Abstract

Climatological characteristics of simulated intense tropical cyclones (TCs) in the western North Pacific were explored with a 20-km-mesh atmospheric general circulation model (AGCM20) and a 5-km-mesh regional atmospheric nonhydrostatic model (ANHM5). From the AGCM20 climate runs, 34 intense TCs with a minimum central pressure (MCP) less than or equal to 900 hPa were sampled. Downscaling experiments were conducted with the ANHM5 for each intense TC simulated by the AGCM20. Only 23 developed into TCs with MCP ≤ 900 hPa. Most of the best-track TCs with an MCP ≤ 900 hPa underwent rapid intensification (RI) and attained maximum intensities south of 25°N. The AGCM20 simulated a similar number of intense TCs as the best-track datasets. However, the intense AGCM20 TCs tended to intensify longer and more gradually; only half of them underwent RI. The prolonged gradual intensification resulted in significant northward shifts of the location of maximum intensity compared with the location derived from two best-track datasets. The inner-core structure of AGCM20 TCs exhibited weak and shallow eyewall updrafts with maxima below an altitude of 6 km, while downscaling experiments revealed that most of the intense ANHM5 TCs underwent RI with deep and intense eyewall updrafts and attained their maximum intensity at lower latitudes. The altitudes of updraft maxima simulated by the AGCM20 descended rapidly during the phase of greatest intensification as midlevel warming markedly developed. The change in major processes responsible for precipitation in AGCM20 TCs before and after maximum intensification suggests close relationships between the large-scale cloud scheme and midlevel warming and prolonged gradual intensification.

1. Introduction

Tropical cyclones (TCs) are among the most extreme weather events on Earth. Intense TCs, such as the extremely destructive Supertyphoon Haiyan (Lin et al. 2013), have had significant negative impacts on coastal countries in the Asia–Pacific region (e.g., Peduzzi et al. 2012; Lander et al. 2014). Recent studies of long-term changes of TCs over East Asia have indicated that TC activity in the region has been enhanced by relatively warm sea surface temperature (SST) and weak vertical wind shear (VWS) (Li et al. 2010; Park et al. 2011). Kossin et al. (2014) found a statistically significant poleward migration of the location of TC maximum intensities in both the Northern and Southern Hemispheres and suggested that this migration was linked to marked changes in potential intensity closely related to SST (Emanuel 1986; Holland 1997) and to changes in VWS.

A number of studies have attempted to project future climatological characteristics of TCs by using state-of-the-art general circulation models (GCMs) (e.g., Stowasser et al. 2007; Murakami et al. 2012). However, the representation of TC structure and evolution is highly sensitive to model resolution; in the case of intense TCs in particular (e.g., Bengtsson et al. 2007; Walsh et al. 2007; Roberts et al. 2015), even a GCM with a horizontal resolution of approximately 0.50° underestimates TC intensity (Murakami and Sugi 2010). In addition, although Murakami et al. (2012) have found that a 20-km-mesh atmospheric GCM (AGCM20) substantially improves the simulated annual mean number and maximum intensity of TCs, Kanada et al. (2013) have reported that the minimum central pressure locations of AGCM20-simulated intense TCs show a significant poleward bias.

Similar poleward biases appear in atmospheric reanalysis datasets. For example, Schenkel and Hart (2012) have reported that all TCs derived from five reanalysis datasets underestimated the rate of intensification as well as the maximum intensity of intense category3–5 TCs on the Saffir–Simpson hurricane wind scale (http://www.nhc.noaa.gov/aboutsshws.php) during the life cycle of the TCs (see Schenkel and Hart 2012, their Fig. 7). In addition, more gradual intensification and a delay in the time to maximum intensity relative to the best-track datasets (Schenkel and Hart 2012) can result in a poleward bias, because TCs generally move poleward. Schenkel and Hart (2012) have suggested that a possible reason for more gradual intensification might be the inability of reanalyses to resolve the inner-core structure. According to sensitivity experiments with high-resolution models, a horizontal resolution of a few kilometers is required to resolve the inner-core structure and associated processes of intense TCs (e.g., Braun and Tao 2000; Gentry and Lackmann 2010). Indeed, the results of Fierro et al. (2009) implied that there are discrepancies in the maximum intensity and kinematic structure of intense TCs between horizontal resolutions that are coarser than 10 km and those that are finer than 5 km.

Although the rate of intensification of TCs reproduced in reanalysis datasets tends to be underestimated, intense TCs usually undergo intensification at a high rate (e.g., Holliday and Thompson 1979; Kaplan and DeMaria 2003). The rate of intensification is closely related to the inner-core structure. For example, an early study by Holliday and Thompson (1979) revealed a definite decrease in eye diameter during rapid intensification (RI), which is defined as a central pressure drop greater than 42 hPa within 24 h. Recent observational studies have reported that, during intensification, the inflow layer of TCs is deeper and stronger, and the TCs have a stronger axisymmetric eyewall with deep, intense updrafts in the inner core (e.g., Rogers et al. 2013). These findings are consistent with the results of theoretical studies, which have suggested that diabatic heating near the storm center induces increases in tangential wind velocity owing to the efficient conversion to kinetic energy (e.g., Hack and Schubert 1986; Vigh and Schubert 2009). Meanwhile, Montgomery and Smith (2014) have proposed a new paradigm for TC intensification, called the “boundary layer spin-up mechanism,” that sheds light on planetary boundary layer (PBL) processes in the inner core. A gradient wind imbalance in the TC PBL (Kepert 2006a,b; Bryan and Rotunno 2009) plays a key role in this mechanism.

Whereas a number of studies have emphasized the relationship between inner-core structure and TC intensity, environmental factors such as SST and VWS are also known to be important for TC development (e.g., Gray 1968; Emanuel 2003). For example, SST plays a crucial role in the potential intensity theory of Emanuel (1986). More recently, Tang and Emanuel (2010, 2012) have suggested a new perspective that treats the interaction between VWS and TC intensity and develops a ventilation index. However, how the simulated inner-core structures and processes respond to these environmental factors and modulate the climatological characteristics of TCs in simulations by climate models remains unknown.

The goal of this study is to clarify the inner-core structures and processes that impact the representation of the climatological characteristics of simulated intense TCs. For this purpose, we used the results of climate runs by AGCM20. To compare the simulated inner-core structures, we also conducted downscaling experiments by using a regional atmospheric nonhydrostatic model with a horizontal resolution of 5 km (ANHM5).

The remainder of this paper is organized as follows. Section 2 describes the models and methodology. Section 3 compares the climatological characteristics of intense TCs between simulations with the AGCM20 and ANHM5 and best-track datasets. In addition, the inner-core structures of simulated TCs are examined. Section 4 discusses the causes of the shallow eyewall updrafts simulated by the AGCM20 and the role of air–sea interactions in simulating TCs. Section 5 provides a summary.

2. Models and methodology

a. Model descriptions

We used two AGCM20 versions, MRI-AGCM3.1 (Mizuta et al. 2006) and MRI-AGCM3.2 (Mizuta et al. 2012), each with a horizontal resolution of approximately 20 km [corresponding to a triangular truncation at wavenumber 959 with a linear Gaussian grid (TL959)]. MRI-AGCM3.1 is a climate model based on the Japan Meteorological Agency (JMA) operational numerical weather prediction model and includes a prognostic Arakawa–Schubert cumulus convection scheme (Arakawa and Schubert 1974; Randall and Pan 1993). The MRI-AGCM3.2 was tuned from the MRI-AGCM3.1 to better capture the climatological characteristics of TCs (Sugi et al. 2009; Murakami et al. 2012). In the MRI-AGCM3.2, the prognostic Arakawa–Schubert cumulus convection scheme has been replaced with a new cumulus convection scheme (Yoshimura et al. 2015). The new scheme (Yoshimura scheme) introduces the entraining and detraining effects of clouds for each model level based on the work of Tiedtke (1989). The replacement scheme substantially improves the reproducibility of TC climatology, including the frequency and maximum intensity of TCs (Murakami et al. 2012). A large-scale cloud scheme similar to that proposed by Smith (1990) in the MRI-AGCM3.1 has been replaced with the Tiedtke cloud scheme (Tiedtke 1993) in the MRI-AGCM3.2. The large-scale cloud schemes are very simple microphysics schemes for the GCMs; they represent the cloud and precipitation processes associated with grid-scale convection. Both models apply the level-2 turbulence closure scheme of Mellor and Yamada (1974) as the PBL scheme. In the vertical direction, there are 60 layers (top at 0.1 hPa) in the MRI-AGCM3.1 and 64 layers (top at 0.01 hPa) in the MRI-AGCM3.2.

We also conducted downscaling experiments with the ANHM5, a nonhydrostatic atmospheric model based on the JMA operational nonhydrostatic mesoscale model (Saito et al. 2007) with a horizontal resolution of 5 km. In accord with Saito et al. (2007), the ANHM5 uses a level-3 Mellor–Yamada–Nakanishi–Niino closure PBL scheme (Nakanishi and Niino 2004) and bulk-type cloud microphysics (Murakami 1990) along with a Kain–Fritsch cumulus parameterization scheme (KF scheme; Kain and Fritsch 1993). Because most inner-core precipitation of an intense TC is produced by the microphysics scheme, there is no large difference in the temporal evolution and the maximum intensity of the intense TC simulated by the ANHM5 with and without the KF scheme (Kanada and Wada 2016). In addition, the ANHM5 applies the spectral nudging method of Nakano et al. (2012). The number of vertical levels is set to 55 (lowest layer is 20 m; top height is approximately 27 km). Table 1 summarizes the specifications of the MRI-AGCM3.1, MRI-AGCM3.2, and ANHM5.

Table 1.

Specifications of the models. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)

Specifications of the models. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)
Specifications of the models. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)

b. Experimental design and methodology for AGCM20 runs

Using a monthly mean SST database (HadISST1; Rayner et al. 2003), we conducted two climate simulations for the period 1979–2003: the first simulation, AS, was conducted with the MRI-AGCM3.1 (AGCM20_3.1 in Murakami et al. 2012), and the second simulation, YS, was conducted with the MRI-AGCM3.2 (AGCM20_3.2).

We used the following criteria to detect TCs simulated by the AGCM20 (Murakami et al. 2012) every 6 h of integration:

  1. maximum relative vorticity at 850 hPa exceeding 1.2 × 10−4 in AS and 2.0 × 10−4 s−1 in YS (see below for details);

  2. maximum wind speed at 850 hPa exceeding 17.0 m s−1;

  3. a warm-core temperature anomaly, defined as the sum of the temperature deviations at 300, 500, and 700 hPa, exceeding 2.0 K, with the temperature deviation at each level computed by subtracting the maximum temperature from the mean temperature over a 10° × 10° grid box centered nearest to the location of maximum vorticity at 850 hPa; and

  4. simulated TC duration exceeding 36 consecutive hours.

Following Murakami et al. (2012), the first criterion was tuned in each AGCM20 run to make the mean number of globally simulated TCs match the number of best-track TCs during a given period (84 TCs yr−1 during 1979–2003). Such tuning may cause difficulty in interpreting the representation of the number of TCs in each category. However, the main purpose of this study was to elucidate the characteristics of “intense” TCs simulated by the AGCM20 and ANHM5 in the climate runs. We therefore used the direct values from the model output. We focused mainly on the evolution of intense TCs until they reached their maximum intensity; extratropical transitions were not considered.

The storm center of each simulated AGCM20 TC was determined as the location of the minimum central pressure every 6 h of integration. We presented all times relative to a reference time: namely, the time when the TC first reached a central pressure less than 950 hPa.

c. Experimental design and methodology for ANHM5

The ANHM5 experiments were performed for all intense YS TCs, which were defined as YS TCs with a minimum central pressure less than or equal to 900 hPa that were simulated in the western North Pacific (WNP) region (0°–40°N, 100°E –180°) during 1979–2003. The total number of such TCs was 34. The ANHM5 experiments started from a relative time of −6 h (−6 h) and ended at least 18 h after YS TCs could no longer be detected. The initial and 6-hourly lateral boundary conditions and SST initial conditions used in the ANHM5 experiments were provided by the YS results. The computational domain of the ANHM5 was 3600 km × 3600 km. To cover the life cycle of each TC, including its intensification, maturation, and decay phases, an additional domain of 4300 km × 4100 km was used for TCs that moved out of the predesignated domain. In the downscaling experiments with ANHM5, 23 of the 34 intense YS TCs developed into TCs with a minimum central pressure less than or equal to 900 hPa. The remaining 11 TCs exhibited a minimum central pressure greater than 900 hPa. We designated the 23 ANHM5 TCs with minimum central pressures less than or equal to 900 hPa and the corresponding 23 YS TCs as L900. The 11 ANHM5 TCs with minimum central pressures greater than 900 hPa and the corresponding 11 YS TCs were designated as H900.

Because the pressure field in a fine-mesh nonhydrostatic model appears noisy around vigorous convection in the eyewall, we defined the storm center as the approximate geometric center (centroid) of each ANHM5 TC, which was determined from the horizontal distribution of sea level pressure (Braun 2002). The centroid was calculated at radial intervals of 5 km and summed within radius R = 100 km for every grid point (Xcp±20grid, Ycp±20grid) around the grid point location of the central pressure (Xcp, Ycp). The initial-guess position was moved to within 40 km of the pressure center. The grid point (X, Y) at which the summation was a minimum was selected as the storm center. In this study, the storm center thus defined was generally within 20 km of the location of the minimum central pressure. Radial Vr and tangential Vt wind velocities determined relative to the storm center were calculated for each Cartesian grid point.

d. Environmental factors

We explored the relationships between environmental conditions and the evolution of simulated intense TCs. For environmental factors, we selected VWS and SST. We defined VWS as the vector difference in the mean horizontal winds in YS between altitudes of 1.5 km (approximately 850 hPa) and 12.5 km (approximately 200 hPa). Mean horizontal winds and mean SST were calculated over a circular area with a radius of 500 km. In addition, we calculated the maximum potential intensity based on the Emanuel (1986, 1995, 2006) model using the atmospheric conditions (pressure, temperature, and mixing ratio) and SST for each grid point. We designated the minimum potential central pressure within a radius of 200 km from the storm center as the Emanuel MPI (E-MPI). Because the ANHM5 experiments were nested in the results of the YS simulation, environmental factors for the corresponding TCs in both the ANHM5 and YS simulations were comparable.

e. Best-track analyses

To validate the climatological characteristics of the simulated TCs, we used two best-track datasets for the WNP region: the Regional Specialized Meteorological Center (RSMC) Tokyo best-track dataset and the Joint Typhoon Warning Center (JTWC) WNP best-track dataset. During the analysis period (1979–2003), the RSMC best-track datasets provided TC center locations and the intensities of both the minimum central pressure and the maximum 10-min mean sustained 10-m wind speed at 6-h intervals. The 10-min sustained maximum wind speed in the RSMC datasets was converted to an equivalent 1-min sustained surface wind speed using a factor of 1.14, following Murakami (2014). The JTWC best-track datasets contain maximum 1-min mean sustained 10-m wind speeds, but they contained no minimum central pressure data until 2001. We therefore used mainly the minimum central pressure in the RSMC dataset, but we used both the RSMC and JTWC datasets when we analyzed the maximum wind speeds. Although Knaff and Zehr (2007) have proposed the new wind–pressure relationship, the original data were used for both the RSMC and JTWC datasets. This study focused on TCs from 1979 to 2003 for which 1-min sustained surface winds were greater than 33 m s−1.

We used two kinds of maximum wind speeds in the simulations: the maximum 10-m wind speed (VMX) and the maximum azimuthal-mean 10-m wind speed (). VMX was used when we validated the results of the simulations with the best-track datasets. We mainly used when we investigated the results of the simulations because intense TCs are known to have axisymmetric structures (e.g., Rogers et al. 2013).

3. Results

a. Climatological characteristics of TCs in the best-track datasets and the AGCM20 simulations

We first evaluated the climatological characteristics of AGCM20 TCs in the WNP. The annual numbers of YS TCs were comparable to those in the two best-track datasets (Table 2). In contrast, the AGCM20 with AS could not simulate category-4 and category-5 TCs. The difference between these results is likely due to the different cumulus parameterization schemes used by the two versions of the AGCM20 (Murakami et al. 2012).

Table 2.

Annual mean numbers of TCs in the analysis region between 1979 and 2003 in the RSMC and JTWC datasets and in the YS and AS simulations with the AGCM20. The TCs were classified based on the maximum 1-min sustained surface wind speeds in the RSMC and JTWC datasets and the maximum 10-m wind speeds in the AGCM20 climate runs, corresponding to categories 1–5, respectively, on the Saffir–Simpson hurricane wind scale, as follows: Cat1, 33–43 m s−1; Cat2, 43–49 m s−1; Cat3, 49–59 m s−1; Cat4, 59–70 m s−1; and Cat5, >70 m s−1.

Annual mean numbers of TCs in the analysis region between 1979 and 2003 in the RSMC and JTWC datasets and in the YS and AS simulations with the AGCM20. The TCs were classified based on the maximum 1-min sustained surface wind speeds in the RSMC and JTWC datasets and the maximum 10-m wind speeds in the AGCM20 climate runs, corresponding to categories 1–5, respectively, on the Saffir–Simpson hurricane wind scale, as follows: Cat1, 33–43 m s−1; Cat2, 43–49 m s−1; Cat3, 49–59 m s−1; Cat4, 59–70 m s−1; and Cat5, >70 m s−1.
Annual mean numbers of TCs in the analysis region between 1979 and 2003 in the RSMC and JTWC datasets and in the YS and AS simulations with the AGCM20. The TCs were classified based on the maximum 1-min sustained surface wind speeds in the RSMC and JTWC datasets and the maximum 10-m wind speeds in the AGCM20 climate runs, corresponding to categories 1–5, respectively, on the Saffir–Simpson hurricane wind scale, as follows: Cat1, 33–43 m s−1; Cat2, 43–49 m s−1; Cat3, 49–59 m s−1; Cat4, 59–70 m s−1; and Cat5, >70 m s−1.

Although the annual mean number of TCs was simulated well in YS, the horizontal distributions of the minimum central pressure locations of the simulated intense TCs showed significant northward shifts compared with the locations in the RSMC best-track dataset (Figs. 1a,b). Most of the best-track TCs were located south of 25°N, whereas in YS about half of the TCs were located north of 25°N. The northward shift of YS TCs was also apparent in the locations of lifetime maximum of VMX (Figs. 1d,e).

Fig. 1.

Track of intense TCs, defined as TCs with MCP ≤ 900 hPa, from (a) RSMC best-track, (b) YS, and (c) the ANHM5 experiments between 1979 and 2003. Small dots represent 6-h tracks of intense TCs during which MCP ≤ 950 hPa. Red and blue colors correspond to RI and non-RI TCs in YS, respectively. Black color indicates H900 TCs. (d) As in (a), but of TCs with a maximum 1-min sustained surface wind speed ≥70 m s−1 in the JTWC best-track dataset. (e),(f) As in (b),(c), but with VMX ≥ 70 m s−1 in YS and the ANHM5 experiments, respectively. Large dots represent the locations of maximum intensity. Black horizontal lines indicate 25°N. Only the locations before peak intensity are shown.

Fig. 1.

Track of intense TCs, defined as TCs with MCP ≤ 900 hPa, from (a) RSMC best-track, (b) YS, and (c) the ANHM5 experiments between 1979 and 2003. Small dots represent 6-h tracks of intense TCs during which MCP ≤ 950 hPa. Red and blue colors correspond to RI and non-RI TCs in YS, respectively. Black color indicates H900 TCs. (d) As in (a), but of TCs with a maximum 1-min sustained surface wind speed ≥70 m s−1 in the JTWC best-track dataset. (e),(f) As in (b),(c), but with VMX ≥ 70 m s−1 in YS and the ANHM5 experiments, respectively. Large dots represent the locations of maximum intensity. Black horizontal lines indicate 25°N. Only the locations before peak intensity are shown.

The northward shift in YS was particularly evident in the case of the more intense TCs (Fig. 2a). In addition, the mean maximum central pressure drop of the simulated intense TCs was small relative to that of the RSMC best-track dataset (Fig. 2b). The mean RSMC best-track minimum central pressure location of category-5 TCs was 16.7°N, whereas the latitude was 23.6°N in YS. Mean maximum central pressure changes from the preceding 24 h (mdCP24) for category-5 TCs were −60 and −45 hPa in the RSMC best-track dataset and YS, respectively. Both the difference between mean simulated and best-track minimum central pressure locations and the difference between the mean simulated and best-track mdCP24 for category-4 and category-5 TCs were statistically significant at p = 0.01 based on a two-sided Welch’s t test.

Fig. 2.

(a) Mean MCP latitudes and (b) mdCP24 from RSMC (black dots), YS (red diamonds), AS (green diamonds), and the ANHM5 experiments (blue diamonds) in the analysis region between 1979 and 2003. Symbols not filled with color indicate mean values that were significantly different from those of the RSMC at p = 0.01 based on a two-sided Welch’s t test. Error bars indicate standard deviations.

Fig. 2.

(a) Mean MCP latitudes and (b) mdCP24 from RSMC (black dots), YS (red diamonds), AS (green diamonds), and the ANHM5 experiments (blue diamonds) in the analysis region between 1979 and 2003. Symbols not filled with color indicate mean values that were significantly different from those of the RSMC at p = 0.01 based on a two-sided Welch’s t test. Error bars indicate standard deviations.

b. Comparisons of simulated intense TCs in YS and ANHM5 experiments

Using the results of YS, we conducted downscaling experiments with the ANHM5 for 34 intense YS TCs; 23 of them developed into L900 ANHM5 TCs.

Most of the minimum central pressure locations of L900 ANHM5 TCs appeared south of 25°N (Fig. 1c). Furthermore, both the mean minimum central pressure location and the mdCP24 after the time when the central pressure was less than 950 hPa in the L900 ANHM5 TCs fell within plus or minus one standard deviation of the corresponding best-track value (Fig. 2 and Table 3). Meanwhile, ANHM5 TCs could not attain a central pressure less than 900 hPa north of 25°N.

Table 3.

Total number of intense TCs (total No.), percentage of total TCs with RI (RI ratio), mean MCP, latitude at which MCP was reached (Latmcp), latitude at which t = 0 h (Latt=0), and mdCP24 after the time central pressure became <950 hPa for intense TCs in the analysis region between 1979 and 2003 in the RSMC best-track dataset and in YS and the ANHM5 experiments. L900 indicates TCs with MCP ≤ 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. H900 indicates TCs with MCP > 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. Numbers in parentheses indicate standard deviations. Mean values in boldface differ significantly from L900 YS TCs (p = 0.1, two-sided Welch’s t test).

Total number of intense TCs (total No.), percentage of total TCs with RI (RI ratio), mean MCP, latitude at which MCP was reached (Latmcp), latitude at which t = 0 h (Latt=0), and mdCP24 after the time central pressure became <950 hPa for intense TCs in the analysis region between 1979 and 2003 in the RSMC best-track dataset and in YS and the ANHM5 experiments. L900 indicates TCs with MCP ≤ 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. H900 indicates TCs with MCP > 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. Numbers in parentheses indicate standard deviations. Mean values in boldface differ significantly from L900 YS TCs (p = 0.1, two-sided Welch’s t test).
Total number of intense TCs (total No.), percentage of total TCs with RI (RI ratio), mean MCP, latitude at which MCP was reached (Latmcp), latitude at which t = 0 h (Latt=0), and mdCP24 after the time central pressure became <950 hPa for intense TCs in the analysis region between 1979 and 2003 in the RSMC best-track dataset and in YS and the ANHM5 experiments. L900 indicates TCs with MCP ≤ 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. H900 indicates TCs with MCP > 900 hPa simulated by the ANHM5 and the corresponding TCs in YS. Numbers in parentheses indicate standard deviations. Mean values in boldface differ significantly from L900 YS TCs (p = 0.1, two-sided Welch’s t test).

Table 3 and Fig. 3 compare the characteristics of the 34 intense YS TCs with the results in the downscaling experiments performed by the ANHM5. In general, intense TCs are characterized by high rates of intensification (e.g., Holliday and Thompson 1979; Kaplan and DeMaria 2003). Indeed, 96% of the L900 ANHM5 TCs underwent RI, defined by mdCP24 ≤ −42 hPa (Holliday and Thompson 1979), whereas only half of the YS TCs underwent RI. Note that only one H900 ANHM5 TC underwent RI.

Fig. 3.

Scatter diagrams for YS TCs and for ANHM5 TCs: (a) MCP (hPa), (b) MCP latitudes (°N), (c) the latitudes (°N), (d) mdCP24 [hPa (24 h)−1], (e) the maximum changes [m s−1 (24 h)−1], and (f) duration (h) between t = 0 h and the time when the TCs attained MCP. Red dots and blue triangles correspond to L900 RI and non-RI TCs in YS. Black crosses correspond to H900 TCs. Solid blue triangles indicate CASE1.

Fig. 3.

Scatter diagrams for YS TCs and for ANHM5 TCs: (a) MCP (hPa), (b) MCP latitudes (°N), (c) the latitudes (°N), (d) mdCP24 [hPa (24 h)−1], (e) the maximum changes [m s−1 (24 h)−1], and (f) duration (h) between t = 0 h and the time when the TCs attained MCP. Red dots and blue triangles correspond to L900 RI and non-RI TCs in YS. Black crosses correspond to H900 TCs. Solid blue triangles indicate CASE1.

The major difference between L900 and H900 TCs was the mean latitude at the beginning of the integration time. The mean latitude at t = 0 h for L900 TCs (16.1°N) was comparable to that in the RSMC best-track dataset (15.0°N), whereas the mean latitude for H900 TCs was shifted northward (23.8°N). Furthermore, the mean latitude of the maximum intensity of the L900 YS TCs (23.4°N) exhibits a significant northward shift compared with that in the RSMC best-track dataset (16.7°N). In contrast, the mean latitude of the maximum intensity of the L900 ANHM5 TCs was 20.2°N.

Figure 3 reveals that almost all YS TCs showed northward shifts of the locations of both their minimum central pressure and (Figs. 3b,c); the intensification periods of these TCs were also longer (Fig. 3f) than those of ANHM5 TCs. The minimum (maximum) central pressure drops ( changes) for most of the YS TCs were smaller than those in the ANHM5 experiments, except for the H900 TCs (Figs. 3d,e).

Most of ANHM5 TCs underwent RI and attained their maximum intensity in a shorter time at lower latitudes than the corresponding YS TCs. The ANHM5 TCs could not develop to a minimum central pressure less than 900 hPa north of 25°N. Thus, significant northward shifts in YS were suppressed in the ANHM5 experiments.

c. Temporal evolutions of TCs in YS and ANHM5 experiments

Figure 4 shows the temporal evolution of a typical TC with a prolonged gradual intensification in YS. We selected a TC for which the discrepancy between the YS and the ANHM5 experiments of the mdCP24 values was a maximum (CASE1). The CASE1 TC moved northward and made landfall in Japan at t = 114 h (Fig. 4a). The mdCP24 values were −35 and −59 hPa in the YS and ANHM5 experiments, respectively.

Fig. 4.

(a) Tracks and (b) the temporal evolutions of and (c) CP for CASE1 in YS (black) and the ANHM5 experiment (red). Radius–time cross sections of azimuthal-mean hourly precipitation (color scale; mm h−1) for CASE1 in (d) YS and (e) ANHM5 experiments. In (d) and (e), black lines with white dots indicate the RMW at an altitude of 10 m. In (b)–(e), the horizontal dashed lines indicate the time that MCP was reached in YS (black) and in the ANHM5 experiment (red).

Fig. 4.

(a) Tracks and (b) the temporal evolutions of and (c) CP for CASE1 in YS (black) and the ANHM5 experiment (red). Radius–time cross sections of azimuthal-mean hourly precipitation (color scale; mm h−1) for CASE1 in (d) YS and (e) ANHM5 experiments. In (d) and (e), black lines with white dots indicate the RMW at an altitude of 10 m. In (b)–(e), the horizontal dashed lines indicate the time that MCP was reached in YS (black) and in the ANHM5 experiment (red).

ANHM5 TCs rapidly intensified; regions with intense eyewall precipitation contracted, and there was a rapid decrease in the radius of (RMW10m) from 45 to 30 km between t = −6 and 12 h. Stern et al. (2015) have also shown in idealized simulations that both the contraction of the radius of maximum wind (RMW) and TC intensification started at the same time, though the contraction halted before the peak intensity was attained. In contrast, this YS TC underwent gradual intensification with a constant RMW10m of 40 km from t = −6 to 66 h. Although the rate of intensification was much smaller in the YS TC than in the ANHM5 TC, the YS TC intensified gradually until t = 90 h and eventually attained a MCP of 874 hPa at 26.0°N. The ANHM5 TC attained the same value (874 hPa) at t = 72 h at 23.0°N.

Figure 5 compares the composite temporal evolutions of intense TCs in the best-track data and L900 and H900 TCs in YS and the ANHM5 experiments. The best-track TCs rapidly intensified until t = 18 h. The rapid intensification during this period was reproduced well by L900 ANHM5 TCs. Mean intensification rates of L900 YS TCs were relatively low compared to the rates of RSMC best-track TCs. Furthermore, the RSMC best-track TCs attained their minimum central pressure at t = 24 h, whereas the L900 YS TCs tended to intensify for a longer time. In particular, L900 non-RI YS TCs were still gradually developing in terms of both central pressure and at t = 72 h. The rates of intensification of the H900 YS TCs were only slightly lower than those for L900 YS TCs. In contrast, the values for the H900 ANHM5 TCs were substantially lower.

Fig. 5.

Temporal evolution of (a) CP, (b) , and (c) RMW10m of intense TCs composited based on the first time the CP became less than 950 hPa: L900 (green) and H900 (blue) YS TCs and L900 (red) and H900 (orange) ANHM5 TCs. Dashed and dotted lines indicate RI and non-RI YS TCs. In (a), data for intense TCs in RSMC (black squares) are also shown. In (b), the 1-min sustained surface wind speed in RSMC (black squares) and JTWC (gray crosses) are also shown. Error bars indicate standard deviations.

Fig. 5.

Temporal evolution of (a) CP, (b) , and (c) RMW10m of intense TCs composited based on the first time the CP became less than 950 hPa: L900 (green) and H900 (blue) YS TCs and L900 (red) and H900 (orange) ANHM5 TCs. Dashed and dotted lines indicate RI and non-RI YS TCs. In (a), data for intense TCs in RSMC (black squares) are also shown. In (b), the 1-min sustained surface wind speed in RSMC (black squares) and JTWC (gray crosses) are also shown. Error bars indicate standard deviations.

The temporal evolution of the RMW10m revealed that high rates of intensification were associated with a decrease of the RMW10m (Fig. 5c). In YS, the mean initial RMW10m of both L900 and H900 TCs was smaller for non-RI TCs than for RI TCs. According to Carrasco et al. (2014), RI TCs tended to have small initial sizes. However, Carrasco et al. (2014) investigated a relationship between “a change of intensity over 24 h” and “a RMW at the initial time” under the definition of a TC size that RMW ≤ 37 km was defined as a small size, while 37 km < RMW ≤ 89 km as a medium size. In YS, the mean initial RMW10m of both L900 and H900 TCs belonged to the medium size category, and there were few TCs with the RMW ≤ 37 km. Because the horizontal resolution of the AGCM20 was 20 km, the relationship of Carrasco et al. (2014) could not be applied in the present study. In YS, the mean RMW10m of L900 (H900) RI TCs at t = −30 h was 78 (83) km, and that of L900 (H900) non-RI TCs was 54 (64) km. This difference was statistically significant at p = 0.01 based on a two-sided Welch’s t test.

d. Comparison of the inner-core structures of simulated intense TCs during the RI phase

Previous studies have reported that TC inner-core structures differ, depending on the rate of intensification (e.g., Rogers et al. 2013). We therefore compared the composite radius–altitude cross sections at t = 12 h of the L900 TCs in YS and the ANHM5 experiments (Fig. 6). We averaged all values for the ANHM5 TCs at radial intervals of 20 km. At t = 12 h, there was no large difference between L900 ANHM5 TCs and RI L900 YS TCs in their mean intensities and rates of intensification (Fig. 5).

Fig. 6.

Composite radius–altitude cross sections of the azimuthal-mean (a) vertical velocity (color scale; m s−1) and the inertial stability squared (black contour; 10−6 s−1), (b) Vt (m s−1), (c) AAM (106 m2 s−1), and (d) Vr (m s−1) at t = 12 h for intense TCs simulated by ANHM5. All values are averaged over radial intervals of 20 km. (e)–(h) As in (a)–(d), but for L900 RI YS TCs. (i)–(l) As in (a)–(d), but for L900 non-RI YS TCs. Magenta and green lines indicate the w and Vt axes, respectively.

Fig. 6.

Composite radius–altitude cross sections of the azimuthal-mean (a) vertical velocity (color scale; m s−1) and the inertial stability squared (black contour; 10−6 s−1), (b) Vt (m s−1), (c) AAM (106 m2 s−1), and (d) Vr (m s−1) at t = 12 h for intense TCs simulated by ANHM5. All values are averaged over radial intervals of 20 km. (e)–(h) As in (a)–(d), but for L900 RI YS TCs. (i)–(l) As in (a)–(d), but for L900 non-RI YS TCs. Magenta and green lines indicate the w and Vt axes, respectively.

The mean inner-core structure of L900 ANHM5 TCs during RI was characterized by deep and intense eyewall updrafts with a pair of intense outflows greater than 15.0 m s−1 at an altitude of 15 km and intense near-surface inflow less than −20.0 m s−1. Mean updrafts greater than 1 m s−1 reached an altitude of 14 km, and a maximum vertical velocity w greater than 1.4 m s−1 appeared at an altitude of 10 km. The tilts of the w and Vt axes, defined by the radii of the maximum azimuthal-mean updraft and maximum azimuthal-mean Vt (RMW) for each level, were relatively small. These features are consistent with airborne Doppler observations of an intensifying storm (see Fig. 6 in Rogers et al. 2013). In contrast, the azimuthal-mean updrafts for the L900 YS TCs were very shallow and weak for both RI and non-RI TCs (Figs. 6e,i); the maximum values appeared below an altitude of 6 km. The w axes of both the RI and non-RI YS TCs tilted outward as the altitude increased.

In the vicinity of deep and intense updrafts around dense and steep surfaces of constant absolute angular momentum (AAM), an intense mean updraft greater than 1.0 m s−1 formed from the leading edge of deep and intense near-surface inflow less than −20 m s−1 in the ANHM5 experiments. In the eyewall regions, the radial gradient of near-surface inflow was steep, and a region of outflow greater than 1.0 m s−1 formed just above the inflow boundary layer (IBL) in the ANHM5 experiments. This structure is consistent with the observational results shown by Zhang et al. (2011). In addition, regions of the upper-tropospheric outflow greater than 15 m s−1 appeared at an altitude of 15 km outside a radius of 100 km.

In contrast, although the intensity of inflow in the lowermost layers was relatively strong in the L900 RI YS TCs, the depth of near-surface inflow less than −10.0 m s−1 was shallow. In addition, there was no outflow evident above the top of the IBL. The iso-AAM surfaces tilted outward as the altitude increased. An upper-tropospheric outflow greater than 15 m s−1 appeared below 15 km outside a radius of 120 km in L900 RI YS TCs. The altitude of the peak outflow is lower in the YS TCs than in ANHM5 TCs.

The azimuthal-mean changes of the radius–altitude cross sections from t = 6 to 12 h revealed a rapid increase in potential temperature greater than 7 K (6 h)−1 in the upper troposphere, with a moderate increase in potential temperature greater than 2 K (6 h)−1 in the midtroposphere in the ANHM5 experiments (Fig. 7). Indeed, most of the ANHM5 TCs exhibited both mid- and upper-tropospheric warm cores (not shown). In most cases, the midtropospheric warm core appeared first, followed by rapid formation of an upper-tropospheric warm core at altitudes around 16 km. A similar double-warm-core structure has been noted in both observational and modeling studies of intense TCs (e.g., Wang and Wang 2014; Stern and Zhang 2016).

Fig. 7.

Composite radius–altitude cross sections of the azimuthal-mean (a) changes in potential temperature [K (6 h) −1], (b) changes in Vt [m s−1 (6 h)−1], (c) changes in AAM [106 m2 s−1 (6 h)−1], and (d) changes in Vr [m s−1 (6 h)−1] from t = 6 to 12 h for intense TCs simulated by ANHM5. All values are averaged over radial intervals of 20 km. (e)–(h) As in (a)–(d), but for L900 RI YS TCs. (i)–(l) As in (a)–(d), but for L900 non-RI YS TCs.

Fig. 7.

Composite radius–altitude cross sections of the azimuthal-mean (a) changes in potential temperature [K (6 h) −1], (b) changes in Vt [m s−1 (6 h)−1], (c) changes in AAM [106 m2 s−1 (6 h)−1], and (d) changes in Vr [m s−1 (6 h)−1] from t = 6 to 12 h for intense TCs simulated by ANHM5. All values are averaged over radial intervals of 20 km. (e)–(h) As in (a)–(d), but for L900 RI YS TCs. (i)–(l) As in (a)–(d), but for L900 non-RI YS TCs.

In contrast, the major warming in both RI and non-RI YS TCs occurred at altitudes below 10 km (Figs. 7e,i). The major increases in the Vt and AAM in YS TCs also occurred below 6 km. In particular, RI YS TCs exhibited regions with large AAM increases [>4 × 106 m2 s−1 (6 h)−1] in the deep layers below 6 km, with large increases in near-surface inflow less than −4 m s−1 (Figs. 7g,h).

Although changes in the RMW10m were relatively small from t = 6 to 12 h (Fig. 5c), ANHM5 TCs intensified rapidly as intensifying Vt throughout the troposphere around and inside the w axes. Increases in AAM greater than 3 × 106 m2 s−1 (6 h)−1 also dominated throughout the troposphere in the ANHM5 TCs. Mean decreases in the RMW10m from t = 6 to 12 h were 6.0 and 1.0 km in L900 RI YS TCs and L900 ANHM5 TCs, respectively.

e. Inner-core warming and altitudes of maximum updraft

We addressed the temporal evolution of inner-core warming in the context of the altitude of maximum eyewall updraft, which was defined by the mean altitude of the maximum azimuthal-mean updraft within a radius of 200 km (Fig. 8). The results of ANHM5 experiments after the spinup period (t = ~6 h) were presented.

Fig. 8.

Temporal evolution of composite vertical profiles of changes in potential temperature [color scale; K (6 h)−1] and vertical velocity (contours; −0.02 and 0 m s−1) averaged within a radius of 20 km for (a) L900 ANHM5 TCs and (b) RI and (c) non-RI YS TCs. For L900 ANHM5 TCs, the results after the spinup period (t = ~6 h) were presented. Magenta horizontal lines indicate integration times with changes in CP ≤ −9.0 hPa from the preceding 6 h, and the vertical dashed lines indicate the time that MCP was reached. Green lines with dots indicate the mean altitude of the maximum azimuthal-mean updraft within a radius of 200 km. (d)–(f) As in (a)–(c), but for H900 TCs.

Fig. 8.

Temporal evolution of composite vertical profiles of changes in potential temperature [color scale; K (6 h)−1] and vertical velocity (contours; −0.02 and 0 m s−1) averaged within a radius of 20 km for (a) L900 ANHM5 TCs and (b) RI and (c) non-RI YS TCs. For L900 ANHM5 TCs, the results after the spinup period (t = ~6 h) were presented. Magenta horizontal lines indicate integration times with changes in CP ≤ −9.0 hPa from the preceding 6 h, and the vertical dashed lines indicate the time that MCP was reached. Green lines with dots indicate the mean altitude of the maximum azimuthal-mean updraft within a radius of 200 km. (d)–(f) As in (a)–(c), but for H900 TCs.

In the ANHM5 experiments, the warming was very large in the upper troposphere during the RI phase. Warming greater than 5 K (6 h)−1 occurred at altitudes of 16 km in the L900 ANHM5 TCs. Upper-tropospheric warming continued until a minimum central pressure was attained, at which time it started to decrease. The H900 ANHM5 TCs also showed similar behavior, with major warming in the upper troposphere. In contrast, most of the major warming first occurred below an altitude of 10 km during the phase of the greatest intensification in the YS TCs. Then, in the RI YS TCs, warming greater than 4 K (6 h)−1 occurred at altitudes of 10–15 km. The deep and intense updrafts around the eyewall sometimes cause upper-tropospheric warming (e.g., Chen and Zhang 2013; Wang and Wang 2014). In the present study, the maximum eyewall updraft at an altitude of 10 km in the L900 RI YS TCs intensified rapidly from t = 6 h (0.35 m s−1) to t = 18 h (0.49 m s−1) (not shown). In contrast, the maximum eyewall updraft at an altitude of 10 km was always less than 0.38 m s−1 from t = 0 to 72 h in L900 non-RI YS TCs (not shown).

It should be noted that the altitudes of maximum eyewall updraft in the YS TCs changed dramatically before and after t = 0 h. Before t = 0 h, the maximum eyewall updraft was found at altitudes of about 10 km, but, after t = 0 h, those altitudes decreased rapidly to less than 6 km. These decreases were concurrent with major midtropospheric warming. Although the altitudes of the warming corresponded to subsidence of the ANHM5 TCs, there was no apparent relationship between the warming and subsidence in the YS TCs. The lack of subsidence at mid- and low levels in YS may be related to the coarse resolution. Another process that may have impacted the warming profile is latent heat release resulting from condensation (e.g., McFarquhar et al. 2012).

We examined the temporal evolution of the warming and AAM during the intensification phase (Fig. 9). Before t = −6 h, the maximum in both warming and the change in AAM were seen at altitudes around 12 km (Figs. 9a,e). From t = −6 to 0 h, additional warming occurred at altitudes around 5 km, and a local maximum in the AAM tendency started to form below 5 km (Fig. 9b). Rapid increases of midlevel warming and AAM below 6 km occurred from t = 0 to 12 h. Note that the altitudes of maximum updraft also decreased from 10 to 3 km during this period (Fig. 8). In contrast, the major warming during RI appeared in the upper troposphere in the ANHM5 TCs. Increases of the AAM greater than 3.5 × 106 m2 s−1 (6 h)−1 occurred throughout the troposphere around the intense eyewall updrafts. These results suggest the vertical advection of AAM with deep and intense updrafts during RI in L900 ANHM5 TCs. To clarify the details of the mechanism, a budget analysis is required, but such an analysis was beyond the scope of this paper.

Fig. 9.

Composite radius–altitude cross sections of the azimuthal-mean changes in AAM [color scale; 106 m2 s−1 (6 h)−1] and changes in potential temperature [contour; 1, 2, 3, 4, and 5 K (6 h)−1] (a) from t = −12 to −6 h, (b) from t = −6 to 0 h, (c) from t = 0 to 6 h, and (d) from t = 6 to 12 h for L900 RI YS TCs. (e)–(h) As in (a)–(d), but for L900 non-RI YS TCs. (i),(j) As in (c),(d), but for L900 ANHM5 TCs. In (i) and (j), all values are averaged over radial intervals of 20 km. Green lines indicate the Vt axis.

Fig. 9.

Composite radius–altitude cross sections of the azimuthal-mean changes in AAM [color scale; 106 m2 s−1 (6 h)−1] and changes in potential temperature [contour; 1, 2, 3, 4, and 5 K (6 h)−1] (a) from t = −12 to −6 h, (b) from t = −6 to 0 h, (c) from t = 0 to 6 h, and (d) from t = 6 to 12 h for L900 RI YS TCs. (e)–(h) As in (a)–(d), but for L900 non-RI YS TCs. (i),(j) As in (c),(d), but for L900 ANHM5 TCs. In (i) and (j), all values are averaged over radial intervals of 20 km. Green lines indicate the Vt axis.

f. Environmental conditions and evolution of the TCs

On the whole, mean environmental conditions for L900 and H900 TCs were favorable for TC development at the beginning of the integration time (t = −6 h) (Fig. 10): a relatively weak VWS less than 8 m s−1 (Reasor et al. 2013) and warm SST greater than 28°C (Kaplan and DeMaria 2003). Except for the H900 non-RI YS TCs, mean VWS (SST) increased (decreased) as the integration time progressed. The difference between L900 and H900 TCs appeared in the thermodynamic environmental conditions; the mean SST tended to be cooler for the H900 TCs than for the L900 TCs. Mean SSTs at t = −6 h were 29.2° and 28.7°C for the L900 and H900 TCs, respectively. Thus, the E-MPI for the H900 TCs was relatively high. The mean minimum E-MPI at t = −6 h for the H900 TCs was 912 hPa, although the mean minimum central pressure for the YS TCs was 894 hPa.

Fig. 10.

Temporal evolution of composite (a) VWS (m s−1) and (b) mean SST (°C) between t = −6 and 72 h and (c) E-MPI (hPa) between t = −48 and 30 h for L900 (green) and H900 (blue) YS TCs. Dashed and dotted lines in (a)–(c) indicate RI and non-RI YS TCs, respectively. Error bars indicate standard deviations for L900 (green) and H900 (blue) YS TCs. (d) The mean E-MPI between t = −48 and −6 h (dots) and E-MPI at t = −6 h (circles) for L900 (green) and H900 (blue) YS TCs and MCP for L900 (green dots) and H900 (blue dots) YS TCs and ANHM5 TCs (red dot). Error bars indicate standard deviations.

Fig. 10.

Temporal evolution of composite (a) VWS (m s−1) and (b) mean SST (°C) between t = −6 and 72 h and (c) E-MPI (hPa) between t = −48 and 30 h for L900 (green) and H900 (blue) YS TCs. Dashed and dotted lines in (a)–(c) indicate RI and non-RI YS TCs, respectively. Error bars indicate standard deviations for L900 (green) and H900 (blue) YS TCs. (d) The mean E-MPI between t = −48 and −6 h (dots) and E-MPI at t = −6 h (circles) for L900 (green) and H900 (blue) YS TCs and MCP for L900 (green dots) and H900 (blue dots) YS TCs and ANHM5 TCs (red dot). Error bars indicate standard deviations.

The E-MPI is often calculated with a few days lead time. Therefore, the E-MPI between t = −48 and −6 h and that at t = −6 h were compared with the mean simulated minimum central pressures (Fig. 10d). The mean E-MPI by central pressure increased as time progressed so that the mean E-MPI was lower between t = −48 and −6 h than at t = −6 h. Simulated minimum central pressures in the L900 YS TCs (886 hPa) and ANHM5 TCs (884 hPa) fell within plus or minus one standard deviation of the corresponding mean E-MPIs (880 and 888 hPa). However, in H900 TCs, the mean simulated minimum central pressure in the ANHM5 TCs (910 hPa) was similar to the mean E-MPI at t = −6 h (912 hPa), whereas the mean simulated minimum central pressure in the YS TCs (894 hPa) was similar to the mean E-MPI between t = −48 and −6 h (896 hPa).

To further investigate TC responses to environmental factors, we determined 12-h changes in the simulated central pressure (CP) every 6 h between t = 0 and 102 h by subtracting the previous 6-h CP from the subsequent 6-h CP as an index of the TC evolution (dcp12). A 6-h segment with a negative (positive) change was identified as an intensifying (weakening) phase. Figure 11 shows the frequencies of 6-h segments in each joint dcp12 as a function of VWS and SST for L900 and H900 YS TCs and ANHM5 TCs. Figure 11 reveals that the most distinct environmental conditions that simulated dcp12 < −20 hPa (12 h)−1 in ANHM5 TCs were a weak VWS less than 10 m s−1 (Fig. 11e) and warm SST greater than 28.8°C (Fig. 11f). In contrast, in the YS TCs, the frequencies of dcp12 < −20 hPa (12 h)−1 were relatively small, even if the VWS was weaker than 10 m s−1 (Fig. 11a) and the SST was higher than 28.8°C (Fig. 11b).

Fig. 11.

The frequency of TCs in each joint 12-h CP change and (a) VWS and (b) SST bins for L900 YS TCs from t = 0 to 72 h. (c),(d) As in (a),(b), but for H900 TCs. (e)–(h) As in (a)–(d), but for ANHM5 TCs.

Fig. 11.

The frequency of TCs in each joint 12-h CP change and (a) VWS and (b) SST bins for L900 YS TCs from t = 0 to 72 h. (c),(d) As in (a),(b), but for H900 TCs. (e)–(h) As in (a)–(d), but for ANHM5 TCs.

The simulated TCs tended to attain a minimum central pressure that corresponded to the E-MPI during the integration period, irrespective of the choice of models, though there were large differences in the rates of intensification between the models. Recently, Bryan and Rotunno (2009) have shown the effects of gradient wind imbalance in the PBL on maximum intensity. During RI, the radial gradient of near-surface inflow in the eyewall region was steep in ANHM5 TCs (Fig. 6d). Indeed, strong supergradient flow appeared around the leading edge of the near-surface inflow inside the RMW in the ANHM5 TCs, whereas there was no intense unbalanced flow in the YS TCs (not shown). Whether or not a gradient wind imbalance in the TC PBL was crucial for the spinup, the AGCM20 underwent gradual intensification while maintaining gradient wind balances.

4. Discussion

a. The shallow updraft in YS

Composite YS TCs exhibited a slantwise secondary circulation with maximum updrafts below 6 km after the greatest intensifying phase (Figs. 6, 7, and 8), although the observational studies showed that intense TCs were characterized by deep and intense eyewall updrafts and that the maximum updraft occurred above an altitude of 10 km (e.g., Rogers et al. 2013).

The relatively shallow updrafts in YS TCs were most likely due to the representation of precipitation processes. Figure 12 compares the temporal evolutions of mean precipitation within a radius of 200 km between L900 and H900 TCs simulated by the AGCM20 and ANHM5. Mean precipitation was larger for RI TCs than for non-RI TCs in both L900 and H900 TCs. This result suggests that there was a large release of latent heat in RI TCs. Before YS TCs attained a central pressure of 950 hPa (from t = −30 to −6 h), most of the precipitation was produced by the cumulus convection scheme. At that time, the maximum azimuthal-mean updraft appeared at altitudes around 10 km. As YS TCs intensified, however, the proportion of precipitation produced by the cumulus convection scheme relative to the total precipitation decreased as the altitudes of the maximum updraft rapidly decreased. The precipitation produced by large-scale condensation then became dominant when the altitudes of the maximum updraft decreased to less than 5 km. These results indicate a close relationship between large-scale condensation, midlevel warming, and the altitudes of the updraft maxima in YS TCs.

Fig. 12.

Temporal evolution of mean precipitation for (a) L900 RI (green dashed lines with dots) and non-RI (aqua dotted lines with circles) YS TCs and L900 ANHM5 TCs (red lines with dots) and (b) H900 RI (blue dashed lines with dots) and non-RI (purple dotted lines with circles) YS TCs and L900 ANHM5 TCs (orange lines with dots). (c),(d) As in (a),(b), but for ratio of precipitation of the cumulus convection scheme to total precipitation (pconv ratio). (e),(f) As in (a),(b), but for mean altitude of the maximum azimuthal-mean updraft. Error bars indicate standard deviations. The ratio of precipitation of the cumulus convection scheme to total precipitation is only shown for YS TCs.

Fig. 12.

Temporal evolution of mean precipitation for (a) L900 RI (green dashed lines with dots) and non-RI (aqua dotted lines with circles) YS TCs and L900 ANHM5 TCs (red lines with dots) and (b) H900 RI (blue dashed lines with dots) and non-RI (purple dotted lines with circles) YS TCs and L900 ANHM5 TCs (orange lines with dots). (c),(d) As in (a),(b), but for ratio of precipitation of the cumulus convection scheme to total precipitation (pconv ratio). (e),(f) As in (a),(b), but for mean altitude of the maximum azimuthal-mean updraft. Error bars indicate standard deviations. The ratio of precipitation of the cumulus convection scheme to total precipitation is only shown for YS TCs.

Heating in the lower troposphere leads to a shallow secondary circulation according to studies of balanced vortex responses (e.g., Holland and Merrill 1984). In the AGCM20 large-scale cloud scheme that we used, diabatic cooling was limited to greater than or equal to −10 K day−1 to avoid computational instability (H. Kawai 2016, personal communication). In addition, the current scheme treated only cloud water, ice content, and the fraction of cloud cover as prognostic variables. A recent modeling study by Miller et al. (2015) showed that the latent heat of fusion had large positive impacts on the rate of intensification, upper-level warming, and high altitudes of the updraft maxima for Hurricane Wilma (2005). The introduction of more detailed cloud microphysics is important because the distributions of water substances have large impacts on the heating profiles of the TC inner core (McFarquhar et al. 2012).

The configuration of the horizontal resolution of a model that simulates TCs is another important factor that impacts the inner-core structures and rates of intensification. Nakano et al. (2017) have compared TC inner-core structures simulated with a 20-km-mesh GCM and a 7-km-mesh GCM with the same model configuration. They showed that the secondary circulation of the TCs was shallower and weaker when simulated with the 20-km-mesh GCM than with the 7-km-mesh GCM. Kanada and Wada (2016) have reported similar results. They conducted sensitivity experiments by using nonhydrostatic models with horizontal resolutions of 20, 10, 5, and 2 km and found that the eyewall updrafts became more intense and much deeper as the resolution of the model became finer. Fierro et al. (2009) have suggested that a 5-km-mesh model can simulate TC inner-core structures. Although previous studies have suggested that a model with a higher resolution is required to represent the inner-core structure and associated processes of intense TCs (e.g., Braun and Tao 2000; Gentry and Lackmann 2010), the relationship between model resolution and intensification processes is very complicated, owing to the interactions between physics and dynamics.

Whether deep and intense updrafts are essential for triggering RI is still being debated (e.g., Montgomery et al. 2006; Kieper and Jiang 2012; Sanger et al. 2014; Alvey et al. 2015). Nonetheless, our results indicated that AGCM20 TCs had shallow eyewall updrafts, exhibited prolonged gradual intensification, and resulted in significant northward bias. In contrast, ANHM5 TCs had deep and intense updrafts, underwent RI, and simulated the climatological characteristics of intense TCs. Because the simulated heating profile had a large impact on the inner-core structure, better representation of the climatological characteristics of intense AGCM20 TCs will require that the simulation of precipitation processes be improved and that understanding of the role of the heating profiles on rates of intensification be clarified.

b. Air–sea interactions

Our results indicate that a nonhydrostatic atmospheric model with a horizontal resolution of 5 km reduced the significant northward bias of the maximum intensity latitudes in AGCM20 TCs. However, the representation of oceanic effects had crucial impacts on the simulated maximum TC intensity and the latitudes.

The oceanic response is well known to be a major factor controlling TC activity (e.g., Emanuel 2003; Shay et al. 2000). Recently, Ogata et al. (2015) conducted a climate run by using an atmosphere–ocean coupled GCM with a horizontal resolution of 60 km (AOGCM60). They compared the results with an atmospheric GCM with a horizontal resolution of 60 km (AGCM60) and showed that the frequency of intense AOGCM60 TCs decreased around 20°–30°N. The reduction was attributed to TC-induced cooling of the SST and subsurface temperature. Similar sea surface cooling associated with the passage of TCs has been observed with in situ profiling floats at 1-day intervals for three WNP TCs (Wada et al. 2014).

Although the ANHM5 reproduced the inner-core structures of the intensifying TCs, the simulated mean central pressure after the TCs became mature was considerably lower than the mean best-track central pressure (Fig. 5). Similar overdevelopments have appeared in a modeling study with a horizontal resolution less than or equal to 5 km (Fierro et al. 2009). In our study, all simulations were conducted with atmospheric models that were not coupled to an ocean model; the effect of the cold wake on the TCs (e.g., Wada et al. 2014) was therefore not considered. The mean minimum central pressure latitude in YS TCs appeared at 20°–30°N (Figs. 1 and 2), where the warm mixed layer is shallow, and the effect of cooling after TC passage was therefore relatively large. The introduction of ocean coupling could help lessen the overestimation of TC intensity forecasts (Ito et al. 2015).

As a next step, climate runs using an atmosphere–ocean coupled GCM with a horizontal resolution of 20 km (AOGCM20) should be conducted. Intercomparisons between the AGCM20 and AOGCM20 will shed light on the impact of the response of the ocean on TC intensity.

5. Summary

We examined the characteristics of 34 intense tropical cyclones (TCs) simulated by a atmospheric general circulation model with a horizontal resolution of 20 km (AGCM20) and a regional atmospheric nonhydrostatic model with a horizontal resolution of 5 km (ANHM5).

Whereas the AGCM20 simulated the number of intense TCs as the best-track datasets, the horizontal distributions of the simulated maximum intensity locations were shifted significantly northward compared with the locations analyzed in the two best-track datasets. About half of the AGCM20 TCs attained their maximum intensity north of 25°N, whereas most of the best-track maximum intensity locations were south of 25°N. This bias was attributed to the prolonged gradual intensification of the AGCM20 TCs. In contrast, most of the ANHM5 TCs underwent rapid intensification and attained their maximum intensity at lower latitudes. The results of the downscaling experiments with the ANHM5 thus revealed the possibility of reducing the northward bias of the simulated maximum intensity locations.

The inner-core structures of the ANHM5 TCs were characterized by deep and intense updrafts with upper-level warming. When the ANHM5 TCs traveled to regions with warm SSTs and weak VWSs, they began to rapidly intensify as the radius of their maximum wind speed rapidly decreased. Meanwhile, the AGCM20 TCs had very weak and shallow eyewall updrafts with the maxima below an altitude of 6 km. During the phase of greatest intensification, the altitudes of the updraft maxima decreased rapidly, during which time there was considerable midlevel warming. The AGCM20 TCs underwent prolonged gradual intensifications that resulted in significant northward shifts of their maximum intensity locations. The descent of the altitudes of the updraft maxima was associated with a transition of major precipitation processes from cumulus parameterization to a large-scale cloud scheme. The AGCM20 TCs continued to intensify with their large-scale cloud schemes, even though environmental conditions were less favorable for TC development.

Our results indicate that close attention should be paid when the climatological characteristics of intense TCs are simulated with models such as the AGCM20 in climate runs. In particular, attention should be paid to the physics of the model that impact the heating profiles of intense TCs at various horizontal resolutions. More importantly, the impact of the warming profile and secondary circulation on intensification processes should be clarified. The E-MPI is a good metric for examining TC intensity and the process of intensification for each model, but the dynamics, thermodynamics, and inner-core structures of simulated TCs differed greatly between the AGCM20 and ANHM5. Moreover, the development of atmosphere–ocean coupled models will be essential if we are to obtain more reliable information on the climatological characteristics of TCs in climate runs.

Acknowledgments

The authors are grateful to three anonymous reviewers and Dr. S. L. Sessions for instructive comments. This study was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan under the framework of the Sousei Program and the Japan Society for the Promotion of Science KAKENHI Grants 26400466 and 15K05292. Numerical simulations were performed using the Earth Simulator.

REFERENCES

REFERENCES
Alvey
,
G. R.
, III
,
J.
Zawislak
, and
E.
Zipser
,
2015
:
Precipitation properties observed during tropical cyclone intensity change
.
Mon. Wea. Rev.
,
143
,
4476
4492
, doi:.
Arakawa
,
A.
, and
W. H.
Schubert
,
1974
:
Interaction of cumulus cloud ensemble with the large-scale environment, part I
.
J. Atmos. Sci.
,
31
,
674
701
, doi:.
Bengtsson
,
L.
,
K. I.
Hodges
,
M.
Esch
,
N.
Keenlyside
,
L.
Kornblueh
,
J.-J.
Luo
, and
T.
Yamagata
,
2007
:
How may tropical cyclones change in a warmer climate?
Tellus
,
59A
,
539
561
, doi:.
Braun
,
S. A.
,
2002
:
A cloud-resolving simulation of Hurricane Bob (1991): Storm structure and eyewall buoyancy
.
Mon. Wea. Rev.
,
130
,
1573
1592
, doi:.
Braun
,
S. A.
, and
W.-K.
Tao
,
2000
:
Sensitivity of high-resolution simulations of Hurricane Bob (1991) to planetary boundary layer parameterizations
.
Mon. Wea. Rev.
,
128
,
3941
3961
, 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.
,
C.
Landsea
, and
Y.
Lin
,
2014
:
The influence of tropical cyclone size on its intensification
.
Wea. Forecasting
,
29
,
582
590
, doi:.
Chen
,
H.
, and
D.-L.
Zhang
,
2013
:
On the rapid intensification of Hurricane Wilma (2005). Part II: Convective bursts and the upper-level warm core
.
J. Atmos. Sci.
,
70
,
146
162
, doi:.
Emanuel
,
K. A.
,
1986
:
An air–sea interaction theory for tropical cyclones. Part I: Steady-state maintenance
.
J. Atmos. Sci.
,
43
,
585
605
, doi:.
Emanuel
,
K. A.
,
1995
:
Sensitivity of tropical cyclones to surface exchange coefficients and a revised steady-state model incorporating eye dynamics
.
J. Atmos. Sci.
,
52
,
3969
3976
, doi:.
Emanuel
,
K. A.
,
2003
:
Tropical cyclones
.
Annu. Rev. Earth Planet. Sci.
,
31
,
75
104
, doi:.
Emanuel
,
K. A.
,
2006
: Subroutine to calculate the maximum wind speed and minimum central pressure achievable in tropical cyclone, given a sounding and a sea surface temperature. [Available online at ftp://texmex.mit.edu/pub/emanuel/TCMAX/pcmin_revised.f.]
Fierro
,
A. O.
,
R. F.
Rogers
,
F. D.
Marks
, and
D. S.
Nolan
,
2009
:
The impact of horizontal grid spacing on the microphysical and kinematic structures of strong tropical cyclones simulated with the WRF-ARW model
.
Mon. Wea. Rev.
,
137
,
3717
3743
, doi:.
Gentry
,
M. S.
, and
G. M.
Lackmann
,
2010
:
Sensitivity of simulated tropical cyclone structure and intensity to horizontal resolution
.
Mon. Wea. Rev.
,
138
,
688
704
, doi:.
Gray
,
W. M.
,
1968
:
Global view of the origin of tropical disturbances and storms
.
Mon. Wea. Rev.
,
96
,
669
700
, doi:.
Hack
,
J. J.
, and
W. H.
Schubert
,
1986
:
Nonlinear response of atmospheric vortices to heating by organized cumulus convection
.
J. Atmos. Sci.
,
43
,
1559
1573
, doi:.
Holland
,
G. J.
,
1997
:
The maximum potential intensity of tropical cyclones
.
J. Atmos. Sci.
,
54
,
2519
2541
, doi:.
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:.
Holliday
,
C. R.
, and
A. H.
Thompson
,
1979
:
Climatological characteristics of rapidly intensifying typhoons
.
Mon. Wea. Rev.
,
107
,
1022
1034
, doi:.
Ito
,
K.
,
T.
Kuroda
,
K.
Saito
, and
A.
Wada
,
2015
:
Forecasting a large number of tropical cyclone intensities around Japan using a high-resolution atmosphere–ocean coupled model
.
Wea. Forecasting
,
30
,
793
808
, doi:.
JMA
,
2007
: Outline of the operational numerical weather prediction at the Japan Meteorological Agency. Appendix, WMO Numerical Weather Prediction Progress Rep., 194 pp. [Available online at http://www.jma.go.jp/jma/jma-eng/jma-center/nwp/outline-nwp/index.htm.]
Kain
,
J.
, and
J.
Fritsch
,
1993
: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 24, Amer. Meteor. Soc., 165–170, doi:.
Kanada
,
S.
, and
A.
Wada
,
2016
:
Sensitivity to horizontal resolution of the simulated intensifying rate and inner-core structure of Typhoon Ida, an extremely intense typhoon
.
J. Meteor. Soc. Japan
,
94A
,
181
190
, doi:.
Kanada
,
S.
,
A.
Wada
, and
M.
Sugi
,
2013
:
Future changes in structures of extremely intense tropical cyclones using a 2-km mesh nonhydrostatic model
.
J. Climate
,
26
,
9986
10 005
, 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:.
Kepert
,
J. D.
,
2006a
:
Observed boundary layer wind structure and balance in the hurricane core. Part I: Hurricane Georges
.
J. Atmos. Sci.
,
63
,
2169
2193
, doi:.
Kepert
,
J. D.
,
2006b
:
Observed boundary layer wind structure and balance in the hurricane core. Part II: Hurricane Mitch
.
J. Atmos. Sci.
,
63
,
2194
2211
, doi:.
Kieper
,
M.
, and
H.
Jiang
,
2012
:
Predicting tropical cyclone rapid intensification using the 37 GHz ring pattern identified from passive microwave measurements
.
Geophys. Res. Lett.
,
39
,
L13804
, doi:.
Kitagawa
,
H.
,
2000
: Radiation processes (in Japanese). JMA/Numerical Prediction Division Rep. 46, 16–31.
Knaff
,
J. A.
, and
R. M.
Zehr
,
2007
:
Reexamination of tropical cyclone wind–pressure relationships
.
Wea. Forecasting
,
22
,
71
88
, doi:.
Kossin
,
J. P.
,
K. A.
Emanuel
, and
G. A.
Vecchi
,
2014
:
The poleward migration of the location of tropical cyclone maximum intensity
.
Nature
,
509
,
349
352
, doi:.
Lander
,
M.
,
C.
Guard
, and
S. J.
Camargo
,
2014
:
Super Typhoon Haiyan [in “State of the Climate in 2013”]
.
Bull. Amer. Meteor. Soc.
,
95
(
7
),
S112
S114
.
Li
,
T.
,
M.
Kwon
,
M.
Zhao
,
J.-S.
Kug
,
J.-J.
Luo
, and
W.
Yu
,
2010
:
Global warming shifts Pacific tropical cyclone location
.
Geophys. Res. Lett.
,
37
,
L21804
, doi:.
Lin
,
I.-I.
, and Coauthors
,
2013
:
An ocean coupling potential intensity index for tropical cyclones
.
Geophys. Res. Lett.
,
40
,
1878
1882
, doi:.
Louis
,
J. F.
,
M.
Tiedtke
, and
J. F.
Geleyn
,
1982
: A short history of the operational PBL parameterization at ECMWF. Proc. Workshop on Planetary Boundary Layer Parameterization, Reading, United Kingdom, ECMWF, 59–79. [Available online at http://www.ecmwf.int/sites/default/files/elibrary/1982/10845-short-history-pbl-parameterization-ecmwf.pdf.]
McFarquhar
,
G. M.
,
B. F.
Jewett
,
M. S.
Gilmore
,
S. W.
Nesbitt
, and
T.-L.
Hsieh
,
2012
:
Vertical velocity and microphysical distributions related to rapid intensification in a simulation of Hurricane Dennis (2005)
.
J. Atmos. Sci.
,
69
,
3515
3534
, doi:.
Mellor
,
G. L.
, and
T.
Yamada
,
1974
:
A hierarchy of turbulence closure models for planetary boundary layers
.
J. Atmos. Sci.
,
31
,
1791
1806
, doi:.
Miller
,
W.
,
H.
Chen
, and
D. L.
Zhang
,
2015
:
On the intensification of Hurricane Wilma (2005). Part III: Effects of latent heat of fusion
.
J. Atmos. Sci.
,
72
,
3829
3849
, doi:.
Mizuta
,
R.
, and Coauthors
,
2006
:
20-km-mesh global climate simulations using JMA-GSM model—Mean climate states
.
J. Meteor. Soc. Japan
,
84
,
165
185
, doi:.
Mizuta
,
R.
, and Coauthors
,
2012
:
Climate simulations using MRI-AGCM with 20-km grid
.
J. Meteor. Soc. Japan
,
90A
,
233
260
, doi:.
Montgomery
,
M. T.
, and
R. K.
Smith
,
2014
:
Paradigms for tropical cyclone intensification
.
Aust. Meteor. Oceanogr. J.
,
64
,
37
66
.
Montgomery
,
M. T.
,
M. E.
Nicholls
,
T. A.
Cram
, and
A. B.
Saunders
,
2006
:
A vortical hot tower route to tropical cyclogenesis
.
J. Atmos. Sci.
,
63
,
355
386
, doi:.
Murakami
,
H.
,
2014
:
Tropical cyclones in reanalysis data sets
.
Geophys. Res. Lett.
,
41
,
2133
2141
, doi:.
Murakami
,
H.
, and
M.
Sugi
,
2010
:
Effect of model resolution on tropical cyclone climate projections
.
SOLA
,
6
,
73
76
.
Murakami
,
H.
, and Coauthors
,
2012
:
Future changes in tropical cyclone activity projected by the new high-resolution MRI-AGCM
.
J. Climate
,
25
,
3237
3260
, doi:.
Murakami
,
M.
,
1990
:
Numerical modeling of dynamical and microphysical evolution of an isolated convective cloud—The 19 July 1981 CCOPE cloud
.
J. Meteor. Soc. Japan
,
68
,
107
128
.
Nakanishi
,
M.
, and
H.
Niino
,
2004
:
An improved Mellor–Yamada level-3 model with condensation physics: Its design and verification
.
Bound.-Layer Meteor.
,
112
,
1
31
, doi:.
Nakano
,
M.
,
T.
Kato
,
S.
Hayashi
,
S.
Kanada
,
Y.
Yamada
, and
K.
Kurihara
,
2012
:
Development of a 5-km-mesh cloud-system-resolving regional climate model at the Meteorological Research Institute
.
J. Meteor. Soc. Japan
,
90A
,
339
350
, doi:.
Nakano
,
M.
, and Coauthors
,
2017
:
Global 7-km mesh nonhydrostatic Model Intercomparison Project for Improving Typhoon Forecast (TYMIP-G7): Experimental design and preliminary results
.
Geosci. Model Dev. Discuss.
, doi:, in press.
Ogata
,
T.
,
R.
Mizuta
,
Y.
Adachi
,
H.
Murakami
, and
T.
Ose
,
2015
:
Effect of air–sea coupling on the frequency distribution of intense tropical cyclones over the northwestern Pacific
.
Geophys. Res. Lett.
,
42
,
10 415
10 421
, doi:.
Park
,
D.-S. R.
,
C.-H.
Ho
,
J.-H.
Kim
, and
H.-S.
Kim
,
2011
:
Strong landfall typhoons in Korea and Japan in a recent decade
.
J. Geophys. Res.
,
116
,
D07105
, doi:.
Peduzzi
,
P.
,
B.
Chatenoux
,
H.
Dao
,
A.
DeBono
,
C.
Herold
,
J.
Kossin
,
F.
Mouton
, and
O.
Nordbeck
,
2012
:
Global trends in tropical cyclone risk
.
Nat. Climate Change
,
2
,
289
294
, doi:.
Randall
,
D.
, and
D.-M.
Pan
,
1993
: Implementation of the Arakawa–Schubert cumulus parameterization with a prognostic closure. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 145–150.
Rayner
,
N.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
, 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:.
Roberts
,
M. J.
, and Coauthors
,
2015
:
Tropical cyclones in the UPSCALE ensemble of high-resolution global climate models
.
J. Climate
,
28
,
574
596
, doi:.
Rogers
,
R. F.
,
P.
Reasor
, and
S.
Lorsolo
,
2013
:
Airborne Doppler observations of the inner-core structural differences between intensifying and steady-state tropical cyclones
.
Mon. Wea. Rev.
,
141
,
2970
2991
, doi:.
Saito
,
K.
,
J.
Ishida
,
K.
Aranami
,
T.
Hara
,
T.
Segawa
,
M.
Narita
, and
Y.
Honda
,
2007
:
Nonhydrostatic atmospheric models and operational development at JMA
.
J. Meteor. Soc. Japan
,
85B
,
271
304
, doi:.
Sanger
,
N. T.
,
M. T.
Montgomery
,
R. K.
Smith
, and
M. M.
Bell
,
2014
:
An observational study of tropical cyclone spinup in Supertyphoon Jangmi (2008) from 24 to 27 September
.
Mon. Wea. Rev.
,
142
,
3
28
, doi:.
Schenkel
,
B. A.
, and
R. E.
Hart
,
2012
:
An examination of tropical cyclone position, intensity, and intensity life cycle within atmospheric reanalysis datasets
.
J. Climate
,
25
,
3453
3474
, doi:.
Shay
,
L. K.
,
G. J.
Goni
, and
P. G.
Black
,
2000
:
Effects of a warm oceanic feature on Hurricane Opal
.
Mon. Wea. Rev.
,
128
,
1366
1383
, doi:.
Shibata
,
K.
, and
T.
Aoki
,
1989
:
An infrared radiative scheme for the numerical models of weather and climate
.
J. Geophys. Res.
,
94
,
14 923
14 943
, doi:.
Shibata
,
K.
, and
A.
Uchiyama
,
1992
:
Accuracy of the delta-four-stream approximation in inhomogeneous scattering atmospheres
.
J. Meteor. Soc. Japan
,
70
,
1097
1109
.
Smith
,
R. N. B.
,
1990
:
A scheme for predicting layer clouds and their water content in a general circulation model
.
Quart. J. Roy. Meteor. Soc.
,
116
,
435
460
, doi:.
Stern
,
D. P.
, and
F.
Zhang
,
2016
:
The warm-core structure of Hurricane Earl (2010)
.
J. Atmos. Sci.
,
73
,
3305
3328
, doi:.
Stern
,
D. P.
,
J. L.
Vigh
,
D. S.
Nolan
, and
F.
Zhang
,
2015
:
Revisiting the relationship between eyewall contraction and intensification
.
J. Atmos. Sci.
,
72
,
1283
1306
, doi:.
Stowasser
,
M.
,
Y.
Wang
, and
K.
Hamilton
,
2007
:
Tropical cyclone changes in the western North Pacific in a global warming scenario
.
J. Climate
,
20
,
2378
2396
, doi:.
Sugi
,
M.
,
H.
Murakami
, and
J.
Yoshimura
,
2009
:
A reduction in global tropical cyclone frequency due to global warming
.
SOLA
,
5
,
164
167
, doi:.
Tang
,
B.
, and
K.
Emanuel
,
2010
:
Midlevel ventilation’s constraint on tropical cyclone intensity
.
J. Atmos. Sci.
,
67
,
1817
1830
, doi:.
Tang
,
B.
, and
K.
Emanuel
,
2012
:
Sensitivity of tropical cyclone intensity to ventilation in an axisymmetric model
.
J. Atmos. Sci.
,
69
,
2394
2413
, doi:.
Tiedtke
,
M.
,
1989
:
A comprehensive mass flux scheme for cumulus parameterization in large-scale models
.
Mon. Wea. Rev.
,
117
,
1779
1800
, doi:.
Tiedtke
,
M.
,
1993
:
Representation of clouds in large-scale models
.
Mon. Wea. Rev.
,
121
,
3040
3061
, doi:.
Vigh
,
J. L.
, and
W. H.
Schubert
,
2009
:
Rapid development of the tropical cyclone warm core
.
J. Atmos. Sci.
,
66
,
3335
3350
, doi:.
Wada
,
A.
,
T.
Uehara
, and
S.
Ishizaki
,
2014
:
Typhoon-induced sea surface cooling during the 2011 and 2012 typhoon seasons: Observational evidence and numerical investigations of the sea surface cooling effect using typhoon simulations
.
Prog. Earth Planet. Sci.
,
1
,
11
, doi:.
Walsh
,
K. J. E.
,
M.
Fiorino
,
C. W.
Landsea
, and
K. L.
McInnes
,
2007
:
Objectively determined resolution-dependent threshold criteria for the detection of tropical cyclones in climate models and reanalyses
.
J. Climate
,
20
,
2307
2314
, 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:.
Yabu
,
S.
,
S.
Murai
, and
H.
Kitagawa
,
2005
: Clear sky radiation scheme (in Japanese). JMA/Numerical Prediction Division Rep. 51, 53–64.
Yoshimura
,
H.
,
R.
Mizuta
, and
H.
Murakami
,
2015
:
A spectral cumulus parameterization scheme interpolating between two convective updrafts with semi-Lagrangian calculation of transport by compensatory subsidence
.
Mon. Wea. Rev.
,
143
,
597
621
, doi:.
Zhang
,
J. A.
,
R. F.
Rogers
,
D. S.
Nolan
, and
F. D.
Marks
,
2011
:
On the characteristic height scales of the hurricane boundary layer
.
Mon. Wea. Rev.
,
139
,
2523
2535
, doi:.

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).