1. Introduction
Translational typhoons can be significantly influenced by orographic effects, particularly in regions characterized by steep mountains like the Central Mountain Range (CMR) in Taiwan. Numerous studies have investigated track deflections resulting from the topographic influence of the CMR (e.g., Jian and Wu 2008; Lin and Savage 2011; Huang et al. 2011; Chien and Kuo 2011; Hsu et al. 2013, 2018; Wu et al. 2015; Tang and Chan 2014, 2016a,b; Huang et al. 2016a,b, 2019, 2020a,b, 2022b; Lin et al. 2016; Huang and Wu 2018). Huang et al. (2020b) mentioned that about 75% of the historical typhoons passed near Taiwan from the east. As these westbound typhoons move close to the east coast of Taiwan, the topographical effect of the CMR may greatly affect the typhoon circulation and induce intense asymmetric northerly winds in the inner vortex of the typhoon, known as the channeling effect (e.g., Jian and Wu 2008; Chang et al. 2013; Huang et al. 2016a). Hsu et al. (2018) conducted a numerical study of typhoons that make landfall on the east coast of Taiwan and found that the induced asymmetric cloud heating in the typhoon plays a crucial role in the leftward track deflection. Huang et al. (2020b, 2022b) investigated westbound to northwestbound typhoons moving near offshore of northern Taiwan and indicated that the recirculating flow around south Taiwan may cause the northward track deflection.
While the majority of the above studies have focused on westbound or northwest bound typhoons, about 15% of the observed typhoons approaching Taiwan are northward bound, including Typhoon Chanthu (2021). The mechanisms of track deflection for northbound typhoons have not been clearly addressed in previous papers. Figure 1 gives some earlier examples of the typhoons that moved northwestward from the Philippines Sea and subsequently changed their tracks to a northward direction and then passed around east Taiwan, including Dinah (1965), Seth (1994), Zeb (1998), Songda (2011), Fung-wong (2014), and Mitag (2019). Few of the northward typhoons passed west of Taiwan in the historical record (figures not shown). Although Seth and Mitag moved somewhat away from the CMR (Figs. 1b,f), both tracks are associated with rightward deflection offshore southeast of Taiwan and then westward and inland deflection toward Taiwan near northeast Taiwan. Such a rightward deflection is even more apparent in the other three typhoons that approached northward closer to south Taiwan (Figs. 1a,c,e).
Historical typhoons passing near the east Philippines and moving northward toward south Taiwan, including (a) Dinah (1965), (b) Seth (1994), (c) Zeb (1998), (d) Songda (2011), (e) Fung-wong (2014), and (f) Mitag (2019). Typhoon intensities are shown by different colors: red for intense typhoon (wind speed greater than 51 m s−1), green for moderate typhoon (wind speed 32.7–50.9 m s−1), blue for tropical storm (wind speed 17.2–32.6 m s−1), and purple for tropical depression (wind speed less than 17.2 m s−1) according to the CWA.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
To understand the dynamic mechanism of track deflection induced by a mountain terrain, regional models with comprehensive physical processes have been applied (e.g., Chang 1982; Bender et al. 1987; Yeh and Elsberry 1993a,b; Huang et al. 2011; Huang et al. 2016a, 2020b; Lin et al. 2005, 2016; Lin and Savage 2011; Tang and Chan 2016a,b; Hsu et al. 2013, 2018; Huang and Wu 2018). These studies have illustrated the evolution of the cyclones’ track and structure during their passage over a real or idealized mountain range. For tropical cyclones (TCs) past a mesoscale mountain range, like the CMR, their induced track deflection is primarily controlled by the nondimensional parameter R/Ly, where R is the cyclone vortex size (radius of maximum tangential wind) and Ly is the length of terrain normal to the basic flow (e.g., Lin et al. 1999, 2002, 2005; Huang et al. 2016a). The northward or southward deflection of a westward-approaching vortex toward the central terrain is determined by R/Ly, where a smaller ratio tends to incur a southward deflection. Huang et al. (2022b) further considered the meridional departure distance between the vortex and the terrain center (LD) and the orientation of terrain so that the parameter R/Ly is modified as R/LE, where LE is the effective terrain scale perpendicular to the basic flow. For northward cyclones, the value of R/LE becomes larger and will incur a rightward track deflection. The parameter of LD accounts for the upstream departure from the main body of the mountain so that the track deflection will decrease as LD becomes larger.
Our earlier simulations have taken the advantage of a global model, specifically the Model for Prediction Across Scales (MPAS; Skamarock et al. 2012), at the full coverage of the large-scale environment and the embedded typhoon, as well as at variable resolution to resolve the inner core of the typhoon vortex (e.g., Huang et al. 2017, 2019, 2022a; Chen et al. 2021). The greatly enhanced resolution of MPAS is comparable to the high resolution used in regional models. Such grid refinement in the global model may enhance the forecasts of typhoon intensity and track at largely reduced computational costs (Zarzycki and Jablonowski 2015; Sakaguchi et al. 2015).
This numerical study for Typhoon Chanthu (2021) was motivated by its special track deflection during the offshore movement around east Taiwan. However, the observed track deflection was not well captured by the forecasts at various operational centers, including the Joint Typhoon Warning Center (JTWC) and the Central Weather Administration (CWA) (figures not shown). For improving the simulation of Chanthu passing around the high CMR, we have relied on MPAS with a multiresolution of 60–15–1 km in this study, where 1-km resolution is centered over Taiwan to investigate Chanthu’s track deflection. Considering the complex steering environment associated with the real case, idealized experiments with a CMR-like mountain range are also conducted in this study using the regional WRF to help identify the track deflection mechanism for mainly northward typhoons. These idealized experiments also consider TCs at somewhat different departing positions and directions to further illustrate the variabilities and uncertainties of the track deflection. Both global MPAS and idealized WRF simulations provide a unique understanding of the track deflection mechanisms for mainly northward typhoons that are not thoroughly investigated in earlier papers. The direction and magnitude of the track deflection for northward typhoons are connected with the two controlling parameters R/LE and LD proposed in our earlier work for westward–northwestward typhoons.
This paper is organized as follows. The MPAS model configurations are provided in section 2. An overview of Chanthu and the experimental designs are given in section 3. Section 4 describes the results of the MPAS simulations and the evolving structure of Chanthu with and without Taiwan terrain. The wavenumber-1 analyses of the potential vorticity (PV) budget were conducted to explore the physical mechanisms for the induced track deflection near Taiwan. Simulations of idealized experiments using the regional WRF are presented in section 5 for comparisons with the MPAS simulations. Conclusions are given in section 6.
2. Model configurations
The MPAS model, version 7.0, is used in this study, which is a nonhydrostatic global model developed at NCAR (Skamarock et al. 2012). An unstructured centroidal Voronoi mesh is utilized for MPAS so that the resolution can be gradually increased into higher resolution at a specific region. This gradual change may reduce the problems in lateral boundary conditions of regional models with nested meshes (Skamarock et al. 2012; Hagos et al. 2013; Park et al. 2014).
Figure 2 shows the grid mesh of MPAS for the numerical experiments in this study with the variable-resolution mesh (60–15–1 km). A resolution (cell-center spacing) of roughly 62 km exists outside the Southeast Asia region and then is gradually increased toward the domain center (Taiwan) to about 1-km resolution. Note that there are two resolution transition zones, from 60 to 15 km and from 15 to 1 km, for double zooming. Such a telescoped zoom will be helpful for resolving the structural changes of the inner typhoon vortex in the vicinity of the high CMR. There are 55 vertical layers in the experiments, mostly stretched in the troposphere, with a model top at 30-km height.
The variable-resolution (60–15–1 km) mesh used in the MPAS simulation at a contour interval of 3 km with the highest resolution of 1 km centered at Taiwan.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
The physical parameterization schemes used in the numerical experiments are those in the MPAS constituting the mesoscale-reference suite (denoted as the M-suite). The M-suite includes the new Tiedtke convective cumulus parameterization, WRF single‐moment 6‐class microphysics scheme (WSM6), Noah land surface model, Yonsei University (YSU) planetary boundary layer (PBL) parameterization, Monin–Obukhov surface layer parameterization, Rapid Radiative Transfer Model for General Circulation Model (RRTMG) longwave (LW) and shortwave (SW) scheme, and Xu–Randall cloud fraction parameterization. The references of these schemes are given in Skamarock et al. (2021).
The initial conditions were taken from the National Centers for Environmental Prediction (NCEP) Global Data Assimilation System Final (GDAS/FNL) operational global analysis on 0.25° × 0.25° grids. The simulations start from 1200 UTC 8 September 2021 to 0000 UTC 13 September 2021, focusing on the track change of Chanthu in the vicinity of Taiwan. The sea surface temperature remains unchanged during the simulation.
3. Typhoon Chanthu and experiments
a. Typhoon Chanthu (2021)
Chanthu formed east of the Philippines and became a tropical storm at 0800 UTC 7 September 2021. During its westward movement, it underwent rapid intensification, as evidenced by a daily change of 60 hPa in the central sea level pressure between 1200 UTC 7 September and 1200 UTC 8 September, according to the best track estimates from CWA (referred to the CWA Typhoon Database). Chanthu reached its peak intensity with a central sea level pressure of 915 hPa and the maximum wind speed of 58 m s−1 at 0600 UTC 10 September. The synoptic maps of CWA indicate that Chanthu moved northward near the northern Philippines under the southwest extension of the subtropical ridge, which provided a northward component in the typhoon’s steering flow (figures not shown). Meanwhile, a midlatitudinal synoptic trough near north China moved eastward and approached the China coast. Consequently, Chanthu moved along the western periphery of the subtropical high, offshore near southeast of Taiwan as shown in Fig. 3a. As it neared south Taiwan, Chanthu deflected rightward and then moved northward offshore along the eastern Taiwan coast. A leftward inland deflection was actually induced as Chanthu departed away northeast of Taiwan, which brought in intense rainbands with more impacts on north Taiwan.
(a) The CWA best track (black), the simulated tracks for CTL (60–15–3-km mesh) (red), and the simulated tracks for CTL (60–15–1-km mesh) (green), and noTW (blue) in the forecast period of 1200 UTC 8 Sep–0000 UTC 13 Sep; (b) as in (a), but for the simulated track errors (km); and (c) as in (a), but for maximum 10-m wind speed (m s−1) of the simulated typhoons as well as the maximum wind speed averaged in 10 and 1 min for the best track data from CWA and JTWC, respectively. All the tracks are marked with solid circles every 6 h. The Taiwan terrain (with a maximum height of 3705 m) is shown in shaded colors (m) in (a), which is resolved by the MPAS 1-km resolution. The best track intensity (orange) from JTWC is also given in (c) for reference. In (a), the dashed pink circle indicates the leftward track deflection focused in this study.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
b. Numerical experiments
The control experiment (denoted by CTL) uses the M-suite with the 60–15–1-km resolution mesh. To investigate the effect of Taiwan’s terrain on the typhoon track, the Taiwan area is changed to ocean in the experiment (noTW). The experiment with the 60–15–3-km resolution mesh is denoted by CTL (60–15–3), in which the configurations are identical to CTL except for the different resolutions. The 60–15–3-km mesh (Huang et al. 2019) is similar to 60–15–1-km mesh, except with a reduced resolution of 3 km in the target region, and this allows the comparison of the new simulation results with the 60–15–3-km mesh that was applied in our earlier simulations (Huang et al. 2019, 2022a).
4. Results of MPAS simulations
a. Simulations with and without Taiwan terrain
Figure 3 shows the simulation results from 1200 UTC 8 September 2021 for the CTL experiment and terrain-sensitivity experiments using the 60–15–1-km mesh. The typhoon center is determined based on the maximum wind speed of the typhoon circulation at 10-m height. The simulated track for the CTL experiment using 60–15–1-km mesh has smaller track errors (less than 120 km in 108 forecast hours) at most of the forecast times as compared to those using the 60–15–3-km mesh, especially when the typhoon is in the vicinity of south Taiwan (Fig. 3b). Consequently, the sharp turn in the earlier track is better simulated by the higher-resolution experiment. Furthermore, a leftward inland track deflection after 1800 UTC 11 September is also produced in CTL, which is also found in the observed track but present further offshore at a later time (0300 UTC 12 September). In this study, we attempt to illustrate the robustness of the 60–15–1-km resolution that allows a global grid mesh down to very fine scale, which is useful for study of the vortex–terrain interaction. Removal of Taiwan’s terrain in CTL results in a track in noTW similar to CTL before 0300 UTC 11 September and after 0600 UTC 12 September, but without a sharp turn when the typhoon moves closer to the southern tip of Taiwan. This track comparison suggests that the topographical influence of Taiwan’s terrain is important in the rightward deflection of the typhoon. The CTL (60–15–3 km) and CTL simulations give similar typhoon intensities but with some underestimate on the near-surface wind speed in comparison to the best track data from JTWC (Fig. 3c). The simulated typhoon without Taiwan terrain is stronger after 1500 UTC 10 and further intensifies after 1500 UTC 11 since it is not negatively impacted by interaction with Taiwan’s terrain compared to CTL.
b. Typhoon circulation and vertical convection
Figures 4a, 4c, 4e, and 4g show the horizontal wind for CTL and noTW at 850 hPa during the track deflection near Taiwan. As shown in Fig. 4a, there is recirculating flow toward the Taiwan Strait, stemming from the outer circulation of Chanthu near northwest Taiwan. The deflection occurs as the recirculating flow rejoins Chanthu’s inner core from the west. This westerly flow converges with the southwesterly flow in the southwest quadrant of the typhoon. As the typhoon moves northward offshore along east Taiwan, stronger southerlies are present east of the typhoon center at 78 h (Fig. 4c). Near this stage, the stronger flow in the vortex rotates to the northeast of the typhoon center, leading to a leftward inland track. As the typhoon moves northward near the central east coast of Taiwan, the flow shading zone is smaller to allow for more recirculating flow over south Taiwan to provide enhanced southerly flow east of the vortex center. Unlike the presence of the intense northerly flow caused by channeling effects for westbound typhoons, the vortex flow becomes relatively weaker near the central east coast for this northward typhoon. Thus, the inner vortex flow can be much stronger to the east than the west of the typhoon center. When the Taiwan island is removed in noTW, there is no flow shading in the vicinity of Taiwan and thus enables a more symmetric vortex during the northward movement (Fig. 4e). Consequently, the cyclonic flow is much stronger and more easterly compared to that in the presence of Taiwan terrain. At this earlier stage, the typhoon is driven in a consistent direction without an eastward shifting of the track as in CTL. At later stages, the typhoon translation is similar to the best track, but lacking the inland turning near northeast Taiwan (Fig. 4g). To illustrate the wind asymmetry in the typhoon’s inner core, the contoured frequency by altitude diagrams (CFADs) (Yuter and Houze 1995) of zonal wind velocity within 100-km radius at 69 and 78 h for CTL and noTW are shown in Figs. 4b, 4d, 4f, and 4h. At the earlier stage, the percentage of westerly wind is larger than that of easterly wind (Fig. 4b) as the recirculating flow has spiraled into the typhoon’s inner core (Fig. 4a), resulting in an eastward (rightward) deflection. On the contrary, westward wind components are more dominant above 3-km height than eastward wind components in the typhoon circulation at the later stage (Fig. 4d), resulting in a westward (leftward) deflection. Figures 4f and 4h show that the eastward and westward wind components are more comparable at both stages in the absence of Taiwan terrain.
The horizontal wind field (m s−1) at 850 hPa for CTL at (a) 69 and (c) 78 h. (e),(g) As in (a) and (c), respectively, but for noTW. The wind speed is also shown in shaded colors (m s−1), and a reference vector (m s−1) is provided at the upper right of each panel. The simulated tracks (red) are shown for CTL or noTW, and the CWA best track (purple) is provided for reference. (b),(d),(f),(h) As in (a), (c), (e), and (g), but for CFADs (%) of zonal wind velocity (m s−1) within the radius of 100 km of the typhoon inner core.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
Figure 5 compares the observed and simulated maximum radar reflectivities when the observed and simulated typhoon centers are close. As the observed Chanthu moved northward offshore along the southeastern coast of Taiwan, the associated stronger reflectivity within the inner vortex core is rotated counterclockwise over time from the southeast (figures not shown) to the northern semicircle of the typhoon (Fig. 5a). The observed maximum reflectivity within the inner core is enhanced with the contracting eyewall when Chanthu further migrated offshore near central Taiwan (Fig. 5b). At both observation times, there are stronger outer rainbands on the southern semicircle of the vortex connecting into the inner eyewall on the eastern semicircle and weaker reflectivity within the eyewall is produced to the west and south of the inner core. As the observed typhoon moved further northward near northeast Taiwan, more intense rainbands appear mainly on the eastern and northern semicircles of the vortex that are much less blocked by the CMR (Fig. 5c). Consequently, a slightly inland track deflection is induced as shown in Fig. 3a. After this short-time leftward deflection, Chanthu continues to move northward but with more symmetric cloud convection around the inner vortex (Fig. 5d) as compared to the previous several hours (Figs. 5b,c).
The observed maximum radar reflectivity (shaded colors; dBZ) at (a) 1130 UTC 11 Sep 2021, (b) 0106 UTC 12 Sep 2021, (c) 0336 UTC 12 Sep 2021, and (d) 0900 UTC 12 Sep 2021 (available from the CWA). (e),(f) As in (a)–(d), but for the CTL simulation at 0900 UTC 11 Sep (69 h) and 2100 UTC 11 Sep (81 h), respectively.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
The simulated radar reflectivity captures some important features of the observed reflectivity at the early stage but generally is stronger in the eyewall and outer spiral rainband (Fig. 5e). Note that the observed reflectivities from the S-band radars of CWA displayed near the east coast of Taiwan do not receive significant attenuation and ground clutter for nearby Chanthu (Fang et al. 2024). The simulated eye at a diameter of about 40 km appears to be considerably larger than the observed. The simulated reflectivity also compares well with the observations at the later stage when the typhoon is offshore near central Taiwan (Fig. 5f), except for stronger rainbands on the northern and eastern semicircles of the inner vortex. The simulated eye has significantly shrunk with enhanced reflectivity in the eyewall consistent with the intensification of the vortex at this time (see Fig. 3). However, it is noted that the simulated strong convection in the inner vortex is somewhat displaced more inland closer to the terrain, compared to the observed. The major differences in the simulated and observed reflectivities may result from the track and intensify deviations of the simulated vortex from the observed. As explained later, this asymmetric cloud convection tends to induce an inland track with stronger deflection than that observed. Despite the inland track deflection induced earlier, the simulation has reasonably captured the major distribution of the observed rainfall with a daily maximum of about 240 mm over northeast Taiwan (figures not shown).
The associated vertical motions at 3-km height at 69 and 78 h for the CTL and noTW simulations are shown in Figs. 6a–d. At earlier stages, stronger vertical motions are primarily present to the southeast of the typhoon center for CTL, coincident with the convergence zone between the typhoon circulation and the large-scale southerly wind (figures not shown). As the typhoon moves northward, stronger upward motions rotate counterclockwise in the eyewall to the north-northeast of the typhoon center (Fig. 6a). Meanwhile, a convergence zone persists to the east and south of the typhoon center, where updrafts are present. When the typhoon approaches closer to Taiwan terrain at 78 h, the horizontal wind at the west quadrant of the outer typhoon vortex is strongly blocked by the terrain, leading to enhanced vertical motions induced between the typhoon center and the topography (Fig. 6b). However, the inner vortex core is still located away from the steep CMR, and this provides for relatively stronger flow near the coast which induces intense upward motions in the inner eyewall to the south and east of the typhoon center due to the enhanced convergence with the recirculating flow. At 81 h, the stronger upward motions are mainly induced at the southeast quadrant of the inner vortex as the flow near the coast is weakened by the terrain blocking (figures not shown). Without Taiwan terrain, stronger vertical motions also occur to the north and southeast of the typhoon center at 69 h for noTW (Fig. 6c), similar to CTL. However, in the absence of Taiwan terrain, the updrafts are not intensified in the more symmetric vortex core (Fig. 6d). At this stage, the typhoon movement in CTL is more westward and inland toward Taiwan terrain as compared to the northward movement in noTW. For understanding the developed vertical structure of the convection in the typhoon’s inner core, Figs. 6e and 6f show the rate of diabatic heating at vertical cross section along the selected transects at 69 and 78 h. At the earlier stage of 69 h, the stronger cloud heating rate occurs in the mid- to upper troposphere northeast of the typhoon center (Fig. 6e). As the typhoon approaches closer to the terrain at 78 h, the vertical motions intensify and produce stronger mid- to lower-tropospheric cloud heating, but now to the southeast of the typhoon center (Fig. 6f).
Vertical velocity (shaded colors; m s−1) and horizontal wind (vectors in m s−1) at 3-km height for CTL at (a) 69 and (b) 78 h. (c),(d) As in (a) and (b), respectively, but for noTW in which Taiwan terrain (green) is removed. A reference vector (m s−1) is provided at the upper right of each panel. The simulated track (red) is overlapped with the typhoon center marked by a solid circle at the time shown. The green line indicates the coastlines. (e),(f) The diabatic heating rate (shaded colors; K h−1) at the vertical cross sections along the pink dashed line indicated in (a) and (b) at 69 and 78 h, respectively.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
To explore the vertical structure associated with the simulated typhoons, the azimuthal-mean tangential wind velocity and transverse circulation at 63, 69, and 78 h for CTL and noTW are shown in Fig. 7. For CTL, the strongest tangential wind is produced at a radius of approximately 0.25° and the low-level inflow is confined below about 2-km height at 63 and 69 h (Figs. 7a,c). The tangential wind develops to higher levels associated with stronger updrafts but with weakened outflow in the inner upper-tropospheric vortex as the typhoon moves closer to south Taiwan at 69 h. However, the typhoon vortex is slightly intensified but with much stronger upward motions in the midlower tropospheric eyewall as well as deeper and stronger inflow at 78 h when the typhoon center is near the position of an inland track deflection (Fig. 7e). As seen in Fig. 4c, the intense inflow results from the large wind asymmetry in the vortex near Taiwan as the vortex is strongly blocked by the CMR. In the absence of Taiwan terrain, the typhoon vortex develops deeper with stronger updrafts in the eyewall at earlier stages (Figs. 7b,d). The northward moving typhoon is somewhat weakened at 78 h (Fig. 7f) as the typhoon center is near the central east coast of Taiwan (see Fig. 4g).
(a) Azimuthal-mean radial and vertical wind components (vectors; m s−1) and tangential velocity (shaded colors; m s−1) in the vertical cross section of height (km) and radius (°) for CTL at (a) 63, (c) 69, and (e) 78 h. (b),(d),(f) As in (a), (c), and (e), respectively, but for noTW. Reference vectors for the wind components are provided at the lower right of the figure.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
The time evolutions of radial and tangential wind components of the typhoon vortex in CTL and noTW are shown in Fig. 8 to explore the topographic effects of the CMR from 12 to 81 h on the development of the recirculating flow that is important for the induced track deflection. Both wind components are averaged in azimuth and 0–2-km height (Figs. 8a,b). There are no large differences in both wind components prior to the stage of the vortex approaching the CMR before 48 h for both runs; however, an increase in radial inflow intensity outside the radius of maximum wind (RMW) (about 0.25°) has emerged from 36 h for CTL. This remote influence of the CMR on the typhoon evolution is evident even when the typhoon is 400 km away from the southern tip of the CMR. After 48 h, the tangential wind weakens in the inner vortex core, while the radial inflow is intensified as compared to that in noTW, particularly at later stages after 66 h when the recirculating flow of the typhoon has converged with the inner vortex flow (see Fig. 4a). Both wind components averaged in 0.125°–0.5° and 0–2-km height for both experiments are shown in Figs. 8c and 8d. At the early stage before 36 h, weaker tangential wind is mainly present to the southwest of the vortex center for CTL, corresponding to the evolution of Chanthu that moves northwestward. The wind differences in both experiments are more pronounced after about 42 h. For CTL, stronger (weaker) tangential winds are produced mainly to the east (west) of the inner core during the period of 12–66 h due to the fact that the vortex circulation at the east quadrant is less affected by the topography. However, the strong tangential wind to the east is weakened after about 51 h for CTL, while still intense for noTW. During 48–72 h, the weak zone of tangential wind is shifted from the northwest to the southwest for noTW in contrast to a shifting from the west to north for CTL. After 72 h, the stronger radial inflow and updrafts are mainly present to the west and south of the vortex center for CTL.
Hovmöller plots of azimuthal-mean radial velocity (contours with an interval of 1 m s−1) and tangential velocity (shaded colors; m s−1) averaged in 0–2-km height from 0000 UTC 9 Sep to 2100 UTC 11 Sep for (a) CTL and (b) noTW. (c),(d) The Hovmöller plots of radial velocity (contours with an interval of 3 m s−1) and tangential velocity (shaded colors; m s−1) averaged in 0–2-km height and the annulus of 0.125°–0.5°.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
Previous typhoon evolution has illustrated the topographic influence of the CMR focusing on low levels. Figure 9 shows the difference in the azimuthal wavenumber-1 flow obtained by the Fourier decomposition that is averaged in 3–9-km height as well as in 1 h at several forecast times (66, 69, 72, and 78 h) between CTL and noTW. At 66 h, there is mostly westerly flow in the vicinity of the inner typhoon vortex. An anticyclonic gyre forms east of the typhoon center at 69 h where the southwesterly is prevailing and crossing the typhoon center (Fig. 9b). In the subsequent time of 72 h, a cyclonic gyre is induced west of the typhoon center, thus consisting of a pair of one cyclonic and one anticyclonic gyre aligned mainly east–west (Fig. 9c). Most of the flow in the gyres during penetration into the typhoon center is southerly to southwesterly. This flow difference may explain why the simulated typhoon track for CTL is more eastward than noTW. This pair of gyres rotates counterclockwise with time at a southwest–northeast alignment, resulting in the primarily southeasterly flow across the typhoon center (Fig. 9d). Consequently, an inland typhoon track deflection is produced at later times for CTL, compared to the alongshore track for noTW. The pair of gyres shows an asymmetric-flow difference with a counterclockwise rotation for this northbound typhoon. This flow behavior is also exhibited in the westbound–northwestbound cyclones moving toward the CMR-like mountain (e.g., Tang and Chan 2016b; Huang et al. 2020b).
The difference (colored vectors) in wavenumber-1 flow averaged in 3–9-km height and in 1 h between CTL and noTW at (a) 66, (b) 69, (c) 72, and (d) 78 h, corresponding to 0600, 0900, 1200, and 1800 UTC 11 Sep 2021, respectively. The simulated tracks are shown for CTL (green) and noTW (blue), marked with a cross sign to indicate the typhoon center at the time. The colors of the vectors indicate their wind speed with a color bar (m s−1) at the right of the figure.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
c. PV tendency budget analysis
TCs tend to move toward the region of maximum wavenumber-1 PV tendency (Wu and Wang 2000; Yu et al. 2007). The translation velocity of typhoons induced by different physical processes, including horizontal PV advection (HADV), vertical PV advection (VADV), differential diabatic heating (HDIA), solenoidal effects, and turbulent mixing effects, can be estimated using a regression of the wavenumber-1 PV budget terms (e.g., Wu and Wang 2000; Huang et al. 2019). In this study, we have calculated the regression velocity for the simulated typhoon within the radius of 400 km from the center and averaged in 3–9-km height and 1 h. Turbulent mixing effects in the PV budget can be neglected due to the performed average in 3–9-km height above the boundary layer. Solenoidal effects are also much smaller in typhoon cases and are not shown.
Figure 10 shows the evolutions of the regressed translation induced by different wavenumber-1 PV budget terms during 63–81 h for CTL and noTW. For CTL, the translation induced by HADV is dominant, except for the later stage when the typhoon moves inland and is strongly affected by Taiwan terrain (Fig. 10a). At this time, HADV contributes to a northeastward movement due to the larger positive wavenumber-1 PV advected to the northeast of the typhoon center (figures not shown). Contribution from VADV increases with time in response to stronger upward motions, particularly after 75 h (see Figs. 6b and 7e). The earlier translation induced by HDIA becomes more southward at 72 h as stronger convection rotates counterclockwise to the north of the typhoon center, which tends to retard the northward translation from HADV. However, HDIA induces a northwestward translation at 78 and 81 h and will help drive the vortex inland. This induced inland translation is in response to negative vertical differential heating southeast of the typhoon center due to the presence of stronger convective heating in the lower troposphere when the typhoon is offshore near the central Taiwan (see Fig. 6f). VADV also contributes some westward or southward translation as a result of stronger upward motions west of the typhoon center. At both times, HDIA and VADV act together to offset the eastward component of the large northeastward translation induced by HADV. This may explain why a small inland (westward) movement of the typhoon is induced at 81 h as seen in the net budget. The estimated translation from the net budget (TSUM) is about 5.39, 5.04, and 3.85 m s−1 as compared to the true translation of 5.59, 5.22, and 5.23 m s−1 at 72, 75, and 78 h, respectively. Overall, the estimated translation is consistent with the WN-1 flow around the vortex and agrees well with the observed moving speed, except at the later stages when the vortex is near landfall. However, both translational velocities have similar directions. In the absence of Taiwan terrain, HADV is also dominant at most of the times and gives a larger eastward component before 75 h than CTL. However, this increased eastward tendency by HADV is greatly offset by both VADV and HDIA with westward components, thus resulting in a more westward track at early stages before 72 h. Both VADV and HDIA become much weakened at later stages, which should be related to internal variations of the typhoon without Taiwan terrain. At the later stage of 81 h, HADV gives a major northeastward translation that is significantly less affected by weak contributions from both VADV and HDIA, unlike in CTL. This weakened offset by both VADV and HDIA is the main reason why a more offshore vortex movement is allowed in the absence of Taiwan terrain as illustrated in Fig. 10c.
The translation velocities (vectors; m s−1) in the period of 0300–2100 UTC 11 Sep contributed by different PV budget terms involving net budget (TSUM), HDIA, HADV, and VADV for (a) CTL and (b) noTW. A reference vector (m s−1) for the induced translation velocity is given at the lower right of (a) and (b). (c) The simulated tracks for CTL (green) and noTW (blue) marked with solid cycles every 3 h and forecast times every 6 h.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
d. Complementary experiments
We have noticed that the leftward track at later stages occurs somewhat earlier and is more inland than the observed. This discrepancy is reasonably anticipated for a long track forecast. Two complementary experiments with and without Taiwan terrain have been conducted for an initial cyclone that departs later from 1800 UTC 9 September, denoted by CTL_0918 and noTW_1118, respectively. Figure 11a shows that CTL_0918 further improves the typhoon track, except for the later leftward deflection that occurs slightly earlier than the observed. The rightward and leftward track deflections are apparently induced by the topographic effect of Taiwan terrain as compared with noTW_0918. The stronger convection in the inner typhoon is located to the southeast of the typhoon center in the midlower troposphere (Fig. 11b), providing the differential cloud heating to drive the typhoon inland as in CTL experiment.
(a) The CWA best track (black), the simulated tracks for CTL_0918 (green), and noTW_0918 (blue) in the forecast period of 1800 UTC 9 Sep–1200 UTC 12 Sep. The tracks are marked with solid circles every 3 h. (b) The diabatic heating rate (shaded colors; K h−1) at the cross section along the pink dashed line indicated in (a).
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
The experiment CTL_0918 has shown different track deflections, but the later inland deflection is induced somewhat earlier. Another complementary experiment thus is conducted for a cyclone departing further later from 1800 UTC 11 September when Chanthu already moved closer to east coast of Taiwan, denoted by CTL_1118. CTL_1118 exhibits an inland track deflection near northeast Taiwan (Fig. 12a), which is closer to the observed than CTL_0918. For this simulation, the corresponding cloud convection has also illustrated stronger development at the offshore quadrant of the typhoon when the leftward deflection just occurs (Fig. 12b). Similar to CTL and CTL_0918, the weaker upward motions in the typhoon’s inner core for CTL_1118 are located in the northwest quadrant as the typhoon approaches northeast Taiwan, in response to the topographic effect of the CMR (figures not shown). The stronger cloud heating to the east of the vortex center can also be found in radar observational synthetics (Fang et al. 2024) and the maximum radar reflectivity of CWA shown in Fig. 5. Thus, such a physically based deflection mechanism may exist in the observed track shifting of Chanthu.
As in Fig. 11, but for CTL_1118 in the forecast period of 1800 UTC 11 Sep–1800 UTC 12 Sep.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
5. Results of idealized experiments
a. The idealized WRF and numerical experiments
Since the real Chanthu case and simulated track exhibit large track variabilities near Taiwan terrain, we have employed idealized experiments to identify and explain the track variabilities under different steering directions and intensities. In this study, idealized experiments were conducted using the nonhydrostatic regional WRF Model, version 3.4.1 (Skamarock et al. 2008). The model setup includes two nested domains with horizontal resolutions of 15 and 3 km, respectively. The outer and inner domains have grid points of 501 × 501 and 1001 × 801, respectively, as shown in Fig. 13. There are 41 vertical layers with the topmost layer at a height of 20 km. An idealized mountain range, resembling the CMR, is constructed using a modified Gaussian function with a narrower tail. The mountain range is 500 km in width and 150 km in length and has a peak height of 3500 m. The mountain range is embedded within an isolated island that mimics the Taiwan island. The mountain peak is located at grid (x, y) of (1950, 1200 km) in the inner domain. The sea surface temperature is kept unchanged at 300.15 K during the entire simulation time. The physical parameterization schemes used in the idealized simulations include the New-Tiedtke convective cumulus parameterization, WSM6 cloud microphysics scheme, and YSU PBL parameterization. The idealized simulations are conducted on an f plane at 23.5°N.
The nested domains used in the idealized experiments. The sizes of the outer domain (D01) and inner domain (D02) are 7500 km × 7500 km and 3000 km × 2400 km at 15- and 3-km resolution, respectively. The coastline (bold contour) and terrain heights (contours at an interval of 1000 m) of the idealized terrain are shown in D02 where the initial vortex center is marked by the typhoon sign as in one example (A25Y400).
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
The initial vortex is at gradient-wind balance, with a maximum tangential wind speed of 55 m s−1 at a radius of 40 km from the vortex center. The initial condition for the environment of basic flow is determined by a prescribed tropical sounding with a uniform relative humidity of 90% below 1-km height, following Huang et al. (2016a, 2020b). The basic flow that steers the vortex is uniform and westerly at a wind speed of 4 m s−1, where the speed is based on the observed Chanthu’s translation near south Taiwan. At constant f, these simulations with a vortex moving eastward may resemble the scenario of northward typhoons toward Taiwan. In this study, we have employed the dynamic vortex initialization (DVI) (Huang et al. 2022a) to obtain a quasi-steady typhoon vortex in the presence of the Ekman boundary layer. Group experiments of A0 are for a mountain range and island that are east-westward oriented, and group experiments of A25 denote the experiment where the mountain range is rotated clockwise by 25°. With the same RMW of the idealized vortices in this study, the nondimensional parameter R/LE is controlled by the rotation of terrain. Terrain-sensitivity tests where the entire terrain is removed (including the mountain and island) were also conducted for group experiments (denoted as noT). The initial vortex center is located at x grid of 1050 km, which is 900 km away from the mountain peak at x grid of 1950 km. After the DVI, the vortex is relocated to the prescribed different meridional departures at y grids of 1350, 1275, 1200, 1125, and 900 km in all the group experiments (A0, A25, and noT) indicated by Y450, Y425, Y400, Y375, and Y300, respectively, where 450, 425, 400, 375, and 300 correspond to their gridpoint numbers with a uniform grid interval of 3 km. The idealized experiments herein designed for an eastbound vortex at different departure points can provide a dynamic investigation for the topographic effects on track deflection and help explain the track deflection of Chanthu under varying synoptic steering.
b. Simulation results with and without terrain
The simulated tracks and intensities for the group experiments of A0 and A25 with and without the terrain are shown in Fig. 14. The track is defined as the position of minimum pressure of the vortex at the sea level surface. For the A0 experiments, the tracks are quite straight and deviate only slightly from the steering direction during the early stages before the first 2 days (Fig. 14a). The vortex deflects gradually rightward (facing downstream) when approaching the coastline within a distance less than 600 km. Noticeable rightward deflections occur near 60 h in these three simulations when the vortex is about 60 km from the coastline. As the vortex deviates right to the south of the terrain center, maximum deflection is induced with a meridional distance of about 138 km in A0Y425 compared to the same simulation but without the terrain (noTY425). The vortex track for A0Y425 is quite similar to the observed track of Typhoon Fung-wong (2014) in Fig. 1. The vortices deflect leftward (i.e., northward) and rebound back toward the terrain for both A0Y425 and A0Y400. However, the leftward deflection does not occur in A0Y375 with a persistent eastward movement about 160 km south of the southern coastline. The earlier rightward deflection is still produced, but considerably reduced, in A0Y300 (not shown in Fig. 14a) when the initial vortex departs 300 km southward of the central terrain latitude. A later leftward inland deflection, as shown for A0Y375, is also not induced for A0Y300.
The simulated tracks with terrain (solid) and without terrain (dashed) marked every 6 h for (a) A0Y425 (red), A0Y400 (orange), and A0Y375 (yellow); (b) A25Y450 (green), A25Y425 (blue), A25Y400 (purple), and A25Y375 (pink). The thick dashed curve in (b) represents the vortex track of A25Y450 at the height of 500 hPa. The thick contour represents the coastline and the thin contours with an interval of 500 m indicate the terrain height. The x axis and y axis show the grids (km) in the inner domain. (c) The time series of simulated maximum 10-m wind speed (m s−1) for these experiments.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
Similar deflections take place for A25 experiments as compared to A0 experiments (Fig. 14b), and all have southward deflections near the same time as in A0. The magnitudes of the deflections at early stages are comparable for both vortices in A0 and A25 experiments when departing at the same latitude. However, the ensuing leftward deflection at later stages is more pronounced in A25. The vortex in A25Y400 experiences a significantly large leftward deflection such that its track almost overlaps with the A25Y425 track at the later time. For A25Y450 with the initial vortex toward the northwestern coastline, the approaching vortex is significantly distorted by the topographic effect of the terrain with stronger blocking on the outer vortex circulation. Consequently, this vortex disorganization leads to a discontinuous track (dashed green in Fig. 14b) at low levels. On the other hand, at the most southern departure point, the vortex takes a leftward inland deflection at later stages for A25Y375 that appears to be much weaker without landfall. The weaker track deflection (near x grid of 2100 km) at 96 h is similar to the observed leftward deflection of the real Chanthu at later stages.
Without the terrain, these vortices are quasi steady in the first half day and gradually intensify to about 80 m s−1 near 48 h and then keep nearly the same intensity with a highly symmetric core afterward (Fig. 14c). All these vortices weaken as they approach the terrain, and the reduction on vortex intensity is related to their distances from the terrain. The aforementioned vortex with large distortion after landfall for A25Y450 has been significantly weakened so that the vortex intensity is not well defined (thus not shown in Fig. 14c), while the vortices passing around the terrain at larger offshore distances are less affected with a quasi-steady intensity of about 60–70 m s−1 for both A25Y375 and A0Y375. Nevertheless, the idealized experiments with the intense vortex reasonably illustrate the track variabilities of real northbound typhoons. The mechanism of track deflection will be focused on A25Y425 where the simulated track and intensity approximate the CTL simulation of MPAS as shown in Fig. 3.
Considering the track changes associated with weaker vortices, we have conducted one sensitivity experiment that reduces the sea surface temperature by 1 K to 299.15 K and is denoted as A25Y425_T26. With the reduced sea surface temperature for A25Y425_T26, the peak intensity of the vortex after DVI may only reach about 70 m s−1 before closing to the terrain. However, the track for A25Y425_T26 is close to A25Y425, indicating that the track deflection for this case is not greatly affected by the initial intensity of the vortex approaching the terrain (figures not shown). On the other hand, doubling the initial intensity of the basic flow in A25 does not change the direction of track deflection but significantly reduces the magnitude of the deflection as shown in other idealized simulations but for westbound or northwestbound TCs (e.g., Bender et al. 1987; Huang et al. 2011; Huang et al. 2016a; Tang and Chan 2016b).
Figure 15 shows the horizontal wind at 850 hPa and the diabatic heating rate at 3-km height for A25Y425 at 60, 84, and 96 h. At 60 h, when the southward deflection takes place, the recirculating flow rejoins the outer vortex flow from its northern quadrant (Fig. 15a). The recirculating flow forms as a rather elongated circulation northeast of the island due to the mountain blocking on the outer skirting flow of the vortex. The distributions of higher diabatic heating rates at this time are rather symmetric, surrounding the inner core with concentric rings (Fig. 15b). As the vortex moves along the southern coastline, the region where the recirculating flow rejoins the vortex flow gradually shifts to the west and southwest quadrants at 84 h (Fig. 15c) and is collocated with stronger convection (Fig. 15d). The vortex then weakens and becomes highly asymmetric at this time when it is significantly distorted by the terrain. This weakened vortex almost loses the feature of a closed circulation when further moving to the southeastern coastline at 96 h (Fig. 15e). Stronger convection now is closer to the terrain base and surrounds the inner vortex core, except for the northern quadrant with weaker flow (Fig. 15f). The outer vortex flow passes over the lower section of the mountain range to converge with the alongshore recirculating flow, resulting in intense convection with a maximum heating rate over 0.05 K s−1 at 3-km height at the west quadrant of the inner vortex. On the other hand, the recirculating flow to the south provides the inner vortex with strong inflow components, thereby facilitating the cloud convection at low levels to the east of the vortex center. Consequently, the inland quadrant of the inner vortex, roughly leftward of the track, is associated with the weakest cloud development, which is also found in the real-case simulation (see Fig. 7b). In the absence of the terrain, such asymmetry in the vortex circulation and cloud convection are not strongly induced in both idealized simulations (figures not shown) and real-case simulation (see Fig. 7d). The presence of the maximum cloud heating near 3-km height to the right of the track in both idealized and real cases may drive the vortex leftward and inland, which will be verified by the PV budget analysis later.
Horizontal wind vectors (m s−1) and speeds (shaded colors; m s−1) at 850 hPa for A25Y425 at (a) 60, (c) 84, and (e) 96 h. (b),(d),(f) As in (a), (c), and (e), but for the diabatic heating rate (shaded colors; 0.01 K s−1) at 3-km height. The simulated track is shown in red and marked by solid circles every 6 h. A reference vector (m s−1) for the wind is given at the upper-right corner of each panel. The coastline and height of the terrain are indicated by a thick purple contour and thin purple contours (at an interval of 1000 m), respectively. The x axis and y axis show the horizontal grids (km) in the inner domain.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
c. PV budget analysis
The vortex movement in the idealized simulations is also diagnosed using the regression of wavenumber-1 PV budget for contributions from different physical processes. The regressed translation velocities and simulated track during 60–102 h for A25Y425 are shown in Fig. 16. In the presence of the terrain in A25Y425, HADV provides the major contribution to the rightward (southward) deflection by 90 h, but considerably offset by both HDIA and VDAD at later stages after 90 h (Fig. 16a). As shown in Fig. 15, the asymmetry in wind and cloud convection of the vortex becomes more pronounced as it moves along the nearshore terrain, leading to larger translation velocity changes induced by VADV and HDIA after 90 h. The translation velocities induced by HDIA and HADV at 96 h are northward (i.e., inland) toward the terrain and eastward along the terrain, respectively, which is similar to the real-case PV budget analysis at 78 and 81 h when the simulated Chanthu moves away from the northeast Taiwan coast. The large northward translation at about 7 m s−1 from HDIA at 96 h has overwhelmed the southward translation component of HADV, resulting in a net vortex translation slightly northward. At 102 h, the induced northeastward translation by HDIA has dominated the considerably weaker westward translation by HADV, which may explain why the vortex rebounds northward to the eastern shore of the terrain as seen in its later track (see Fig. 15). The translation induced by VADV is considerably slower compared to HDIA at the later stages, due to the fact that larger positive PV is not well collocated with stronger upward motions (figures not shown). The sensitivity tests on the initial departure position and intensity of the vortex (A25Y400 and A25Y425_T26) also show that HDIA is the dominant mechanism of the inland track deflection of the vortex with similar landfall as in A25Y425 (figures not shown).
(a),(b) As in Figs. 10a and 10b, but from 60 to 102 forecast hours for A25Y425 and noTY425, respectively. (c) The simulated track for A25Y425 marked with solid cycles every 6 h and forecast times every 12 h. The coastline and height of the terrain are indicated by a thick contour and thin contours (at an interval of 500 m), respectively, in the inner domain.
Citation: Monthly Weather Review 152, 9; 10.1175/MWR-D-23-0237.1
In the absence of the terrain, the intense vortex in noTY425 is quite symmetric with the induced eastward translation completely dominated by HADV during all the simulation time (Fig. 16b). In the presence of the terrain, the topographic effect strongly affects the asymmetric development of vertical motions and convection that plays a pronounced role in the track evolution of the vortex as closer to the southeastern coastline. The underlying mechanism, identified through the wavenumber-1 PV budget analysis, illustrates the pivotal role of the asymmetric cloud heating in the track deflection of the migrating vortex at later stages for both idealized and real-case simulations. In particular, we notice that the simulated track deflections at later stages for A25Y425 and CTL (MPAS simulation) are stronger and more inland as compared to the observed Chanthu with weaker and further offshore deflection. The PV budget for A25Y375 with a more offshore track, however, also indicates that the leftward deflection 96 h is contributed by HDIA as well, but at relatively reduced magnitudes by about 30% (figures not shown).
6. Conclusions
Typhoon Chanthu (2021) moved northwestward toward south Taiwan but experienced a rightward deflection when approaching the Taiwan terrain. As it continued to move offshore along east Taiwan, a leftward (inland) track deflection is induced at later stages when the typhoon moves offshore near northeast Taiwan. This study utilizes the global model MPAS with multiple resolutions to simulate the track deflection of Chanthu. The MPAS model exhibits the capability to accurately capture the track deflection of Chanthu with the high-resolution region centered on Taiwan. Use of the enhanced 60–15–1-km resolution helps capture the detailed typhoon evolution, providing a more reasonable simulation as compared to the simulation with 60–15–3-km resolution.
The simulations with and without Taiwan terrain highlight the topographic influence of the CMR on the track deflection of Chanthu near Taiwan. As the typhoon moves northward closer to south Taiwan, flow recirculating around Taiwan from the outer vortex forms an elongated cyclonic circulation with weaker shaded flow embedding the Taiwan Island. The elongation of the shading zone is gradually reduced as the typhoon moves northward offshore along east Taiwan. Stemming from the elongated vortex circulation, the recirculating flow, mainly westerly to southwesterly south of Taiwan, tends to converge with the inner typhoon vortex on the southern semicircle of the vortex, leading to a rightward (eastward) deflection. Subsequently, the inner typhoon vortex on the western semicircle is affected by stronger blocking effects of the CMR, thus exhibiting highly asymmetric flow associated with stronger recirculating southerly–easterly flow that is much less affected by the terrain. This flow asymmetry thus facilitates a leftward (inland) deflection of the typhoon vortex toward northeast Taiwan.
The earlier rightward and later leftward track deflection is accompanied by the deep-layer mean wavenumber-1 flow difference between the simulations with and without Taiwan terrain that exhibits a cyclonic gyre and an anticyclonic gyre, leftward and rightward of the track, respectively. This pair of gyres rotates counterclockwise with time during the offshore vortex movement along east Taiwan, which helps induce a leftward deflection of the northbound Chanthu. Such counterclockwise rotating gyres for the northbound typhoon are also exhibited in the simulations of west-northwestbound Maria (2018) and northwestbound Lekima (2019) that moved toward Taiwan (e.g., Huang et al. 2020b, 2022b), as well as in idealized simulations of northwestbound cyclones (e.g., Tang and Chan 2016b).
Idealized experiments are also conducted using the regional model WRF to explore the variabilities of track deflection under different vortex departures incidental to an elongated mountain range. These idealized experiments, similar to the real case, explore the spectrum of track deflections associated with an intense cyclone moving around the elongated mountain range. There are similar track deflections in the idealized experiments, as in the MPAS simulations and in the observations, in response to the orographic influence when an eastward moving vortex is approximately 600 km away from the mountain range. The idealized experiments indicate that meridional deflection of the vortex track is related to the parameter of LD where LD is the meridional departure distance from the latitude of the mountain range. The track deflection is reduced by increasing LD, which, however, is still induced even for LD greater than 300 km. As R/LE, where R is the vortex size and LE is the effective length scale of the mountain range perpendicular to the basic flow, becomes smaller with a larger LE in the case when the mountain range is slightly rotated clockwise, a similar rightward deflection remains induced as well. This scenario accounts for the incidental vortex from south-southeast for real cases. In contrast to the leftward deflection of westbound vortices ahead of a wide mountain range with much reduced R/LE, the rightward deflection for mainly northbound vortices will prevail due to the larger value of R/LE in this case. As the vortex moves further downstream along the mountain range, strong topographic blocking occurs at the leftward quadrant of the moving vortex to enhance the asymmetry in wind and convection favorable for an inland deflection of the track.
Analysis of wavenumber-1 PV budget also aids a dynamic interpretation for the mechanisms of track deflection for both real-case simulations and idealized simulations. For the idealized vortex, similar to Chanthu, the rightward deflection is mainly dominated by horizontal PV advection in response to the recirculating flow around the leading terrain at earlier stages. As the vortex moves further downstream, the leftward inland track can be induced owing to the stronger cloud heating at low levels at the offshore quadrant of the vortex. Similar features of PV budget terms are also present in the real-case simulation. The variations in the intensity and structure of the departing vortex for both cases have also affected the auxiliary impacts from vertical PV advection, but the leftward track deflection at later stages appears to be a common consequence of the asymmetric cloud heating associated with the evolving translational vortex. The idealized experiments also show that this induced later leftward track deflection occurs regardless of a small rotation of the mountain range since the terrain blocking remains sufficiently strong to suppress the convection at the inland quadrant of the vortex. The simulated leftward track deflection for all the MPAS experiments is somewhat earlier and more upstream than the observed, which motivates an aid of performing the systematically configured idealized experiments to verify the dynamical mechanism for the later track deflection. The mechanisms of the inland track deflection addressed in this study are physically based for northward typhoons and may provide a possible dynamic explanation for the leftward track deflection of Chanthu. The effect of asymmetric cloud heating leading to a leftward deflection of primarily northward typhoons is also illustrated in primarily westward typhoons (e.g., Hsu et al. 2018). The difference between both typhoon cases is that the asymmetric cloud heating is a consequence of the recirculating flow from the terrain that converges with the inner vortex circulation for northward typhoons, while it is the subsidence to the southwest of the typhoon center for westbound typhoons that produces the asymmetry. On the other hand, the leftward deflection is also induced by the stronger recirculating flow at the downshear quadrant of the vortex near departure for northward typhoons, while by the stronger channeling flow at the downshear quadrant of the vortex near landfall at the terrain for some westward typhoons approaching the central to northern portions of the terrain (Wu et al. 2015; Huang et al. 2016a). This study has provided a unique dynamic interpretation of the intriguing track deflection associated with northbound typhoons past an elongated mountain range, like the CMR. Understanding of the track deflection at two different stages for similar typhoons is important for scientific insights and forecast warning.
Acknowledgments.
This study was supported by the Ministry of Science and Technology (MOST) in Taiwan. The authors thank the National Center for High-Performance Computing (NCHC) for providing computational and storage resources. Support for author Skamarock was provided by the National Center for Atmospheric Research through support from the National Science Foundation under Cooperative Support Agreement AGS-0856145.
Data availability statement.
The FNL data used for the model initial conditions were obtained from the website of the NCEP, and the best track data were obtained from the JTWC and CWA. The forecasts of Chanthu from several operational centers were obtained from http://www.typhoon2000.ph/multi/log.php?name=CHANTHU_2021. All the model setups and simulation results in this study are available from the corresponding author, Dr. C.-Y. Huang (hcy@atm.ncu.edu.tw), at National Central University.
REFERENCES
Bender, M. A., R. E. Tuleya, and Y. Kurihara, 1987: A numerical study of the effect of island terrain on tropical cyclones. Mon. Wea. Rev., 115, 130–155, https://doi.org/10.1175/1520-0493(1987)115<0130:ANSOTE>2.0.CO;2.
Chang, C.-P., Y.-T. Yang, and H.-C. Kuo, 2013: Large increasing trend of tropical cyclone rainfall in Taiwan and the roles of terrain. J. Climate, 26, 4138–4147, https://doi.org/10.1175/JCLI-D-12-00463.1.
Chang, S. W.-J., 1982: The orographic effects induced by an island mountain range on propagating tropical cyclones. Mon. Wea. Rev., 110, 1255–1270, https://doi.org/10.1175/1520-0493(1982)110%3C1255:TOEIBA%3E2.0.CO;2.
Chen, S.-Y., C.-P. Shih, C.-Y. Huang, and W.-H. Teng, 2021: An impact study of GNSS RO data on the prediction of Typhoon Nepartak (2016) using a multiresolution global model with 3D-hybrid data assimilation. Wea. Forecasting, 36, 957–977, https://doi.org/10.1175/WAF-D-20-0175.1.
Chien, F.-C., and H.-C. Kuo, 2011: On the extreme rainfall of Typhoon Morakot (2009). J. Geophys. Res., 116, D05104, https://doi.org/10.1029/2010JD015092.
Fang, W.-T., P.-L. Chang, and M.-J. Yang, 2024: An observational study on the rapid intensification of Typhoon Chanthu (2021) near the complex terrain of Taiwan. Mon. Wea. Rev., 152, 769–791, https://doi.org/10.1175/MWR-D-23-0155.1.
Hagos, S., R. Leung, S. A. Rauscher, and T. Ringler, 2013: Error characteristics of two grid refinement approaches in aquaplanet simulations: MPAS-A and WRF. Mon. Wea. Rev., 141, 3022–3036, https://doi.org/10.1175/MWR-D-12-00338.1.
Hsu, L.-H., H.-C. Kuo, and R. G. Fovell, 2013: On the geographic asymmetry of typhoon translation speed across the mountainous island of Taiwan. J. Atmos. Sci., 70, 1006–1022, https://doi.org/10.1175/JAS-D-12-0173.1.
Hsu, L.-H., S.-H. Su, R. G. Fovell, and H.-C. Kuo, 2018: On typhoon track deflections near the east coast of Taiwan. Mon. Wea. Rev., 146, 1495–1510, https://doi.org/10.1175/MWR-D-17-0208.1.
Huang, C.-Y., C.-A. Chen, S.-H. Chen, and D. S. Nolan, 2016a: On the upstream track deflection of tropical cyclones past a mountain range: Idealized experiments. J. Atmos. Sci., 73, 3157–3180, https://doi.org/10.1175/JAS-D-15-0218.1.
Huang, C.-Y., I.-H. Wu, and L. Feng, 2016b: A numerical investigation of the convective systems in the vicinity of southern Taiwan associated with Typhoon Fanapi (2010): Formation mechanism of double rainfall peaks. J. Geophys. Res. Atmos., 121, 12 647–12 676, https://doi.org/10.1002/2016JD025589.
Huang, C.-Y., Y. Zhang, W. C. Skamarock, and L.-F. Hsu, 2017: Influences of large-scale flow variations on the track evolution of Typhoons Morakot (2009) and Megi (2010): Simulations with a global variable-resolution model. Mon. Wea. Rev., 145, 1691–1716, https://doi.org/10.1175/MWR-D-16-0363.1.
Huang, C.-Y., C.-H. Huang, and W. C. Skamarock, 2019: Track deflection of Typhoon Nesat (2017) as realized by multiresolution simulations of a global model. Mon. Wea. Rev., 147, 1593–1613, https://doi.org/10.1175/MWR-D-18-0275.1.
Huang, C.-Y., C.-W. Chou, S.-H. Chen, and J.-H. Xie, 2020a: Topographic rainfall of tropical cyclones past a mountain range as categorized by idealized simulations. Wea. Forecasting, 35, 25–49, https://doi.org/10.1175/WAF-D-19-0120.1.
Huang, C.-Y., T.-C. Juan, H.-C. Kuo, and J.-H. Chen, 2020b: Track deflection of Typhoon Maria (2018) during a westbound passage offshore of Northern Taiwan: Topographic influence. Mon. Wea. Rev., 148, 4519–4544, https://doi.org/10.1175/MWR-D-20-0117.1.
Huang, C.-Y., J.-Y. Lin, W. C. Skamarock, and S.-Y. Chen, 2022a: Typhoon forecasts with dynamic vortex initialization using an unstructured mesh global model. Mon. Wea. Rev., 150, 3011–3030, https://doi.org/10.1175/MWR-D-21-0235.1.
Huang, C.-Y., S.-H. Sha, and H.-C. Kuo, 2022b: A modeling study of Typhoon Lekima (2019) with the topographic influence of Taiwan. Mon. Wea. Rev., 150, 1993–2011, https://doi.org/10.1175/MWR-D-21-0183.1.
Huang, K.-C., and C.-C. Wu, 2018: The impact of idealized terrain on upstream tropical cyclone track. J. Atmos. Sci., 75, 3887–3910, https://doi.org/10.1175/JAS-D-18-0099.1.
Huang, Y.-H., C.-C. Wu, and Y. Wang, 2011: The influence of island topography on typhoon track deflection. Mon. Wea. Rev., 139, 1708–1727, https://doi.org/10.1175/2011MWR3560.1.
Jian, G.-J., and C.-C. Wu, 2008: A numerical study of the track deflection of Supertyphoon Haitang (2005) prior to its landfall in Taiwan. Mon. Wea. Rev., 136, 598–615, https://doi.org/10.1175/2007MWR2134.1.
Lin, Y.-L., and L. C. Savage III, 2011: Effects of landfall location and the approach angle of a cyclone vortex encountering a mesoscale mountain range. J. Atmos. Sci., 68, 2095–2106, https://doi.org/10.1175/2011JAS3720.1.
Lin, Y.-L., J. Han, D. W. Hamilton, and C.-Y. Huang, 1999: Orographic influence on a drifting cyclone. J. Atmos. Sci., 56, 534–562, https://doi.org/10.1175/1520-0469(1999)056%3C0534:OIOADC%3E2.0.CO;2.
Lin, Y.-L., D. B. Ensley, S. Chiao, and C.-Y. Huang, 2002: Orographic influences on rainfall and track deflection associated with the passage of a tropical cyclone. Mon. Wea. Rev., 130, 2929–2950, https://doi.org/10.1175/1520-0493(2002)130%3C2929:OIORAT%3E2.0.CO;2.
Lin, Y.-L., S.-Y. Chen, C. M. Hill, and C.-Y. Huang, 2005: Control parameters for the influence of a mesoscale mountain range on cyclone track continuity and deflection. J. Atmos. Sci., 62, 1849–1866, https://doi.org/10.1175/JAS3439.1.
Lin, Y.-L., S.-H. Chen, and L. Liu, 2016: Orographic influence on basic flow and cyclone circulation and their impacts on track deflection of an idealized tropical cyclone. J. Atmos. Sci., 73, 3951–3974, https://doi.org/10.1175/JAS-D-15-0252.1.
Park, S.-H., J. B. Klemp, and W. C. Skamarock, 2014: A comparison of mesh refinement in the global MPAS-A and WRF models using an idealized normal-mode baroclinic wave simulation. Mon. Wea. Rev., 142, 3614–3634, https://doi.org/10.1175/MWR-D-14-00004.1.
Sakaguchi, K., and Coauthors, 2015: Exploring a multiresolution approach using AMIP simulations. J. Climate, 28, 5549–5574, https://doi.org/10.1175/JCLI-D-14-00729.1.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., https://doi.org/10.5065/D68S4MVH.
Skamarock, W. C., J. B. Klemp, M. G. Duda, L. D. Fowler, S.-H. Park, and T. D. Ringler, 2012: A multiscale nonhydrostatic atmospheric model using centroidal Voronoi tesselations and C-grid staggering. Mon. Wea. Rev., 140, 3090–3105, https://doi.org/10.1175/MWR-D-11-00215.1.
Skamarock, W. C., and Coauthors, 2021: A description of the Advanced Research WRF Model version 4.3. NCAR Tech. Note NCAR/TN-556+STR, 165 pp., https://doi.org/10.5065/1dfh-6p97.
Tang, C. K., and J. C. L. Chan, 2014: Idealized simulations of the effect of Taiwan and Philippines topographies on tropical cyclone tracks. Quart. J. Roy. Meteor. Soc., 140, 1578–1589, https://doi.org/10.1002/qj.2240.
Tang, C. K., and J. C. L. Chan, 2016a: Idealized simulations of the effect of Taiwan topography on the tracks of tropical cyclones with different sizes. Quart. J. Roy. Meteor. Soc., 142, 793–804, https://doi.org/10.1002/qj.2681.
Tang, C. K., and J. C. L. Chan, 2016b: Idealized simulations of the effect of Taiwan topography on the tracks of tropical cyclones with different steering flow strengths. Quart. J. Roy. Meteor. Soc., 142, 3211–3221, https://doi.org/10.1002/qj.2902.
Wu, C.-C., T.-H. Li, and Y.-H. Huang, 2015: Influence of mesoscale topography on tropical cyclone tracks: Further examination of the channeling effect. J. Atmos. Sci., 72, 3032–3050, https://doi.org/10.1175/JAS-D-14-0168.1.
Wu, L., and B. Wang, 2000: A potential vorticity tendency diagnostic approach for tropical cyclone motion. Mon. Wea. Rev., 128, 1899–1911, https://doi.org/10.1175/1520-0493(2000)128%3C1899:APVTDA%3E2.0.CO;2.
Yeh, T.-C., and R. L. Elsberry, 1993a: Interaction of typhoons with the Taiwan orography. Part I: Upstream track deflections. Mon. Wea. Rev., 121, 3193–3212, https://doi.org/10.1175/1520-0493(1993)121<3193:IOTWTT>2.0.CO;2.
Yeh, T.-C., and R. L. Elsberry, 1993b: Interaction of typhoons with the Taiwan orography. Part II: Continuous and discontinuous tracks across the island. Mon. Wea. Rev., 121, 3213–3233, https://doi.org/10.1175/1520-0493(1993)121<3213:IOTWTT>2.0.CO;2.
Yu, H., W. Huang, Y. H. Duan, J. C. L. Chan, P. Y. Chen, and R. L. Yu, 2007: A simulation study on pre-landfall erratic track of typhoon Haitang (2005). Meteor. Atmos. Phys., 97, 189–206, https://doi.org/10.1007/s00703-006-0252-1.
Yuter, S. E., and R. A. Houze Jr., 1995: Three-dimensional kinematic and microphysical evolution of Florida cumulonimbus. Part II: Frequency distributions of vertical velocity, reflectivity, and differential reflectivity. Mon. Wea. Rev., 123, 1941–1963, https://doi.org/10.1175/1520-0493(1995)123<1941:TDKAME>2.0.CO;2.
Zarzycki, C. M., and C. Jablonowski, 2015: Experimental tropical cyclone forecasts using a variable-resolution global model. Mon. Wea. Rev., 143, 4012–4037, https://doi.org/10.1175/MWR-D-15-0159.1.
