Abstract

The temporally dense geostationary satellite observations made possible by recent technological advances enable atmospheric motion vectors (AMVs) to be derived that are suitable for capturing atmospheric flows even of mesoscale phenomena, for which in situ data are scarce. Tropical cyclone (TC) outflows around the cloud top, reflecting TC secondary circulation, were computed by using AMVs derived from successive Multifunctional Transport Satellite (MTSAT) imagery, and the relationship between TC intensification rate (defined as the change of the best-track maximum sustained wind in the previous 24 h) and the outflow was investigated for 44 TCs occurring during 2011–14. During the TC intensification phase, temporal changes in the outflow were generally synchronous with changes in the cloud-top temperature of TC inner-core convective clouds detected by MTSAT infrared band. It was noteworthy that the intensification rates of 66% of the TCs peaked 0–36 h after outflow maximization and that the intensification rate for TCs with a maximum rate of >15 m s−1 day−1 peaked after the outflow maximum. Furthermore, TCs with a large intensification rate and latent-heat release around the midlevel tended to have a large outflow during the intensification phase. A comparison of TCs with and without convective bursts (CBs) revealed that the correlation between outflow and the TC intensification rate was higher for TCs accompanied by CBs than for those without CBs, implying that a rapid deepening of inner-core convection is important for intensification of a TC’s secondary circulation. The outflow tended to be most correlated with the TC intensification rate 0–6 h earlier.

1. Introduction

A tropical cyclone (TC) is a storm system that is characterized by a low pressure center surrounded by spiral rainbands, which produce strong winds and heavy rain. The primary (i.e., horizontal) circulation of a TC can be depicted as a quasi-axisymmetric vortex of gradient wind and hydrostatic balance. As a TC develops, the secondary (i.e., radial and vertical) circulation intensifies in response to both diabatic heating and momentum forcing together with surface friction. The inward transport of large absolute angular momentum in the boundary layer by the secondary circulation spins up the TC’s primary circulation (Shapiro and Willoughby 1982; Fudeyasu and Wang 2011).

The TC inner core, including eyewalls and inner rainbands, is generally recognized as the area within a radius from the TC center that is 2–3 times the radius of maximum wind (Wang 2009; Li and Wang 2012). In the inner core, the tangential velocity reaches its maximum strength and strong updrafts compose the secondary circulation (Holland and Merrill 1984; Smith et al. 2009; Houze 2010). Diabatic heating (latent heating) associated with deep convection in the TC inner core plays a particularly important role in the intensification of the primary and secondary circulations. Schubert et al. (1999) used the Rossby–Ertel potential vorticity equation to examine the response of TC primary circulation to diabatic heating in the TC inner core. Shapiro and Willoughby (1982) showed by the Sawyer–Eliassen equation that TC secondary circulation is driven by midlevel diabatic heating caused by deep convection within the TC inner core. In a theoretical study, Emanuel (1986) suggested an essential process describing the positive feedback between TC primary and secondary circulations and the increase in the radial gradient of the entropy flux in and above the boundary layer due to the heat and water vapor fluxes from the ocean. In recent years, numerical studies performed with a high-resolution nonhydrostatic model have revealed that upward mass fluxes generated by inner-core convection play an important role in TC intensification (Rogers 2010; Wang and Wang 2014).

The formation and development of the TC inner core are often considered to be a result of consecutive vigorous convections called convective bursts (CBs; Riehl and Malkus 1961; Zehr 1992). Guimond et al. (2010) showed the existence of deep CBs preceding a period of TC rapid intensification for Hurricane Dennis (2005) by analyzing wind-profile data observed by airborne Doppler radar. Hazelton et al. (2017) used composites of TC numerical simulations to indicate that TCs with CBs tend to be accompanied by stronger radial inflow at 2-km height than are those without CBs. CBs are considered to play several essential roles in TC intensification. First, CBs facilitate the exchange of air between the troposphere and the lower stratosphere, and they enhance the upper-tropospheric warm core near the TC center (Elsberry et al. 2013; Ohno and Satoh 2015). Second, the upward motion induced by CBs conveys a large amount of absolute angular momentum from the surface to near the tropopause. This process is essential to develop the TC vortex vertically (Sawada and Iwasaki 2007; Bryan and Rotunno 2009). In terms of the interaction between a TC and the environmental wind field, the presence of an “outflow jet,” a synoptic atmospheric flow from a TC to the environment, can facilitate the enhancement of outflow and lead to rapid intensification of the TC (Rappin et al. 2011).

Although many studies have focused on the role of TC secondary circulation, which is essential for TC intensification, not many studies have observationally verified the relationship between a TC’s secondary circulation and its intensification, because in situ and aircraft observations are spatially and temporally sparse. In recent years, however, temporally dense geostationary satellite observations have been made available by technological advances (Berger et al. 2011; Bessho et al. 2016; Line et al. 2016). Satellite observations at time intervals as short as 5–15 min, often referred to as rapid scans, are useful for deriving spatially and temporally dense atmospheric motion vectors (AMVs; Velden et al. 2005; Oyama 2015). AMVs are wind products that are identified by tracking clouds and water vapor patterns in successive geostationary satellite images, and they are used not only for numerical weather prediction (Warrick 2016) but also for analysis of atmospheric wind fields associated with synoptic (Molinari and Vollaro 1989) and mesoscale phenomena (Apke et al. 2016). For example, the assimilation of AMV data in numerical weather prediction models has been shown to improve TC forecasts (Langland et al. 2009; Yamashita 2012; Wu et al. 2014), and Oyama et al. (2016) showed that upper-tropospheric AMVs could be used to detect characteristic phenomena during intensification of Typhoon Danas (1324)—in particular, increases in radial outflow (i.e., secondary circulation) and tangential winds (i.e., primary circulation) around the cloud top associated with CBs.

The purpose of this study was to verify the relationship between TC intensification and secondary circulation associated with TC inner-core convection by using AMVs derived from observations of the Multifunctional Transport Satellite (MTSAT) and other meteorological information deduced from MTSAT and other satellite observations of TCs in the western North Pacific Ocean basin. To accomplish this objective, the study focuses on the TC intensification phase. Figure 1 summarizes the TC radial outflow near the cloud top that was investigated by using the AMVs in this study and related TC processes. The results of this study are expected to contribute to elucidation of the TC intensification process via the secondary circulation and to improve the diagnosis of TC intensification. Section 2 describes the studied TCs and the data and method used in this study. Section 3 examines the quantitative and temporal relationships between TC inner-core convection and cloud-top outflow and between the outflow and the TC intensification rate, together with the occurrence of CBs during the TC intensification phase. The results are summarized and discussed in section 4.

Fig. 1.

Schematic diagram summarizing the cloud-top outflow of a TC (denoted by “Process D”) that was investigated by using MTSAT upper-tropospheric AMVs in this study, along with related TC processes. Red arrows denote the atmospheric flows composing the secondary circulation, and blue arrows denote the primary circulation. Deep convective clouds within the eyewalls and the warm core are drawn as the light-gray and pink areas, respectively.

Fig. 1.

Schematic diagram summarizing the cloud-top outflow of a TC (denoted by “Process D”) that was investigated by using MTSAT upper-tropospheric AMVs in this study, along with related TC processes. Red arrows denote the atmospheric flows composing the secondary circulation, and blue arrows denote the primary circulation. Deep convective clouds within the eyewalls and the warm core are drawn as the light-gray and pink areas, respectively.

2. Data and method

a. Studied TCs and data

The studied TCs and data are summarized in Tables 1 and 2, respectively. For the TC center positions and intensities—namely, the 10-min-average maximum sustained winds (MSW) and minimum sea level pressure (MSLP) and the smallest radius of a 30-kt wind (R30; 1 kt = 0.51 m s−1), best-track data from the Regional Specialized Meteorological Center Tokyo (RSMC Tokyo; http://www.jma.go.jp/jma/jma-eng/jma-center/rsmc-hp-pub-eg/besttrack.html) were used. For the analysis of cloud-top temperatures of deep convection with hourly MTSAT infrared imagery data, hourly TC positions were estimated from the 6-hourly best-track data by linear interpolation.

Table 1.

Configurations and analysis results for TCs in this study. R30, VWS, r, and NCBs in this table are the smallest radius of 30-kt winds, averaged vertical wind shear (200–850 hPa) within a radius of 600 km from the TC center, the correlation coefficient between the UMaxOutflow and TB_C200 time series, and the number of CBs during the TC intensification phase, respectively.

Configurations and analysis results for TCs in this study. R30, VWS, r, and NCBs in this table are the smallest radius of 30-kt winds, averaged vertical wind shear (200–850 hPa) within a radius of 600 km from the TC center, the correlation coefficient between the UMaxOutflow and TB_C200 time series, and the number of CBs during the TC intensification phase, respectively.
Configurations and analysis results for TCs in this study. R30, VWS, r, and NCBs in this table are the smallest radius of 30-kt winds, averaged vertical wind shear (200–850 hPa) within a radius of 600 km from the TC center, the correlation coefficient between the UMaxOutflow and TB_C200 time series, and the number of CBs during the TC intensification phase, respectively.
Table 2.

Data used in this study.

Data used in this study.
Data used in this study.

The best-track analysis uses the following intensity categories: tropical depression (TD) has MSW < 34 kt (17 m s−1), tropical storm (TS) has MSW from 34 to 48 kt (25 m s−1), severe tropical storm (STS) has MSW = 48–64 kt (33 m s−1), and typhoon (TY) has MSW greater than 64 kt. RSMC Tokyo regards the transitional time period from TD to TS as the “typhoon formation time.” In this study, the change of MSW in the previous 24 h is defined as the TC intensification rate (e.g., Kaplan et al. 2010). Note that during the TD period the best-track data do not include MSWs for every TC but do include the TC center position and MSLP. The “intensification phase” of a TC is defined as the period from the typhoon formation time to the time at which the lifetime maximum MSW is first observed (i.e., the start time of the mature TC phase).

To compute the AMVs and analyze the cloud-top temperatures associated with deep convection, this study used MTSAT imagery data that were obtained by the infrared (IR1; wavelengths of 10.3–11.3 μm) and water vapor (WV; wavelengths of 6.5–7.0 μm) bands. Furthermore, to examine the formation of the TC inner core, the convective rainfall was identified by using the polarization-corrected temperature of the 91-GHz band (PCT91) of the Special Sensor Microwave Imager/Sounder (SSMIS) that is on board the Defense Meteorological Satellite Program (DMSP) satellites (Spencer et al. 1989; Hawkins et al. 2008). Because observations by three DMSP satellites (i.e., F-16, F-17, and F-18) were used in this study, the frequency of the observations was 6 times per day at maximum.

To verify the quality of the MTSAT AMV data, this study used dropsonde wind-profile data obtained by the Dropwindsonde Observations for Typhoon Surveillance near the Taiwan Region (DOTSTAR; Wu et al. 2005). This verification was conducted for only TCs that approached near the Taiwan region where the DOTSTAR data were available.

Vertical wind shear (VWS) in a TC’s environment causes a TC vortex to become tilted and negatively influences TC intensity (Kaplan and DeMaria 2003; Ueno 2007; Reasor et al. 2013). In particular, VWS greater than a threshold of approximately 10 m s−1 may significantly influence TC intensity (Gallina and Velden 2002; Paterson et al. 2005), although it sometimes creates favorable conditions for the outbreak of convection leading to TC intensification (Molinari et al. 2013). To investigate the influence of VWS on the relationship between a TC’s outflow near the cloud top, as estimated by using AMVs, and TC intensification, VWS was computed as the averaged wind difference between the 200- and 850-hPa levels in a circle with a radius of 600 km around the TC center by using the Japanese 55-year Reanalysis dataset (JRA-55; Kobayashi et al. 2015). The grid size of JRA-55 data is 1.25° in coordinates of latitude and longitude.

To examine the relationship between latent heating that occurs within the inner core and intensification of the outflow near the cloud top, latent-heating profiles were evaluated. The data used were the Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) “L2H25” product (version 7) provided by the Japan Aerospace Exploration Agency (JAXA; Shige et al. 2004, 2007).

In this study, TCs occurring in the western North Pacific basin during 2011–14 were investigated, focusing on relatively strong TCs with MSWs of greater than 40 kt (20.6 m s−1). In addition, TCs whose areas were adequately covered by AMV data were selected. The filtering of TCs on the basis of the data coverage of AMVs in the TC areas is described in section 2b. Because the focus was on the TC intensification process, the analysis period for each TC was limited to the TC intensification phase. Of 100 TCs occurring during 2011–14, 44 TCs were selected (Table 1).

b. Derivation of AMVs

Upper-tropospheric AMVs were computed at 0000, 0600, 1200, and 1800 UTC from the successive MTSAT images observed by the IR1 and WV bands by using the AMV derivation technique of the Meteorological Satellite Center (MSC) of the Japan Meteorological Agency (JMA; Oyama 2010). In this technique, the wind vectors were computed by tracking high-level clouds in successive images obtained at intervals of 15 min around the 6-hourly times. The tracking process was performed with a cross-correlation matching technique. The target box, that is, the image segment used for tracking clouds, was a square with a side of 16 image pixels (image pixel size is 4 km at the subsatellite point for the IR1 and WV bands) on a 0.25° latitude × 0.25° longitude grid. The AMVs were allocated at the cloud-top height, which was estimated using data for the brightness temperature TB of the IR1 band in the target box by referring to the first-guess field (6-h forecast) of JMA’s Global Spectral Model (GSM). The horizontal resolution of GSM data is 20 km. The TB data of IR1 used for the height allocation were corrected by the IR–WV intercept method, using the TB data of the IR1 and WV bands (Nieman et al. 1993), to take into account cloud semitransparency. For all TCs, AMV data were computed from the typhoon formation time (the start of TS) to the mature TC phase by taking into account the periods of best-track MSW data.

For the wind field analysis, this study used IR AMVs and WV AMVs assigned to pressure levels between 100 and 300 hPa in a complementary manner to maximize the AMV data coverage. AMV data were screened by using the quality indicator (QI) that was developed by the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT; Holmlund 1998); that is, only AMV data with QI greater than 0.3 were used (EUMETSAT 2015). The QI threshold of 0.3 was determined by referring to the threshold to screen AMVs that is used for producing the divergence product by EUMETSAT (EUMETSAT 2015).

TCs for which the AMV data were sparsely derived were excluded to avoid analysis errors due to small data sampling. For this reason, only TCs for which the AMV data covered more than 50% of the area within a radius of 200 km from the TC center were selected. AMV data for a TC may be sparsely derived 1) because the cloud pattern in the target box used for tracking is vague, as seen for a distinct “Central Dense Overcast,” 2) as a result of navigation errors in the satellite imagery, or 3) because of abrupt changes in the atmospheric flow. After this filtering and the elimination of weak TCs (section 2a), 44 of the total of 100 TCs occurring during 2011–14 in the western North Pacific basin were selected. The tracks of the selected TCs are shown in Fig. 2.

Fig. 2.

Tracks of TCs used in this study: (a) 2011, (b) 2012, (c) 2013, and (d) 2014. Tracks are shown only from the TD period to the start of the mature phase. Black, blue, green, and red lines denote TD, TS, STS, and TY periods, respectively. See Table 1 for the TC data.

Fig. 2.

Tracks of TCs used in this study: (a) 2011, (b) 2012, (c) 2013, and (d) 2014. Tracks are shown only from the TD period to the start of the mature phase. Black, blue, green, and red lines denote TD, TS, STS, and TY periods, respectively. See Table 1 for the TC data.

c. Analysis of TC outflow around the cloud top using AMVs

The radial component of the wind vectors of the upper-tropospheric AMVs was calculated to represent the outflow around the TC cloud top. The maximum radial wind (UMaxOutflow) was derived as the maximum of the azimuthally averaged radial winds on five annuli at radii of 100, 150, 200, 250, and 300 km from the best-track TC center (Fig. 3). These radii were adopted on the basis of the general inner-core size reported by previous studies (e.g., Croxford and Barnes 2002). It is possible that in some cases the TC center, that is, the location of the minimum sea level pressure, which is usually detected as the center of rotation near the surface, might be different from the rotation center around the cloud top. This difference might be due to a tilting of the TC vortex by a large VWS or to the accuracy of the hourly TC center derived from the best-track data. To assess the influence of a position difference, the best-track TC center was compared with the center of rotation around the cloud top, which was determined from upper-tropospheric AMV data, spatially interpolated on a 0.125° latitude × 0.125° longitude grid, as the rotation maximum. This comparison showed that the average position difference during the TC intensification phase was 0.5°–0.6°and that it was slightly dependent on the VWS between 0.5 and 12.8 m s−1 (figure not shown). Moreover, the average difference in UMaxOutflow between the two locations was ~1.0 m s−1 during the TC intensification phase. The influence of this relatively small UMaxOutflow difference, which may be due to the azimuthal averaging of wind components, on the analysis result is considered to be small.

Fig. 3.

UMaxOutflow derivation, with UMaxOutflow being defined as the maximum azimuthally averaged radial wind on annuli (±50 km) at radii of 100, 150, 200, 250, and 300 km (red circles) from the TC center (plus sign). Upper-tropospheric AMVs (blue wind barbs) are displayed against the MTSAT-2 TB (WV band: 6.5–7.0 μm) of Typhoon Nuri (1420) at 0600 UTC 31 Oct 2014, during its intensification phase.

Fig. 3.

UMaxOutflow derivation, with UMaxOutflow being defined as the maximum azimuthally averaged radial wind on annuli (±50 km) at radii of 100, 150, 200, 250, and 300 km (red circles) from the TC center (plus sign). Upper-tropospheric AMVs (blue wind barbs) are displayed against the MTSAT-2 TB (WV band: 6.5–7.0 μm) of Typhoon Nuri (1420) at 0600 UTC 31 Oct 2014, during its intensification phase.

d. Detections of convective bursts and rainfall asymmetry

The characteristics of convective activity within the TCs were examined by using TB (IR1) data of MTSAT, which represent the cloud-top temperature of optically dense clouds such as cumulonimbus clouds and a high-level cirrus canopy. To detect deep convective clouds associated with CBs, the averaged TB in a circle with a radius of 200 km around the TC center (TB_C200) was used. The radius of 200 km was determined as being large enough to cover the TC inner core (e.g., Croxford and Barnes 2002), and it is similar to the radius used by Steranka et al. (1986) to detect CBs.

CBs were defined as a period of consecutive deep convective activity during which TB_C200 decreased by more than 20 K in the 12 h before it reached a minimum value colder than 238 K. The threshold of 12 h was decided by referring to the result of Steranka et al. (1986) that indicated that convective bursts leading to TC intensification continue for 9 h or longer. The threshold of 238 K was empirically determined to select CBs with enough deep convection to cause diabatic heating (Steranka et al. 1986). The duration of each CB was defined as the period during which TB_C200 monotonically decreased between the end of the previous CB and the occurrence of the TB_C200 minimum. An example of CB detection is shown by using the time series of TB_C200 and MSW of Typhoon Danas (1324; Fig. 4). Three CBs were detected during the TD period and TC intensification phase.

Fig. 4.

Time series of TB_C200 and MSW of Typhoon Danas (1324) from the TD period to the decay phase. Three CBs are detected as a period of consecutive convective activity during which TB_C200 dropped at least 20 K in the 12 h before it reached a minimum temperature colder than 238 K (shown by the horizontal thick dashed line).

Fig. 4.

Time series of TB_C200 and MSW of Typhoon Danas (1324) from the TD period to the decay phase. Three CBs are detected as a period of consecutive convective activity during which TB_C200 dropped at least 20 K in the 12 h before it reached a minimum temperature colder than 238 K (shown by the horizontal thick dashed line).

Furthermore, to examine the asymmetry of rainfall area that is associated with the formation and development of the TC inner core during CBs, we introduced the azimuthal wavenumber components of the convective rainfall area near the TC center (Ueno 2007), which were analyzed by using PCT91 data. Low PCT91 represents an area of deep convection since the microwave signals are attenuated by the existence of ice rain particles. This study investigated the wavenumber components within a radius of 150 km from the TC center.

3. Results

a. Verification of AMVs using DOTSTAR data

Oyama (2015) evaluated the quality of the MTSAT upper-tropospheric AMVs against rawinsonde wind data in circles with a radius of 600 km around the centers of 16 TCs that occurred during 2011–13. It was found that the vector difference and bias were 6.7–7.5 m s−1 and from −0.8 to −0.5 m s−1, respectively. This evaluation study indicated that the difference between the AMVs and the rawinsonde winds could come from collocation and derivation errors, which may increase when temporal and spatial variations of the TC wind field are large.

For an additional study, the MTSAT upper-tropospheric AMVs within a radius of 600 km from the TC center were verified with reference to DOTSTAR wind-profile data. In the 44 TCs, DOTSTAR data were available for Typhoons Meari (1105), Roke (1115), Talim (1205), Saola (1209), Jelawat (1217), Kong-rey (1315), Fitow (1323), Matmo (1410), and Fung-wong (1416). The AMV data (0.3 < QI) at a pressure level between 100 and 300 hPa were compared with the DOTSTAR wind, which was vertically interpolated to the AMV’s pressure level within a horizontal distance of 32 km from the location of the AMV data. The 32 km of horizontal distance for the comparison between the AMV and DOTSTAR data was determined as one-half of the target box size that was used in the AMV derivations (section 2b). The verification results showed that the mean vector difference and mean bias for the AMVs were 6.9 and 0.4 m s−1, respectively, for radii of 0–600 km (Table 3), which are comparable to the results with reference to rawinsonde wind data by Oyama (2015). Although the dependencies of the mean vector difference and mean bias on the radius from the TC center are not clear, the mean wind speed for radii of 0–200 km is slightly larger than those for radii of 200–400 km and radii of 400–600 km.

Table 3.

Verification results of MTSAT AMVs at pressure levels of 100–300 hPa (0.3 < QI) with reference to dropsonde wind-profile data observed by DOTSTAR for Typhoons Meari (1105), Roke (1115), Talim (1205), Saola (1209), Jelawat (1217), Kong-rey (1315), Fitow (1323), Matmo (1410), and Fung-wong (1416). Radius denotes the distance between the best-track TC center and the location of AMV data.

Verification results of MTSAT AMVs at pressure levels of 100–300 hPa (0.3 < QI) with reference to dropsonde wind-profile data observed by DOTSTAR for Typhoons Meari (1105), Roke (1115), Talim (1205), Saola (1209), Jelawat (1217), Kong-rey (1315), Fitow (1323), Matmo (1410), and Fung-wong (1416). Radius denotes the distance between the best-track TC center and the location of AMV data.
Verification results of MTSAT AMVs at pressure levels of 100–300 hPa (0.3 < QI) with reference to dropsonde wind-profile data observed by DOTSTAR for Typhoons Meari (1105), Roke (1115), Talim (1205), Saola (1209), Jelawat (1217), Kong-rey (1315), Fitow (1323), Matmo (1410), and Fung-wong (1416). Radius denotes the distance between the best-track TC center and the location of AMV data.

b. Characteristics of the detected convective bursts

During the intensification phase, TCs were accompanied by 0–3 CBs. CBs were detected in ~70% of the TCs (31 of 44; Table 1). This detection rate is comparable to the CB detection rate of ~80% during the TC intensification phase of TCs occurring during 1999–2001 as obtained from satellite observations (Hennon 2006). The CB detection rate during both the TD period and the TC intensification phase in this study was ~91% (40 of 44 TCs).

A scatterplot of the maximum TC intensification rate versus CB duration (Fig. 5a) shows that rapidly developing TCs tended to have short-duration CBs, reflecting the fact that the TC intensification period generally depends on the TC intensification rate. It is not clear whether the CB duration depends on the order of appearance of the CBs.

Fig. 5.

Scatterplots of the lifetime maximum TC intensification rate (the change of MSW in the previous 24 h; m s−1 day−1) against (a) the duration period and (b) the minimum TB_C200 of CBs during the TC intensification phase, and (c) scatterplots between the CB’s duration period and minimum TB_C200 for the 31 TCs with CBs (Table 1).

Fig. 5.

Scatterplots of the lifetime maximum TC intensification rate (the change of MSW in the previous 24 h; m s−1 day−1) against (a) the duration period and (b) the minimum TB_C200 of CBs during the TC intensification phase, and (c) scatterplots between the CB’s duration period and minimum TB_C200 for the 31 TCs with CBs (Table 1).

The minimum TB_C200 for CBs in relation to the maximum TC intensification rate (Fig. 5b) reveals a remarkable feature: the rapidly developing TCs tended to have low TB_C200. This result indicates that convections associated with CBs of the rapidly developing TCs tended to be deeper than those of the moderately developing TCs. It was also noteworthy that the minimum TB_C200 for CBs appeared to be independent of CB duration period (Fig. 5c).

To investigate the TC structural changes related to the occurrences of CBs, rainfall asymmetries of 14 TCs with multiple CBs were examined by using the azimuthal wavenumber-1 component of the rainfall area within the inner core (Table 4). The wavenumber-1 component of the rainfall distribution for each CB occurrence was obtained from PCT91 data within the 12 h nearest to when TB_C200 for the CB reached the minimum. The result showed that the rainfall asymmetry of subsequent CBs tended to be less than that of the preceding CBs, suggesting that the TC inner core developed gradually with the successive occurrence of CBs. These features of axisymmetrization together with the deepening of convection during TC intensification are consistent with the results of idealized numerical experiments (Miyamoto and Takemi 2013). The wavenumber-0 component, indicating the axisymmetry of the TC inner core, tended to increase during the same period, but this tendency was not as clear as that of the wavenumber-1 component.

Table 4.

Azimuthal wavenumber-1 components (K2) of the convective rainfall area within a radius of 150 km from the TC center detected by SSMIS PCT91 around CB detection times of TCs with multiple CBs during the intensification phase. An em dash means that the analysis was missing because of the shortage of SSMIS data coverage within a radius of 150 km from the TC center.

Azimuthal wavenumber-1 components (K2) of the convective rainfall area within a radius of 150 km from the TC center detected by SSMIS PCT91 around CB detection times of TCs with multiple CBs during the intensification phase. An em dash means that the analysis was missing because of the shortage of SSMIS data coverage within a radius of 150 km from the TC center.
Azimuthal wavenumber-1 components (K2) of the convective rainfall area within a radius of 150 km from the TC center detected by SSMIS PCT91 around CB detection times of TCs with multiple CBs during the intensification phase. An em dash means that the analysis was missing because of the shortage of SSMIS data coverage within a radius of 150 km from the TC center.

c. Relationship between TC inner-core convection and cloud-top outflows

In this section, the relationship between inner-core convection and radial outflow near the cloud top, representing the TC secondary circulation, is examined in the TB_C200 and UMaxOutflow data. First, Typhoon Roke (1115) is examined as an example (Fig. 6). Roke (1115) formed near the Mariana Islands and moved northwestward toward Okinawa while gradually intensifying. After staying around Okinawa for several days, Roke (1115) rapidly developed and reached an MSW of 45 m s−1 and an MSLP of 940 hPa in its mature phase (Figs. 2a and 6a). Two CBs were detected: one during the TD period and the other during the intensification phase (Fig. 6b). The first CB occurred between 1700 UTC 11 September and 1300 UTC 12 September, and the second CB occurred between 0000 UTC 17 September (IR image shown in Fig. 7a) and 0900 UTC 17 September (Fig. 7b). The rapid descent of TB_C200 during both the first and second CBs indicates that inner-core convection was extremely active during both CB periods. In addition, TB_C200 minima appeared around 1800 UTC, implying that the deepening convection was associated with the diurnal cycle (Muramatsu 1983; Takeda and Oyama 2003; Dunion et al. 2014). It is noteworthy that UMaxOutflow during the second CB increased by ~4 m s−1 (Fig. 6b), indicating that the secondary circulation intensified greatly during that period. The upper-tropospheric AMVs and the AMV heights (Fig. 8) clearly showed that during the second CB the heights ascended approximately from 200 to 100 hPa, accompanied by an increase in the radial wind component. Another noticeable feature is the increase in UMaxOutflow on 19 September (Fig. 6b), which was caused by outflow intensification over the northern part of Roke and which implies the existence of an outflow jet formed by an interaction with the midlatitude jet. This outflow jet can be subjectively analyzed by using the JRA-55 200-hPa wind field and the MTSAT WV imagery (figures not shown).

Fig. 6.

Time series of (a) MSW and MSLP of the best-track data and (b) TB_C200 and UMaxOutflow for Typhoon Roke (1115). The horizontal thick dashed line in (b) denotes TB_C200 of 238 K.

Fig. 6.

Time series of (a) MSW and MSLP of the best-track data and (b) TB_C200 and UMaxOutflow for Typhoon Roke (1115). The horizontal thick dashed line in (b) denotes TB_C200 of 238 K.

Fig. 7.

MTSAT IR-band (10.3–11.3 μm) TB images in an area of 14° longitude × 14° latitude centered on Typhoon Roke (1115) at (a) 0000 and (b) 0900 UTC 17 Sep 2011, bracketing the period of the CB that occurred during the TC intensification phase.

Fig. 7.

MTSAT IR-band (10.3–11.3 μm) TB images in an area of 14° longitude × 14° latitude centered on Typhoon Roke (1115) at (a) 0000 and (b) 0900 UTC 17 Sep 2011, bracketing the period of the CB that occurred during the TC intensification phase.

Fig. 8.

Spatial distributions of MTSAT AMVs (vectors) between 100 and 300 hPa and the AMV heights (color scale) in an area of 8° longitude × 8° latitude centered on Typhoon Roke (1115) at times around the beginning and end of the CB occurring during the TC intensification phase: (a) 1800 UTC 16 Sep and (b) 0600 UTC 17 Sep 2011.

Fig. 8.

Spatial distributions of MTSAT AMVs (vectors) between 100 and 300 hPa and the AMV heights (color scale) in an area of 8° longitude × 8° latitude centered on Typhoon Roke (1115) at times around the beginning and end of the CB occurring during the TC intensification phase: (a) 1800 UTC 16 Sep and (b) 0600 UTC 17 Sep 2011.

Next, the statistical relationships between the inner-core convection and outflows around the cloud top of the 44 TCs were investigated by examining lag correlations between TB_C200 and UMaxOutflow time series at lag times of −12, −6, 0, +6, and +12 h during the TC intensification phase (Fig. 9). The magnitude of the average lag correlation between TB_C200 and UMaxOutflow during the TC intensification phase of all of the 44 TCs was largest (correlation coefficient r = −0.56) when the lag time is 0. This result means that the change in UMaxOutflow was generally synchronous with the change in TB_C200 within a time difference of ±3 h.

Fig. 9.

Average lag correlations between UMaxOutflow and TB_C200 for lag times of −12, −6, 0, +6, and +12 h for the 44 TCs (Table 1) during the intensification phase. The error bars denote the standard deviation.

Fig. 9.

Average lag correlations between UMaxOutflow and TB_C200 for lag times of −12, −6, 0, +6, and +12 h for the 44 TCs (Table 1) during the intensification phase. The error bars denote the standard deviation.

Examination of the dependence of the correlation between TB_C200 and UMaxOutflow with a lag time of 0 on R30, VWS, and intensification period (Table 5) showed a slight dependence on R30 (Table 5, top rows) in that the magnitude of the correlation for TCs with small R30 (<250 km) was smaller than that for TCs with large R30 (≥250 km). For example, correlation magnitudes were small in the case of Mawar (1203) and Guchol (1204), which had a small R30 during TC the intensification phase (Table 1). This dependence of the correlation magnitude on R30 implies that atmospheric flows with a large radius of curvature in a large TC make it possible for AMVs to be derived more accurately than do those associated with a small TC. It may also indicate that the influence of VWS on the TC wind structure might differ depending on TC size (Wong and Chan 2004). The magnitude of the average correlation of TCs with averaged VWS < 5 m s−1 during the TC intensification phase was slightly larger than that for TCs with VWS > 5 m s−1 (Table 5, middle rows), suggesting that large VWS can distort a TC’s axisymmetric wind structure. The dependence of the correlation magnitude on VWS appears to be smaller than the dependence on R30. The magnitude of the average correlation for TCs with a long intensification period tended to be larger than that for TCs with a short intensification period (Table 5, bottom rows); this result is an indication that the correlation magnitude is influenced by the sample size. For example, Toraji (1317), which had the smallest correlation magnitude between TB_C200 and UMaxOutflow, had a 30-h intensification period that included just six AMV observation times.

Table 5.

Dependencies of the correlation between TB_C200 and UMaxOutflow (r) on smallest radius of 30-kt winds, averaged vertical wind shear within a radius of 600 km from the TC center, and the intensification period for the 44 TCs in 2011–14 (Table 1). The largest value of the average correlations in each examination is shown in boldface text.

Dependencies of the correlation between TB_C200 and UMaxOutflow (r) on smallest radius of 30-kt winds, averaged vertical wind shear within a radius of 600 km from the TC center, and the intensification period for the 44 TCs in 2011–14 (Table 1). The largest value of the average correlations in each examination is shown in boldface text.
Dependencies of the correlation between TB_C200 and UMaxOutflow (r) on smallest radius of 30-kt winds, averaged vertical wind shear within a radius of 600 km from the TC center, and the intensification period for the 44 TCs in 2011–14 (Table 1). The largest value of the average correlations in each examination is shown in boldface text.

In addition to R30, VWS, and the TC intensification period, it is likely that topography can distort the axisymmetry of a TC’s wind structure, which would also lead to a small correlation magnitude between TB_C200 and UMaxOutflow. This situation was seen in the case of Vicente (1208).

d. Relationship between TC intensification rate and cloud-top outflow

This section examines the relationship between TC intensity (MSW)/intensification rate (the change of MSW in the previous 24 h) and UMaxOutflow, which reflects the TC secondary circulation. UMaxOutflow values of strong TCs with large lifetime maximum MSW tended to be larger than those of weaker TCs (Fig. 10a). The correlation coefficient between the lifetime maximum MSW and the maximum UMaxOutflow during the TC intensification phase was 0.52, and r between the lifetime maximum MSW and the average UMaxOutflow was 0.55. The magnitude of the correlation between the lifetime maximum MSW and the average cloud-top height (CTH) within a radius of 200 km from the TC center during the TC intensification phase was also relatively large (Fig. 10b); that is, r was −0.64 for the maximum CTH and −0.46 for the mean CTH. The relatively high correlation between the lifetime maximum MSW and CTH suggests that strong TCs tended to have deeper convection in the inner core than weak TCs. This interpretation is supported by the results for the relationship between lifetime maximum MSW and UMaxOutflow (Fig. 10a), and it is also consistent with the findings of previous studies (Steranka et al. 1986; Schubert et al. 1999). The scatterplot of maximum TC intensification rate versus UMaxOutflow (Fig. 10c) shows features similar to that between lifetime maximum MSW and UMaxOutflow (Fig. 10a), and the scatterplot of TC intensification rate versus CTH (Fig. 10d) also has features similar to that of lifetime maximum MSW versus CTH (Fig. 10b). These similarities (cf. Figs. 10a and 10c and cf. Figs. 10b and 10d) indicate that strong TCs generally experience rapid intensification (Lee et al. 2016).

Fig. 10.

Scatterplots of (a) lifetime maximum MSW (i.e., during the mature phase) vs UMaxOutflow, (b) lifetime maximum MSW vs CTH [averaged AMV height (hPa) within a radius of 200 km from the TC center], (c) the maximum TC intensification rate vs UMaxOutflow, and (d) the maximum TC intensification rate vs CTH during the TC intensification phase for the 44 TCs in Table 1.

Fig. 10.

Scatterplots of (a) lifetime maximum MSW (i.e., during the mature phase) vs UMaxOutflow, (b) lifetime maximum MSW vs CTH [averaged AMV height (hPa) within a radius of 200 km from the TC center], (c) the maximum TC intensification rate vs UMaxOutflow, and (d) the maximum TC intensification rate vs CTH during the TC intensification phase for the 44 TCs in Table 1.

To investigate the importance of CBs in the development of the secondary circulation, the relationship between the maximum TC intensification rate and UMaxOutflow was examined for TCs with and without CBs during the TC intensification phase. These results showed remarkable differences between TCs with CBs and those without CBs (Fig. 11); namely, the correlation between the TC maximum intensification rate and the maximum or average UMaxOutflow for TCs with CBs was much larger than that for TCs without CBs. This difference between the two groups implies that CBs may enhance TC secondary circulation and create favorable conditions for TC intensification.

Fig. 11.

Scatterplots of the maximum TC intensification rate vs UMaxOutflow during the TC intensification phase for (a) 31 TCs with CBs and (b) 13 TCs without CBs.

Fig. 11.

Scatterplots of the maximum TC intensification rate vs UMaxOutflow during the TC intensification phase for (a) 31 TCs with CBs and (b) 13 TCs without CBs.

To investigate large TC intensification rates associated with large outflow near the cloud top, average latent-heat profiles within a radius of 200 km from the TC center were examined with reference to the TC intensification rate for the intensification phase of the 44 TCs (Fig. 12). The result shows that TCs with rapid intensification tended to generate a large amount of latent heat within the inner core, especially around the middle troposphere, an indication that midlevel diabatic heating released by inner-core convection likely plays an important role in the intensification of the TC secondary circulation (Shapiro and Willoughby 1982).

Fig. 12.

Vertical profiles of the latent-heating rate within a radius of 200 km from the TC center for various categories of TC intensification rate (the change of MSW in the previous 24 h; m s−1 day−1). Blue, green, orange, and red curves denote the average latent-heating profiles of TCs with maximum intensification rates of ≥15, 10–15, 5–10, and 0–5 m s−1 day−1, respectively.

Fig. 12.

Vertical profiles of the latent-heating rate within a radius of 200 km from the TC center for various categories of TC intensification rate (the change of MSW in the previous 24 h; m s−1 day−1). Blue, green, orange, and red curves denote the average latent-heating profiles of TCs with maximum intensification rates of ≥15, 10–15, 5–10, and 0–5 m s−1 day−1, respectively.

e. Lag time between TC intensification and increases in cloud-top outflow

TCs are accompanied by CBs prior to TC rapid intensification (Steranka et al. 1986; Ebert and Holland 1992; Guimond et al. 2010; Rogers 2010; Hazelton et al. 2017). For the lag time between the occurrence of CBs and the occurrence of TC intensification, on the basis of satellite IR observations, Steranka et al. (1986) showed that the onset of CBs tended to occur about 24 h before the increase of MSW in the previous 24 h of greater than 5 m s−1 was observed for TCs occurring in the Atlantic Ocean. Guimond et al. (2010) revealed that the onset of CBs in Hurricane Dennis (2005) occurred about 12 h before the TC experienced rapid intensification, which was associated with the structural change of the warm core from asymmetric to axisymmetric. These previous studies suggest that it is meaningful to estimate the average time difference between the observation time of TC intensification rate and the cloud-top outflow peak that corresponds to the occurrence of CB for the elucidation of the TC intensification process.

Table 6 shows the frequencies of TCs with regard to the maximum TC intensification rate (the change of MSW in the previous 24 h) and the time difference T1 between the occurrence of maximum UMaxOutflow and the typhoon formation time during the TC intensification phase for the 44 TCs. It is noteworthy that the time difference T1 tended to be shorter for rapidly developing TCs with a maximum intensification rate of ≥15 m s−1 day−1. The value separating short and long time differences, 15 m s−1 day−1, is almost same as the threshold used by Kaplan and DeMaria (2003) to discriminate TC rapid intensification, that is, a 1-min average MSW of 30 kt (15.4 m s−1) day−1.

Table 6.

Frequencies of TCs with regard to the maximum TC intensification rate [IntRate: the change of MSW in the previous 24 h (m s–1 day–1); rows] and the time difference T1 between the occurrence time of maximum UMaxOutflow and the typhoon formation time (columns) during the TC intensification phase for the 44 TCs in Table 1.

Frequencies of TCs with regard to the maximum TC intensification rate [IntRate: the change of MSW in the previous 24 h (m s–1 day–1); rows] and the time difference T1 between the occurrence time of maximum UMaxOutflow and the typhoon formation time (columns) during the TC intensification phase for the 44 TCs in Table 1.
Frequencies of TCs with regard to the maximum TC intensification rate [IntRate: the change of MSW in the previous 24 h (m s–1 day–1); rows] and the time difference T1 between the occurrence time of maximum UMaxOutflow and the typhoon formation time (columns) during the TC intensification phase for the 44 TCs in Table 1.

With respect to the time difference T2 between the occurrence of maximum UMaxOutflow and the observation time of maximum TC intensification rate (Table 7), the observation time of maximum TC intensification rate tended to follow the occurrence of the maximum outflow for TCs with a maximum intensification rate of >15 m s−1 day−1, implying that strong updrafts within the TC inner core, indicated by a strong secondary circulation, are important for TC rapid intensification. In addition, the occurrence of maximum UMaxOutflow tended to precede the observation time of maximum TC intensification rate by 0–36 h in 66% of the TCs (29 of 44 TCs; Table 7). It is noteworthy that the lag time of 0–36 h is comparable to the lag time between the occurrence of CBs and a future MSW increase reported by Steranka et al. (1986; ~24 h), and Hennon (2006; 18–24 h).

Table 7.

Frequencies of TCs with regard to the maximum TC IntRate and the time difference T2 between the occurrence of maximum TC intensification rate and the occurrence of maximum UMaxOutflow during the TC intensification phase for the 44 TCs in Table 1.

Frequencies of TCs with regard to the maximum TC IntRate and the time difference T2 between the occurrence of maximum TC intensification rate and the occurrence of maximum UMaxOutflow during the TC intensification phase for the 44 TCs in Table 1.
Frequencies of TCs with regard to the maximum TC IntRate and the time difference T2 between the occurrence of maximum TC intensification rate and the occurrence of maximum UMaxOutflow during the TC intensification phase for the 44 TCs in Table 1.

The same relationships were examined in the 31 TCs with CBs (CB group) and the 13 TCs without CBs (non-CB group) (Table 8). Noticeable is that the time difference T2 for the non-CB group tended to be shorter than that for the CB group, a suggestion being that the non-CB group tended to develop more quickly after the occurrence of the maximum outflow than the CB group did. Since a TC with small size tends to experience rapid intensification (Chen et al. 2011; Carrasco et al. 2014), the difference of TC size between the CB and non-CB groups was examined using R30 of the best-track data (Table 9). One statistical difference between the CB and non-CB groups is in the mean R30 of TCs: the CB group had 289.3 km of mean R30, and the non-CB group had 269.2 km of mean R30; the other is the percentage of TCs with R30 < 350 km: the values for the CB and non-CB groups are 71% and 85%, respectively. These results imply a possibility that the difference of TC size between the CB and non-CB groups caused the difference of the TC intensification rate between the CB and non-CB groups.

Table 8.

As in Table 7, but for TCs with CBs (CB group) and without CBs (non-CB group) during the TC intensification phase.

As in Table 7, but for TCs with CBs (CB group) and without CBs (non-CB group) during the TC intensification phase.
As in Table 7, but for TCs with CBs (CB group) and without CBs (non-CB group) during the TC intensification phase.
Table 9.

Comparison of averaged best-track R30 during the TC intensification phase between 31 TCs with CBs (CB group) and 13 TCs without CBs (non-CB group) in 2011–14 (Table 1).

Comparison of averaged best-track R30 during the TC intensification phase between 31 TCs with CBs (CB group) and 13 TCs without CBs (non-CB group) in 2011–14 (Table 1).
Comparison of averaged best-track R30 during the TC intensification phase between 31 TCs with CBs (CB group) and 13 TCs without CBs (non-CB group) in 2011–14 (Table 1).

f. Potential of TC outflow for diagnosing the TC intensification rate

In this section, the possibility that TC outflows near the cloud top, derived by using AMVs, can be used to diagnose TC intensity changes is examined. Figure 13 shows the correlations between the 6-hourly UMaxOutflows (m s−1) during the TC intensification phase and the TC intensification rates (the change of MSW in the previous 24 h; m s−1 day−1) at lag times (the observation time of TC intensification rate − the observation time of UMaxOutflow) for 44 TCs in 2011–14. UMaxOutflow was positively correlated with the TC intensification rates at times between −18 and +18 h from the UMaxOutflow observation time, a suggestion being that the secondary circulation represented by the UMaxOutflow made a key role in TC intensification. The correlation values at lag times between +6 and +18 h were relatively small partly because parts of the values were computed using the TC intensification rates during the mature or decay phase. The relatively small number of observations at lag times between −18 and −6 h indicates that MSW data of the best-track data were available for the times at and after the typhoon formation time. It is noteworthy that the correlation was largest at a lag time of −6 h. This result suggests that a value of UMaxOutflow during TC intensification phase could be a predictor to diagnose TC intensification around the time at which the UMaxOutflow is observed, but continuing study using more TC cases is necessary to obtain more robust results because the sample size of TCs in this study was relatively small.

Fig. 13.

Correlations (solid line with filled circles) between UMaxOutflows observed at all of the 6-hourly times during the TC intensification phase and TC intensification rates (IntRate: the change of MSW in the previous 24 h; m s−1 day−1) with lag times (time of TC intensification rate − time of UMaxOutflow). Also shown are the numbers of observational data used for computing the correlations (solid line with times signs).

Fig. 13.

Correlations (solid line with filled circles) between UMaxOutflows observed at all of the 6-hourly times during the TC intensification phase and TC intensification rates (IntRate: the change of MSW in the previous 24 h; m s−1 day−1) with lag times (time of TC intensification rate − time of UMaxOutflow). Also shown are the numbers of observational data used for computing the correlations (solid line with times signs).

4. Summary and discussion

To verify the role of TC secondary circulation in TC intensification by using observational data, this study examined the relationship between TC intensification and cloud-top outflows, which reflect the secondary circulation, by using 6-hourly upper-tropospheric AMVs derived from successive MTSAT imagery for 44 TCs occurring in the western North Pacific basin during 2011–14. The time variation of the cloud-top outflows, defined as the maximum radial wind (UMaxOutflow) detected by using the upper-tropospheric AMVs, during the TC intensification phase was well correlated with the time variation of cloud-top temperature of convective clouds within the TC inner core, defined as the average IR TB within a radius of 200 km from the TC center (TB_C200). The magnitude of the correlation between UMaxOutflow and TB_C200 was largest when the lag time was zero, indicating that deep convection in the TC inner core is essential for the intensification of the TC secondary circulation.

With respect to the relationship between the TC intensification rate (the change of MSW in the previous 24 h) and UMaxOutflow, the magnitude of the correlation between the maximum TC intensification rate and UMaxOutflow for TCs with CBs was larger than that for TCs without CBs, implying that rapid deepening of convection associated with CBs may facilitate the intensification of the secondary circulation via latent-heat release, mainly around middle levels.

The lag time T2 of the maximum TC intensification rate relative to the occurrence of the maximum UMaxOutflow was 0–36 h in 66% of the TCs, which is comparable to the lag time between the occurrence of CBs and a future TC intensification obtained by previous studies. In almost all TCs with a TC intensification rate of greater than 15 m s−1, the time of maximum TC intensification tended to follow the occurrence of maximum UMaxOutflow. Further, it is noteworthy that the time difference T2 for TCs with CBs tended to be longer than that for TCs without CBs, which may be partly explained by the difference of TC size.

The positive correlations between UMaxOutflow at all of the 6-hourly times during the TC intensification phase and the TC intensification rates at times between −18 and +18 h from the occurrence time of UMaxOutflow implied that UMaxOutflow can be a predictor to diagnose TC intensification in, for example, statistical dynamic models such as the Statistical Hurricane Intensity Prediction Scheme (DeMaria and Kaplan 1994) and the rapid intensity index (Kaplan and DeMaria 2003; Kaplan et al. 2010).

This study obtained several noticeable results about the relationship between TC intensification and cloud-top outflow. To obtain more robust results, it is necessary to continue the study using more TC cases and other data sources of observations, for example, the Joint Typhoon Warning Center TC best-track data. For the TC process, it is necessary to investigate the influence of environmental atmospheric flows, that is, outflow jets, on the outcomes obtained by this study. In this context, it will be important to take account of a TC’s inertial stability (Schubert and Hack 1982), which can indicate the robustness of the TC vortex. Inertial stability computed from upper-tropospheric AMVs could contribute to such an analysis (Oyama et al. 2016). Addressing the challenge of deriving and verifying the TC environmental VWS using a high-resolution numerical model is also necessary to evaluate the influence of VWS on TC intensity and structure more accurately. Furthermore, the verification of AMV height (Velden and Bedka 2009) is another important task to obtain more robust results on the cloud-top outflow from AMVs.

For future work regarding technological advancement in satellite observations, Himawari-8, which took over for MTSAT from 7 July 2015, has the Advanced Himawari Imager (Bessho et al. 2016), which can observe clouds with higher temporal (2.5–10 min) and spatial (2 km for IR and WV bands) resolution than the MTSAT Japanese Advanced Meteorological Imager does. In addition, the number of observation bands increased from 5 to 16 in the switch of the satellite. These significant upgrades from MTSAT to Himawari-8 will make it possible to derive AMV data with greater temporal and spatial density and higher quality (Shimoji 2014). Following Himawari-8, GOES-R was launched on 19 November 2016 and started its observations to cover the Pacific and Atlantic Ocean basins with the Advanced Baseline Instrument, which has 16 observation bands and a function of rapid-scan observations (Schmit et al. 2017). The GOES-R observations will upgrade the AMV products and contribute not only to TC studies but also to operational TC analysis and forecasting (Daniels et al. 2016). These improvements of AMVs by introducing the next-generation geostationary satellites such as Himawari-8/9 and GOES-R will lead to further insights into TC intensification processes that are related to TC secondary circulation. Moreover, analyses that are based on temporally dense AMV data will also provide information on the processes of a TC’s diurnal cycle and phenomena that occur on a short time scale.

Because this study was conducted under limitations on the sample size of TCs and the TC basin, additional future studies targeting more TCs and other regions from the western North Pacific basin are important to obtain robust results on the relationship between TC intensification and cloud-top outflow. The author hopes that this study will contribute to the TC science community as a start to observational verification studies on the relationship between TC intensification and TC secondary circulation.

Acknowledgments

This study used MTSAT imagery and ancillary data from the Japan Meteorological Agency Meteorological Satellite Center for computing AMVs. The author is grateful to Kazuki Shimoji and Masahiro Hayashi of JMA MSC for providing these data. The author also extends the gratitude to staff members at JMA Headquarters for providing the SSMIS data. The TRMM PR latent-heating-rate data (L2H25, version 7) were obtained online from the JAXA G-Portal (https://www.gportal.jaxa.jp/gp/top.html). DOTSTAR data to verify the AMV data were obtained from the website of the National Taiwan University (http://typhoon.as.ntu.edu.tw/DOTSTAR/en/). In the preparation of the manuscript, Kozo Okamoto and Akiyoshi Wada of the Meteorological Research Institute of JMA gave valuable comments and suggestions. The author also expresses gratitude to the reviewers for their many valuable comments and suggestions that helped to improve this paper.

REFERENCES

REFERENCES
Apke
,
J. M.
,
J. R.
Mecikalski
, and
C. P.
Jewett
,
2016
:
Analysis of mesoscale atmospheric flows above mature deep convection using super rapid scan geostationary satellite data
.
J. Appl. Meteor. Climatol.
,
55
,
1859
1887
, doi:.
Berger
,
H.
,
R.
Langland
,
C. S.
Velden
,
C. A.
Reynolds
, and
P. M.
Pauley
,
2011
:
Impact of enhanced satellite-derived atmospheric motion vector observations on numerical tropical cyclone track forecasts in the western North Pacific during TPARC/TCS-08
.
J. Appl. Meteor. Climatol
,
50
,
2309
2318
, doi:.
Bessho
,
K.
, and Coauthors
,
2016
:
An introduction to Himawari-8/9—Japan’s new-generation geostationary meteorological satellites
.
J. Meteor. Soc. Japan
,
94
,
151
183
, doi:.
Bryan
,
G. H.
, and
R.
Rotunno
,
2009
:
Evaluation of an analytical model for the maximum intensity of tropical cyclones
.
J. Atmos. Sci.
,
66
,
3042
3060
, doi:.
Carrasco
,
C. A.
,
C. W.
Landsea
, and
Y.-L.
Lin
,
2014
:
The influence of tropical cyclone size on its intensification
.
Wea. Forecasting
,
29
,
582
590
, doi:.
Chen
,
D. Y.-C.
,
K. K. W.
Cheung
, and
C.-S.
Lee
,
2011
:
Some implications of core regime wind structures in western North Pacific tropical cyclones
.
Wea. Forecasting
,
26
,
61
75
, doi:.
Croxford
,
M.
, and
G. M.
Barnes
,
2002
:
Inner core strength of Atlantic tropical cyclones
.
Mon. Wea. Rev.
,
130
,
127
139
, doi:.
Daniels
,
J.
,
W.
Bresky
,
A.
Bailey
,
A.
Allegrino
,
S.
Wanzong
, and
C.
Velden
,
2016
: Use of GOES-R Advanced Baseline Imager (ABI) proxy data to assess the performance of the GOES-R winds algorithm. Proc. 13th Int. Winds Workshop, Monterey, CA, EUMETSAT, http://cimss.ssec.wisc.edu/iwwg/iww13/proceedings_iww13/papers/session2/IWW13_Session2_3_Daniels_final.pdf.
DeMaria
,
M.
, and
J.
Kaplan
,
1994
:
A Statistical Hurricane Intensity Prediction Scheme (SHIPS) for the Atlantic basin
.
Wea. Forecasting
,
9
,
209
220
, doi:.
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
, doi:.
Ebert
,
E. E.
, and
G. J.
Holland
,
1992
:
Observations of record cold cloud-top temperatures in Tropical Cyclone Hilda (1990)
.
Mon. Wea. Rev.
,
120
,
2240
2251
, doi:.
Elsberry
,
R. L.
,
L.
Chen
,
J.
Davidson
,
R.
Rogers
,
Y.
Wang
, and
L.
Wu
,
2013
:
Advances in understanding and forecasting rapidly changing phenomena in tropical cyclones
.
Trop. Cyclone Res. Rev.
,
2
,
13
24
.
Emanuel
,
K. A.
,
1986
:
An air–sea interaction theory for tropical cyclones. Part I: Steady-state maintenance
.
J. Atmos. Sci.
,
43
,
585
604
, doi:.
EUMETSAT
,
2015
: Divergence product: Product guide. EUMETSAT Tech. Doc. EUM/TSS/MAN/15/795706, version 1C, 10 pp., http://www.eumetsat.int/website/wcm/idc/idcplg?IdcService=GET_FILE&dDocName=PDF_DIV_FACTSHEET&RevisionSelectionMethod=LatestReleased&Rendition=Web.
Fudeyasu
,
H.
, and
Y.
Wang
,
2011
:
Balanced contribution to the intensification of a tropical cyclone simulated in TCM4: Outer-core spinup process
.
J. Atmos. Sci.
,
68
,
430
449
, doi:.
Gallina
,
G. M.
, and
C. S.
Velden
,
2002
: Environmental vertical wind shear and tropical cyclone intensity change utilizing enhanced satellite derived wind information. 25th Conf. on Hurricanes and Tropical Meteorology, San Diego, CA, Amer. Meteor. Soc., 3C.5, https://ams.confex.com/ams/pdfpapers/35650.pdf.
Guimond
,
S. R.
,
G. M.
Heymsfield
, and
F. J.
Turk
,
2010
:
Multiscale observations of Hurricane Dennis (2005): The effects of hot towers on rapid intensification
.
J. Atmos. Sci.
,
67
,
633
654
, doi:.
Hawkins
,
J. D.
,
F. J.
Turk
,
T. F.
Lee
, and
K.
Richardson
,
2008
:
Observations of tropical cyclones with the SSMIS
.
IEEE Trans. Geosci. Remote Sens.
,
46
,
901
912
, doi:.
Hazelton
,
A. T.
,
R. F.
Rogers
, and
R. E.
Hart
,
2017
:
Analyzing simulated convective bursts in two Atlantic hurricanes. Part I: Burst formation and development
.
Mon. Wea. Rev.
,
145
,
3073
3094
, doi:.
Hennon
,
P. A.
,
2006
: The role of the ocean in convective burst initiation: Implications for tropical cyclone intensification. Ph.D. thesis, The Ohio State University, 185 pp., http://rave.ohiolink.edu/etdc/view?acc_num=osu1135914268.
Holland
,
G. J.
, and
R. T.
Merrill
,
1984
:
On the dynamics of tropical cyclone structural changes
.
Quart. J. Roy. Meteor. Soc.
,
110
,
723
745
, doi:.
Holmlund
,
K.
,
1998
:
The utilization of statistical properties of satellite-derived atmospheric motion vectors to derive quality indicators
.
Wea. Forecasting
,
13
,
1093
1104
, doi:.
Houze
,
R. A.
, Jr.
,
2010
:
Clouds in tropical cyclones
.
Mon. Wea. Rev.
,
138
,
293
344
, doi:.
Kaplan
,
J.
, and
M.
DeMaria
,
2003
:
Large-scale characteristics of rapidly intensifying tropical cyclones in the North Atlantic basin
.
Wea. Forecasting
,
18
,
1093
1108
, doi:.
Kaplan
,
J.
,
M.
DeMaria
, and
J. A.
Knaff
,
2010
:
A revised tropical cyclone rapid intensification index for the Atlantic and eastern North Pacific basins
.
Wea. Forecasting
,
25
,
220
241
, doi:.
Kobayashi
,
S.
, and Coauthors
,
2015
:
The JRA-55 Reanalysis: General specifications and basic characteristics
.
J. Meteor. Soc. Japan
,
93
,
5
48
, doi:.
Langland
,
R. H.
,
C.
Velden
,
P. M.
Pauley
, and
H.
Berger
,
2009
:
Impact of satellite-derived rapid-scan wind observations on numerical model forecasts of Hurricane Katrina
.
Mon. Wea. Rev.
,
137
,
1615
1622
, doi:.
Lee
,
C.-Y.
,
M. K.
Tippett
,
A. H.
Sobel
, and
S. J.
Camargo
,
2016
:
Rapid intensification and the bimodal distribution of tropical cyclone intensity
.
Nat. Commun.
,
7
,
10625
, doi:.
Li
,
Q.-Q.
, and
Y.
Wang
,
2012
:
A comparison of inner and outer spiral rainbands in a numerically simulated tropical cyclone
.
Mon. Wea. Rev.
,
140
,
2782
2805
, doi:.
Line
,
W. E.
,
T. J.
Schmit
, and
S. J.
Goodman
,
2016
:
Use of geostationary super rapid scan satellite imagery by the Storm Prediction Center
.
Wea. Forecasting
,
31
,
483
494
, doi:.
Miyamoto
,
Y.
, and
T.
Takemi
,
2013
:
A transition mechanism for the spontaneous axisymmetric intensification of tropical cyclones
.
J. Atmos. Sci.
,
70
,
112
129
, doi:.
Molinari
,
J.
, and
D.
Vollaro
,
1989
:
External influences on hurricane intensity. Part I: Outflow layer eddy angular momentum fluxes
.
J. Atmos. Sci.
,
46
,
1093
1105
, doi:.
Molinari
,
J.
,
J.
Frank
, and
D.
Vollaro
,
2013
:
Convective bursts, downdraft cooling, and boundary layer recovery in a sheared tropical storm
.
Mon. Wea. Rev.
,
141
,
1048
1060
, doi:.
Muramatsu
,
T.
,
1983
:
Diurnal variations of satellite-measured TBB areal distribution and eye diameter of mature typhoons
.
J. Meteor. Soc. Japan
,
61
,
77
90
, doi:.
Nieman
,
S. J.
,
J.
Schmetz
, and
W. P.
Menzel
,
1993
:
A comparison of several techniques to assign heights to cloud tracers
.
J. Appl. Meteor.
,
32
,
1559
1568
, doi:.
Ohno
,
T.
, and
M.
Satoh
,
2015
:
On the warm core of a tropical cyclone formed near the tropopause
.
J. Atmos. Sci.
,
72
,
551
571
, doi:.
Oyama
,
R.
,
2010
: Upgrade of Atmospheric Motion Vector derivation algorithms at JMA/MSC. Meteorological Satellite Center Tech. Note 54, 32 pp., http://www.data.jma.go.jp/mscweb/technotes/msctechrep54-1.pdf.
Oyama
,
R.
,
2015
:
Characteristics of upper-tropospheric atmospheric motion vectors (AMV) in tropical cyclone areas derived using MTSAT rapid-scan observation data (in Japanese with English abstract and figure and table captions)
.
Tenki
,
62
,
881
894
, http://www.metsoc.jp/tenki/pdf/2015/2015_10_0015.pdf.
Oyama
,
R.
,
A.
Wada
, and
M.
Sawada
,
2016
:
Intensification of Typhoon Danas (1324) captured by MTSAT upper tropospheric atmospheric motion vectors
.
SOLA
,
12
,
135
139
, doi:.
Paterson
,
L. A.
,
B. N.
Hanstrum
,
N. E.
Davidson
, and
H. C.
Weber
,
2005
:
Influence of environmental vertical wind shear on the intensity of hurricane-strength tropical cyclones in the Australian region
.
Mon. Wea. Rev.
,
133
,
3644
3660
, doi:.
Rappin
,
E. D.
,
M. C.
Morgan
, and
G. J.
Tripoli
,
2011
:
The impact of outflow environment on tropical cyclone intensification and structure
.
J. Atmos. Sci.
,
68
,
177
194
, doi:.
Reasor
,
P. D.
,
R.
Rogers
, and
S.
Lorsolo
,
2013
:
Environmental flow impacts on tropical cyclone structure diagnosed from airborne Doppler radar composites
.
Mon. Wea. Rev.
,
141
,
2949
2969
, doi:.
Riehl
,
H.
, and
J.
Malkus
,
1961
:
Some aspects of Hurricane Daisy, 1958
.
Tellus
,
13
,
181
213
, doi:.
Rogers
,
R.
,
2010
:
Convective-scale structure and evolution during a high-resolution simulation of tropical cyclone rapid intensification
.
J. Atmos. Sci.
,
67
,
44
70
, doi:.
Sawada
,
M.
, and
T.
Iwasaki
,
2007
:
Impacts of ice phase processes on tropical cyclone development
.
J. Meteor. Soc. Japan
,
85
,
479
494
, doi:.
Schmit
,
T. J.
,
P.
Griffith
,
M. M.
Gunshor
,
J. M.
Daniels
,
S. J.
Goodman
, and
W. J.
Lebair
,
2017
:
A closer look at the ABI on the GOES-R series
.
Bull. Amer. Meteor. Soc.
,
98
,
681
698
, doi:.
Schubert
,
W. H.
, and
J. J.
Hack
,
1982
:
Inertial stability and tropical cyclone development
.
J. Atmos. Sci.
,
39
,
1687
1697
, doi:.
Schubert
,
W. H.
,
M. T.
Montgomery
,
R. K.
Taft
,
T. A.
Guinn
,
S. R.
Fulton
,
J. P.
Kossin
, and
J. P.
Edwards
,
1999
:
Polygonal eyewalls, asymmetric eye contraction, and potential vorticity mixing in hurricanes
.
J. Atmos. Sci.
,
56
,
1197
1223
, doi:.
Shapiro
,
L. J.
, and
H. E.
Willoughby
,
1982
:
The response of balanced hurricanes to local sources of heat and momentum
.
J. Atmos. Sci.
,
39
,
378
394
, doi:.
Shige
,
S.
,
Y. N.
Takayabu
,
W.-K.
Tao
, and
D. E.
Johnson
,
2004
:
Spectral retrieval of latent heating profiles from TRMM PR data. Part I: Development of a model-based algorithm
.
J. Appl. Meteor. Climatol.
,
43
,
1095
1113
, doi:.
Shige
,
S.
,
Y. N.
Takayabu
,
W.-K.
Tao
, and
C.-L.
Shie
,
2007
:
Spectral retrieval of latent heating profiles from TRMM PR data. Part II: Algorithm improvement and heating estimates over tropical ocean regions
.
J. Appl. Meteor. Climatol.
,
46
,
1098
1124
, doi:.
Shimoji
,
K.
,
2014
: Motion tracking and cloud height assignment methods for Himawari-8 AMV. Proc. 12th Int. Winds Workshop, Copenhagen, Denmark, EUMETSAT, https://www.eumetsat.int/website/wcm/idc/idcplg?IdcService=GET_FILE&dDocName=PDF_CONF_P61_S2_06_SHIMOJI_V&RevisionSelectionMethod=LatestReleased&Rendition=Web.
Smith
,
R. K.
,
M. T.
Montgomery
, and
V. S.
Nguyen
,
2009
:
Tropical cyclone spin-up revisited
.
Quart. J. Roy. Meteor. Soc.
,
135
,
1321
1335
, doi:.
Spencer
,
R. W.
,
H. M.
Goodman
, and
R. E.
Hood
,
1989
:
Precipitation retrieval over land and ocean with the SSM/I: Identification and characteristics of the scattering signal
.
J. Atmos. Oceanic Technol.
,
6
,
254
273
, doi:.
Steranka
,
J.
,
E. B.
Rodgers
, and
R. C.
Gentry
,
1986
:
The relationship between satellite measured convective bursts and tropical cyclone intensification
.
Mon. Wea. Rev.
,
114
,
1539
1546
, doi:.
Takeda
,
T.
, and
R.
Oyama
,
2003
:
Periodic time variation of low-TBB cloud area in typhoon
.
J. Meteor. Soc. Japan
,
81
,
1497
1503
, doi:.
Ueno
,
M.
,
2007
:
Observational analysis and numerical evaluation of the effects of vertical wind shear on the rainfall asymmetry in the typhoon inner-core region
.
J. Meteor. Soc. Japan
,
85
,
115
136
, doi:.
Velden
,
C. S.
, and
K. M.
Bedka
,
2009
:
Identifying the uncertainty in determining satellite-derived atmospheric motion vector height attribution
.
J. Appl. Meteor. Climatol.
,
48
,
450
463
, doi:.
Velden
,
C. S.
, and Coauthors
,
2005
:
Recent innovations in deriving tropospheric winds from meteorological satellites
.
Bull. Amer. Meteor. Soc.
,
86
,
205
223
, doi:.
Wang
,
H.
, and
Y.
Wang
,
2014
:
A numerical study of Typhoon Megi (2010). Part I: Rapid intensification
.
Mon. Wea. Rev.
,
142
,
29
48
, doi:.
Wang
,
Y.
,
2009
:
How do outer spiral rainbands affect tropical cyclone structure and intensity?
J. Atmos. Sci.
,
66
,
1250
1273
, doi:.
Warrick
,
F.
,
2016
: NWP SAF AMV monitoring: The 7th Analysis Report (AR7). NWP SAF Doc. NWPSAF-MO-TR-032, 35 pp., http://nwpsaf.eu/monitoring/amv/nwpsaf_mo_tr_032.pdf.
Wong
,
M. L.
, and
J. C. L.
Chan
,
2004
:
Tropical cyclone intensity in vertical wind shear
.
J. Atmos. Sci.
,
61
,
1859
1876
, doi:.
Wu
,
C.-C.
, and Coauthors
,
2005
:
Dropwindsonde Observations for Typhoon Surveillance near the Taiwan Region (DOTSTAR): An overview
.
Bull. Amer. Meteor. Soc.
,
86
,
787
790
, doi:.
Wu
,
T.-C.
,
H.
Liu
,
S. J.
Majumdar
,
C. S.
Velden
, and
J. L.
Anderson
,
2014
:
Influence of assimilating satellite-derived atmospheric motion vector observations on numerical analyses and forecasts of tropical cyclone track and intensity
.
Mon. Wea. Rev.
,
142
,
49
71
, doi:.
Yamashita
,
K.
,
2012
: An observing system experiment of MTSAT rapid scan AMV using JMA meso-scale operational NWP system. Proc. 11th Int. Winds Workshop, Auckland, New Zealand, EUMETSAT, https://www.eumetsat.int/website/wcm/idc/idcplg?IdcService=GET_FILE&dDocName=PDF_CONF_P60_S4_15_YAMASHIT_V&RevisionSelectionMethod=LatestReleased&Rendition=Web.
Zehr
,
R. M.
,
1992
: Tropical cyclogenesis in the western North Pacific. NOAA Tech. Rep. NESDIS 61, 181 pp., https://docs.lib.noaa.gov/noaa_documents/NESDIS/TR_NESDIS/TR_NESDIS_61.pdf.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).