1. Introduction
Many studies have reported significant diurnal variations in clouds and precipitation of tropical cyclones (TCs). The diurnal cycle may play an important role in changes of TC internal structures and associated dynamical processes (Lajoie and Butterworth 1984; Kossin 2002; Yaroshevich and Ingel 2013; Dunion et al. 2014; Tang and Zhang 2016; Zhang et al. 2020). For example, nighttime radiative cooling may impact the genesis and intensification of TCs through increasing the instability and relative humidity of the troposphere (Melhauser and Zhang 2014; Tang and Zhang 2016). Upper-level clouds in mature TCs tend to expand and contract diurnally as indicated by satellite infrared (IR) imagery (Kossin 2002; Dunion et al. 2014), which may in turn influence TC intensity. TCs’ internal structures are also subject to influence from diurnally varying low-level atmospheric conditions (Zhang et al. 2020); for example, boundary layer inflow and moist entropy in TCs are stronger in nighttime than in the afternoon. In short, the diurnal signal is an important indicator for the change of internal structure and cyclone intensity of TCs.
TC precipitation over the ocean tends to maximize in the early morning (Jiang et al. 2011; Shu et al. 2013; Bowman and Fowler 2015; Wu et al. 2015; Leppert and Cecil 2016), consistent with the diurnal cycle of the overall oceanic convection (Collier and Bowman 2004; Serra and McPhaden 2004; Bowman et al. 2005). The early morning peak of TC rainfall over the ocean is possibly due to the radiation–convection interaction; that is, the nighttime cloud-top radiative cooling destabilizes the troposphere and then enhances deep convection (Miller and Frank 1993; Xu and Randall 1995; Tao et al. 1996). However, a secondary convective peak in the afternoon has also been reported (Lajoie and Butterworth 1984; Kossin 2002; Leppert and Cecil 2016). Kossin (2002) hypothesized that this semidiurnal signal is associated with the solar semidiurnal atmospheric tide, which induces oscillations in the atmospheric lapse rate, leading to maximum destabilization in the early morning and a secondary maximum in the afternoon. The semidiurnal cycle of TC convection may also be attributed to the boundary layer convective downdrafts (Lajoie and Butterworth 1984; Li and Wang 2012). Specifically, peak convection in the overnight produces strong downdraft and suppresses subsequent convection. As the boundary layer gradually recovers in the next several hours through air–sea interaction, convection redevelops and reaches a secondary peak in the afternoon.
TCs may display a strong diurnal pulse in the cirrus canopy and some degree of outward propagation signal in the precipitation. Many studies demonstrated that the diurnal signal of IR brightness temperature, which is the coldest during midnight in the TC’s inner core, propagates radially outward overtime and reaches a TC’s most outer regions by the afternoon (Lajoie and Butterworth 1984; Kossin 2002; Dunion et al. 2014; Wu and Ruan 2016, Ditchek et al. 2019a,b). A somewhat weaker outward propagating diurnal signal was also reported in TC precipitation (Wu et al. 2015), with a 3–6-h phase delay in peak precipitation between the inner core (0–100 km) and outer rainband (300–400 km). However, the outward propagating signal of TC precipitation was not at all observed by spaceborne precipitation radar and microwave imagers, suggesting that the diurnal pulse may manifest near the cloud top instead of deep convective column (Leppert and Cecil 2016). Leppert and Cecil (2016) argued that the outward propagating signal reported by Wu et al. (2015) may actually come from cloud top due to the inclusion of IR information in the precipitation estimates they used. Interestingly, Ditchek et al. (2019b) recently reported that lightning (a proxy of intense deep convection) shows a diurnal pulse in certain TCs. They found that cold cloud-top pulse and lightning pulse coincide 61% of the time. Overall, the possibility of diurnal pulse occurrence in TC precipitation and corresponding conditions are not clear and require more in-depth investigation.
The diurnal cycle of internal distribution (structure) of TC precipitation is less known, although it has been examined as a function of radial distribution (e.g., diurnal pulse). It is well known that the distribution of TC precipitation has an asymmetric structure, which is impacted by multiple factors but dominated by the vertical wind shear (Lonfat et al. 2004; Chen et al. 2006; Cecil 2007; Wingo and Cecil 2010; Xu et al. 2014; Yu et al. 2015, 2017). For example, the heaviest precipitation within a TC is located in the downshear to downshear-left regions (Chen et al. 2006; Cecil 2007; Wingo and Cecil 2010; Yu et al. 2015, 2017). TC precipitation tends to be broader and more asymmetric in high shear conditions (Cecil 2007; Wingo and Cecil 2010). Shear plays a dominant role in precipitation distribution even for landfalling and post-landfall TCs (Xu et al. 2014; Yu et al. 2015, 2017). Shear may show significant diurnal cycle due to diurnally changing low-level winds over certain regions, such as the South China Sea (SCS) (Jiang et al. 2017; Du and Chen 2019; Huang et al. 2010; Huang and Chan. 2011). It is intriguing whether there is significant diurnal variation on the asymmetric distribution of TC precipitation over regions where environmental shear presents with (e.g., SCS) and without diurnal variability.
While the diurnal cycle of TC convection over the ocean has been extensively investigated, it has received less attention over land. Certain studies reported that TC precipitation over land displays similar diurnal behaviors as over the ocean. Using global satellite precipitation data, Bowman and Fowler (2015) found that the average TC rainfall with a land fraction less than 70% consistently peaks near 0600 local time (LT), but the diurnal signal weakens with the increase of land fraction. Hu et al. (2017) demonstrated a similar diurnal pattern (i.e., early morning peak) for overland TC precipitation in China based on rain gauge observations. However, this overland diurnal signal is only significant for precipitation at least 400 km away from the TC center, and inner-core convection shows merely a flat diurnal pattern. They hypothesized that a TC’s interaction with land may lead to high-frequency oscillations in TC precipitation that weakens the diurnal signal. It is unknown why this effect is the strongest in the TC inner core rather than the outer rainband. Nevertheless, Hu et al. (2017) examined TC rainfall diurnal cycle in terms of conditional rain rate (>0.1 mm), which accounts for only the intensity of TC precipitation, not its occurrence frequency. It is possible that TC precipitation intensity is enhanced in the morning due to radiative cooling destabilization, but precipitation frequency may maximize in the afternoon when the environment is more favorable for convective occurrence due to surface heating over land. In addition, though gauges provide the best accuracy of rain, they have limitations in capturing details of TC rainfall distribution due to coarse spatial resolution [e.g., most gauges are located 50–100 km apart in Hu et al. (2017)]. On the other hand, Tang et al. (2019) conducted ideal simulations to investigate TC rainfall diurnal cycle, so that influences from environmental factors other than land and ocean are excluded. Their simulations showed an afternoon peak for TC precipitation over land, which is induced by the combination of direct (increase in low-level buoyancy) and indirect effects (low-level convergence generated by sea breeze) due to increasing surface temperature in the afternoon. Nevertheless, these results are based on idealized simulation, yet need to be validated by observations.
Overall, the diurnal cycle of TC precipitation has not been fully understood and warrants further investigation. This study further investigates the above-mentioned issues regarding the diurnal cycle of TC precipitation, including land–sea contrasts, diurnal pulse, or phase delay between radial distances, and asymmetric distribution. Our study focuses on the northwest Pacific (NWP) region, as the NWP has the largest amount and greatest land–sea contrasts of TC-induced total and extreme precipitation (Zhang et al. 2009; Jiang and Zipser 2010; Chang et al. 2012; Mori and Takemi 2016; Zhang et al. 2017). Also, the NWP shows marked regional variations on large-scale environments, such as for the SCS versus the open ocean of the NWP (referred to herein as the OWP). Basically, the OWP is dominated by the subtropical high, while the SCS is characterized by the southwesterly monsoon flow during the summer. As a result, the OWP has higher sea surface temperatures and lower vertical wind shear compared to the SCS. Also, the SCS is enclosed by landmasses and therefore may have greater impacts from land (Chen et al. 2006; Hendricks et al. 2010; Jiang and Zipser 2010).
2. Data and methods
a. TC best-track data and satellite rainfall product
TC information including the center location, intensity, motion direction, and translation speed is extracted from the International Best Track Archive for Climate Stewardship (IBTrACS; Knapp et al. 2010) dataset. The latest version of IBTrACS data is used, which has a temporal resolution of 3 h. IBTrACS combines a global collection of TC track records from up to 25 meteorological agencies all over the world. This study uses the best available track records in the database during 2001–18, following the selection procedure in Bowman and Fowler (2015, their Table 2).
TC precipitation is derived from the Integrated Multi-satellitE Rainfall from GPM (IMERG) product (Hou et al. 2014; Huffman et al. 2020). IMERG combines multisatellite passive microwave (PMW) precipitation estimates, microwave-calibrated IR rain estimates, and rain gauge measurements to produce rain estimates with fine spatiotemporal resolution. PMW precipitation estimates are first combined through a morphing process (Joyce et al. 2004; Joyce and Xie 2011), and microwave-calibrated IR precipitation estimates are included when PMW data are too sparse. In addition, rain gauge data are used to provide monthly and regional constraints (bias corrections) on the satellite estimates. This study adopted the research version 06 (final run) of IMERG, with a spatial resolution of 0.1° and temporal resolution of 30 min. The IMERG dataset provides several precipitation products. This study utilizes the microwave–IR combined (with gauge calibration) precipitation estimates (precipitationcal) during 2001–18. Precipitation data are extracted with a temporal interval of 3 h (UTC), which are temporally collocated with the interpolated TC best track data. Note that the IMERG precipitation is not accumulated in a 3-h interval, but only the first-hour rainfall within the 3-h period is used.
IMERG is based on an improved precipitation estimation algorithm and includes more PMW observations (superior to IR technique), compared to previous global satellite rainfall products (Huffman et al. 2020). IMERG shows significant improvement and unprecedented capability in representing the spatial and temporal variations of precipitation globally (Kim et al.2017; Tan et al. 2019; Ma et al. 2021). However, IMERG also has limitations (Huffman et al. 2020). For example, relationships between the surface precipitation and PMW brightness temperature of various microwave frequencies depend upon different surface classes (ocean or land), which may cause uncertainty of the precipitation estimation near the coastline or over complex terrain (Kim et al.2017; Huffman et al. 2020). A slight lag (~0.59 h) in the diurnal phase of IMERG precipitation over land was reported (Tan et al. 2019), which needs cautious attention for a diurnal cycle study.
b. TC precipitation extraction and definition
This study targets TC-associated precipitation over the NWP oceans, and adjacent tropical and subtropical Asian landmasses (10°–35°N, 100°–140°E; Fig. 1a). Precipitation within a 500-km radius from the TC center is classified as TC rainfall, based on the popularly accepted definition (Jiang et al. 2011; Xu et al. 2014; Bowman and Fowler 2015; Wu et al. 2015). Only TCs reaching at least tropical storm intensity [maximum sustained wind Vmsw ≥ 33 kt (1 kt ≈ 0.51 m s−1)] are considered. Six-hourly TC locations derived from the IBTrACS are linearly interpolated into 3-hourly (UTC) data. As a result, a total of 16 377 TC samples (3-hourly events) from 637 storms are identified (Fig. 1a). These 3-hourly events in UTC are further adapted into LT events, based on the longitude of the TC center.
Distributions of IMERG 3-h TC samples during 2001–18 and the TC impacting time. (a) Location of weak (33 < Vmsw < 82 kt; green dots) and strong (Vmsw > 82 kt; yellow dots) TCs; the black box and red box mark the SCS and the OWP region respectively. (b),(c) Mean annual TC impacting time (h yr−1) over the ocean and land, respectively. Note that different color scales are applied for ocean and land.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
To examine TC rainfall diurnal variations, TC rainfall is further classified as oceanic precipitation and land-based precipitation based on topography information. We use the topography data provided by the National Geophysical Data Center (NGDC). The NGDC topography data are based on the ETOPO1 Global Relief Model, a 1 arc-minute global relief model that integrates land topography and ocean bathymetry (Amante and Eakins 2009). TC-induced precipitation is also stratified with respect to the TC intensity and distance from the TC center. TC events (3 hourly) are categorized into weak storms (tropical storms, category 1; 33 ≤ Vmsw < 83 kt) and strong storms (categories 2–5; Vmsw ≥ 83 kt), consistent with the definition by Wu et al. (2015).
Samples (number) of TCs (ns), accumulated 3-h TC periods (np), and TC-associated IMERG grid times (ng; number divided by 104) over ocean areas of the study box (10°–35°N, 100°–140°E). Samples are listed as a function of LT and TC annular distance (R).
c. TC rainfall asymmetry
Diurnal variations of TC internal precipitation structure (asymmetry),such as rainfall distribution relative to vertical wind shear, are examined. Analyses of TC rainfall asymmetry are conducted over two different regions for comparison: 1) a box covering the northern SCS and coastal Southern China defined as the SCS region (black parallel box in Fig. 1a) and 2) the open ocean in NWP defined as the OWP region (red box in Fig. 1a). The SCS region is enclosed by landmass, while the OWP is an open ocean area. Also, the large-scale conditions are significantly different between the SCS and the OWP; for example, the OWP has higher sea surface temperature and lower vertical wind shear than the SCS (Chen et al. 2006; Hendricks et al. 2010; Jiang and Zipser 2010). Meanwhile, TCs in the SCS box are divided into two groups depending on their land coverage fraction to examine land–sea contrasts, ocean-dominated (land fraction < 1/2) and land-dominated (land fraction > 1/2) TCs. TCs are assigned to a specific box if their storm centers are located within the box with > 10 km from the domain lateral boundary. Some TCs may have a large portion of precipitation located outside the box, but their entire precipitation and wind fields are still used for compositing.
Vertical wind shear is computed by subtracting the mean wind at 850 hPa from the averaged wind at 200 hPa (V200 − V850) within an annulus between 500 and 750 km from the TC center (Hence and Houze 2011; Tao and Jiang 2015). Note that the SCS and OWP winds are constrained to the west and east of 122°E, respectively, as their flow patterns are apparently different. Wind vectors are derived from the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis (ERA5) data (Hersbach et al. 2020). TC precipitation fields are composited in the shear-relative framework. A mean shear vector is first derived by the mean zonal and meridional component of shear for each compositing bin. Individual TC precipitation fields are rotated with respect to their environmental shear pointing to the mean shear’s direction. Rotated precipitation fields are eventually composited in four 6-h bins, including overnight (0000–0600 LT), morning (0600–1200 LT), afternoon (1200–1800 LT), and evening (1800–2400 LT). An asymmetric index is also provided for measuring the magnitude of TC precipitation asymmetry. Specifically, it is defined as (Rmax − Rmin)/Rmean. Here, Rmax and Rmin represent the mean rain rate of the maximum and minimum precipitation quadrants within an annulus, while Rmean stands for mean rain rate of the particular annular region. The greater the asymmetric index, the more asymmetric the TC precipitation distribution.
3. Results
This study examines 637 TCs (at least tropical storm intensity), including a total of 16 377 TC periods (3-h composites; Fig. 1a). Most of the strong TCs (>category 2 intensity) occur over the open ocean, and overland samples are dominated by weak TCs. The TC impact times over ocean and land are also provided (Figs. 1b,c), which represent the accumulated times that a specific grid (location) spent within a 500-km radius of TC centers. The TC impact time reaches more than 250 h yr−1 in the NWP between 10° and 25°N, and gradually decreases toward the equator and midlatitudes to 50–100 h yr−1 (Fig. 1b). Over land, TC impact time maximizes over major islands and coastal areas of continents (Fig. 1c), such as the northern Philippines, Taiwan, southern China, and eastern Indochina. TCs impact the coast of southern China for more than 130 h yr−1, and the impact time decreases to less than 50 h yr−1 farther inland (Fig. 1c). Although the TC impact time shows significant spatial variability, TC samples (grid scale) do not exhibit any noticeable difference diurnally (Tables 1 and 2). In other words, there will be little influence on TC rainfall diurnal cycle due to TC sample bias (vs time of the day).
Figure 2 shows the geographical distribution of annual TC rainfall and its contribution to total annual rainfall. Obviously, annual TC rainfall is largely determined by TC impact time (frequency of TC occurrence), as their spatial patterns match well. TC rainfall in the tropical NWP exceeds 200 mm yr−1 (Fig. 2a), contributing more than 30% to the total precipitation (Fig. 2c), which is consistent with Jiang and Zipser (2010). Maximum TC rainfall is located in the NWP to the east of the Philippines with an annual amount of ~300 mm (Fig. 2a), which contributes 35%–40% of the total annual precipitation (Fig. 2c). Similarly, the SCS receives ~30% of its annual rainfall from TCs (>200 mm yr−1). Over the East Asian continent, TC rainfall maximizes (70–80 mm yr−1) along areas within less than 200 km from the coast and decreases inland rapidly; for example, there is less than 20 mm yr−1 over areas 300–500 km from the coast (Fig. 2b). TC rainfall accounts for roughly 10% of the total precipitation along coastal regions, and only 1%–2% over farther inland areas (Fig. 2d).
(a),(b) Mean annual TC rainfall (mm yr−1) over the ocean and land and (c),(d) contribution (%) of TC rainfall to total annual rainfall over the ocean and land, respectively. Green, orange, and red contours in (a) and (c) represent 500-, 750-, and 1000-m elevations, respectively. Note that different color scales are applied for ocean and land.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
a. Diurnal cycles of mean rain rate
Figures 3 and 4 display the composited TC rainfall maps over 6-h periods, namely, the overnight (0000–0600 LT), morning (0600–1200 LT), afternoon (1200–1800 LT) and evening (1800–2400 LT) periods. Over the ocean (Fig. 3), TC rainfall is generally enhanced (larger areas with high rain rate) through the overnight to morning (0000–1200 LT), consistent with previous studies (Jiang et al. 2011; Bowman and Fowler 2015; Wu et al. 2015). The diurnal cycle of TC rainfall over the ocean also shows some spatial variability. For example, TC precipitation over the tropical OWP (east of the Philippines) peaks in the overnight hours (0000–0600 LT; Fig. 3a), whereas in the SCS TC rainfall maximizes in the morning (0600–1200 LT; Fig. 3b). This phase delay between the open ocean and the SCS is possibly due to more significant remote influence from land when TCs translate from the OWP to the land-enclosed water (SCS; Silva Dias et al. 1987). Similarly, previous studies reported ±3-h fluctuations in diurnal phases of TC rainfall among different basins (Bowman and Fowler 2015; Wu et al. 2015).
TC rainfall (mm yr−1) over the ocean composited in 6-h periods: (a) overnight (0000–0600 LT), (b) morning (0600–1200 LT), (c) afternoon (1200–1800 LT), and (d) evening (1800–2400 LT). Green, orange, and red contours over land represent 500-, 750-, and 1000-m elevations, respectively.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
As in Fig. 3, but for TC rainfall over land. Note that the color scale is different from Fig. 3.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
TC rainfall over land exhibits diurnal behaviors different from that over the ocean (Fig. 4). TC precipitation over the coast of southern China maximizes in the afternoon (1200–1800 LT), with both the total and heavy precipitation areas extending farther inland (Fig. 4c). During the nocturnal period (1800–0600 LT), heavy TC rainfall (e.g., >50 mm yr−1) concentrates just along the coast (within 100 km) and relatively weak TC precipitation dominates the inland areas (>100 km from the coastline). Compared to coastal areas in southern China, major islands such as the Philippines still show significant diurnal variations on TC precipitation, which is markedly enhanced from the afternoon to evening (Figs. 4c,d). However, diurnal signals of TC precipitation over smaller islands like Taiwan and Hainan Island (China) are not clear, possibly due to limited TC samples over these regions. Interestingly, TC precipitation over the eastern Indochina Peninsula shows little diurnal variability, even though it located over a landmass larger than the Philippines. This is possibly caused by the TC’s interaction with steep terrains (>750 m; contours in Figs. 2 and 3) near the coast of eastern Indochina; for example, precipitation could be enhanced due to convergence between nocturnal downslope winds and the TC low-level onshore flows. In other words, lower terrain (<300 m; contours in Figs. 2 and 3) near the coast of southern China may have minimum influence on TC precipitation, at least without significant nocturnal enhancement.
Figure 5 shows the diurnal cycle of TC and non-TC (unconditional) mean rain rate over both land and ocean. Note that non-TC rainfall is calculated by removing the TC precipitation from total precipitation in the TC season, following Jiang and Zipser (2010). In general, the primary TC rainfall peak occurs in the early morning (0600 LT) and the afternoon (1500 LT) over the ocean and land, respectively (Figs. 5a,b). Surprisingly, Hu et al. (2017) reported an opposite peak time (0300–0600 LT) of TC precipitation over land, possibly because they only examined the conditional rain rate. TC rainfall diurnal patterns are generally consistent with those of non-TC rainfall (Figs. 5c,d), although the strong TC category does differ more. This may suggest that the main mechanisms responsible for rainfall peaks may be similar between TC and non-TC precipitation (e.g., mechanisms of nocturnal cloud-top radiative cooling over the ocean and afternoon surface heating over land). Oceanic TC precipitation also has secondary peaks at 1800 LT. In contrast, non-TC rainfall over the ocean solely maximizes at 0600 LT and minimizes at 1800 LT (Fig. 5c). Although strong TCs induce a higher mean rain rate than weak TCs, they show similar rainfall diurnal patterns.
Areal mean rain rate as a function of local time: (a) precipitation over ocean, (b) precipitation over land, and (c),(d) non-TC rainfall over the ocean and land, respectively. Different colors in (a) and (b) represent weak TCs (33 < Vmsw < 82 kt; blue bars), strong TCs (Vmsw > 82 kt; red bars), and all TCs (gray bars). Amplitudes of diurnal cycles are marked by percentages in the upper left of each plot. Italic values indicate that the diurnal variance is statistically significant above the 95% level.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
TC rainfall is further separated by annular areas of radius of 0–150 km (inner region), 150–300 km (intermediate region), and 300–500 km (outer rainband). Their diurnal cycles are shown in Fig. 6. Generally, precipitation in most TC annular regions (either over ocean or land) shows significant diurnal variation (passing the 95% confidence level), in contrast to the previous finding that diurnal cycle of land-based precipitation is only significant for radial distance greater than 400 km (Hu et al. 2017). Within the inner region (0–150 km), TC rainfall over the ocean displays a double-peak diurnal pattern, including a major peak at 0600 LT and the secondary one at 1800 LT. While the primary rainfall peak over the ocean becomes more dominant in intermediate-to-outer regions (150–500 km), the semidiurnal cycle weakens with the increase of distance from the TC center (Fig. 6c) and eventually diminishes in the most outer regions (300–500 km; Fig. 6e). Interestingly, TC inner-region (<150 km) rainfall over land shows a similar diurnal cycle as its counterpart over the ocean (i.e., double peaks at 0600 and 1800 LT), especially for strong TCs (Fig. 6b). Diurnal peaks (phases) of TC rainfall over the ocean consistently occur at 0600 LT and do not delay with the increase of radial distance, consistent with Leppert and Cecil (2016). In contrast, land-based precipitation from TC outer region (300–500 km) solely peaks in the afternoon (1500 LT) for both strong and weak TCs, opposite to the peak time at 0300–0600 LT reported by Hu et al. (2017). Since the outer-rainband rainfall dominates the total TC rainfall over land (e.g., far more overland samples from the outer-rainband than other TC regions; Table 2), it explains why regional TC mean rain rate maximizes in the afternoon over land in this study (Figs. 4 and 5). In the intermediate regions (150–300 km), the afternoon rainfall peak is evident in weak TCs but not in strong TCs. Note that strong TCs have a limited sample size for an individual annular region. In short, the primary diurnal phase of TC rainfall over the ocean is fairly consistent throughout the TC, but it shifts significantly over land from the early morning to the afternoon with the increase of the distance from the TC center. However, Hu et al. (2017) reported that TC rainfall over land dominantly peaks in the early morning, and the diurnal cycle is only evident in the most outer rainband (>450 km). This is possibly because Hu et al. (2017) only examined the conditional rain rate (data points with rain rate > 0.1 mm), while the above results show the diurnal patterns of unconditional mean rain rate. Basically, mean rain rate over a region is determined not only by precipitation intensity [conditional rain rate as used in Hu et al. (2017)] but also the precipitation occurrence frequency. Therefore, diurnal cycles of both conditional rain rate and precipitation occurrence frequency are further examined in this study.
Areal mean rain rate as a function of LT in annular areas between (a),(b) 0- and 150-km, (c),(d) 150- and 300-km, and (e),(f), 300- and 500-km radii from the TC center, for (left) precipitation over the ocean and (right) land-based precipitation. Different colors represent various TC intensities. Italic values indicate that the diurnal variance is statistically significant above the 95% level.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
b. Precipitation occurrence frequency and intensity
Figure 7 shows the occurrence frequency and conditional rain rate of TC precipitation as a function of local time. Here, the conditional rain rate is calculated by averaging data points with rain rate > 0.1 mm h−1 as in Hu et al. (2017). TC precipitation frequency exhibits little diurnal variability (<10%) over the ocean (Fig. 7a) but evidently peaks in the afternoon (1500 LT) over land with the magnitude of ~50% for weak TCs and ~20% for strong TCs (Fig. 7b). However, TC precipitation intensity (conditional rain rate) displays clear diurnal variations over land and ocean, both with a primary peak in the early morning (0300–0600 LT) and a subpeak in the afternoon (1500–1800 LT; Figs. 7c,d). This is consistent with previous studies (Jiang et al. 2011; Shu et al. 2013; Bowman and Fowler 2015; Wu et al. 2015; Leppert and Cecil 2016; Hu et al. 2017). Since TC precipitation over the ocean only shows significant diurnal variability on conditional rain rate but not precipitation frequency, the nocturnal peak of total TC rainfall over the ocean is likely determined by the enhancement of deep convection due to nocturnal radiative cooling at the cloud top (Miller and Frank 1993; Xu and Randall 1995; Tao et al. 1996). On the other hand, the diurnal cycle of total TC rainfall over land is impacted by diurnal variations of both occurrence frequency and conditional intensity. Nevertheless, precipitation frequency shows a much stronger diurnal cycle amplitude (~50%) than that of the conditional rain rate (~25%), indicating that precipitation frequency is more dominant in the diurnal cycle of TC rainfall. As a result, the mean rain rate (Fig. 5b) over land shares the same diurnal cycle patterns as the precipitation frequency (Fig. 7b), both strongly peaking in the afternoon (1500 LT).
Diurnal cycles of TC precipitation: (a),(b) occurrence frequency and (c),(d) conditional rain rate (>0.1 mm h−1) for (left) precipitation over the ocean and (right) land-based precipitation. Different colors represent various TC intensities. Italic values indicate that the diurnal variance is statistically significant above the 95% level.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
Diurnal cycles of TC precipitation frequency and intensity are further analyzed as a function of annular distance of 0–150, 150–300, and 300–500 km from the TC center (Figs. 8 and 9 ). Precipitation frequency in the TC’s inner region (0–150 km) is the highest but shows virtually no diurnal variation, either over land or ocean (magnitude < 5%; Figs. 8a,b). Frequency diurnal variations of land-based precipitation are significant in the intermediate-to-outer TC regions (>150 km), and their amplitudes increase outward from the TC center. The exception is strong TCs, whose land-based precipitation frequency in the intermediate region (150–300 km) has an only marginal peak at 0600 LT. Over the ocean, TC precipitation frequency only shows some diurnal variability (magnitude of 8%–15%) over the outer regions (300–500 km) with a small peak occurring at 1500 LT. On the other hand, the TC shows a significant double-peak diurnal pattern (primary one at 0300–0600 LT and subpeak at 1500–1800 LT), for conditional rain rates over all regions regardless of distance from the TC center (Fig. 9). However, the distribution of precipitation in the outer rainband over land is not as smooth as other regions, possibly due to their stronger interactions with land. Overall, the conditional rain rate persistently maximizes in the early morning for precipitation over either the ocean or land, consistent with results based on rain gauge measurement in China (Hu et al. 2017). In contrast, the occurrence frequency of TC rainfall over land peaks in the afternoon especially for outer rainband regions, which had not been considered or reported by previous studies (Hu et al. 2017).
As in Fig. 6, but for TC precipitation occurrence frequency.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
As in Fig. 6, but for TC conditional rain rate.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
Diurnal cycles of TC precipitation frequency are also broken down to various rate rates: weak (0.1–2 mm h−1), moderate (2–10 mm h−1), and intense (>10 mm h−1) precipitation (Fig. 10). Over the ocean, TC precipitation shows a vague diurnal pattern for weak precipitation (multiple peaks; Fig. 10a) but an obvious peak at 0600 LT for intense precipitation (Fig. 10e), although the ranges of the y axes differ for the frequencies with rain rates. Intense TC precipitation over the ocean also has a secondary peak at 1800 LT (Fig. 10e). Over land, frequencies of weak and moderate TC precipitation both maximize in the afternoon (1200–1500 LT), possibly due to the increased surface heating during this period (Figs. 10b,d). Interestingly, moderate TC precipitation shows nearly a flipped diurnal pattern between ocean (Fig. 10c) and land (Fig. 10d) with maximum frequency occurring at 0600 LT and 1200–1500 LT, respectively. The frequency of intense TC precipitation over land exhibits a similar diurnal pattern as over ocean, peaking in the early morning (0600 LT) and afternoon (1500–1800 LT), especially for strong TCs.
Diurnal cycle of the occurrence frequency of TC precipitation: (a),(b), weak (0.1–2 mm h−1), (c),(d), moderate (2–10 mm h−1), and (e),(f) intense (>10 mm h−1), for (left) precipitation over the ocean and (right) precipitation over land. Different colors represent various TC intensities. Italic values indicate that the diurnal variance is statistically significant above the 95% level.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
c. Diurnal variations on TC rainfall asymmetry
While the last section analyzes the diurnal cycle of azimuthally averaged TC precipitation, this section examines whether the asymmetric distribution of TC rainfall also varies diurnally. The asymmetric distribution of TC precipitation for the SCS and OWP regions (marked in Fig. 1) are composited relative to environmental vertical wind shear, which is a dominant factor for TC rainfall distribution (Chen et al. 2006; Cecil 2007; Xu et al. 2014; Yu et al. 2015). Also, shear is first analyzed and followed by analyses on TC precipitation asymmetric distribution. Note that TCs over the SCS region are further classified into ocean-dominated (land coverage fraction < 1/2) and land-dominated (land coverage fraction > 1/2) TCs.
1) Environmental vertical wind shear
Figure 11 shows the frequency distribution of TC environmental shear, as well as mean shear vectors. It is noted that mean shear vectors (white arrow) are derived from mean zonal and meridional components of shear, and therefore their magnitudes are lower than mean magnitudes of shears. In general, the shear over the SCS (both ocean and land) is stronger and has greater diurnal variability than that over the OWP. Specifically, the mean shear over the SCS ocean and land regions increases from the afternoon/evening to overnight/morning by 25% (or 1.5 m s−1) and 50% (or 2.5 m s−1), respectively (Figs. 11a–d). Although the mean shear over the SCS oceanic areas is just moderate (6–7 m s−1), there is also a significant fraction (~30%) of strong shear greater than 10 m s−1. Frequency of strong shear (>10 m s−1) is also elevated during the overnight-to-morning hours (Figs. 11a,b). While the shear over the SCS land region is generally smaller (mean of 3–5 m s−1) than that over the ocean possibly due to surface friction (Figs. 11e–h), it is markedly enhanced during the overnight-to-morning periods (Figs. 11e,f). Over the OWP region, most TC environmental shear is just weak (0–5 m s−1; 35%) to moderate (5–10 m s−1; 40%) with a mean shear of 3–4 m s−1, and it is slightly enhanced (<10%) during the overnight-to-morning times (Figs. 11i,j). On the other hand, the shear values of TC and non-TC environments during the TC season show nearly the same magnitudes and diurnal variations (see Fig. S1 in the online supplemental material), suggesting there is little, if any, influence from the TC’s secondary circulation on the calculated vertical wind shear.
Mean vertical wind shear and frequency distribution of shear as a function of direction and magnitude for (a)–(d) ocean-dominated TCs over the SCS, (e)–(h) land-dominated TCs over the SCS, and (i)–(l) TCs over the OWP. Frequencies are represented by the shaded contours, and mean shear vectors are marked by the white arrows. Accumulated frequencies (%) of shear magnitudes of 0–5, 5–10, 10–15, and 15–20 m s−1 are marked by red fonts on the axis. Sample size (n) in each 6-h bin is marked at the upper-left corner of each panel.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
Horizontal winds at 850 and 200 hPa for TC environments (500–750-km radial distance) are examined separately to determine their contributions to diurnal variations of vertical wind shear (Fig. 12). Over oceanic area of the SCS, the upper troposphere (200 hPa) is dominated by northeasterly winds, as it is located at the southeast edge of the South Asian upper-level high pressure (anticyclone) system (Fig. 12a). The low-level (850 hPa) is characterized by prevailing southwesterly monsoon flows that maximize in the overnight hours (0000–0600 LT; Fig. 12b; Chen 2003; Ciesielski and Johnson 2006; Chen et al. 2017). The land region of the SCS is closer to the inner area of the South Asia high at upper levels and subtropical high at low levels, leading to weaker northeasterly/easterly winds and southerly to southeasterly flows at 200- and 850-hPa levels, respectively (Figs. 12c,d). Both the 200- and 850-hPa winds over the SCS maximize overnight, which is possibly explained by two mechanisms. First of all, the low-level anomalous winds over southern China (SCS land areas) exhibits clockwise rotation diurnally due to the inertial oscillation (Coriolis force) and vertical mixing or friction (Jiang et al. 2017; Du and Chen 2019). Second, the diurnal atmospheric pressure tide induces a planetary-scale land–sea breeze circulation spanning about 1000 km over the SCS/NWP, whose time phase is several hours later than that of local land–sea breeze (Huang et al. 2010; Huang and Chan 2011; Du and Chen 2019). During nighttime, a planetary-scale sea breeze occurs and enhances the prevailing low-level southerly/southwesterly winds over the SCS. These factors contribute to the strengthening of nocturnal winds at low levels and the diurnal variations of wind direction over the SCS region. In contrast, the OWP region has minimum impact from land and is characterized by weak northeasterly flow at 200 hPa and southeasterly flow at 850 hPa (Figs. 12e,f), which exhibits no significant diurnal variation.
The 6-hourly composited horizontal wind vectors for (a),(b) ocean-dominated TCs over the SCS, (c),(d) land-dominated TCs over the SCS, and (e),(f) TCs over the OWP, for the (left) 200- and (right) 850-hPa pressure levels.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
2) Asymmetric distribution of TC precipitation
Figure 13 shows the asymmetric distribution of TC precipitation for various regions and time periods, which is composited relative to the shear (white solid arrow). In general, TCs either over the ocean (OWP and SCS ocean) or making landfall (SCS land) produce the heaviest precipitation in the downshear to downshear-left regions, consistent with previous studies (Chen et al. 2006; Cecil 2007; Xu et al. 2014; Yu et al. 2015). Based on the asymmetric index (yellow digits in Fig. 13), TCs over the SCS ocean, where the shear is the strongest, show the greatest asymmetries in precipitation distribution. Interestingly, land-dominated TCs over the SCS have the smallest precipitation asymmetries, even though the shear over the SCS land is stronger than over the OWP. This is possibly because the dominance of shear on the precipitation distribution of landfalling TCs may reduce due to TCs’ interaction with land. The largest asymmetry with the TC precipitation field occurs in the intermediate-to-outer region (150–400 km). Over the SCS, maximum precipitation of ocean-dominated TCs occurs in the morning (0600–1200 LT) when the shear maximizes with an increase of ~25% (or 1.5 m s−1) from the afternoon/evening (Figs. 13a–d). TC rainy areas (e.g., rain rate > 5 mm h−1) are much more extensive in the downshear-left quadrant in the overnight/morning and substantially shrink in the afternoon/evening, consistent with the diurnal variations of shear. As a result, TC precipitation asymmetries maximize during these times, especially over the intermediate and outer-rainband regions. However, precipitation asymmetries of the inner region (<150 km) show little diurnal variation.
The 6-hourly composited precipitation field relative to the vertical wind shear for (a)–(d) ocean-dominated TCs over the SCS, (e)–(h) land-dominated TCs over the SCS, and (i)–(l) TCs over the OWP. The mean shear (white solid vectors) and averaged storm motions (hollow vectors) are overplotted for reference. Dashed straight lines separate shear-relative quadrants, and dashed circles mark the 150- and 300-km radial distance from the TC center. The asymmetry index of each quadrant within annular rings of 0–150, 150–300, and 300–400 km are marked by yellow numbers.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0795.1
For land-dominated TCs over the SCS region (Figs. 13e–h), their rainband is also more asymmetric and extensive in the downshear-left quadrant in the overnight/morning (0000–1200 LT). However, the location of maximum precipitation in land-dominated TCs shifts from downshear-left to downshear in the afternoon (Figs. 13g,h). It should be noted that the downshear-left quadrant of land-dominated TCs may be largely over the ocean, as it is in the rear-left region of the TC motion for TCs landfalling in southern China. While precipitation downshear-left (mostly over the water) experiences enhancement in the morning, precipitation on the downshear-right region (mostly over land) tends to expand farther inland in the afternoon (Fig. 13g). Therefore, diurnal changes on the TC precipitation distribution of land-dominated TCs are likely induced by the combination of shear effect (e.g., weakening of the environmental wind shear from the morning to the afternoon) and impact from land (e.g., increasing instability over land due to stronger surface heating in the afternoon).
Precipitation asymmetries of OWP TCs are comparable to those of ocean-dominated TCs over the SCS, although the shear is much weaker over the OWP (Figs. 13i–l). Again, the maximum precipitation of OWP TCs is located in the downshear-left quadrant. Obviously, there is little diurnal variation on rainfall distribution of OWP TCs, and their precipitation asymmetries remain nearly the same throughout the day. This is well consistent with the fact that the shear over this region has virtually no diurnal variation, suggesting the dominant role of the shear in determining whether TC precipitation distribution will change diurnally over open ocean areas.
4. Conclusions and discussion
a. Conclusions
Although the TC rainfall diurnal cycle has been investigated by many studies, it is still not fully understood, especially for TC precipitation over land. Previous studies well agreed that TC rainfall over the ocean peaks in the early morning (Jiang et al. 2011; Shu et al. 2013; Bowman and Fowler 2015; Wu et al. 2015; Leppert and Cecil 2016), which is consistent with the diurnal cycle of general oceanic convection possibly due to nighttime cloud-top radiative cooling (Collier and Bowman 2004; Serra and McPhaden 2004; Bowman et al. 2005). However, there is a divergence of results on diurnal variations of TC precipitation over land, suggesting that TC interaction with land may have complicated impacts on precipitation diurnal cycle (Jiang et al. 2011; Bowman and Fowler 2015; Hu et al. 2017; Tang et al. 2019). Also, it is unknown whether there are diurnal variations on the internal structure of TC precipitation, which shows significant asymmetry relative to vertical wind shear (Chen et al. 2006; Cecil 2007; Wingo and Cecil 2010). Using the state-of-the-art satellite rainfall product IMERG, this study exhaustively investigates diurnal variations of TC precipitation and their land–sea contrasts in the NWP, in terms of mean rain rate, precipitation frequency, conditional rain intensity, and precipitation asymmetry.
In conclusion, major results from this study are summarized as follows:
There are obvious land–sea contrasts on diurnal variations of TC precipitation, such as that precipitation maxima occur in the early morning (~0600 LT) and the afternoon (~1500 LT) over the ocean and land, respectively. TC precipitation over the ocean also shows an evident semidiurnal cycle, with a secondary precipitation peak at 1800 LT.
Over the ocean, the primary peak of TC precipitation consistently occurs at 0600 LT regardless of the radial distance, and the secondary semidiurnal peak at 1800 LT vanishes at large distances from the storm center. Over land, the innermost core of TCs (<150 km) has a similar diurnal cycle as over the ocean (double peaks); however, beyond 150 km the early morning peak vanishes and the afternoon peak dominates.
Occurrence frequency and intensity (conditional rain rate) of TC precipitation show different diurnal patterns. TC precipitation frequency exhibits little diurnal variability over the ocean, but strongly peaks in the afternoon (~1500 LT) over land, especially in the outer TC regions (150–500 km). In contrast, conditional rain rate over the ocean (land) obviously maximizes in the early morning (afternoon), regardless of distance from the TC center.
TC rainfall diurnal cycle also depends on precipitation magnitude. Over the ocean, weak TC precipitation (0.1–2 mm h−1) presents a virtually weak diurnal cycle, but intense precipitation (>10 mm h−1) shows an obvious peak at 0600 LT. Over land, frequencies of weak and moderate precipitation (0.1–10 mm h−1) both maximize in the afternoon (1200–1500 LT), whereas intense precipitation surprisingly peaks in the early morning (0300 LT).
The asymmetric distribution of TC precipitation over the SCS also varies diurnally especially in the intermediate to outer regions (>150 km), consistent with diurnal variations of environmental vertical wind shear (e.g., maximizing in the nighttime and minimizing in the afternoon). In contrast, the shear over the NWP open ocean areas (OWP) is weaker and exhibits little diurnal variation compared to the SCS; therefore, the precipitation distribution of TCs over the OWP shows virtually no diurnal cycle.
b. Discussion
It is well accepted that TC precipitation over the ocean shares a similar diurnal cycle as non-TC oceanic convection, both peaking in the early morning (Jiang et al. 2011; Shu et al. 2013; Bowman and Fowler 2015; Leppert and Cecil 2016). However, previous studies reported that TC precipitation over land still maximizes in the morning (Bowman and Fowler 2015; Hu et al. 2017), which is in contrast to the diurnal behavior (afternoon peak) of non-TC convection over land (Dai 2001; Nesbitt and Zipser 2003). The underlying mechanisms behind this finding are not yet clear, and working hypotheses are not available (Bowman and Fowler 2015; Hu et al. 2017). In contrast, a model simulation by Tang et al. (2019) suggested that TC precipitation over land may peak in the afternoon due to both the increase in low-level buoyancy (increasing surface temperature) and the enhancement of low-level convergence generated by sea breeze. Observational results from this study actually support the model simulation by Tang et al. (2019), showing that land-based TC precipitation maximizes in the afternoon, especially in the outer TC regions (>150 km from the TC center).
This study further finds that diurnal cycles of precipitation frequency and intensity (conditional rain rate) markedly differ from each other, which may provide some insights into mechanisms responsible for diurnal variations on TC precipitation. First of all, the nocturnal peak (0300–0600 LT) of TC rainfall over ocean is manifested in terms of precipitation intensity, while precipitation frequency has virtually no diurnal variation. This suggests that the increase of overnight TC precipitation over ocean is due to enhancement of existing convective clouds instead of new convective development. The enhancement is likely induced by elevated instability within the existing cloud column due to nighttime cloud-top radiative cooling, as well as low-level convergence caused by differential radiative cooling at night between cloudy and cloud-free areas (Melhauser and Zhang 2014; Nicholls 2015; Tang and Zhang 2016; Ruppert and Hohenegger 2018). These are named “radiative cooling” mechanisms. However, the afternoon precipitation peak of land-based TCs is dominated by the marked increase (by 50%) of precipitation frequency. In contrast, precipitation intensity of land-based TCs shows a similar diurnal cycle as oceanic convection, maximizing in the early morning (increase by 20%). The higher frequency of land-based TC precipitation in the afternoon may indicate an increase of new convection, which could be triggered by elevated low-level buoyancy over cloud-free areas due to stronger surface heating there, (i.e., the “thermal-driven” mechanism) (Dai 2001; Nesbitt and Zipser 2003). On the other hand, nocturnal enhancement in precipitation intensity of land-based TC precipitation may share the same so-called radiative cooling mechanisms as those over ocean (i.e., nighttime cloud-top radiative cooling and differential radiative cooling effects). In short, the radiative cooling mechanism plays a dominant role in the nocturnal maximum of TC precipitation over the ocean. Although radiative cooling and thermal-driven mechanisms both play important roles in diurnal variations of land-based TC precipitation, the thermal-driven mechanism may be more dominant.
Acknowledgments
This research was supported by the National Key R&D Program of China (2019YFC1510400), National Natural Science Foundation of China (41975053), and the Guangdong Provincial Department of Science and Technology, China (Grants 2019QN01G107 and 2019ZT08G090). All three anonymous reviewers are much appreciated for their constructive comments and suggestions in improving this paper.
Data availability statement
The IMERG rainfall product data can be downloaded from NASA Goddard Earth Sciences Data and Information Services Center (https://disc.gsfc.nasa.gov/datasets/GPM_3IMERGHH_06/summary), the IBTrACS data can be accessed from the NOAA National Climatic Data Center (https://climatedataguide.ucar.edu/climate-data/ibtracs-tropical-cyclone-best-track-data), and the ERA5 data can be accessed from the ECMWF (https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5).
REFERENCES
Amante, C., and B. Eakins, 2009: ETOPO1 1 arc-minute global relief model: Procedures, data sources, and analysis. NOAA Tech. Memo. NESDIS NGDC-24, 25 pp., https://www.ngdc.noaa.gov/mgg/global/relief/ETOPO1/docs/ETOPO1.pdf.
Bowman, K. P., and M. D. Fowler, 2015: The diurnal cycle of precipitation in tropical cyclones. J. Climate, 28, 5325–5334, https://doi.org/10.1175/JCLI-D-14-00804.1.
Bowman, K. P., J. C. Collier, G. R. North, Q. Wu, E. Ha, and J. Hardin, 2005: Diurnal cycle of tropical precipitation in Tropical Rainfall Measuring Mission (TRMM) satellite and ocean buoy rain gauge data. J. Geophys. Res., 110, D21104, https://doi.org/10.1029/2005JD005763.
Cecil, D., 2007: Satellite-derived rain rates in vertically sheared tropical cyclones. Geophys. Res. Lett., 34, L02811, https://doi.org/10.1029/2006GL027942.
Chang, C.-P., Y. Lei, C.-H. Sui, X. Lin, and F. Ren, 2012: Tropical cyclone and extreme rainfall trends in East Asian summer monsoon since mid-20th century. Geophys. Res. Lett., 39, L18702, https://doi.org/10.1029/2012GL052945.
Chen, S. S., J. A. Knaff, and F. D. Marks, 2006: Effects of vertical wind shear and storm motion on tropical cyclone rainfall asymmetries deduced from TRMM. Mon. Wea. Rev., 134, 3190–3208, https://doi.org/10.1175/MWR3245.1.
Chen, T.-C., 2003: Maintenance mechanism of the summer monsoon circulation: A planetary-scale perspective. J. Climate, 16, 2022–2037, https://doi.org/10.1175/1520-0442(2003)016<2022: MOSMCA>2.0.CO;2.
Chen, T.-C., J.-D. Tsay, and J. Matsumoto, 2017: Interannual variation of the summer rainfall center in the South China Sea. J. Climate, 30, 7909–7931, https://doi.org/10.1175/JCLI-D-16-0889.1.
Ciesielski, P., and R. Johnson, 2006: Contrasting characteristics of convection over the northern and southern South China Sea during SCSMEX. Mon. Wea. Rev., 134, 1041–1062, https://doi.org/10.1175/MWR3113.1.
Collier, J. C., and K. P. Bowman, 2004: Diurnal cycle of tropical precipitation in a general circulation model. J. Geophys. Res., 109, D17105, https://doi.org/10.1029/2004JD004818.
Dai, A., 2001: Global precipitation and thunderstorm frequencies. Part II: Diurnal variations. J. Climate, 14, 1112–1128, https://doi.org/10.1175/1520-0442(2001)014<1112:GPATFP>2.0.CO;2.
Ditchek, S. D., K. L. Corbosiero, R. G. Fovell, and J. Molinari, 2019a: Electrically active tropical cyclone diurnal pulses in the Atlantic basin. Mon. Wea. Rev., 147, 3595–3607, https://doi.org/10.1175/MWR-D-19-0129.1.
Ditchek, S. D., J. Molinari, K. L. Corbosiero, and R. G. Fovell, 2019b: An objective climatology of tropical cyclone diurnal pulses in the Atlantic basin. Mon. Wea. Rev., 147, 591–605, https://doi.org/10.1175/MWR-D-18-0368.1.
Du, Y., and G. Chen, 2019: Climatology of low-level jets and their impact on rainfall over southern China during the early-summer rainy season. J. Climate, 32, 8813–8833, https://doi.org/10.1175/JCLI-D-19-0306.1.
Dunion, J. P., C. D. Thorncroft, and C. S. Velden, 2014: The tropical cyclone diurnal cycle of mature hurricanes. Mon. Wea. Rev., 142, 3900–3919, https://doi.org/10.1175/MWR-D-13-00191.1.
Fisher, R., 1971: Statistical methods for research workers. Biometrics, 27, 1106, https://doi.org/10.2307/2528855.
Hence, D. A., and R. A. Houze, 2011: Vertical structure of hurricane eyewalls as seen by the TRMM precipitation radar. J. Atmos. Sci., 68, 1637–1652, https://doi.org/10.1175/2011JAS3578.1.
Hendricks, E. A., M. S. Peng, B. Fu, and T. Li, 2010: Quantifying environmental control on tropical cyclone intensity change. Mon. Wea. Rev., 138, 3243–3271, https://doi.org/10.1175/2010MWR3185.1.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement Mission. Bull. Amer. Meteor. Soc., 95, 701–722, https://doi.org/10.1175/BAMS-D-13-00164.1.
Hu, H., Y. Duan, Y. Wang, and X. Zhang, 2017: Diurnal cycle of rainfall associated with landfalling tropical cyclones in China from rain gauge observations. J. Appl. Meteor. Climatol., 56, 2595–2605, https://doi.org/10.1175/JAMC-D-16-0335.1.
Huang, W.-R., and J. C. L. Chan, 2011: Maintenance mechanisms for the early-morning maximum summer rainfall over southeast China. Quart. J. Roy. Meteor. Soc., 137, 959–968, https://doi.org/10.1002/qj.815.
Huang, W.-R., J. C. L. Chan, and S.-Y. Wang, 2010: A planetary-scale land–sea breeze circulation in East Asia and the western North Pacific. Quart. J. Roy. Meteor. Soc., 136, 1543–1553, https://doi.org/10.1002/qj.663.
Huffman, G., and Coauthors, 2020: Integrated Multi-satellite Retrievals for the Global Precipitation Measurement (GPM) Mission (IMERG). Satellite Precipitation Measurement, V. Levizzani et al., Eds., Springer, 343–353, https://doi.org/10.1007/978-3-030-24568-9_19.
Jiang, H., and E. J. Zipser, 2010: Contribution of tropical cyclones to the global precipitation from eight seasons of TRMM data: Regional, seasonal, and interannual variations. J. Climate, 23, 1526–1543, https://doi.org/10.1175/2009JCLI3303.1.
Jiang, H., C. Liu, and E. J. Zipser, 2011: A TRMM-based tropical cyclone cloud and precipitation feature database. J. Appl. Meteor. Climatol., 50, 1255–1274, https://doi.org/10.1175/2011JAMC2662.1.
Jiang, Z., D. Zhang, R. Xia, and T. Qian, 2017: Diurnal variations of presummer rainfall over southern China. J. Climate, 30, 755–773, https://doi.org/10.1175/JCLI-D-15-0666.1.
Joyce, R., and P. Xie, 2011: Kalman filter-based CMORPH. J. Hydrometeor., 12, 1547–1563, https://doi.org/10.1175/JHM-D-11-022.1.
Joyce, R., J. Janowiak, P. Arkin, and P. Xie, 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydrometeor., 5, 487–503, https://doi.org/10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2.
Kim, K., J. Park, J. Baik, and M. Choi, 2017: Evaluation of topographical and seasonal feature using GPM IMERG and TRMM 3B42 over Far-East Asia. Atmos. Res., 187, 95–105, https://doi.org/10.1016/j.atmosres.2016.12.007.
Knapp, K., M. C. Kruk, D. H. Levinson, H. J. Diamond, and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTrACS). Bull. Amer. Meteor. Soc., 91, 363–376, https://doi.org/10.1175/2009BAMS2755.1.
Kossin, J. P., 2002: Daily hurricane variability inferred from GOES infrared imagery. Mon. Wea. Rev., 130, 2260–2270, https://doi.org/10.1175/1520-0493(2002)130<2260:DHVIFG>2.0.CO;2.
Lajoie, F., and I. Butterworth, 1984: Oscillation of high-level cirrus and heavy precipitation around Australian region tropical cyclones. Mon. Wea. Rev., 112, 535–544, https://doi.org/10.1175/1520-0493(1984)112<0535:OOHLCA>2.0.CO;2.
Leppert, K. D., and D. J. Cecil, 2016: Tropical cyclone diurnal cycle as observed by TRMM. Mon. Wea. Rev., 144, 2793–2808, https://doi.org/10.1175/MWR-D-15-0358.1.
Li, Q., and Y. Wang, 2012: Formation and quasi-periodic behavior of outer spiral rainbands in a numerically simulated tropical cyclone. J. Atmos. Sci., 69, 997–1020, https://doi.org/10.1175/2011JAS3690.1.
Lonfat, M., F. D. Marks, and S. S. Chen, 2004: Precipitation distribution in tropical cyclones using the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager: A global perspective. Mon. Wea. Rev., 132, 1645–1660. https://doi.org/10.1175/1520-0493(2004)132<1645:PDITCU>2.0.CO;2.
Ma, Q., Y. Li, H. Feng, Q. Yu, Y. Zou, F. Liu, and B. Pulatov, 2021: Performance evaluation and correction of precipitation data using the 20-year IMERG and TMPA precipitation products in diverse subregions of China. Atmos. Res., 249, 105304, https://doi.org/10.1016/j.atmosres.2020.105304.
Melhauser, C., and F. Zhang, 2014: Diurnal radiation cycle impact on the pregenesis environment of Hurricane Karl (2010). J. Atmos. Sci., 71, 1241–1259, https://doi.org/10.1175/JAS-D-13-0116.1.
Miller, R., and W. Frank, 1993: Radiative forcing of simulated tropical cloud clusters. Mon. Wea. Rev., 121, 482–498, https://doi.org/10.1175/1520-0493(1993)121<0482:RFOSTC>2.0.CO;2.
Mori, N., and T. Takemi, 2016: Impact assessment of coastal hazards due to future changes of tropical cyclones in the North Pacific Ocean. Wea. Climate Extremes, 11, 53–69, https://doi.org/10.1016/j.wace.2015.09.002.
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, https://doi.org/10.1175/1520-0442-16.10.1456.
Nicholls, M. E., 2015: An investigation of how radiation may cause accelerated rates of tropical cyclogenesis and diurnal cycles of convective activity. Atmos. Chem. Phys., 15, 9003–9029, https://doi.org/10.5194/acp-15-9003-2015.
Ruppert, J. H., and C. Hohenegger, 2018: Diurnal circulation adjustment and organized deep convection. J. Climate, 31, 4899–4916, https://doi.org/10.1175/JCLI-D-17-0693.1.
Serra, Y., and M. McPhaden, 2004: In situ observations of diurnal variability in rainfall over the tropical Pacific and Atlantic Oceans. J. Climate, 17, 3496–3509, https://doi.org/10.1175/1520-0442(2004)017<3496:ISOODV>2.0.CO;2.
Shu, H., Q. Zhang, and B. Xu, 2013: Diurnal variation of tropical cyclone rainfall in the western North Pacific in 2008–2010. Atmos. Oceanic Sci. Lett., 6, 103–108, https://doi.org/10.1080/16742834.2013.11447064.
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, https://doi.org/10.1175/1520-0493(1987)115<1465:DFTTCO>2.0.CO;2.
Tan, J., G. J. Huffman, D. T. Bolvin, and E. J. Nelkin, 2019: Diurnal cycle of IMERG V06 precipitation. Geophys. Res. Lett., 46, 13 584–13 592, https://doi.org/10.1029/2019GL085395.
Tang, X., and F. Zhang, 2016: Impacts of the diurnal radiation cycle on the formation, intensity, and structure of Hurricane Edouard (2014). J. Atmos. Sci., 73, 2871–2892, https://doi.org/10.1175/JAS-D-15-0283.1.
Tang, X., Q. Cai, J. Fang, and J. M. Tan, 2019: Land–sea contrast in the diurnal variation of precipitation from landfalling tropical cyclones. J. Geophys. Res. Atmos., 124, 12 010–12 021, https://doi.org/10.1029/2019JD031454.
Tao, C., and H. Jiang, 2015: Distributions of shallow to very deep precipitation–convection in rapidly intensifying tropical cyclones. J. Climate, 28, 8791–8824, https://doi.org/10.1175/JCLI-D-14-00448.1.
Tao, W. K., S. Lang, J. Simpson, C. Sui, B. Ferrier, and M.-D. Chou, 1996: Mechanisms of cloud–radiation interaction in the tropics and midlatitudes. J. Atmos. Sci., 53, 2624–2651, https://doi.org/10.1175/1520-0469(1996)053<2624:MOCRII>2.0.CO;2.
Wingo, M. T., and D. J. Cecil, 2010: Effects of vertical wind shear on tropical cyclone precipitation. Mon. Wea. Rev., 138, 645–662, https://doi.org/10.1175/2009MWR2921.1.
Wu, Q., and Z. Ruan, 2016: Diurnal variations of the areas and temperatures in tropical cyclone clouds. Quart. J. Roy. Meteor. Soc., 142, 2788–2796, https://doi.org/10.1002/qj.2868.
Wu, Q., Z. Ruan, D. Chen, and T. Lian, 2015: Diurnal variations of tropical cyclone precipitation in the inner and outer rainbands. J. Geophys. Res., 120 (1), 1–11, https://doi.org/10.1002/2014JD022190.
Xu, K.-M., and D. Randall, 1995: Impact of interactive radiative transfer on the macroscopic behavior of cumulus ensembles. Part II: Mechanisms for cloud–radiation interactions. J. Atmos. Sci., 52, 800–817, https://doi.org/10.1175/1520-0469(1995)052<0800:IOIRTO>2.0.CO;2.
Xu, W., H. Jiang, and X. Kang, 2014: Rainfall asymmetries of tropical cyclones prior to, during, and after making landfall in South China and southeast United States. Atmos. Res., 139, 18–26, https://doi.org/10.1016/j.atmosres.2013.12.015.
Yaroshevich, M., and L. Ingel, 2013: Diurnal variations in the intensity of tropical cyclones. Izv. Atmos. Ocean. Phys., 49, 375–379, https://doi.org/10.1134/S0001433813040117.
Yu, Z., Y. Wang, and H. Xu, 2015: Observed rainfall asymmetry in tropical cyclones making landfall over China. J. Appl. Meteor. Climatol., 54, 117–136, https://doi.org/10.1175/JAMC-D-13-0359.1.
Yu, Z., and Coauthors, 2017: On the relationship between intensity and rainfall distribution in tropical cyclones making landfall over China. J. Appl. Meteor. Climatol., 56, 2883–2901, https://doi.org/10.1175/JAMC-D-16-0334.1.
Zhang, J. A., J. P. Dunion, and D. S. Nolan, 2020: In situ observations of the diurnal variation in the boundary layer of mature hurricanes. Geophys. Res. Lett., 47, 2019GL086206, https://doi.org/10.1029/2019GL086206.
Zhang, Q., L. Wu, and Q. Liu, 2009: Tropical cyclone damages in China 1983–2006. Bull. Amer. Meteor. Soc., 90, 489–496, https://doi.org/10.1175/2008BAMS2631.1.
Zhang, Q., and Coauthors, 2017: Impact of tropical cyclones on flood risk in southeastern China: Spatial patterns, causes, and implications. Global Planet. Change, 150, 81–93, https://doi.org/10.1016/j.gloplacha.2017.02.004.
