these processes are found off of (i) the southeast coast of China in summer, (ii) the east coasts of both South Africa and the continental United States during winter, and (iii) the southwest coast of India during summer. Higher spatial resolution data would be required to detect the boundary layer wind oscillation (BLWO), LSH–static destabilization (SD), and anvil gravity wave trapping (AGWT) because they are strictly mesoscale processes. In essence, diurnal variability over planet Earth is heterogeneous but regionally coherent and arising from multiple mechanisms competing for dominance. At no time in seeking to decipher the cause of diurnal signals can the elements of “time to fruition,” “downstream propagation,” “available moisture,” and “radiative– dynamic interactions” be overlooked because all these processes are physically related to the diurnal variability of rainfall and all underlie the specific amplitude and exact timing of a phase peak. Diurnal rainfall variability is not a simple process and defies simple explanations. Observational analyses such as presented here cannot unravel all the underlying physics producing variations, but they do help quantify the strength, phasing, loca-

VOLUME 19

tions, and regional- to large-scale coherency of the variations. It remains for modeling studies to reproduce the complexities found in the observations before this subject can be put to rest. This has been recently stressed by Duynkerke and Teixeira (2001), Davis et al. (2003), and Dai and Trenberth (2004), all of whom find that modern numerical weather prediction and global climate models cannot adequately reproduce the complexity of diurnal processes that occur in the atmosphere. Until such a time as this takes place, the topic deserves more attention, particularly insofar as studies to corroborate and explain observed multimode phenomena. Acknowledgments. The authors gratefully acknowledge the assistance of Prof. Gregory Tripoli from the University of Wisconsin in preparing those portions of the appendix concerning explanations of the land surface heating mechanisms producing diurnal variability in rainfall, particularly the slope flow and MCS mechanisms. We also extend our appreciation to Dr. Kwo-Sen Kuo at NASA Goddard Space Flight Center (GSFC) for assisting with the graphical development of Figure 9 and to Dr. Erich Stocker and his staff at the TRMM Data Information System (TSDIS) of NASA GSFC for supporting the production and reprocessing of the TRMM level-2 precipitation datasets. This research has been supported by NASA TRMM and GWEC Grants NAG5-4752 and NAG5-12064, respectively, and the GPM Project at GSFC.

APPENDIX Summary of Principal Mechanisms for Diurnal Rainfall Variability This appendix provides explanations for a number of mechanisms that have been introduced in the published literature as possible causes for diurnal variability in precipitation. Table 1 summarizes the salient material concerning these mechanisms.

a. LSH mechanisms It is generally accepted that an afternoon rainfall maximum over continents is due to the land surface heating cycle (i.e., the diurnal solar heating cycle of the relatively low heat capacity land surface). Diurnally modulated LSH has two major effects on the atmosphere, both of which can lead to the diurnal variation of rainfall: (i) the static destabilization of the atmosphere both within the planetary boundary layer (PBL) and the free troposphere (referred to as the LSH–SD mechanism) and (ii) the formation of regional meso-

15 OCTOBER 2006

YANG AND SMITH

scale circulations driven by differential heating, that is, by horizontal variations in land surface heating. Static destabilization occurs both as a result of sensible and latent heating (evaporation and evapotranspiration) from the surface. The ratio of sensible to latent heat fluxes, called the Bowen ratio, varies greatly from region to region depending on soil moisture, vegetation cover, and time of day. Typically in highly vegetated regions, the Bowen ratio will approach 0.1, while in arid regions and regions of stressed vegetation, the Bowen ratio will approach 1.0. Often, the Bowen ratio is less than 1, suggesting a strong importance of diurnal variations in moisture on diurnal rainfall. Dry and moist static stability variations associated with diurnal thermal and moisture oscillations lead to the growth of dry and moist convective overturnings that are created as static stability lowers to a critical point. Dry convective overturning is the more likely circulation contained in the PBL, a diurnal circulation typically not associated with precipitating clouds. Thunderstorms (i.e., deep moist convection) generally arise from conditional instability with the most significant contribution stemming from surface moisture fluxes. Such overturnings will not take place until the dry convective overturning process at least penetrates the PBL. Recent research is showing that these penetrations normally occur within regions where convective roll circulations forming in the planetary boundary layer are being lifted. The structure and intensity of these rolls depend on the relationship between local wind shear and static stability. Deep moist convective overturnings, once initiated, can reach as high as the tropopause and are typically associated with afternoon rain showers (Byers and Braham 1948; Byers and Rodebush 1948; Braham 1952; Gentry and Moore 1954; Ogura and Takahashi 1971). Within environments of weak wind shear, these showers take the form of “air mass thunderstorms” (ATMS), also called “pulse” thunderstorms. When more significant shear is present, the associated momentum transport by the showers can lead to meso-␥, meso-␤, and even meso-␣ organizations of the convective overturnings over time (e.g., supercells). Such organizations can prolong the period of afternoon rain into the evening hours. In addition to initiation by coherent structures of the PBL, deep convection can be initiated by diurnally regulated thermally driven mesoscale circulations. These circulations are composed of two major categories: (i) those driven by diurnally occurring surface temperature variations associated with the differential heating of the atmosphere by the surface (referred to as the LSH–DH mechanism) and (ii) those forced by el-

5215

evation variations (discussed in section b of the appendix). Both types of these solenoidal circulations tend to have vertical lifting depths on the order of the PBL depth, which is often too shallow to initiate deep convection except under certain conditions. The most heralded of the thermal contrast-driven circulations (although not the most important for precipitation generation) is the land–sea breeze circulation (LSBC), which has been studied extensively since the early twentieth century (Clowes 1917; van Bemmelen 1922; McAuliffe 1922; Kobayasi and Sasaki 1932; Arakawa and Utsugi 1937; Koschmieder 1941; Hornickle 1942; Wexler 1946; Byers and Rodebush 1948; Leopold 1949; Defant 1951; Fisher 1960; Wallington 1960; Yu and Wagner 1970; Chen and Nash 1994; Simpson 1996; Carbone et al. 2000; Wang et al. 2000; Mori et al. 2004). Notably, detailed theoretical formulations and steadily improving computer simulations of the LSBC represented the first success story in the field of meteorology insofar as numerical prediction of a diurnally regulated precipitation process (see Haurwitz 1947; Schmidt 1947; Pearce 1955; Fisher 1961; Estoque 1961, 1962; Magata 1965; McPherson 1970; Neumann and Mahrer 1971; Pielke 1974; Saito et al. 2001). Diurnal land–sea temperature variations occur as a result of the low heat capacity of land compared to that of water. When the water temperature is near the diurnal mean of the land temperature, the land will tend to be warmer than water in daytime and cooler during nighttime. This will excite a solenoidal circulation between the land and water that will drive a frontogenesis process concentrating the upward lifting along a “sea breeze front” or a “land breeze front” depending on whether it is day or night. Over the water during the afternoon or over the land during the evening, sinking motion in compensation for the lifting predominates, thus suppressing rain. Hence, the solenoidal LSBC naturally focuses the energy of afternoon heating into convective storms located along a narrow mesoscale frontal line. These regions of lifting tend to be too shallow to initiate deep convection except in regions of the Tropics where high maritime humidities feature low cloud bases and low levels of free convection. As noted earlier, the combination of boundary layer roll structures superimposed upon regions of concentrated sea breeze lifting can also explain how generally marginal mesoscale circulations can initiate deep afternoon rain showers. A similar type of solenoidal circulation also occurs inland across persistent differential land surface heating boundaries stemming from strong soil moisture gradients, vegetation cover gradients, or striated land–water surface patterns (e.g., Smith et al. 1994; Wai and Smith

5216

JOURNAL OF CLIMATE

1998). As in the case of the LSBC, there is an immense literature on this topic, although the physics of the atmospheric circulation are virtually the same. This type of inland circulation is also found in the vicinity of semipermanent drylines (Sun and Wu 1992), although Savijärvi (1991) suggests that in this case, the process is more likely driven by an inertial oscillation resulting from the nighttime frictional decoupling of boundary layer air from the surface.

b. MTFPS mechanism A number of studies have noted that an earlymorning precipitation maximum can result from a timed progression process related to events that take place during the prior afternoon’s solar heating period, regardless of whether the environment is oceanic or continental (see, e.g., Mapes and Houze 1993; Chen and Houze 1997; Sherwood and Wahrlich 1999). In fact, this process explains a regular 12–18-h eastward progression of a summertime precipitation maximum in the central United States, going from an afternoon maximum near the high-elevation peaks of the Rocky Mountains to an early-morning maximum within the Mississippi–Ohio valley. This is due to a combination of slope flow dynamics and MCS dynamics. In this paper, the process is given the name MTFPS. This is a complex process requiring a lengthy explanation. Over various continents, thermally driven slope flow circulations are an important source of mesoscale forcing of deep precipitating convection, and ultimately may be the most important source of precipitation between the Rocky Mountain and Appalachian divides. Solenoidal slope flow circulations occur because of the difference in elevation of surface heating leading to horizontal mesoscale pressure gradients. These circulations produce upslope flows in the afternoon and more shallow downslope (katabatic) flows at night. Near the peak of a mountain, an ensemble of afternoon thermal upslope flows will converge, producing a “chimney effect” (Henz 1974), whereas converging daytime slope flows will produce concentrated lifting analogous to sea breeze frontal lifting. Once solar heating begins to intensify during midmorning over a mountain range with plains to the east, a westward directed upslope flow commences, initiating a process involving the formation and eastward propagation of thunderstorms, as explained below. When a large-scale geostrophic wind blows across a mountain ridge (such as the Rockies or Andes) in the presence of daytime upslope flow, a morning leeside convergence zone develops along the axis of the nighttime “dryline” (to be defined below). Away from the mountains out over the plains, a sinking motion occurs

VOLUME 19

into the afternoon from the return flow of the solenoidal circulation set up by the daytime slope flow and from compensation for radiative cooling. This causes warming and drying, a stabilizing effect on the inversion above the PBL, which ultimately suppresses the formation of convective storms over the plains in the afternoon. As a result, convection and rain is focused on the leeside convergence zone of the mountains. Warming over the mountain slopes themselves is partially suppressed by adiabatic expansion cooling produced by the upslope winds, resulting in a strengthened PBL inversion along the slope. Upslope of the leeside convergence zone, the slope flow inversion is eventually broken, creating a deep mixed-layer connecting the cross ridge flow aloft with the surface, allowing downward flux of westerly momentum (this is only possible when the upper levels are well mixed and the westerly flow can descend). The result is a warm dry sinking motion upslope of the convergence zone, as well as convergence with the cool moist upslope flow downslope of the convergence zone (the development of the moisture source is discussed below). The converging flows maximize the moisture and thermal gradients along the convergence axis, creating a near-zerothorder discontinuity in moisture leading to the term dryline, as alluded to above. The position of the dryline is assigned to the axis of the leeside convergence zone. The daytime inversion enables daytime moisture fluxes from the surface to heavily moisten a shallow atmospheric layer and not allow this PBL air to mix with the deep troposphere. As a result, much like the sea breeze, the slope flow system leads to the production of afternoon showers along a relatively narrow convergence line (i.e., the dryline, in which the upslope side is dry and the downslope side is moist). Drylines occur on all of the different scales of mountain slopes. They always begin early at higher elevations and propagate eastward into the upslope toward lower elevations as the daytime wears on. This occurs as a result of the protected PBL deepening and breaking the inversion at progressively lower and lower elevations. This erosion process is driven by dry or moist (once thunderstorms form) convection focused on the dryline, devouring the inversion like a buzz saw and enabling the dryline to propagate eastward according to the mean midlevel westerly steering flow. During the evening, a new dryline forms back toward the mountains as a new nocturnal boundary layer inversion forms. The air beneath this inversion then remoistens in the morning, beginning after sunrise when vegetation and irrigated fields out on the plains begin their intense phase of evaporation and transpiration. This is the important moisture source alluded to above that feeds the leeside conver-

15 OCTOBER 2006

YANG AND SMITH

gence zone later in the day, as opposed to transported Gulf of Mexico moisture long believed to represent the principal Great Plains moisture source. The dryline represents the axis where the PBL inversion is broken when moving upslope from the plains. Along this line, the convergence is often sufficiently strong and deep to push the moist slope flow into free convection and to form deep convective thunderstorms that tilt back toward the west. The dry and deeply mixed descending air upslope of the dryline fuels the downdrafts within the convection line by providing a mixing source of dry air and insertion of dry air near the surface, all of which aids the evaporative cooling process needed for downdraft production. Gust fronts from downdrafts traveling east further support uplift within the leeside convergence zone. As a result of these processes, almost all surface heating and moistening is focused into deep convection along the dryline while being suppressed elsewhere. As the dryline propagates eastward according to the steering current, eroding its way downslope, the deep precipitating convection propagates with it, representing the onset of a phase-shifted diurnal rainfall maximum. Because of the shallow depth of the nocturnal PBL and its inability to lift air to thecloud base, there is no maximum of thunderstorms associated with a similar surface-based convergence line over the plains arising from nighttime downslope flow. Nevertheless, thunderstorms and MCS formation develop. This storm phase is a result of the differing depths of the daytime and nocturnal PBLs leading to the formation of a low-level nocturnal jet on the lee slopes referred to as the nocturnal southerly low-level jet. This feature is driven by an inertial oscillation set off by decoupling the upper daytime boundary layer beneath the slope flow inversion from surface friction (i.e., in the layer between the nighttime and daytime PBL tops) and by the loss of the daytime mesoscale pressure gradient acceleration directed upslope (Lettau 1956; Blackadar 1957). The maximum intensity of the jet occurs at the top of the nocturnal PBL, typically about 100–300 m above the surface, and to the east of the daytime dryline position. As the friction and mesoscale pressure gradient forces are released, the ageostrophic component of the wind, that is, the vector wind connecting the southerlysouthwesterly geostrophic flow vector to the westward directed daytime upslope flow vector, swings clockwise. After approximately 6 h (⬃one-third of a halfpendulum day at 35°N), the ageostrophic component is directed to the north-northwest. This jet then rapidly transports moisture stored in the upper daytime PBL parallel to the mountain ridge (i.e., southerlysoutheasterly flow along the Rockies).

5217

The combination of these two slope flow mechanisms extends the daytime forcing of convection by the daytime dryline into the evening. This occurs as the evening progresses and dryline convection decouples from surface heating and moistening, instead coupling with the higher moisture and instability present above the developing nocturnal PBL. The suppression of daytime slope flow by the daytime inversion downslope of the dryline then weakens and the storms move over the top of the nocturnal inversion, coupled into the moist layer between the daytime and nighttime inversion bases. Since the friction is cut off, the inertial oscillation of this layer leads to the supergeostrophic low-level jet, which supplies abundant and rapidly transported sensible heat and moisture into the nocturnal convective system; in fact, the moist entropy transport by the nocturnal lowlevel jet more than compensates for any heat and moisture being supplied from solar heating and moistening. Once this convective spreading process takes place, the nature of the convection is transformed from a squall line–type line structure to an expanding bloblike structure, fueled by the rapidly developing moisture supply. At this juncture, a massive deep convection system has been created that is coupled to the upper troposphere, spawning new cells from the increasing moisture supply and continuous gravity wave transmission. It is at this stage that the deep vertical circulation organizes a supporting quasigeostrophic circulation aloft through the formation of a mobile shortwave and associated jet streak at upper levels, and thus the formation of an MCS capable of eastward propagation through quasigeostrophic dynamics. Tripoli and Cotton (1989a,b) first linked the daytime (slope flow) and nighttime (MCS formation) processes using a dynamical model. They showed that the effect of the mesoscale forcing of the Rocky Mountain slope flow system is to manufacture MCSs with quasigeostrophic components. Maddox et al. (1981) and Maddox and Doswell (1982) confirmed earlier with synoptic mesoscale data analysis how quasigeostrophic circulations are strengthened or created by MCSs. Insofar as precipitation, when the MCSs become well organized, they produce a late-evening precipitation burst, while the associated deep convection up through the troposphere is initiating the formation of the mobile baroclinic shortwave that propagates eastward, continuously phase shifting the precipitation maximum. On some occasions, a preexisting shortwave perturbation arrives on the scene, couples with the MCS’s deep convective towers, and greatly enhances the MCS intensity as the available latent heating energizes the captured shortwave through its jet core entrance. It is this initiation and/or amplification of shortwaves in the presence

5218

JOURNAL OF CLIMATE

of MCSs that explains why jet streaks are almost always observed with these systems. Finally, as the MCS propagates eastward, its evolution and propagation is supported by gravity wave dynamics that produce surface density currents arising from evaporatively cooled downdrafts. These cooled downdrafts and cold pools trigger new cell growth to the east. By early morning, the MCS precipitation maximum is often found over the central United States within the Mississippi valley. Hence the slope flow mechanism leads to a focused afternoon maximum of convection over higher terrain, which then systematically progresses toward a nocturnal maximum at lower elevations along the slope. Later on, once the MCS has been formed, quasigeostrophic dynamics take over and the precipitation maximum is steadily shifted eastward. A number of studies since the early twentieth century have presented observational evidence of this process (e.g., Kincer 1916; Means 1944, 1952; Bonner 1963; Crow 1969; Wallace 1975; Riley et al. 1987; Maddox 1981a,b; Carbone et al. 2002). Because all but the latter study focused on limited aspects of the diurnal precipitation process though, one finds ambiguities in the reported times of precipitation maxima along the zonal axis between the Rockies and into the Mississippi–Ohio valley. The location of the afternoon precipitation maximum depends greatly on the wind and stability regime present and where the principal dryline is established. In the Rockies, under moderate westerly flow aloft, the principal dryline tends to be situated most often in western Texas, western Oklahoma, and western Kansas. However, during special situations such as the North American monsoon, the dryline may be situated much farther west near the Continental Divide in central Colorado. When the principal dryline is in western Texas, secondary drylines may also occur near the Continental Divide, although this situation is not likely to lead to a continuous coupling of high mountain convection with nocturnal plains circulations. Over a 2-yr study period, Maddox (1981a) found that about 15% of nocturnal mesoscale convective systems over the Great Plains form over the Rockies while others tend to be initiated on drylines or other diurnal rainfall systems. Other study periods exhibited significant variations from that statistic. Also, a study by Hu (2003) found that the well-established LE–EM precipitation maximum over the central United States is controlled on the decadal scale by the strength of the lowlevel southerly jet. This would likely be due to the mean position of the Atlantic subtropical high that drives the background southerly geostrophic low-level flow over the Great Plains. Its position is likely related to multiyear climate oscillations such as the changing phase of

VOLUME 19

ENSO, the North Atlantic Oscillation (NAO), and/or factors affecting the strength of the North American Monsoon. Besides the central United States, these systems have been found downwind of (i) the Andes in South America, (ii) the Himalayas in China, and (iii) the Ethiopian highlands in Africa. Recent modeling studies suggest that the mechanism by which these propagating systems (waves) develop in the barotropically unstable flow of the Sahel, previously thought to result from barotropic instability, is tied to the effects of diurnal slope flows off the lee (western) side of the Ethiopian highlands (trade flow regime). Significantly, the downstream effects of these slope flow–driven, diurnally propagating Sahel disturbances may serve as the principal genesis mechanism for Atlantic hurricanes. In the United States, Cotton et al. (1983) and Wetzel et al. (1983) followed one such disturbance eastward from the Rockies all the way to the United Kingdom, where it ultimately sank the Queen Elizabeth ocean liner anchored in London. In essence, and as noted at the beginning of the section and as supported by the study of Carbone et al. (2002), the precipitation produced by the MTFPS mechanism is arguably the most important source of summertime precipitation between the Rocky Mountain and Appalachian cordilleras.

c. BLWO, LSCH, and SDTM mechanisms Additional mechanisms affecting continental precipitation have been studied to explain diurnal rainfall variability, including (i) the BLWO, a meso-␣-scale phenomenon that involves a solar-forced convective boundary layer cycle that enhances local convergence, mechanical convection, Richardson number instability, and the strength of the near-surface winds, ultimately producing a delayed precipitation maximum (Blackadar 1957; Holton 1967; Wallace 1975); (ii) LSCH, a large-scale phenomenon that involves day–night differences in the modulation of upper-tropospheric divergence for which the divergence anomaly aloft can propagate and produce a downstream phase shift in the precipitation maximum (Silva Dias et al. 1987); and (iii) the planetary-scale SDTM or S2 pressure wave oscillation (Brier 1965; Brier and Simpson 1969; Haurwitz and Cowley 1973; Hamilton 1981), which involves the effects of tidal forces on surface pressure oscillations and the concomitant influence of the surface mass convergence oscillations on precipitation. In general, the latter mechanism is too weak to explain the large-scale pervasive diurnal variability observed in precipitation, including over oceans, and moreover, it is important to recognize that Lindzen (1978) found that the diurnally modulated latent heating stemming from the precipita-

15 OCTOBER 2006

YANG AND SMITH

tion itself must be accounted for in modifying the phasing of pure S2 tidal oscillations to match the times of observed tidal extremes.

d. SRC mechanism Two separate mechanisms have been touted to explain the morning rainfall maximum over oceans. The first may be called the SRC interaction, a synoptic- to planetary-scale mechanism that presumes that enhanced cloud-top IR cooling at night (stemming from the lack of cloud-top solar absorption) and a consequent increase in the thermal lapse rate (e.g., Kraus 1963; Lavoie 1963; Ramage 1971) lead to stronger convection and nighttime rainfall (tropical forecasters often refer to this as the “early-morning precipitation max”). This mechanism suggests that radiative forcing tends to favor more intense rainfall during the latenight period, provided the preexistence of cloudiness. However, there are problems with this mechanism in explaining diurnal variability in deep convective environments because significant differences between daytime and nighttime oceanic lapse rates are not observed (e.g., Betts 1982; Emanuel 1986, 1994; Xu and Emanuel 1989), and the troposphere may actually become stabilized from deep cumulus layer overturning (e.g., Ruprecht and Gray 1976a,b; Gray and Jacobson 1977). Nonetheless, this is the favored mechanism proffered in a comprehensive general circulation model (GCM) study on cloudiness by Randall et al. (1991), noting that in a later related study, Lin et al. (2000) reported that diurnal phasing in that model is very sensitive to a specified parameter that links cumulus kinetic energy to cloud mass flux.

e. DRC mechanism The second oceanic mechanism of interest may be called the DRC interaction, which is based on the assumption that day–night differences persist in radiative cooling over deep convection regions in contrast with surrounding clear-air regions. This process suppresses convection during daytime, resulting in more rainfall at night. Daytime suppression results from the outer clear regions undergoing less subsidence warming in response to ongoing radiative cooling because of increased daytime radiative heating from water vapor absorption, thus reducing convergence into the convection region and inhibiting daytime convective growth. This mechanism was first proposed by Ruprecht and Gray (1976a,b) and Gray and Jacobson (1977), and later supported by Foltz and Gray (1979), McBride and Gray (1980), and Ackerman and Cox (1981). However, this process is only appropriate to extended organized

5219

convection (such as meso-␣-scale mesoscale convective systems and tropical cyclones, and synoptic-scale convergence zones) where background subsidence can be altered at regional scales. In any case, and regardless of the fact that both SRC and DRC diurnal interactions are rooted in daytime–nighttime radiative cooling differences, the two mechanisms are diametrically opposed in explaining which portion of the diurnal cycle is actually perturbed by the underlying forcing mechanism. In essence, the SRC interaction favors nighttime enhancement through increased thermodynamic instability while the DRC interaction favors daytime suppression through decreased daytime convergence into the convective region.

f. AGWT mechanism A third dynamically related, oceanic meso-␣/␤-scale mechanism has been suggested from a modeling study that considers the often noted tendency for out-to-sea hurricanes to undergo nighttime intensity amplifications. The study by Tripoli (1992), using a fully nonhydrostatic cloud resolving mesoscale model, found that enhanced nighttime cloud-top infrared cooling destabilizes the cloudy outflow layers and anvil portions of a hurricane, leading to gravity wave trapping within the cyclone anvil (thus the AGWT mechanism) stemming from gravity wave energy generated in the lower troposphere in and around the eyewall and transmitted upward into the anvil. This excess trapped energy then serves to increase circulation efficiency, thus intensifying the storm and enhancing the precipitation rate. This explanation supersedes the more prosaic explanation offered by Hobgood (1986), which was based on invoking the SRC radiative mechanism within a relatively crude three-dimensional hurricane model [an environmentally isolated, primitive equation/sigma-coordinate, three-dimensional model developed by Anthes et al. (1971) and Anthes (1972) was used]. Notably, the mechanism of gravity wave dynamics influencing diurnal cloud and rainfall variability has attracted much attention as of late (e.g., Mapes et al. 2003b; Warner et al. 2003). Various ongoing studies concerning ducted gravity waves modulating diurnal rainfall are just in the process of being published.

g. OSH mechanism On top of these primary mechanisms and as noted in sections 5 and 6, various observational studies have found that a secondary rainfall maximum occurs during the afternoon over oceans (e.g., Gray and Jacobson 1977; McGarry and Reed 1978; Reed and Jaffe 1981; Augustine 1984; Fu et al. 1990; Serra and McPhaden

5220

JOURNAL OF CLIMATE

2004). The modeling results of Anderson et al. (1996), Weller and Anderson (1996), and Sui et al. (1997, 1998) all support the argument for this secondary feature. The latter study emphasizes that boundary layer forcing from diurnal variations of sea surface temperature (SST) gives rise to an afternoon maximum if convective conditions are undisturbed, but transform to an LE– EM maximum under disturbed convective conditions. This is referred to as the OSH mechanism, analogous to the LSH–SD mechanism but taking place over oceans at planetary scales.

h. SRCM mechanism The Sui et al. (1997, 1998) studies also draw attention to two factors that should be considered in defining the timing controls on diurnal cloud precipitation variability, that is, the ambient precipitable water (PW) and the cloud storm life cycle that microphysically evolves over time to produce precipitation. Based on Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) data, they found that a nocturnal precipitation mode can be explained by relatively more available condensed moisture at night due to diurnally varying radiative cooling, with the resultant change in tropospheric moisture stimulating condensation and precipitation. This mechanism, referred to as SRCM and supported in earlier modeling studies (Tao et al. 1993, 1996), is in contrast with the Randall et al. (1991) SRC explanation that suggests that the key diurnal control leading to early-morning oceanic precipitation maxima is nighttime thermodynamic destabilization following suppressed afternoon convective conditions from clouds absorbing solar radiation, enabling destabilized nighttime lapse rates developing over time to produce enhanced early-morning precipitation.

i. SRCL mechanism Outside of continental regions unaffected by the MTFPS process, and notwithstanding the Dai (2001) study that found a morning maximum in drizzle and light rain over land arising from enhanced nighttime radiative cooling and consequent increased relative humidities (a meteorological state that slowly evolves to weak condensation, referred to as the SRCL mechanism), there is no widely accepted explanation for an LE–EM mode over continental regions. Regardless, such secondary maxima have been reported in observational studies. For example, Belden (1936), Ramage (1952), Schwartz and Bosart (1979), Oki and Musiake (1994), and Chen et al. (1999) have all noted LE–EM land maxima based on rain gauge measurements.

VOLUME 19

j. CGWM mechanism Another process that has been suggested as a mechanism for producing oceanic diurnal rainfall variability is large-scale vertical motion (e.g., McBride and Gray 1980). Since the variability of PW is generally associated with large-scale vertical motion, a combination of vertical motion and ambient moisture conditions can produce diurnal variability, presuming the large-scale vertical motion is itself diurnally modulated. However, it is the release of latent heat from precipitation that generally drives vertical overturning. That is, the diurnal cycle of large-scale vertical motion is the result of diurnal rainfall variability rather than the cause, notwithstanding the case of large-scale continental processes inducing propagating vertical motions that can reach the ocean by remote forcing, as studied by Silva Dias et al. (1987). However, remote forcing appears to be more relevant along a coast rather than from a deep continental interior. As mentioned earlier, the recent studies of Mapes et al. (2003a,b) and Warner et al. (2003) have investigated how horizontally and vertically propagating gravity waves excited at the diurnal period by direct thermal, coastal interior surface forcing (i.e., the LSH– DH mechanism producing an onshore afternoon precipitation maximum) radiate diurnal oscillations offshore, delayed to an LE–EM mode by the phase speed of the gravity waves as they move over the coast. This meso-␣-scale mechanism is referred to as CGWM. The explanation of this mechanism in the Mapes et al. and Warner et al. studies corroborates earlier results provided by Yang and Slingo (2001), who did not present any detailed analyses but recognized the existence of the CGWM process. REFERENCES Ackerman, S. A., and S. K. Cox, 1981: GATE phase III mean synoptic-scale radiative convergence profiles. Mon. Wea. Rev., 109, 371–383. Albright, M. D., E. E. Recker, R. J. Reed, and R. Dang, 1985: The diurnal variation of deep convection and inferred precipitation in the central tropical Pacific during January–February 1979. Mon. Wea. Rev., 113, 1663–1680. Anderson, S. P., R. A. Weller, and R. B. Lukas, 1996: Surface buoyancy forcing and the mixed layer of the western Pacific warm pool: Observation and one-dimensional model results. J. Climate, 9, 3056–3085. Andersson, T., 1970: The diurnal variation of precipitation frequency over Weather Ship M. J. Appl. Meteor., 9, 17–19. Anthes, R. A., 1972: Development of asymmetries in a threedimensional numerical model of the tropical cyclone. Mon. Wea. Rev., 100, 461–476. ——, S. L. Rosenthal, and J. W. Trout, 1971: Preliminary results from an asymmetric model of the tropical cyclone. Mon. Wea. Rev., 99, 744–758.

15 OCTOBER 2006

YANG AND SMITH

Arakawa, H., and M. Utsugi, 1937: Theoretical investigation on land and sea breezes. Geophys. Mag., 11, 97–104. Asai, T., S. Ke, and Y.-M. Kodama, 1998: Diurnal variation of cloudiness over East Asia and the Western Pacific Ocean as revealed by GMS during the warm season. J. Meteor. Soc. Japan, 76, 675–684. Augustine, J. A., 1984: The diurnal variation of large-scale inferred rainfall over the tropical Pacific Ocean during 1979. Mon. Wea. Rev., 112, 1745–1751. Ballard, J. C., 1933: The diurnal variation of free-air temperature and of the temperature lapse rate. Mon. Wea. Rev., 61, 61–80. Belden, W. S., 1936: Diurnal distribution of rainfall at St. Joseph, Mo., April to October. Mon. Wea. Rev., 64, 228–230. Bell, G. D., M. S. Halpert, C. F. Ropelewski, V. E. Kousky, A. V. Douglas, R. C. Schnell, and M. E. Gelman, 1999: Climate assessment for 1998. Bull. Amer. Meteor. Soc., 80, S1–S48. Bell, T. L., and N. Reid, 1993: Detecting the diurnal cycle of rainfall using satellite observations. J. Appl. Meteor., 32, 311–322. Benton, G. S., and M. S. Estoque, 1954: Water vapor transfer over the North American continent. J. Meteor., 11, 462–477. Bergman, J. W., and M. L. Salby, 1997: The role of cloud diurnal variations in the time-mean energy budget. J. Climate, 10, 1114–1124. Betts, A. K., 1982: Saturation point analysis of moist convective overturning. J. Atmos. Sci., 39, 1484–1505. Blackadar, A. K., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversions. Bull. Amer. Meteor. Soc., 38, 283–290. Blaskovic, M., R. Davies, and J. B. Snider, 1991: Diurnal variation of marine stratocumulus over San Nicolas Island during July 1987. Mon. Wea. Rev., 119, 1469–1478. Boccippio, D. J., W. Koshak, and R. J. Blakeslee, 2002: Performance assessment of the optical transient detector and lightning imaging sensor. Part I: Predicted diurnal variability. J. Atmos. Oceanic Technol., 19, 1318–1332. Bonner, W. D., 1963: Thunderstorms and the low-level Jet. Mesometeorology Research Paper 22, University of Chicago, Chicago, IL, 21 pp. ——, 1968: Climatology of the low-level jet. Mon. Wea. Rev., 96, 833–850. Bougeault, P., 1985: The diurnal cycle of the marine stratocumulus layer: A higher-order model study. J. Atmos. Sci., 42, 2826–2843. Braham, R. B., Jr., 1952: The water and energy budgets of the thunderstorm and their relation to thunderstorm development. J. Atmos. Sci., 9, 227–242. Brier, G. W., 1965: Diurnal and semidiurnal atmospheric tides in relation to precipitation variation. Mon. Wea. Rev., 93, 93– 100. ——, and J. S. Simpson, 1969: Tropical cloudiness and precipitation related to pressure and tidal variations. Quart. J. Roy. Meteor. Soc., 95, 120–147. Brill, K., and B. Albrecht, 1982: Diurnal variation of the tradewind boundary layer. Mon. Wea. Rev., 110, 601–613. Browner, S. P., W. L. Woodley, and C. G. Griffith, 1977: Diurnal oscillation of the area of cloudiness associated with tropical storms. Mon. Wea. Rev., 105, 856–864. Byers, H. R., and R. R. Braham Jr., 1948: Thunderstorm structure and circulation. J. Atmos. Sci., 5, 71–86. ——, and H. R. Rodebush, 1948: Causes of thunderstorms of the Florida peninsula. J. Meteor., 5, 275–280. Camp, J. P., A. I. Watson, and H. E. Fuelberg, 1998: The diurnal

5221

distribution of lightning over north Florida and its relation to the prevailing low-level flow. Wea. Forecasting, 13, 729–739. Carbone, R. E., J. W. Wilson, T. D. Keenan, and J. M. Hacker, 2000: Tropical island convection in the absence of significant topography. Part I: Life cycle of diurnally forced convection. Mon. Wea. Rev., 128, 3459–3480. ——, J. D. Tuttle, D. A. Ahijevych, and S. B. Trier, 2002: Inferences of predictability associated with warm season precipitation episodes. J. Atmos. Sci., 59, 2033–2056. Chang, A. T. C., L. S. Chiu, and G. Yang, 1995: Diurnal cycle of oceanic precipitation from SSM/I data. Mon. Wea. Rev., 123, 3371–3380. Chang, E. K. M., S. Lee, and K. L. Swanson, 2002: Storm track dynamics. J. Climate, 15, 2163–2183. Chapman, S., and R. S. Lindzen, 1970: Atmospheric Tides. Reidel, 200 pp. Chen, S. S., and R. A. Houze Jr., 1997: Diurnal variation and life-cycle of deep convective systems over the Pacific warm pool. Quart. J. Roy. Meteor. Soc., 123, 357–388. Chen, T.-C., and K. Takahashi, 1995: Diurnal variation of outgoing longwave radiation in the vicinity of the South China Sea: Effect of intraseasonal oscillation. Mon. Wea. Rev., 123, 566– 577. ——, M.-C. Yen, J.-C. Hsieh, and R. W. Arritt, 1999: Diurnal and seasonal variations of the rainfall measured by the automatic rainfall and meteorological telemetry system in Taiwan. Bull. Amer. Meteor. Soc., 80, 2299–2312. 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. ——, and J.-J. Wang, 1994: Diurnal variation of surface thermodynamic fields on the island of Hawaii. Mon. Wea. Rev., 122, 2125–2138. Ciesielski, P. E., W. H. Schuibert, and R. H. Johnson, 2001: Diurnal variability of the marine boundary layer during ASTEX. J. Atmos. Sci., 58, 2355–2376. Clowes, E. S., 1917: Sea breeze on eastern long island. Mon. Wea. Rev., 45, 345–346. Cook, A. W., 1939: Diurnal variation of summer rainfall at Denver. Mon. Wea. Rev., 67, 95–98. Cooper, H. J., M. Garstang, and J. Simpson, 1982: The diurnal interaction between convection and peninsular-scale forcing over south Florida. Mon. Wea. Rev., 110, 486–503. Cotton, W. R., R. L. George, P. J. Wetzel, and R. L. McAnelly, 1983: A long-lived mesoscale convective complex. Part I: The mountain-generated component. Mon. Wea. Rev., 111, 1893– 1918. Crow, L. W., 1969: Relationships between hail and rain in Kansas, Nebraska and Eastern Colorado. LWC Rep. 76, Final Report on Contract NSF C522, National Science Foundation, Arlington, VA, 34 pp. Dai, A., 2001: Global precipitation and thunderstorm frequencies. Part II: Diurnal variations. J. Climate, 14, 1112–1128. ——, and J. Wang, 1999: Diurnal and semidiurnal tides in global surface pressure fields. J. Atmos. Sci., 56, 3874–3891. ——, and K. E. Trenberth, 2004: The diurnal cycle and its depiction in the Community Climate System Model. J. Climate, 17, 930–951. ——, ——, and T. R. Karl, 1999: Effects of clouds, soil moisture, precipitation, and water vapor on diurnal temperature range. J. Climate, 12, 2451–2473. Davis, C. A., K. W. Manning, R. E. Carbone, S. B. Trier, and J. D. Tuttle, 2003: Coherence of warm-season continental rainfall

5222

JOURNAL OF CLIMATE

in numerical weather prediction models. Mon. Wea. Rev., 131, 2667–2679. Defant, F., 1951: Local winds. Compendium of Meteorology, T. F. Malone, Ed., Amer. Meteor. Soc., 655–672. Deschamps, P. Y., and R. Frouin, 1984: Large diurnal heating of the sea surface observed by the HCMR experiment. J. Phys. Oceanogr., 14, 177–184. Deser, C., and C. A. Smith, 1998: Diurnal and semidiurnal variations of the surface wind field over the tropical Pacific Ocean. J. Climate, 11, 1730–1748. Dinku, T., and E. N. Anagnostou, 2005: Regional differences in overland rainfall estimation from PR-calibrated TMI algorithm. J. Appl. Meteor., 44, 189–205. Douglas, M. W., and S. Li, 1996: Diurnal variation of the lowertropospheric flow over the Arizona low desert from SWAMP-1993 observations. Mon. Wea. Rev., 124, 1211–1224. ——, A. Valdez-Manzanilla, and R. Garcia Cueto, 1998: Diurnal variation and horizontal extent of the low-level jet over the northern Gulf of California. Mon. Wea. Rev., 126, 2017–2025. Durre, I., and J. M. Wallace, 2001a: The warm season dip in diurnal temperature range over the eastern United States. J. Climate, 14, 354–360. ——, and ——, 2001b: Factors influencing the cold-season diurnal temperature range in the United States. J. Climate, 14, 3263– 3278. Duvel, J. P., and R. S. Kandel, 1985: Regional-scale diurnal variations of outgoing inferred radiation observed by METEOSAT. J. Climate Appl. Meteor., 24, 335–349. Duynkerke, P. G., 1989: The diurnal variation of a marine stratrocumulus layer: A model sensitivity study. Mon. Wea. Rev., 117, 1710–1725. ——, and P. Hignett, 1993: Simulation of diurnal variation in a stratocumulus-capped marine boundary layer during FIRE. Mon. Wea. Rev., 121, 3291–3300. ——, and J. Teixeira, 2001: Comparison of the ECMWF reanalysis with FIRE I observations: Diurnal variation of marine stratocumulus. J. Climate, 14, 1466–1478. Easterling, D. R., and P. J. Robinson, 1985: The diurnal variation of thunderstorm activity in the United States. J. Climate Appl. Meteor., 24, 1048–1058. Ebert, E. E., and M. J. Manton, 1998: Performance of satellite rainfall estimation algorithms during TOGA COARE. J. Atmos. Sci., 55, 1537–1557. Ellingson, R. G., and M. B. Ba, 2003: A study of diurnal variation of OLR from the GOES Sounder. J. Atmos. Oceanic Technol., 20, 90–98. Emanuel, K., 1986: An air–sea interaction theory of tropical cyclones. Part I: Steady state maintenance. J. Atmos. Sci., 43, 584–604. ——, 1994: Atmospheric Convection. Oxford University Press, 580 pp. Estoque, M. A., 1961: A theoretical investigation of the sea breeze. Quart. J. Roy. Meteor. Soc., 87, 136–146. ——, 1962: The sea breeze as a function of the prevailing synoptic situation. J. Atmos. Sci., 19, 244–250. Faysash, D. A., and E. A. Smith, 2000: Simultaneous retrieval of diurnal to seasonal surface temperatures and emissivities over SGP ARM-CART site using GOES split window. J. Appl. Meteor., 39, 971–982. Fisher, E. L., 1960: An observational study of the sea breeze. J. Atmos. Sci., 17, 645–660. ——, 1961: A theoretical study of the sea breeze. J. Atmos. Sci., 18, 216–233.

VOLUME 19

Foltz, G. S., and W. M. Gray, 1979: Diurnal variations in the troposphere’s energy balance. J. Atmos. Sci., 36, 1450–1466. Fritz, S., 1963: The diurnal variation of ground temperature as measured from TIROS II. J. Appl. Meteor., 2, 645–648. Fu, R., A. D. Del Genio, and W. B. Rossow, 1990: Behavior of deep convective clouds in the tropical Pacific from ISCCP radiances. J. Climate, 3, 1129–1152. Gallo, K. P., D. R. Easterling, and T. C. Peterson, 1996: The influence of land use/land cover on climatological values of the diurnal temperature range. J. Climate, 9, 2941–2944. Garreaud, R. D., and J. M. Wallace, 1997: The diurnal march of convective cloudiness over the Americas. Mon. Wea. Rev., 125, 3157–3171. ——, and R. Muñoz, 2004: The diurnal cycle in circulation and cloudiness over the subtropical southeast Pacific. J. Climate, 17, 1699–1710. Gentry, R. C., and P. L. Moore, 1954: Relation of local and general wind interaction near the sea coast to time and location of air mass showers. J. Meteor., 11, 507–511. Gray, W. M., and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105, 1171–1188. Haddad, Z. S., E. A. Smith, C. D. Kummerow, T. Iguchi, M. R. Farrar, S. L. Durden, M. Alves, and W. S. Olson, 1997: The TRMM “Day-1” radar/radiometer combined rain-profiling algorithm. J. Meteor. Soc. Japan, 75, 799–809. Hamilton, K., 1981: A note on the observed diurnal and semidiurnal rainfall variations. J. Geophys. Res., 86, 12 122–12 126. ——, 1984: Detection of the lunar diurnal atmospheric tide. Mon. Wea. Rev., 112, 1620–1625. Hann, J., 1901: Lehrbuch der Meteorologie. C. H. Tauchnitz, 805 pp. Hartmann, D. L., K. J. Kowalewsky, and M. L. Michelsen, 1991: Diurnal variations of outgoing longwave radiation and albedo from ERBE Scanner data. J. Climate, 4, 598–617. Haurwitz, B., 1947: Comments on the sea breeze circulation. J. Meteor., 4, 1–8. ——, and A. D. Cowley, 1973: The diurnal and semidiurnal barometric oscillations, global distribution, and annual variation. Pure Appl. Geophys., 102, 193–222. Hays, P. B., D. L. Wu, and The HRDI Science Team, 1994: Observations of the diurnal tide from space. J. Atmos. Sci., 51, 3077–3093. Henz, J., 1974: Colorado High Plains thunderstorm systems—A descriptive radar-synoptic climatology. M.S. thesis, Dept. of Atmospheric Sciences, Colorado State University, 82 pp. Hirose, M., and K. Nakamura, 2005: Spatial and diurnal variation of precipitation systems over Asia observed by the TRMM Precipitation Radar. J. Geophys. Res., 110, D05106, doi:10.1029/2004JD004815. Hitschfeld, W., and J. Bordan, 1954: Errors inherent in the radar measurement of rainfall at attenuating wavelengths. J. Meteor., 11, 58–67. Hobgood, J. S., 1986: A possible mechanism for the diurnal oscillations of tropical cyclones. J. Atmos. Sci., 43, 2901–2922. Hollingsworth, A., 1971: The effect of ocean and Earth tides on the semi-diurnal lunar air tide. J. Atmos. Sci., 28, 1021–1044. Holton, J. R., 1967: The diurnal boundary layer wind oscillation above sloping terrain. Tellus, 19, 199–205. Hong, Y., K.-L. Hsu, S. Sorooshian, and X. Gao, 2005: Improved representation of diurnal variability of rainfall retrieved from the Tropical Rainfall Measurement Mission Microwave Imager adjusted Precipitation Estimation From Remotely Sensed Information Using Artificial Neural Networks

15 OCTOBER 2006

YANG AND SMITH

(PERSIANN) system. J. Geophys. Res., 110, D06102, doi:10.1029/2004JD005301. Hornickle, K., 1942: Danziger Seewindstudien III. Dan. Meteor. Forschungsarb., 11, 1–52. Hou, A. Y., S. Q. Zhang, A. M. da Silva, and W. S. Olson, 2000: Improving assimilated global datasets using TMI rainfall and columnar moisture observations. J. Climate, 13, 4180–4195. Hu, Q., 2003: A multidecadal variation in summer season diurnal rainfall in the central United States. J. Climate, 16, 174–178. Huffman, G. J., R. F. Adler, M. Morrissey, D. T. Bolvin, S. Curtis, R. Joyce, B. McGavock, and J. Susskind, 2001: Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeor., 2, 36–50. Ignatov, I., and G. Gutman, 1999: Monthly mean diurnal cycles in surface temperatures over land for global climate studies. J. Climate, 12, 1900–1910. Iguchi, T., and R. Meneghini, 1994: Intercomparison of singlefrequency methods for retrieving a vertical rain profile from airborne or spaceborne radar data. J. Atmos. Oceanic Technol., 11, 1507–1511. ——, T. Kozu, R. Meneghini, J. Awaka, and K. Okamoto, 2000: Rain-profiling algorithm for the TRMM precipitation radar. J. Appl. Meteor., 39, 2038–2052. Ikai, J., and K. Nakamura, 2003: Comparison of rain rates over the ocean derived from TRMM Microwave Imager and Precipitation Radar. J. Atmos. Oceanic Technol., 20, 1709–1726. Imaoka, K., and R. W. Spencer, 2000: Diurnal variation of precipitation over the tropical oceans observed by TRMM/TMI combined with SSM/I. J. Climate, 13, 4149–4158. Israël, H., 1952: The diurnal variation of atmospheric electricity as a meteorological–aerological phenomenon. J. Atmos. Sci., 9, 328–332. Janowiak, J. E., P. A. Arkin, and M. Morrissey, 1994: An examination of the diurnal cycle in oceanic tropical rainfall using satellite and in situ data. Mon. Wea. Rev., 122, 2296–2311. Jones, C., L. M. V. Carvalho, R. W. Higgins, D. E. Waliser, and J. K. E. Schemm, 2004: Climatology of tropical intraseasonal convective anomalies: 1979–2002. J. Climate, 17, 523–539. Jordan, C. L., 1980: Diurnal variation of precipitation in the eastern tropical Atlantic. Mon. Wea. Rev., 108, 1065–1069. Karl, T. R., G. Kukla, and J. Gavin, 1984: Decreasing diurnal temperature range in the United States and Canada from 1941 through 1980. J. Climate Appl. Meteor., 23, 1489–1504. Kincer, J. B., 1916: Daytime and nighttime precipitation and their economic significance. Mon. Wea. Rev., 44, 628–633. King, T. S., and R. C. Balling Jr., 1994: Diurnal variations in Arizona monsoon lightning data. Mon. Wea. Rev., 122, 1659– 1664. Kishtawal, C. M., and T. N. Krishnamurti, 2001: Diurnal variation of summer rainfall over Taiwan and its detection using TRMM observations. J. Appl. Meteor., 40, 331–344. ¨ ber land- und seewinde. Beitr. Kobayasi, T., and T. Sasaki, 1932: U Phys. Atmos., 19, 17–21. Koschmieder, H., 1941: Danziger Seewindstudien I. Dan. Meteor. Forschungsarb., 10, 1–39. Kousky, V. E., 1980: Diurnal rainfall variation in Northeast Brazil. Mon. Wea. Rev., 108, 488–498. Kraus, E. B., 1963: The diurnal precipitation change over the sea. J. Atmos. Sci., 20, 546–551. Krishnamurti, T. N., and C. M. Kishtawal, 2000: A pronounced continental-scale diurnal mode of the Asian summer monsoon. Mon. Wea. Rev., 128, 462–473. Kummerow, C., W. S. Olson, and L. Giglio, 1996: A simplified

5223

scheme for obtaining precipitation and vertical hydrometeor profiles from passive microwave sensors. IEEE Trans. Geosci. Remote Sens., 34, 1213–1232. ——, W. Barnes, T. Kozu, J. Shiue, and J. Simpson, 1998: The tropical rainfall measuring mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 809–817. ——, and Coauthors, 2000: The status of the Tropical Rainfall Measuring Mission (TRMM) after two years in orbit. J. Appl. Meteor., 39, 1965–1982. Landin, M. G., and L. F. Bosart, 1985: Diurnal variability of precipitation in the northeastern United States. Mon. Wea. Rev., 113, 989–1014. Lavoie, R. L., 1963: Some aspects of meteorology of the tropical Pacific viewed from an atoll. Rep. 27, Institute of Geophysics, University of Hawaii, 76 pp. Leopold, L. B., 1949: The interaction of the trade wind and sea breeze, Hawaii. J. Meteor., 6, 312–320. Lettau, H., 1956: Note on the structure of the atmospheric surface layer. J. Atmos. Sci., 13, 507–509. Lewis, J. S., 1971: Venus: Surface temperature variations. J. Atmos. Sci., 28, 1084–1085. Lieberman, R. S., 1991: Nonmigrating diurnal tides in the equatorial middle atmosphere. J. Atmos. Sci., 48, 1112–1123. ——, and C. B. Leovy, 1995: A numerical model of nonmigrating diurnal tides between the surface and 65 km. J. Atmos. Sci., 52, 389–409. Liebmann, B., and A. Gruber, 1988: The annual variation of the diurnal cycle of outgoing longwave radiation. Mon. Wea. Rev., 116, 1659–1670. Lin, X., D. A. Randall, and L. D. Fowler, 2000: Diurnal variability of the hydrologic cycle and radiative fluxes: Comparisons between observations and a GCM. J. Climate, 13, 4159–4179. Lindzen, R., 1978: Effect of daily variations of cumulonimbus activity on the atmospheric semidiurnal tide. Mon. Wea. Rev., 106, 526–533. Liu, C., and M. W. Moncrieff, 1998: Diurnal variations in the troposphere’s energy balance. J. Atmos. Sci., 55, 2329–2344. Lopez, R. E., and R. L. Holle, 1986: Diurnal and spatial variability of lightning activity in northeastern Colorado and central Florida during the summer. Mon. Wea. Rev., 114, 1288–1312. Machado, L. A. T., J.-Ph. Duvel, and M. Desbois, 1993: Diurnal variations and modulation by easterly waves of the size distribution of convective cloud clusters over West Africa and the Atlantic Ocean. Mon. Wea. Rev., 121, 37–49. Maddox, R. A., 1981a: The structure and life-cycle of midlatitude mesoscale convective complexes. Ph.D. dissertation, Colorado State University, 311 pp. ——, 1981b: Satellite depiction of the life cycle of a mesoscale convective complex. Mon. Wea. Rev., 109, 1583–1586. ——, and C. A. Doswell III, 1982: An examination of jet stream configurations, 500 mb vorticity advection and low-level thermal advection patterns during extended periods of intense convection. Mon. Wea. Rev., 110, 184–197. ——, D. J. Perkey, and J. M. Fritsch, 1981: Evolution of upper tropospheric features during the development of a mesoscale convective complex. J. Atmos. Sci., 38, 1664–1674. Magata, M., 1965: A study of the sea breeze by numerical experimentation. Pap. Meteor. Geophys., 16, 23–36. Maier, L. M., E. P. Krider, and M. W. Maier, 1984: Average diurnal variation of summer lightning over the Florida peninsula. Mon. Wea. Rev., 112, 1134–1140. Mapes, B. E., and R. A. Houze Jr., 1993: Cloud clusters and su-

5224

JOURNAL OF CLIMATE

perclusters over the oceanic warm pool. Mon. Wea. Rev., 121, 1398–1415. ——, T. T. Warner, and M. Xu, 2003a: Diurnal patterns of rainfall in northwestern South America. Part I: Observations and context. Mon. Wea. Rev., 131, 799–812. ——, ——, and ——, 2003b: Diurnal patterns of rainfall in northwestern South America. Part III: Diurnal gravity waves and nocturnal convection offshore. Mon. Wea. Rev., 131, 830–844. Markowski, P. M., and D. J. Stensrud, 1998: Mean monthly diurnal cycles observed with PRE-STORM surface data. J. Climate, 11, 2995–3009. McAuliffe, J. P., 1922: Cause of the accelerated sea breeze over Corpus Christi, Tex. Mon. Wea. Rev., 50, 581–582. McBride, J. L., and W. M. Gray, 1980: Mass divergence in tropical weather systems. Part I: Diurnal variation. Quart. J. Roy. Meteor. Soc., 106, 501–516. 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. McNider, R. T., and R. A. Pielke, 1981: Diurnal boundary-layer development over sloping terrain. J. Atmos. Sci., 38, 2198– 2212. McPherson, R. D., 1970: A numerical study of the effect of a coastal irregularity on the sea breeze. J. Appl. Meteor., 9, 767–777. Means, L. L., 1944: The nocturnal maximum occurrance of thunderstorms in the midwestern states. Misc. Rep. 16, Dept. of Meteorology, University of Chicago, 37 pp. ——, 1952: On thunderstorm forecasting in the central United States. Mon. Wea. Rev., 80, 165–189. Meisner, B. N., and P. A. Arkin, 1987: Spatial and annual variations in the diurnal cycle of large-scale tropical convective cloudiness and precipitation. Mon. Wea. Rev., 115, 2009– 2032. Meneghini, R., and T. Kozu, 1990: Spaceborne Weather Radar. Artech House, 199 pp. ——, T. Iguchi, T. Kozu, L. Liao, K. Okamoto, J. Jones, and J. Kwiatkowski, 2000: Use of the surface reference technique for path attenuation estimates from the TRMM Precipitation Radar. J. Appl. Meteor., 39, 2053–2070. Menzel, W. P., D. P. Wylie, and K. I. Strabala, 1992: Seasonal and diurnal changes in cirrus clouds as seen in four years of observations with the VAS. J. Appl. Meteor., 31, 370–385. Miller, M. A., M. P. Jensen, and E. E. Clothiaux, 1998: Diurnal cloud and thermodynamic variations in the stratocumulus transition regime: A case study using in situ and remote sensors. J. Atmos. Sci., 55, 2294–2310. Mohr, K. I., 2004: Interannual, monthly, and regional variability in the wet season diurnal cycle of precipitation in sub-Saharan Africa. J. Climate, 17, 2441–2453. Morcrette, J.-J., 1991: Evaluation of model-generated cloudiness: Satellite-observed and model-generated diurnal variability of brightness temperature. Mon. Wea. Rev., 119, 1205–1224. 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. Murakami, M., 1983: Analysis of the deep convective activity over the western Pacific and Southeast Asia. Part 1: Diurnal cycle. J. Meteor. Soc. Japan, 61, 60–75. Nakamura, K., K. Okamoto, T. Ihara, J. Awaka, J. Kozu, and T. Manabe, 1990: Conceptual design of rain radar for the Tropi-

VOLUME 19

cal Rainfall Measuring Mission. Int. J. Satell. Commun., 8, 257–268. Negri, A. J., T. L. Bell, and L. Xu, 2002: Sampling of the diurnal cycle of precipitation using TRMM. J. Atmos. Oceanic Technol., 19, 1333–1344. Nesbitt, S. W., and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16, 1456–1475. Neumann, J., and Y. Mahrer, 1971: A theoretical study of the land and sea breeze circulation. J. Atmos. Sci., 28, 532–542. Nicolini, M., K. M. Waldron, and J. Paegle, 1993: Diurnal oscillations of low-level jets, vertical motion, and precipitation: A model case study. Mon. Wea. Rev., 121, 2588–2610. Nitta, T., and S. Esbensen, 1974: Diurnal variations in the western Atlantic trades. J. Meteor. Soc. Japan, 52, 254–257. Ogura, Y., and T. Takahashi, 1971: Numerical simulation of the life cycle of a thunderstorm cell. Mon. Wea. Rev., 99, 895–911. Okamoto, K., Ed., 1988: A feasibility study of rain radar for the Tropical Rainfall Measuring Mission. J. Commun. Res. Lab., 35, 109–208. Oki, T., and K. Musiake, 1994: Seasonal change of the diurnal cycle of precipitation over Japan and Malaysia. J. Appl. Meteor., 33, 1445–1463. Olson, W. S., and Coauthors, 2006: Precipitation and latent heating distributions from satellite passive microwave radiometry. Part I: Improved method and uncertainties. J. Appl. Meteor., 45, 702–720. Pearce, R. P., 1955: The calculation of the sea breeze circulation in terms of the differential heating across the coastline. Quart. J. Roy. Meteor. Soc., 81, 351–381. Petenko, I. V., and S. Argentini, 2002: The annual behavior of the semidiurnal and diurnal pressure variations in East Antarctica. J. Appl. Meteor., 41, 1093–1100. Pielke, R. A., 1974: A three-dimensional numerical model of the sea breezes over south Florida. Mon. Wea. Rev., 102, 115– 139. ——, 2002: Mesoscale Meteorological Modeling. 2d ed. Academic Press, 676 pp. Pierce, L. T., 1934: Diurnal variation in the dew-point temperature at Asheville, N.C. Mon. Wea. Rev., 62, 289–293. Ramage, C. S., 1952: Diurnal variation of summer rainfall over east China, Korea and Japan. J. Atmos. Sci., 9, 83–86. ——, 1971: Monsoon Meteorology. Academic Press, 295 pp. Randall, D. A., Harshvardhan, and D. A. Dazlich, 1991: Diurnal variability of the hydrologic cycle in a general circulation model. J. Atmos. Sci., 48, 40–62. Rasmusson, E. M., 1968: Atmospheric water vapor transport and the water balance of North America: Part II. Large-scale water balance investigations. Mon. Wea. Rev., 96, 720–734. Ray, C. L., 1928: Diurnal variation of rainfall at San Juan, P.R. Mon. Wea. Rev., 56, 140–141. Reed, R. J., and K. D. Jaffe, 1981: Diurnal variation of summer convection over West Africa and the tropical eastern Atlantic during 1974 and 1978. Mon. Wea. Rev., 109, 2527–2534. Reiter, E. R., and M. Tang, 1984: Plateau effects on diurnal circulation patterns. Mon. Wea. Rev., 112, 638–651. Rickenbach, T. M., 2004: Nocturnal cloud systems and the diurnal variation of clouds and rainfall in southwestern Amazonia. Mon. Wea. Rev., 132, 1201–1219. Rife, D. L., T. T. Warner, F. Chen, and E. G. Astling, 2002: Mechanisms for diurnal boundary layer circulations in the great basin desert. Mon. Wea. Rev., 130, 921–938. Riley, G. T., M. G. Landin, and L. F. Bosart, 1987: The diurnal

15 OCTOBER 2006

YANG AND SMITH

variability of precipitation across the central Rockies and adjacent Great Plains. Mon. Wea. Rev., 115, 1161–1172. Rogers, D. P., X. Yang, P. M. Norris, D. W. Johnson, G. M. Martin, C. A. Friehe, and B. W. Berger, 1995: Diurnal evolution of the cloud-topped marine boundary layer. Part I: Nocturnal stratocumulus development. J. Atmos. Sci., 52, 2953–2966. Roll, H. U., 1965: Physical Climatology. University of Chicago Press, 272 pp. Rosenthal, S. L., 1960: The aperiodic diurnal range of temperature over the north Atlantic ocean. J. Atmos. Sci., 17, 78–83. ——, and W. A. Baum, 1956: Diurnal variation of surface pressure over the north Atlantic ocean. Mon. Wea. Rev., 84, 379–387. Rozendaal, M. A., C. B. Leovy, and S. B. Klein, 1995: An observational study of diurnal variations of marine stratiform cloud. J. Climate, 8, 1795–1809. Ruprecht, E., and W. M. Gray, 1976a: Analysis of satelliteobserved cloud clusters. Part I: Wind and dynamic fields. Tellus, 28, 391–413. ——, and ——, 1976b: Analysis of satellite-observed cloud clusters. Part II: Thermal, moisture and precipitation. Tellus, 28, 414–426. Saito, K., T. Keenan, G. Holland, and K. Puri, 2001: Numerical simulation of the diurnal evolution of tropical island convection over the Maritime Continent. Mon. Wea. Rev., 129, 378– 400. Salby, M. L., and P. Callaghan, 1997: Sampling error in climate properties derived from satellite measurements: Consequences of undersampled diurnal variability. J. Climate, 10, 18–36. Saltzman, B., and J. A. Pollack, 1977: Sensitivity of the diurnal surface temperature range to changes in physical parameters. J. Appl. Meteor., 16, 614–619. Sato, T., and F. Kimura, 2003: A two-dimensional numerical study on diurnal cycle of mountain lee precipitation. J. Atmos. Sci., 60, 1992–2003. Savijärvi, H., 1991: The United States Great Plains diurnal ABL variation and the nocturnal low-level jet. Mon. Wea. Rev., 119, 833–840. Schmidt, F. H., 1947: An elementary theory of the land- and seabreeze circulation. J. Atmos. Sci., 4, 9–20. Schwartz, B. E., and L. F. Bosart, 1979: The diurnal variability of Florida rainfall. Mon. Wea. Rev., 107, 1535–1545. Serra, Y. L., and M. J. McPhaden, 2004: In situ observations of diurnal variability in rainfall over the tropical Pacific and Atlantic Oceans. J. Climate, 17, 3496–3509. Sharma, A. K., A. T. C. Chang, and T. T. Wilheit, 1991: Estimation of the diurnal cycle of oceanic precipitation from SSM/I data. Mon. Wea. Rev., 119, 2168–2175. Sherwood, S. C., and R. Wahrlich, 1999: Observed evolution of tropical deep convective events and their environment. Mon. Wea. Rev., 127, 1777–1795. Shi, S. F., and E. Chen, 1984: On the use of GOES thermal data to study effects of land use on diurnal temperature fluctuation. J. Appl. Meteor., 23, 426–433. Short, D. A., and J. M. Wallace, 1980: Satellite-inferred morning to evening cloudiness changes. Mon. Wea. Rev., 108, 1160– 1169. ——, and K. Nakamura, 2000: TRMM radar observations of shallow precipitation over the tropical oceans. J. Climate, 13, 4107–4124. Silva Dias, P. L., J. P. Bonatti, and V. E. Kousky, 1987: Diurnally forced tropical tropospheric circulation over South America. Mon. Wea. Rev., 115, 1465–1478.

5225

Simpson, J. E., 1996: Diurnal changes in sea-breeze direction. J. Appl. Meteor., 35, 1166–1169. Simpson, J., C. Kummerow, W.-K. Tao, and R. Adler, 1996: On the Tropical Rainfall Measuring Mission (TRMM). Meteor. Atmos. Phys., 60, 19–36. ——, J. Halverson, H. Pierce, C. Morales, and T. Iguchi, 1998: Eyeing the eye: Exciting early stage science results from TRMM. Bull. Amer. Meteor. Soc., 79, 1711. Skaggs, R. H., 1969: Analysis and regionalization of the diurnal distribution of tornadoes in the United States. Mon. Wea. Rev., 97, 103–115. Slingo, A., R. C. Wilderspin, and S. J. Brentnall, 1987: Simulation of the diurnal cycle of outgoing longwave radiation with an atmospheric GCM. Mon. Wea. Rev., 115, 1451–1457. Smith, E. A., M. M.-K. Wai, H. J. Cooper, M. T. Rubes, and A. Hsu, 1994: Linking boundary circulations and surface processes during FIFE 89. Part I: Observational analysis. J. Atmos. Sci., 51, 1497–1529. ——, F. J. Turk, M. R. Farrar, A. Mugnai, and X. Xiang, 1997: Estimating 13.8-GHz path-integrated attenuation from 10.7GHz brightness temperatures for TRMM combined PR–TMI precipitation algorithm. J. Appl. Meteor., 36, 365–388. ——, and Coauthors, 1998: Results of WetNet PIP-2 project. J. Atmos. Sci., 55, 1483–1536. Smith, G. L., and D. A. Rutan, 2003: The diurnal cycle of outgoing longwave radiation from Earth Radiation Budget Experiment measurements. J. Atmos. Sci., 60, 1529–1542. Smith, W. S., and C.-Y. J. Kao, 1996: Numerical simulations of the marine stratocumulus-capped boundary layer and its diurnal variation. Mon. Wea. Rev., 124, 1803–1816. Soman, V. V., J. B. Valdés, and G. R. North, 1995: Satellite sampling and the diurnal cycle statistics of Darwin rainfall data. J. Appl. Meteor., 34, 2481–2490. Sorooshian, S., X. Gao, K. Hsu, R. A. Maddox, Y. Hong, H. V. Gupta, and B. Imam, 2002: Diurnal variability of tropical rainfall retrieved from combined GOES and TRMM satellite information. J. Climate, 15, 983–1001. Stramma, L., P. Cornillon, R. A. Weller, J. F. Price, and M. G. Briscoe, 1986: Large diurnal sea surface temperature variability: Satellite and in situ measurements. J. Phys. Oceanogr., 16, 827–837. Sui, C.-H., K.-M. Lau, Y. N. Takayabu, and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE. J. Atmos. Sci., 54, 639–655. ——, ——, and X. Li, 1998: Convective–radiative interaction in simulated diurnal variations of tropical cumulus ensemble. J. Atmos. Sci., 55, 2345–2357. Sun, W.-Y., and C.-C. Wu, 1992: Formation and diurnal variation of the dryline. J. Atmos. Sci., 49, 1606–1619. Sutton, J. L., C. B. Leovy, and J. E. Tillman, 1978: Diurnal variations of the Martian surface layer meteorological parameters during the first 45 sols at two Viking Lander sites. J. Atmos. Sci., 35, 2346–2355. Tao, W.-K., C.-H. Sui, B. Ferrier, S. Lang, J. Scala, M.-D. Chou, and K. Pickering, 1993: Heating, moisture, and water budgets of tropical and midlatitude squall lines: Comparisons and sensitivity to longwave radiation. J. Atmos. Sci., 50, 673–690. ——, S. Lang, J. Simpson, C.-H. Sui, B. Ferrier, and M.-D. Chou, 1996: Mechanisms of cloud–radiation interaction in the Tropics and midlatitudes. J. Atmos. Sci., 53, 2624–2651. ——, and Coauthors, 2006: Retrieval of latent heating from TRMM measurements. Bull. Amer. Meteor. Soc., in press.

5226

JOURNAL OF CLIMATE

Taylor, G. I., 1929: Waves and tides in the atmosphere. Proc. Roy. Soc. London, A156, 318–326. Thiao, W., and O. M. Turpeinen, 1992: Large-scale diurnal variations of tropical cold cloudiness based on a simple cloud indexing method. J. Climate, 5, 173–180. Thiele, O. W., 1966: Observed diurnal oscilliations of pressure and density in the upper stratosphere and lower mesosphere. J. Atmos. Sci., 23, 424–430. Trenberth, K. E., J. M. Caron, D. P. Stepaniak, and S. Worley, 2002: Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, doi:10.1029/2000JD000298. Tripoli, G. J., 1986: A numerical investigation of an orogenic mesoscale convective system. Ph.D. dissertation, Colorado State University, 290 pp. ——, 1992: An explicit three-dimensional nonhydrostatic simulation of a tropical cyclone. Meteor. Atmos. Phys., 49, 229–254. ——, and W. R. Cotton, 1989a: Numerical study of an observed orogenic mesoscale convective system. Part I: Simulated genesis and comparison with observations. Mon. Wea. Rev., 117, 273–304. ——, and ——, 1989b: Numerical study of an observed orogenic mesoscale convective system. Part II: Analysis of governing dynamics. Mon. Wea. Rev., 117, 305–328. Turk, F. J., G. D. Rohaly, J. Hawkins, E. A. Smith, F. S. Marzano, A. Mugnai, and V. Levizzani, 1999: Meteorological applications of precipitation estimation from combined SSM/I, TRMM and infrared geostationary satellite data. Microwave Radiometery and Remote Sensing of the Earth’s Surface and Atmosphere, P. Pampaloni and S. Paloscia, Eds., VSP, 353– 363. Turton, J. D., and S. Nicholls, 1987: A study of the diurnal variations of stratocumulus using a multiple mixed layer model. Quart. J. Roy. Meteor. Soc., 113, 969–1009. van Bemmelen, W., 1922: Land- und Seebrise in Batavia. Beitr. Phys. Atmos., 10, 169–177. Vincent, D. G., 1994: The South Pacific convergence zone (SPCZ): A review. Mon. Wea. Rev., 122, 1949–1970. Wai, M. M.-K., and E. A. Smith, 1998: Linking boundary circulations and surface processes during FIFE 89. Part II: Maintenance of secondary circulation. J. Atmos. Sci., 55, 1260–1276. Waliser, D. E., and C. Gautier, 1993: A satellite-derived climatology of ITCZ. J. Climate, 6, 2162–2174. Wallace, J. M., 1975: Diurnal variations in precipitation and thunderstorm frequency over the conterminous United States. Mon. Wea. Rev., 103, 406–419. ——, and F. R. Hartranft, 1969: Diurnal wind variations: Surface to 30 kilometers. Mon. Wea. Rev., 97, 446–455. ——, and D. B. Patton, 1970: Diurnal temperature variations: Surface to 25 kilometers. Mon. Wea. Rev., 98, 548–552. ——, and R. F. Todd, 1974: Some further results concerning the vertical structure of atmospheric tidal motions within the lowest 30 kilometers. Mon. Wea. Rev., 102, 795–803. ——, and P. V. Hobbs, 1977: Atmospheric Science: Introductory Survey. Academic Press, 457 pp. Wallington, C. E., 1960: Convection at the sea breeze front. Cu-

VOLUME 19

mulus Dynamics, C. E. Anderson, Ed., Pergamon Press, 119– 125. Wang, J.-J., R. M. Rauber, H. T. Ochs III, and R. E. Carbone, 2000: The effects of the island of Hawaii on offshore rainband evolution. Mon. Wea. Rev., 128, 1052–1069. Warner, T. T., R. E. Mapes, and M. Xu, 2003: Diurnal patterns of rainfall in northwestern South America. Part II: Model simulations. Mon. Wea. Rev., 131, 813–829. Warren, S. G., C. J. Hahn, J. London, R. M. Chervin, and R. L. Jenne, 1986: Global distribution of total cloud cover and cloud-type amounts over land. Tech. Note NCAR/TN273⫹STR, National Center for Atmospheric Research, Boulder, CO, 29 pp. Watson, A. I., R. E. López, and R. L. Holle, 1994: Diurnal cloudto-ground lightning patterns in Arizona during the southwest monsoon. Mon. Wea. Rev., 122, 1716–1725. Webster, P. J., C. A. Clayson, and J. A. Curry, 1996: Clouds, radiation, and the diurnal cycle of sea surface temperature in the tropical western Pacific. J. Climate, 9, 1712–1730. Weller, R. A., and S. P. Anderson, 1996: Surface meteorology and air–sea fluxes in the western equatorial Pacific warm pool during TOGA Coupled Ocean–Atmosphere Response Experiment. J. Climate, 9, 1959–1990. Wetzel, P. J., W. R. Cotton, and R. L. McAnelly, 1983: A longlived mesoscale convective complex. Part II: Evolution and structure of the mature complex. Mon. Wea. Rev., 111, 1919– 1937. Wexler, R., 1946: Theory and observations of land and sea breezes. Bull. Amer. Meteor. Soc., 27, 272–287. Wilheit, T., and Coauthors, 1994: Algorithms for the retrieval of rainfall from passive microwave measurements. Remote Sens. Rev., 11, 163–194. Wylie, D. P., and H. M. Woolf, 2002: The diurnal cycle of uppertropospheric clouds measured by GOES-VAS and the ISCCP. Mon. Wea. Rev., 130, 171–179. Xu, K., and K. A. Emanuel, 1989: Is the tropical atmosphere conditionally unstable? Mon. Wea. Rev., 117, 1471–1479. Yang, G.-Y., and J. Slingo, 2001: The diurnal cycle in the Tropics. Mon. Wea. Rev., 129, 784–801. Yang, S., and E. A. Smith, 1999: Four-dimensional structure of monthly latent heating derived from SSM/I satellite measurements. J. Climate, 12, 1016–1037. ——, W. S. Olson, J.-J. Wang, T. L. Bell, E. A. Smith, and C. D. Kummerow, 2006: Precipitation and latent heating distributions from satellite passive microwave radiometry. Part II: Evaluation of estimates using independent data. J. Appl. Meteor., 45, 721–739. Yu, T.-W., and N. Y. Wagner, 1970: Diurnal variation of onshore wind speed near a coastline. J. Appl. Meteor., 9, 760–766. Zhang, D.-L., and W.-Z. Zheng, 2004: Diurnal cycles of surface winds and temperatures as simulated by five boundary layer parameterizations. J. Appl. Meteor., 43, 157–169. Zhao, Y., and B. C. Weare, 1994: The effect of diurnal variation of cumulus convection on large-scale low-frequency oscillations in the Tropics. J. Atmos. Sci., 51, 2653–2663.