1. Introduction
Tropical cyclone (TC) intensity and intensity change has been a major research focus in recent years. However, the size or structure of a TC is also an important issue. Lessons learned from Typhoon Toraji (2001) in the Taiwan area indicate that a compact typhoon, which is characterized by small radius of maximum wind (RMW) and large rate of decrease of wind speed outside the core (Chen et al. 2011), may be accompanied by heavy rainfall over a small region that makes quantitative precipitation forecast a challenging issue. Note that the direct hit by a TC with intense structure can result in serious disasters but the impact scenario depends on the landfall position very much. Following the preliminary investigation in Chen et al. (2011), this study focuses on the synoptic-dynamical characteristics of TCs in the western North Pacific (WNP) with compact structure, which differs substantially from the majority of other TC vortices with extended rainband structures.
A better prediction of the TC structure especially in size and strength is crucial for disaster mitigation strategy. A good example is with respect to aviation safety. A crosswind of only 20–30 kt (1 kt = 0.514 m s−1) causes not much damage by itself but may be enough to prevent aircraft from taking off and escaping the much stronger winds to come (Holland, 2009). To the airport manager, the forecast of the extent of 30-kt winds around a TC and their anticipated time of arrival at the runway area is important. Another example is the westward- or northwestward-moving TCs that pass through the region north of Taiwan but without making landfall. The radius of gale-force wind (RGFW) forecast is quite important for this type of TCs. The cases with smaller RGFW may not cause damage at all whereas those with larger RGFW may affect Taiwan’s northern coastal areas.
Furthermore, it is now realized that the Saffir–Simpson Hurricane Scale (SSHS) may not be fully adequate to estimate the magnitude of possible disasters, instead new methodologies have been developed such as the integrated kinetic energy (Kantha 2006; Powell and Reinhold 2007; Maclay et al. 2008). These methods can be used to estimate the potential level of disasters in most cases, but the studies of Powell and Reinhold (2007) and Maclay et al. (2008) showed that even the integrated kinetic energy is inadequate to represent the destructive potential of small and intense storms. Such a unique structure of compact TCs is the focus in this study.
Knaff et al. (2003, 2008) analyzed another group of TCs, termed annular hurricanes (AHs), which share some common features (e.g., few or no rainband structures) with compact TCs. In contrast to compact TCs, AHs have eye sizes larger than the average value with symmetrically distributed eyewall convection, which implies that they have a fairly large radius of maximum wind. Statistically, AHs tend to maintain their peak intensity longer and weaken more slowly than the average hurricanes. Similar to the AHs, the compact TCs are seldom accompanied by active rainbands, but at the same time they possess relatively small eye radii. Compact TCs and AHs have average intensities closer to the average maximum potential intensities (MPIs) during intensification to typhoons or supertyphoons compared to those for incompact cases (Chen et al. 2011), which imply a higher efficiency of intensification in compact TCs and AHs. The large-eye structure in AHs can be reproduced in numerical simulations (Zhang et al. 2005; Wang 2008). In particular, Wang (2008) showed that the formation of AH was closely related to the interaction between the inner spiral rainbands and the eyewall convection.
A number of recent studies that were based on idealized simulations examined the sensitivity of TC size to environmental humidity, initial vortex size, and surface entropy flux (Hill and Lackmann 2009; Xu and Wang 2010a,b). The results from these studies generally showed that moist environment and large initial vortex size favor the development of larger TCs. However, not much has been explored on the impacts from dynamic external forcing on TC structure, which is a point to address in this study. In the following, section 2 describes the datasets used in this study, the numerical model used for simulations, and reviews the structural parameter S. Section 3 describes the verification and analyses of model simulations. Several sensitivity experiments that investigate the dependence of the compact TC structure on environmental conditions are presented in section 4. Section 5 provides a summary and discussion.
2. Data and methodology
a. Data
The positions and intensities of the involved TCs at 6-h intervals are obtained from the best-track dataset of the Joint Typhoon Warning Center (JTWC). The Final Operational Global Analysis (FNL) of the National Centers for Environmental Prediction (NCEP) is used as the initial field of model simulation. This dataset is adopted from the Global Forecast System (GFS) that is operationally run 4 times per day in near–real time at NCEP. The NCEP GFS dataset starts from 1999 July on a 1° latitude–longitude grid continuously at every 6 h. The analyses are available on the surface as well as 26 vertical levels from 1000 to 10 mb.
To compare with the simulation results, the Quick Scatterometer (QuikSCAT) oceanic winds, which are derived from microwave scatterometer data, are used (Boukabara et al. 2002; Weissman et al. 2002). The wind speed and direction at 10-m height are retrieved from measurements of the power of the backscattered radiation. The QuikSCAT wind data correspond to surface winds with a time-averaging period of 8–10 min. The QuikSCAT data are available twice a day starting from June 1999 to November 2009 on a 25-km grid. The QuikSCAT winds generally agree well with those taken by ocean buoys (Ebuchi et al. 2002; Pickett et al. 2003; Sharma and D’Sa 2008). Validation of QuikSCAT wind vector by dropwindsonde data showed that the QuikSCAT wind vectors (non-rain-flagged data) below 17.2 m s−1 are accurate enough for determining the wind radius of 34-kt wind, while the error bound of the QuikSCAT wind estimate in high-wind regimes (above 17.2 m s−1) near a TC is suggested to be about 4 m s−1 (Chou et al. 2010).
b. Model
This study uses the Advanced Research Weather Research and Forecasting Model (WRF) model version 3.1.1 (ARW). The equation set of the WRF is fully compressible, Eulerian, and nonhydrostatic with a run-time hydrostatic option. During the model integration, the scalar variables are conservative. The model uses terrain-following, hydrostatic-pressure vertical coordinates with the top of the model being a constant pressure surface. A detailed description of the WRF has been provided by Skamarock et al. (2005). For the simulations in this study, three nested domains, which are fixed geographically, are employed in the model with two-way interaction. The domains have horizontal resolutions of 45, 15, and 5 km, and grid dimensions of 140 × 140, 283 × 256, and 520 × 589, respectively. There are 28 vertical levels from the surface to 50 hPa. Initial and lateral boundary conditions for the model are from the NCEP FNL. A 4-day simulation initialized at 1200 UTC 17 May is performed for Typhoon Yutu, while a 5-day simulation initialized at 1200 UTC 8 July is performed for Typhoon Manyi.
The precipitation physics schemes used in this study include explicit moisture calculation and cumulus parameterization. The sophisticated microphysics scheme of Lin et al. (1983) that includes ice, snow, and graupel processes is suitable for real data high-resolution simulations. The Kain–Fritsch convective parameterization scheme (Kain and Fritsch 1993) is used to represent the effects of subgrid-scale convection in the two coarser-resolution domains, but is omitted in the fine domain. In all domains, the Yonsei University (YSU) planetary boundary layer (PBL) scheme (Noh et al. 2003) is used to calculate the vertical fluxes of sensible heat, moisture, and momentum in the lower boundary. Regarding radiative processes, the Dudhia shortwave radiation scheme (Dudhia 1989) and the Rapid Radiative Transfer Model (RRTM) longwave radiation scheme (Mlawer et al. 1997) are used. The model is initialized about 72 h before the TC reaches its maximum intensity. No data assimilation and vortex bogussing are adopted during the control simulation.
c. Parameters representing the compactness of a TC
Chen et al. (2011) showed that the compact TC structure is favorable in early intensification stages and the mature stages have more cases with an incompact structure. A relatively higher intensification rate within 24 h can be identified in most of the compact TCs. The composite results of cloud-top temperature also showed that compact TCs have a fairly axisymmetric convection characteristic, while their wind speed increases primarily at the inner-core region as indicated from the QuikSCAT winds. Moreover, it was found in Chen et al. (2011) that the role of low-level environmental forcing is more important in the development of incompact TCs, while mainly internal dynamics are crucial to the development of compact TCs.
A specific note here is that the design of the S parameter is to distinguish the conventional structural parameters such as size and strength, the latter two being a local measure of wind speed or an average wind speed at the outer core. In contrast, the S parameter compares the relative wind structures in the inner-core and outer-core regions (Fig. 2 in Chen et al. 2011), and thus the relative contributions of dynamical processes to inner-core intensification to outer-core strengthening. Consequently, the S parameter represents a new measure of TC structure that is distinguishable from size or strength. There are TC cases with similar size or strength parameter, but with largely different S parameter. Some of these cases have been documented in Chen et al. (2011).
3. Numerical simulations
To help further understand the physical processes responsible for the evolution and maintenance of the compact TC structure, selected cases are simulated using the WRF. In particular, the following discussion will focus on the simulations for Typhoon Yutu (2007) that represents the compact case, and Typhoon Manyi (2007) that is an incompact case.
a. Verification
The simulated tracks and the best tracks for these two TC cases are shown in Figs. 1a,b. As seen, the simulated tracks agree quite well with the best tracks especially in regard to the timing of recurvature. The track errors during the entire simulation period (period before maximum intensity) are 139.5 (111.2) km and 265.3 (126.0) km for Yutu and Manyi, respectively. In addition, the simulated intensities agree with those in the best tracks throughout the simulation period, although the time of attaining maximum intensity for Typhoon Manyi is later than that observed (Figs. 1c,d). The intensification of Typhoon Yutu occurs mainly during the period of 1200 UTC 18 May–1200 UTC 20 May (24–72 h in the control simulation) when the minimum mean sea level pressure (MSLP) of Yutu decreases from 985 to 924 hPa. For Typhoon Manyi, the main intensification period is from 1200 UTC 10 July–1200 UTC 12 July (48–96 h in the control simulation) when the MSLP decreases from 986 to 938 hPa. The analyses in the following sections focus on these intensification periods.
Best (black) and simulated (blue) track of (a) Typhoon Yutu (2007) and (b) Typhoon Manyi (2007). Time series of the minimum central pressure from JTWC (hPa; blue solid line), simulated minimum central pressure (hPa; blue dashed line), simulated S parameter (red dashed line), strength (m s−1; black dashed line) and size (km; green dashed line) for (c),(e) Yutu and (d),(f) Manyi. Dots, triangles, and vertical bars are computed S parameter, strength, and size using QuikSCAT oceanic winds, respectively.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Best (black) and simulated (blue) track of (a) Typhoon Yutu (2007) and (b) Typhoon Manyi (2007). Time series of the minimum central pressure from JTWC (hPa; blue solid line), simulated minimum central pressure (hPa; blue dashed line), simulated S parameter (red dashed line), strength (m s−1; black dashed line) and size (km; green dashed line) for (c),(e) Yutu and (d),(f) Manyi. Dots, triangles, and vertical bars are computed S parameter, strength, and size using QuikSCAT oceanic winds, respectively.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Best (black) and simulated (blue) track of (a) Typhoon Yutu (2007) and (b) Typhoon Manyi (2007). Time series of the minimum central pressure from JTWC (hPa; blue solid line), simulated minimum central pressure (hPa; blue dashed line), simulated S parameter (red dashed line), strength (m s−1; black dashed line) and size (km; green dashed line) for (c),(e) Yutu and (d),(f) Manyi. Dots, triangles, and vertical bars are computed S parameter, strength, and size using QuikSCAT oceanic winds, respectively.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
The S parameter, strength (defined as the average tangential wind speed within the range of 100–300 km from the TC center), and size (defined as the radius of 15 m s−1 near-surface wind as in Lee et al. 2010) are calculated based on the simulations. These simulated values are then compared with those derived from the QuikSCAT oceanic winds whenever the scatterometer data are available. For Typhoon Yutu, the simulated S parameter, strength, and size are all nearly identical to those observed (Figs. 1c,e), which indicate that the structure evolution of this typhoon is well reproduced in the model. On the other hand, the simulated S parameter for Typhoon Manyi fluctuates with time, but the slight increasing trend until about 0000 UTC 13 July matches with that shown in the QuikSCAT data (Fig. 1d). As in the case of Yutu, the simulated strength of Manyi agrees with observations, which indicates that the outer-core wind structures of both compact and incompact TCs are well simulated by WRF with such horizontal resolution. During early development, the simulated size of Typhoon Manyi, which originated from the vortex in the GFS analysis, is too small. However, the simulated vortex increases in size rapidly after 0000 UTC 10 July, and matches with the QuikSCAT-derived size in the subsequent times.
b. Analyses of axisymmetric structures
The radial variations of the azimuthally averaged 850-hPa tangential wind and the associated relative vorticity at different simulation periods for both TC cases are shown in Fig. 2. For Yutu, a rapid intensification occurs from 24 to 48 h (Fig. 2a). Although wind speed increases over a large radial extent up to about 300–400 km, the major increase occurs within a radius of 100 km where the relative vorticity also increases substantially but not outside the RMW (Fig. 2c). For Manyi, the tangential wind increases over a large radial band from 48 to 72 h, with the largest increase within a radius of 300 km (Fig. 2b). Although the major increase in relative vorticity in the same period occurs within the RMW, vorticity outside the RMW also increases (Fig. 2d).
Azimuthally averaged (a),(b) 850-hPa tangential wind (m s−1) and (c),(d) relative vorticity (10−4 s−1) for (a),(c) Typhoon Yutu and (b),(d) Typhoon Manyi. Lines with different colors represent different simulated times as labeled.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Azimuthally averaged (a),(b) 850-hPa tangential wind (m s−1) and (c),(d) relative vorticity (10−4 s−1) for (a),(c) Typhoon Yutu and (b),(d) Typhoon Manyi. Lines with different colors represent different simulated times as labeled.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Azimuthally averaged (a),(b) 850-hPa tangential wind (m s−1) and (c),(d) relative vorticity (10−4 s−1) for (a),(c) Typhoon Yutu and (b),(d) Typhoon Manyi. Lines with different colors represent different simulated times as labeled.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Within the next 24 h (48–72-h simulation for Yutu and 72–96-h for Manyi), the intensities of both TC cases continue to increase while showing different structure changes. The radial profile of tangential wind and relative vorticity for Yutu remain almost unchanged except for an increase in intensity. On the other hand, for Manyi there is a contraction of the RMW from 140 to 100 km when the maximum tangential wind speed increases to about 57 m s−1. Similar contrast in evolution occurs when the two simulated TCs weaken: Yutu maintains its tangential wind profile, while the RMW of Manyi stops contracting and increases. These differences in structural changes between the compact and incompact TCs motivate further analysis of several physical parameters that are related to internal structural evolution.
c. Stability analyses
Holland and Merrill (1984) used composite data to calculate the inertial and static stabilities of TCs. Their results showed that the inner-core inertial stability at the mid- to low level is three orders of magnitude larger than that of the outer region. Low-level inertial stability often increases in the inner core and at the later stages of development. The large radial variation in inertial stability represents the main constraint on the response of TCs to different forcing mechanisms. On the other hand, static stability changes only slightly with radius. However, it has a much larger magnitude at upper-level levels, where the inertial stability is much weaker and more favorable for horizontal movement of air parcels. Owing to the relative distributions of inertial stability and static stability, inner-core convective heating can directly change the intensity of TCs but may influence strength and size only indirectly. Comparatively, the effect of upper-level forcing is easier to reach the inner region of a TC and cause an intensity change. Because of the strong inertial stability at the mid- to low levels of the eye region, it is very difficult for a low-level environmental forcing to affect the inner-core region. Nevertheless, low-level environmental forcing may effectively change the size and strength of TCs.
Following this approach of Holland and Merrill (1984), the inertial and static stabilities are calculated for both TC cases here. The low-level inflow in Typhoon Yutu concentrates at a radius of about 300 km at 24 h of simulation (not shown). From 24 to 48 h, which is the end of the typhoon’s rapid intensification period, inflow increases significantly within 250 km from the center with maximum speed up to 28 m s−1 at 950 hPa (Fig. 3a). Similar to the composite results in Holland and Merrill, the inertial stability is very large near the center at 48 h. Strong outflow occurs at upper levels with a maximum speed of 24 m s−1 at 150 hPa, and a secondary outflow maximum occurs at midlevels with maximum speed of 8 m s−1 at 650 hPa. Within 100 km, the radial winds change from inward below 800 hPa to outward aloft. At the same time, the region with maximum relative vorticity larger than 5 × 10−3 s−1 is located within 50 km and below 800 hPa, which is consistent with the maximum inflow position (Fig. 3b). Furthermore, the divergence field shows a typical pattern of typhoon structure with strong convergence at the lower levels and strong divergence at the upper levels (Fig. 3c).
Radial-pressure (hPa) distribution of the simulated (a) azimuthally averaged radial wind (blue contours; interval: 4 m s−1; dashed lines for inflow), static stability (red contours), inertial stability (shaded), and radial-vertical wind field (arrows) for Typhoon Yutu at 48 h. Simulated azimuthally averaged (b) relative vorticity (10−4 s−1) and (c) divergence (10−4 s−1).
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Radial-pressure (hPa) distribution of the simulated (a) azimuthally averaged radial wind (blue contours; interval: 4 m s−1; dashed lines for inflow), static stability (red contours), inertial stability (shaded), and radial-vertical wind field (arrows) for Typhoon Yutu at 48 h. Simulated azimuthally averaged (b) relative vorticity (10−4 s−1) and (c) divergence (10−4 s−1).
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Radial-pressure (hPa) distribution of the simulated (a) azimuthally averaged radial wind (blue contours; interval: 4 m s−1; dashed lines for inflow), static stability (red contours), inertial stability (shaded), and radial-vertical wind field (arrows) for Typhoon Yutu at 48 h. Simulated azimuthally averaged (b) relative vorticity (10−4 s−1) and (c) divergence (10−4 s−1).
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
The inertial stability for Typhoon Manyi (Fig. 4a) is much weaker than that for Yutu. The inertial stability increases from 72 to 96 h of simulation. The low-level inflow maintains a magnitude of about 4 m s−1 from a radius of 500 km inward to about 100 km at 48 h (not shown). This inflow pattern below 850 hPa persists up to 96 h and is then followed by an inward shift of the maximum wind speed (Fig. 4a). At 48 h, the area with vorticity larger than 2 × 10−4 s−1 extends to well beyond a radius of 160 km. Later on at 72 h, the region with relative vorticity larger than 8 × 10−4 s−1 is located at a radius of 80 km and below 800 hPa, which is associated with a shallow and weak convergence at the low levels (not shown). When the maximum intensity is reached at 96 h, the maximum relative vorticity is about 10 × 10−4 s−1 (Figs. 4b,c).
As in Fig. 3, but for Typhoon Manyi at 96 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 3, but for Typhoon Manyi at 96 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 3, but for Typhoon Manyi at 96 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Comparing the radial flow patterns of these two TC cases to the normalized radial wind responses to different forcings in Holland and Merrill (1984) suggests that the combination of inner-core heating and upper-level forcing plays an important role in Typhoon Yutu’s development. For instance, when Yutu is traveling northwestward in the first 48 h of simulation, it is approaching the upper-level trough to the north where strong west-southwesterlies are found (Fig. 5a). Comparatively, the Yutu vortex is only under the influence of a high pressure system at the low levels but quite far away from the trough to the north (Fig. 5b). For Typhoon Manyi, however, it is more likely that an external lower-level forcing occurs at its outer-core region. A likely candidate is the strong southwesterlies and southerlies that are entering the vortex of Manyi (Fig. 5d). At the upper levels, Manyi seems not to interact with the trough to its north much and thus upper-level forcing is weaker than in the case of Yutu. In short summary, the stability analyses together with examination of synoptic patterns suggest that the intensification of Typhoon Yutu from 24 to 48 h is highly related to internal dynamics, while for Typhoon Manyi the low-level environmental influences play a more important role during the TC’s development.
Simulated (a),(c) 300- and (b),(d) 850-hPa geopotential height (m; contours) and wind barbs for (a),(b) Typhoon Yutu at 48 h and (c),(d) Typhoon Manyi at 72 h of model integration.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Simulated (a),(c) 300- and (b),(d) 850-hPa geopotential height (m; contours) and wind barbs for (a),(b) Typhoon Yutu at 48 h and (c),(d) Typhoon Manyi at 72 h of model integration.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Simulated (a),(c) 300- and (b),(d) 850-hPa geopotential height (m; contours) and wind barbs for (a),(b) Typhoon Yutu at 48 h and (c),(d) Typhoon Manyi at 72 h of model integration.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
4. Model sensitivity tests
To better understand the physical processes responsible for the evolution and maintenance of compact or incompact typhoons, several sensitivity experiments are conducted. These experiments help to address the respective roles of internal and external dynamics, and also environmental humidity in the structure changes of TCs. Previous studies such as Hill and Lackmann (2009) and Xu and Wang (2010a,b) examined the sensitivity of the TC inner-core size to the surface entropy flux, environmental humidity, and initial vortex size based on idealized simulations. Since the structural evolutions of Typhoon Yutu and Manyi have been well simulated by WRF, our sensitivity experiments are designed for the same TC cases.
a. Experimental design
To identify the role of initial vortex structure under different environments, experiments are designed to generate different synthetic vortices, both compact and incompact, in the associated environments of Typhoon Yutu and Manyi. In this regard, Yutu’s environment is considered to be more favorable for evolving to compact structure, whereas Manyi’s environment is more favorable for incompact structure. The built-in bogus scheme of the WRF was used to perform the experiments. For the compact vortex experiments, the intensity (Vmax) of the initial vortex is set as 25, 35, and 45 m s−1 for the V25, V35, and V45 experiments, respectively. The same RMW of 80 km and the same α value of 0.75 are used in each vortex (Table 1). For the incompact vortex experiments, the RMW is increased to 100, 200, and 300 km, but the Vmax and α remain as 35 m s−1 and 0.75, respectively. Thus, there are four possible sets of experiments, each of which combines one of the environmental flows and one of the synthetic vortices. These experiments are termed EYutu–Vcompact, EYutu–Vincompact, EManyi–Vcompact and EManyi–Vincompact (E for environment and V for vortex).
Model design of the sensitivity experiments with respect to (top) environmental flow and (bottom) environmental humidity.
To better understand the influence of environmental humidity to the TC structure, the RH at a specified radial domain is changed in additional sensitivity experiments. The model initial RH is shown in Fig. 6 for Yutu and Manyi, respectively. It can be seen that the humid area of Yutu from 500 to 700 hPa (model levels 11–14) is confined at the inner-core region. However, the area with RH greater than 80% in Manyi occupies a large domain at midlevels especially south of the center. These results indicate that the environment associated with Yutu’s development is relatively dry compared to that of Manyi’s.
(a) 700- and (b) 500-hPa relative humidity (%, shaded) at 1200 UTC 17 May 2007 in the initial conditions for simulating Typhoon Yutu. (c),(d) As in (a),(b), but for Typhoon Manyi at 1200 UTC 8 Jul 2007. Axes shown are grid points of the domain, which is 2700 × 2700 km2 in size.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
(a) 700- and (b) 500-hPa relative humidity (%, shaded) at 1200 UTC 17 May 2007 in the initial conditions for simulating Typhoon Yutu. (c),(d) As in (a),(b), but for Typhoon Manyi at 1200 UTC 8 Jul 2007. Axes shown are grid points of the domain, which is 2700 × 2700 km2 in size.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
(a) 700- and (b) 500-hPa relative humidity (%, shaded) at 1200 UTC 17 May 2007 in the initial conditions for simulating Typhoon Yutu. (c),(d) As in (a),(b), but for Typhoon Manyi at 1200 UTC 8 Jul 2007. Axes shown are grid points of the domain, which is 2700 × 2700 km2 in size.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
The humidity-related sensitivity experiments are designed as following. The +20L (L for large region) experiment increases the RH of the control experiment by 20% within 180 km. The RH then decreases linearly from 180 to 720 km where the RH value is the same as in the control experiment. There is no change of RH outside a radius of 720 km. In the same manner, the RH of the −20S (S for small region) experiment is reduced by 20% within 90 km and then increases linearly with radius to 360 km where the RH is the same as in the control run. The modification of humidity is performed for each level according to the original humidity such that the vertical distribution remains largely unchanged. If the modified humidity is larger than 100%, it is set to 100%. Thus, each TC case has four combinations of experiments that increase or decrease the RH within a large or small region, plus the amount of RH perturbation can be changed.
b. Sensitivity to environment condition
The simulated tracks in the sensitivity experiments for Typhoon Yutu show that the moving speed is proportional to the Vmax or RMW of the initial vortex in both compact vortex and incompact vortex experiments (Figs. 7a,b). For the environment associated with Typhoon Manyi, the movement speeds of the compact vortices are all similar (Fig. 7c), while those of the incompact vortices depend on the RMW (Fig. 7d). Since the incompact TC is also large in size, these results indicate that relatively large TC vortices may move with larger deviations from the steering flow especially during recurvature (e.g., by more interaction with midlatitude troughs), and consequently their moving speeds become faster.
Best-track (black solid dot) and simulated tracks in the control (black open dot) and sensitivity experiments (red, green, and blue) of (a) EYutu–Vcompact, (b) EYutu–Vincompact, (c) EManyi–Vcompact, and (d) EManyi–Vincompact.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Best-track (black solid dot) and simulated tracks in the control (black open dot) and sensitivity experiments (red, green, and blue) of (a) EYutu–Vcompact, (b) EYutu–Vincompact, (c) EManyi–Vcompact, and (d) EManyi–Vincompact.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Best-track (black solid dot) and simulated tracks in the control (black open dot) and sensitivity experiments (red, green, and blue) of (a) EYutu–Vcompact, (b) EYutu–Vincompact, (c) EManyi–Vcompact, and (d) EManyi–Vincompact.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
The time series of the structure parameters Pmin, S, and size for these experiments are examined. As expected, the intensification of the compact initial vortices within Yutu’s environment depends on the initial Vmax (Fig. 8a). As in the control experiment, the S and size values remain small throughout the simulations (i.e., the compact vortices remain compact; Figs. 8c,e). On the other hand, among the EYutu–Vincompact experiments the initial vortices with larger RMW value have larger S and size values. The corresponding simulations show that these vortices can remain large in S and size even under an environment not favorable for growth in S and size (Figs. 8d,f). That is, although there is no low-level forcing like that during the evolution of Typhoon Manyi, these vortices can still maintain their outer-core structures. The likely reason is that these synthetic vortices have large RMW values, and thus convective heating is available in the near outer-core region to maintain the wind speeds over there.
Time series of the simulated structure parameters (a),(b) Pmin (hPa); (c),(d) S parameter; and (e),(f) size (km) in the (a),(c),(e) EYutu–Vcompact and (b),(d),(f) EYutu–Vincompact sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Time series of the simulated structure parameters (a),(b) Pmin (hPa); (c),(d) S parameter; and (e),(f) size (km) in the (a),(c),(e) EYutu–Vcompact and (b),(d),(f) EYutu–Vincompact sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Time series of the simulated structure parameters (a),(b) Pmin (hPa); (c),(d) S parameter; and (e),(f) size (km) in the (a),(c),(e) EYutu–Vcompact and (b),(d),(f) EYutu–Vincompact sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
For the EManyi–Vcompact experiments, the vortices in all three experiments of V25, V35, and V45 intensify rapidly in the first 30 h and then weaken in the next 12 h. (Fig. 9a). The two with higher initial intensities then intensify again until almost the end of simulations. During the first 54 h, these vortices remain compact (Fig. 9c). After 54 h, the S value in the V25 experiment increases rapidly similar to the control experiment, whereas in the other two experiments (V35 and V45) the S values remain quite small. That is, there is a critical intensity for an initially compact vortex to remain compact under the influence of an environment more favorable for evolving into incompact structures. Note that in these simulations, all three vortices start to increase in size at about 30–36 h (Fig. 9e). This increase in size is due to the abundant low-level forcing as diagnosed in the evolution of Typhoon Manyi that influence the out-core region. However, when the inertial stability within the inner core is large enough such as convective heating is effective to intensify the vortex, the overall structure can remain to be a compact one as in experiments V35 and V45.
As in Fig. 8, but for Typhoon Manyi.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 8, but for Typhoon Manyi.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 8, but for Typhoon Manyi.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
For the EManyi–Vincompact experiments, the evolution of the S parameter is highly sensitive to the initial RMW of the vortex (Fig. 9d). For example, the S parameter of the vortex in the R300 experiment increases rapidly from 24 h to very large values. On contrary, although the vortices in experiments R100 and R200 are incompact at the beginning, their S parameters only increase gradually in the simulations. Similar to the previous set of experiments, the low-level forcing in the environment is increasing the sizes of the vortices especially those in R100 and R200 (since the initial size of the R300 vortex is already nearly 660 km at the initial time). However, if the initial RMW is small, the inner-core convective heating continues to increase the wind speed over there while the external forcing from the environment is changing the outer-core structure. Consequently, the overall radial profile of wind speed does not change much as revealed in the S parameter.
Examining the actual vortex structures in the above experiments confirms the discussion on relative roles of inner- and outer-core processes. The low-level relative vorticity patterns in the EYutu–Vcompact experiments all exhibit highly axisymmetric structures (Figs. 10a,d,f). In R300 of the EYutu–Vincompact experiments, rainband structures start to emerge when the initial RMW of the vortex is large. Interestingly, the eyewalls in this set of experiments depict elliptical to polygonal structures (Figs. 10b,e,g), which is consistent with the fact that the associated environment in these experiments is quite quiescent and that this kind of eyewall structures can develop by internal vortex dynamics (Schubert et al. 1999). On the other hand, spiral bandlike structures are common in both the EManyi–Vcompact and EManyi–Vincompact sets of experiments, especially to the south and southwest of the vortices (Fig. 11). These bands of high relative vorticity are associated with the strong southwesterlies in the environment of Typhoon Manyi, and the vorticity affects the sizes of these vortices that depend on outer-core wind structures.
Simulated 850-hPa relative vorticity (10−4 s−1) at 48 h of the control simulation (CTL), EYutu–Vcompact (V25, V35, and V45), and EYutu–Vincompact (R100, R200, and R300) sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Simulated 850-hPa relative vorticity (10−4 s−1) at 48 h of the control simulation (CTL), EYutu–Vcompact (V25, V35, and V45), and EYutu–Vincompact (R100, R200, and R300) sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Simulated 850-hPa relative vorticity (10−4 s−1) at 48 h of the control simulation (CTL), EYutu–Vcompact (V25, V35, and V45), and EYutu–Vincompact (R100, R200, and R300) sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Manyi at 72 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Manyi at 72 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Manyi at 72 h.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
c. Sensitivity to environmental humidity
The enhanced-RH experiments for Typhoon Yutu show that the system structure becomes slightly more incompact (Fig. 12a). Increasing RH over a larger domain (+20L and +40L) results in more incompact structure than that over a small domain (+20S and +40S). On the other hand, changes in the S parameter with the reduced RH are small (Fig. 12b). That is, the development of a compact TC such as Yutu is more governed by dynamical factors than by thermodynamics ones.
Time series of the simulated S parameter in the environmental humidity sensitivity experiments for (a),(b) Typhoon Yutu and (c),(d) Typhoon Manyi. The “+” (“−”) symbol means increase (decrease) in RH in a large (L) or small (S) region with the percentage change shown as the number. See the text and Table 1 for details.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Time series of the simulated S parameter in the environmental humidity sensitivity experiments for (a),(b) Typhoon Yutu and (c),(d) Typhoon Manyi. The “+” (“−”) symbol means increase (decrease) in RH in a large (L) or small (S) region with the percentage change shown as the number. See the text and Table 1 for details.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
Time series of the simulated S parameter in the environmental humidity sensitivity experiments for (a),(b) Typhoon Yutu and (c),(d) Typhoon Manyi. The “+” (“−”) symbol means increase (decrease) in RH in a large (L) or small (S) region with the percentage change shown as the number. See the text and Table 1 for details.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
The vortex structures of Yutu in the enhanced-humidity experiments are similar to that in the control experiment except that the sizes of the vorticity rings are slight larger (Figs. 13a–e). The maximum magnitudes of vorticity are similar as well. When humidity is increased in the outer-core region region in the +20L and +40L experiments, vorticity band structures to the east of center develop. For the reduced-RH experiments, the −5S experiment has a similar pattern and magnitudes of vorticity as in the control experiment (Fig. 13h). When the RH is reduced more in the −10S and −20S experiments, the values of maximum vorticity are smaller than that in the control experiment as there is less support of deep convection for spinning up the low-level vorticity (Figs. 13f,g).
As in Fig. 10, but for Typhoon Yutu at 72 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Yutu at 72 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Yutu at 72 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
While the original humidity distribution associated with Typhoon Manyi is extended over quite a large region, modifications of humidity over the small domain in experiments +20S and +40S do not change the compactness of the typhoon much (Fig. 12c). However, the experiments with enhanced RH over the large domain (+20L and +40L) result in much larger values of the S parameter, especially during the first 60 h of simulation. On the other hand, reducing the environmental RH is quite effective in decreasing the Typhoon’s compactness (Fig. 12d), which is distinct with the response of Yutu to reduced humidity.
These effects on Manyi’s compactness are due to modified sizes of the major vorticity ring structures in the Typhoon. The four experiments with enhanced RH all experience higher intensification rates and have larger vorticity ring structures than that in the control (Figs. 14b–e). These results also show that an increase in RH can increase the size of a typhoon. When the RH is reduced over a small area, the vorticity is confined within that area and thus the typhoon becomes more compact (Figs. 14f–h). The much reduced RH within the core region as in experiment −20S also results in a weaker TC.
As in Fig. 10, but for Typhoon Manyi at 96 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Manyi at 96 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
As in Fig. 10, but for Typhoon Manyi at 96 h of simulation in the environmental humidity sensitivity experiments.
Citation: Monthly Weather Review 140, 12; 10.1175/MWR-D-11-00317.1
5. Summary and conclusions
This study furthers the observational analyses in Chen et al. (2011) on compact and incompact TCs, and focuses on the synoptic-dynamical factors that determine the development of the two TC structures. For this purpose, numerical simulations using the WRF are carried out for Typhoons Yutu and Manyi in 2007, which developed to the compact structure and incompact structure, respectively. The simulation results are verified by the JTWC best track and the structure parameters derived from QuikSCAT oceanic winds. In particular, evolutions of the tangential wind speeds and relative vorticity are examined. The simulation results show that for Typhoon Yutu the wind speed increases primarily in the inner-core region especially during the rapid intensification period from 24 to 48 h. The region of strong relative vorticity and high inertial stability is concentrated within a small radius. In contrast, Typhoon Manyi’s tangential wind speeds increases over a much larger radial domain.
The evolutions of low-level inflows are also quite different for these two typhoon cases. The low-level inflow, which is accompanied by low-level convergence, in Typhoon Manyi maintains a magnitude of about 16 m s−1 from 500 km inward to about a 100-km radius. However, the inflow of Typhoon Yutu increases its magnitude from a radius of 250 km almost linearly toward the center region, and thus the low-level convergence is concentrated near the center. Regarding the inertial and static stability, the distributions in Typhoon Yutu are similar to those of the composite results discussed by Holland and Merrill (1984). The reasoning that Holland and Merrill (1984) proposed thus helps to explain how a compact typhoon forms and maintains its structure. Namely, the strong inertial stability at the low-level core region is resistant to influences from low-level forcing. However, the strong static stability and low inertial stability at the upper levels make it more susceptible to influences from upper-level forcing. Consequently, low-level forcing mainly changes the size and strength of TCs. This understanding combined with the examination of the two typhoons’ associated environmental flows show that the compactness of Yutu is highly related to internal dynamics, while for Manyi the low-level environmental influences play a determining role when the typhoon develops to an incompact structure.
A series of sensitivity experiments are conducted to help clarify the relative roles of initial vortex structure and environmental flow to a TC’s structural evolution. A compact synthetic vortex is bogussed into the environment of Typhoons Yutu and Manyi, respectively. The environment flows of Typhoons Yutu and Manyi are considered favorable for forming a compact and incompact TC, respectively. Similar experiments are performed with an initial synthetic vortex with an incompact structure. The results show that the initial vortex largely determines the TC structure throughout the simulation up to 96 h under the Yutu environmental flow. This applies to the initially incompact vortices as well because they can more or less maintain their structures in the simulations. Under Typhoon Manyi’s environmental flow, the initially compact vortices can maintain their structure within the early stage of about 54 h. Unless the initial vortex intensity is high enough, the environmental flow forces the TC structure to become incompact at the later stage. For the initially incompact synthetic vortices, the RMW is the critical factor. Namely, the vortices become more incompact as in the control simulation of Typhoon Manyi. However, if the RMW is initially small (~100 km or less) this process will be much slower. Nevertheless, all initial vortices including the compact and incompact ones under Manyi’s environmental flow increase rapidly in TC size until about 54 h.
This numerical study confirms the observational analyses in Chen et al. (2011) in regard to the distinct roles of internal dynamical processes and external forcing in the development of compact and incompact TCs. However, there are other environment factors that may determine the structural evolution of TCs. One plausible example is midlevel humidity, which is known to affect TC sizes critically (Hill and Lackmann 2009). Chen et al. (2011) compared the 850–700-hPa relative humidity associated with compact and incompact TCs, and found that for incompact TCs it is only marginally larger than that for compact ones. Given the large uncertainty of humidity data in global analyses, in this study we are able to determine by numerical simulations whether this difference is robust. Results of the environmental humidity experiments herein are in general consistent with the results of Hill and Lackmann (2009) and Xu and Wang (2010b), but with interesting distinct responses from compact and incompact TCs, respectively. Overall, a humid environment favors development of more intense and incompact TCs, whereas a relatively dry environment produces more compact TCs. However, the compact TCs show much less sensitivity to humidity than the incompact ones. This is believed to be due to a lack of external forcing in the synoptic environment and that the initial vortex structure of a compact TC largely determines the subsequent development. In contrast, external forcing such as strong southwesterly flows enables incompact TC development. When humidity is enhanced in the outer-core region, more severe convection and thus larger wind speeds are resulted that increase the incompactness of the TC further. Another scenario is that the inner-core convection is enhanced by humidity such that the gradient of wind speeds from the inner to outer core increases. In this situation, the incompact TC also shows quite a large response to changes in humidity.
Acknowledgments
We are grateful to the Data Bank for Atmospheric Project, Taiwan Typhoon and Flood Research Institute, the National Applied Research Laboratories for the support of atmospheric research data, and the Computer and Information Networking Center, National Taiwan University for the support of high-performance computing facilities. This research is supported by the National Science Council of the Republic of China (Taiwan) under Grants NSC 98-2625-M-002-002 and NSC 99-2625-M-002-013-MY3. The participation of the second author (KKWC) in this study is supported by the Macquarie University Research Development Grant 9201000735.
REFERENCES
Boukabara, S.-A., R. N. Hoffman, C. Grassotti, and S. M. Leidner, 2002: Physically-based modeling of QSCAT SeaWinds passive microwave measurements for rain detection. J. Geophys. Res., 107, 4786, doi:10.1029/2001JD001243.
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.
Chou, K.-H., C.-C. Wu, P.-H. Lin, and S. Majumdar, 2010: Validation of QuikSCAT wind vectors by dropwindsonde data from Dropwindsonde Observations for Typhoon Surveillance Near the Taiwan Region (DOTSTAR). J. Geophys. Res., 115, D02109, doi:10.1029/2009JD012131.
Dudhia, J., 1989: Numerical study of convection observed during the winter monsoon experiment using a mesoscale two-dimensional model. J. Atmos. Sci., 46, 3077–3107.
Ebuchi, N., H. C. Graber, and M. J. Caruso, 2002: Evaluation of wind vectors observed by QuikSCAT/SeaWinds using ocean buoy data. J. Atmos. Oceanic Technol., 19, 2049–2062.
Hill, K. A., and G. M. Lackmann, 2009: Influence of environmental humidity on tropical cyclone size. Mon. Wea. Rev., 137, 3294–3315.
Holland, G. J., Ed., 2009: Global guide to tropical cyclone forecasting. Bureau of Meteorology Research Centre, Australia. [Available online at http://cawcr.gov.au/bmrc/pubs/tcguide/globa_guide_intro.htm.]
Holland, G. J., and R. T. Merrill, 1984: On the dynamics of tropical cyclone structural changes. Quart. J. Roy. Meteor. Soc., 110, 723–745.
Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain-Fritcsh scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 246 pp.
Kantha, L., 2006: Time to replace the Saffir-Simpson Hurricane Scale? Eos, Trans. Amer. Geophys. Union, 87, 3–6.
Knaff, J. A., J. P. Kossin, and M. DeMaria, 2003: Annular hurricanes. Wea. Forecasting, 18, 204–223.
Knaff, J. A., T. A. Cram, A. B. Schumacher, J. P. Kossin, and M. DeMaria, 2008: Objective identification of annular hurricanes. Wea. Forecasting, 23, 17–28.
Lee, C.-S., K. K. W. Cheung, W.-T. Fang, and R. L. Elsberry, 2010: Inital maintenance of tropical cyclone size in the western North Pacific. Mon. Wea. Rev., 138, 3207–3223.
Lin, Y.-L., R. D. Farley, and H. D. Orville, 1983: Bulk parameterization of the snow field in a cloud model. J. Climate Appl. Meteor., 22, 1065–1092.
Maclay, K. S., M. DeMaria, and T. H. V. Harr, 2008: Tropical cyclone inner-core kinetic energy evolution. Mon. Wea. Rev., 136, 4882–4898.
Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102 (D14), 16 663–16 682.
Noh, Y., W. G. Cheon, S. Y. Hong, and S. Raasch, 2003: Improvement of the K-profile model for the planetary boundary layer based on large eddy simulation data. Bound.-Layer Meteor., 107, 421–427.
Pickett, M. H., W. Tang, L. K. Rosenfeld, and C. H. Wash, 2003: QuikSCAT satellite comparison with nearshore buoy wind data off the U.S. west coast. J. Atmos. Oceanic Technol., 20, 1869–1879.
Powell, M., and T. Reinhold, 2007: Tropical cyclone destructive potential by integrated kinetic energy. Bull. Amer. Meteor. Soc., 88, 513–526.
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.
Sharma, N., and E. D’Sa, 2008: Assessment and analysis of QuikSCAT vector wind products for the Gulf of Mexico: A long-term and hurricane analysis. Sensors, 8, 1927–1949, doi:10.3390/s8031927.
Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, and J. G. Powers, 2005: A description of the Advanced Research WRF version 2. NCAR Tech. Note TN-468_STR, 88 pp.
Wang, Y., 2008: Structure and formation of an annular hurricane simulated in a fully compressible, nonhydrostatic model—TCM4. J. Atmos. Sci., 65, 1505–1527.
Weissman, D. E., M. A. Bourassa, and J. Tongue, 2002: Effects of rain rate and wind magnitude on SeaWinds scatterometer wind speed errors. J. Atmos. Oceanic Technol., 19, 738–746.
Xu, J., and Y. Wang, 2010a: Sensitivity of tropical cyclone inner-core size and intensity to the radial distribution of surface entropy flux. J. Atmos. Sci., 67, 1831–1852.
Xu, J., and Y. Wang, 2010b: Sensitivity of the simulated tropical cyclone inner-core size to the initial vortex size. Mon. Wea. Rev., 138, 4135–4157.
Zhang, Q.-H., S.-J. Chen, Y.-H. Kuo, K.-H. Lau, and R. A. Anthes, 2005: Numerical study of a typhoon with a large eye: Model simulation and verification. Mon. Wea. Rev., 133, 725–742.
