Next Article in Journal
Trends and Shifts in Blocking Metrics in the Southern Hemisphere Derived from the ERA5 and the NCEP/NCAR Reanalysis for the 1940–2022 Period
Previous Article in Journal
Characterizing TVOC Increases and Terpene-Derived BVOC Contributions in Wood-Finished Public Facilities
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Atmosphere
Volume 17
Issue 9
10.3390/atmos17090888
atmosphere-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Effects of Luzon Topography on Track Deflection During the Early Development of Typhoon Yagi (2024)

by
Yuan Zhu
1,2 and
Weibiao Li
1,2,*
1
School of Atmospheric Sciences, Sun Yat-sen University, Zhuhai 519080, China
2
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519082, China
*
Author to whom correspondence should be addressed.
Atmosphere 2026, 17(9), 888; https://doi.org/10.3390/atmos17090888
Submission received: 16 July 2026 / Revised: 3 September 2026 / Accepted: 3 September 2026 / Published: 11 September 2026
(This article belongs to the Section Meteorology)
Browse Figures
Versions Notes

Abstract

Island terrain poses a significant challenge in tropical cyclone (TC) track prediction, particularly during the early development of TCs. Using the Weather Research and Forecasting (WRF) model, China Meteorological Administration (CMA) best-track data, and ERA5 reanalysis, we conducted control (CTRL), terrain-removal (Te0P), terrain-enhancement (Te2P), and no-land-sensible-heat (HFX0) experiments for Typhoon Yagi (2024) near Luzon during 1–3 September. CTRL showed a pronounced northward track bias. Te0P shifted westward and closer to observations, whereas Te2P further amplified the northward bias; HFX0 remained close to CTRL. Terrain-height changes substantially modified the surrounding mid-tropospheric environmental flow and asymmetric horizontal-advection potential-vorticity tendency (PVTh). Although Te0P produced the closest track, its PVTh orientation deviated most from ERA5, suggesting that the track improvement partly reflects a compensating reduction in the excessive northward response in CTRL rather than a uniformly more realistic dynamical structure. Land sensible heating mainly affected the boundary layer and nearby circulation, with weaker effects on the broader mid-tropospheric flow and track. These results show that terrain dynamics dominate the simulated track response, while land sensible heat provides secondary modulation, and highlights the need to evaluate track, environmental flow, and PVTh jointly near complex islands.

1. Introduction

Tropical cyclone (TC) track forecasts directly affect estimates of landfall location, the spatial extent of hazards, and disaster-preparedness decisions. Over recent decades, advances in observing systems, data assimilation, and numerical weather prediction have substantially improved TC track forecasts over the western North Pacific and extended useful forecast lead times [1,2]. Nevertheless, TC tracks remain difficult to predict under complex circulation regimes, and numerical models can still produce substantially different solutions [3]. Track predictability is closely related to the quality of the forecast environmental steering flow; the relatively low steering-flow forecast skill over the western North Pacific remains an important constraint on track prediction [4].
TC motion is governed primarily by the environmental steering flow, with additional contributions from beta drift, environmental potential-vorticity (PV) anomalies, asymmetric vortex structure, and diabatic processes [5]. Under weak environmental steering, the relative contributions of vortex asymmetries, convective feedbacks, and initial-condition uncertainty increase, and ensemble members may separate rapidly [6]. Recurvature and unusual track changes can also arise from changes in deep-layer steering and interactions with anomalous environmental circulation systems [7,8,9]. Thus, diagnosing track errors requires not only an examination of the local vortex structure but also a quantitative assessment of the surrounding environmental flow and its vertical distribution.
When a TC approaches an island or coastal mountain range, terrain–vortex interaction introduces additional complexity. Idealized and real-case studies have shown that island mountains can alter upstream tracks and translation speeds, or even produce discontinuous center relocations, through dynamical blocking, flow splitting and deflection, terrain-induced pressure anomalies, vortex deformation, lee-vortex formation, and center reorganization [10,11,12,13]. Terrain effects depend not only on mountain height, but also on mountain width and slope, static stability, vortex intensity and size, translation speed, landfall position, and approach angle [14]. Numerical and idealized studies further show that terrain height, island geometry, and surface characteristics can modify upstream deflection and circulation adjustment [15,16,17,18]. Comparisons of Taiwan and Philippine topographies indicate that conclusions derived from Taiwan’s continuous Central Mountain Range cannot be applied uncritically to the multi-island and multi-range environment of the Philippines [16].
Recent high-resolution simulations have emphasized the case dependence of terrain-induced track deflection. Taiwan’s terrain modified the motion, circulation, and intensity evolution of Typhoon Lekima and Tropical Cyclone Atsani [19,20], while terrain effects on precipitation and vorticity evolution have also been documented for landfalling storms [21]. Experiments with different storm tracks relative to Taiwan further demonstrate that the response depends on storm–terrain geometry [22]. Typhoon Chanthu exhibited successive rightward and leftward deflections near Taiwan, showing that terrain does not impose a single preferred deflection direction [23]. Recent dynamical modeling and a WRF study of Typhoon Doksuri across the Luzon Strait likewise indicate that terrain slope, vortex intensity, approach angle, and physical configuration jointly control the track response [24,25].
Potential-vorticity tendency (PVT) diagnostics provide a useful dynamical framework for linking environmental circulation, vortex propagation, and track deflection. A TC may be viewed as a positive PV anomaly embedded in the environment, and its motion can be related to the azimuthal wavenumber-1 component of PVT [26,27]. PVT diagnostics have subsequently been applied to high-resolution simulations, although convective-scale PV anomalies and spatial noise can reduce diagnostic stability and require regridding, smoothing, and azimuthal decomposition [28]. In the present study, we focus specifically on the horizontal-advection contribution to PVT, denoted PVTh, which is interpreted together with the environmental flow and observed motion rather than as a unique predictor of TC motion.
Most terrain–TC studies have focused on mature storms with well-organized structures or on prescribed axisymmetric idealized vortices, whereas terrain responses during early development remain less well understood. Weak or organizing storms often have relatively shallow circulations, and the low-level circulation center, midlevel vorticity center, and deep-convection center may not be vertically aligned. The Kong-Rey case demonstrated that asymmetric latent heating can separate upper- and lower-level centers and substantially affect storm structure and motion [29]. These characteristics imply that terrain-deflection mechanisms derived for mature intense TCs may not apply directly to an early-stage, vertically evolving vortex.
In addition to terrain elevation, surface roughness, soil conditions, and land–atmosphere heat exchange influence the boundary-layer structure of a TC approaching land. Idealized studies have shown that land–sea roughness contrasts generate asymmetric near-surface winds and can influence TC motion, whereas the direct track effect of land sensible heat flux is generally weaker than that of friction and roughness [30]. Landfalling-TC studies also show that terrain and land–sea contrasts can modify circulation, precipitation, and translation speed [20,21,31]. A no-sensible-heat experiment can therefore diagnose sensitivity to land-surface thermal forcing, but it should be interpreted as a perturbation to boundary-layer thermodynamics and asymmetric circulation rather than evidence of a systematic sensible-heat bias in the real atmosphere.
Typhoon Yagi developed east of the Philippines in 2024, crossed Luzon, entered the South China Sea, and subsequently intensified into a super typhoon. Its early track, passage across the Philippines, and later intensification posed substantial challenges for operational models [32]. Observational analysis indicates that Yagi’s subsequent evolution over the South China Sea involved pronounced oceanic and thermodynamic changes [33]. The case combines an organizing vortex, complex Luzon topography, and land-surface thermal exchange, and is therefore suitable for assessing the relative contributions of terrain dynamics and sensible heating to track-simulation errors.
Building on these considerations, this study examines Typhoon Yagi during its early development near Luzon to assess the relative contributions of complex-island terrain dynamics and land sensible heating to simulated track error, and to clarify the mid-tropospheric dynamical link between terrain-height changes and track deflection. In contrast to studies focused primarily on mature tropical cyclones or prescribed axisymmetric vortices, the present analysis targets an organizing storm and separates terrain-dynamic sensitivity from land-surface thermal sensitivity within a common numerical framework.
Three questions are addressed: (1) whether the premature northward turn and northward track bias in CTRL are associated with an overly strong model response to terrain-related blocking and flow-around adjustment over Luzon; (2) how terrain-height changes modify the mid-tropospheric environmental flow and horizontal-advection PVTh asymmetry; and (3) how land sensible heat flux provides secondary modulation of the track and surrounding circulation when terrain dynamics dominate. Addressing these questions provides a process-oriented basis for attributing early-stage TC track errors and interpreting terrain and land-surface sensitivity experiments near complex islands.

2. Materials and Methods

2.1. Data Sources

The CMA tropical cyclone best-track dataset compiled by the Shanghai Typhoon Institute of the China Meteorological Administration was used to represent the observed position of Yagi and to evaluate track and motion-direction errors in the simulations [34,35]. ERA5, the fifth-generation global atmospheric reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF), was used as the reference for the mid-tropospheric environmental field and horizontal-advection potential-vorticity tendency (PVTh) diagnostics [36]. The analyzed ERA5 variables included zonal and meridional wind, potential vorticity (PV), and geopotential height. For direct comparison, the WRF wind and PV fields used in the PVTh analysis were interpolated onto the same regular latitude–longitude grid as ERA5.
Initial and lateral boundary conditions for WRF were obtained from the National Centers for Environmental Prediction Global Forecast System (NCEP GFS) 0.25° historical archive [37]. The GFS data have a temporal interval of 6 h and provide the atmospheric temperature, moisture, wind, pressure, and land-surface states required for model initialization and boundary forcing.
Static terrain and land-surface data were obtained from the geographical datasets distributed with the WRF Preprocessing System (WPS) [38], including terrain elevation, land-use category, soil type, and vegetation information. This study focuses on terrain elevation over Luzon and its surrounding region; terrain-removal and terrain-enhancement experiments were constructed from the control-terrain field to assess the influence of complex topography on tropical cyclone (TC) motion and the mid-tropospheric circulation.

2.2. WRF Model Configuration and Sensitivity Experiments

2.2.1. WRF Model Configuration

The Weather Research and Forecasting Model version 4.3 (WRF v4.3) was used for the case simulation [38]. Three two-way nested domains were configured with horizontal grid spacings of 18, 6, and 2 km and grid dimensions of 199 × 135, 400 × 262, and 706 × 493, respectively. The model employed 34 vertical levels with a model top at 50 hPa. The integration extended from 1800 UTC 1 September to 0600 UTC 3 September 2024, for a total of 36 h. The main model time step was 30 s. Output intervals were 180 min for d01 and d02 and 60 min for d03.
The physical parameterization configuration was based on the WRF Tropical Physics Suite and adjusted according to the resolution of each nested domain. The WRF Single-Moment 6-Class scheme (WSM6) was used for cloud microphysics [39], and the Rapid Radiative Transfer Model for General Circulation Models (RRTMG) was used for both longwave and shortwave radiation [40]. The Kain–Fritsch cumulus parameterization was applied in d01 and d02 [41], whereas cumulus parameterization was disabled in the 2 km d03 domain so that deep convection was represented primarily by the model dynamics and explicit microphysics. The Yonsei University (YSU) scheme was used for the planetary boundary layer [42], the revised MM5 Monin–Obukhov scheme for the surface layer [43], and the Noah land-surface model for land processes [44,45].
The nested-domain configuration is shown in Figure 1.

2.2.2. Sensitivity Experiment Design

One control simulation and three sensitivity experiments were conducted to distinguish the effects of Luzon terrain dynamics and land sensible heat flux. CTRL used the realistic terrain, standard land-surface fluxes, and the common physical parameterizations. In Te0P, terrain elevation within the Luzon target region was set to 0 m while the land-surface properties were retained. In Te2P, terrain elevation within the target region was set to twice that in CTRL. HFX0 retained the realistic terrain but set the sensible heat flux to 0 W m−2 over land. Except for the perturbed variable, all experiments used the same initial and boundary conditions, integration period, model domains, physical parameterizations, and output frequency.
The experiment design is summarized in Table 1.

2.3. Diagnostic Methods

The CMA best-track positions were used as the observed TC track. The WRF TC center was identified from the minimum sea-level pressure and cross-checked against the low-level cyclonic circulation and vortex structure to reduce the possibility of misidentification caused by local pressure perturbations over complex terrain. The temporal continuity of the resulting track was also examined, and no unrealistic center jumps were found during the analyzed period. No spatial or temporal smoothing was applied to the retained TC-center coordinates. Track error was defined as the great-circle distance between the simulated and CMA centers at the same time. TC motion direction was calculated from the displacement between adjacent center positions. Direction angles of 0, 90, 180, and 270° denote motion toward the east, north, west, and south, respectively.
To characterize the mid-tropospheric environmental flow surrounding the TC, the area-weighted mean horizontal wind was calculated separately at 500, 600, and 700 hPa within an annulus 300–550 km from the center, followed by averaging across the three pressure levels. The area weight on the regular latitude–longitude grid was proportional to cos φ, where φ is latitude. ERA5 diagnostics were centered on the contemporaneous CMA best-track position, whereas each WRF experiment was centered on its own simulated TC position. The resulting zonal and meridional components are denoted by u and v. The environmental-flow speed was calculated as V = u 2 + v 2 , and the vector direction was calculated using a t a n 2 ( v , u ) .
This quantity is used as a proxy for the mid-tropospheric environmental steering flow. The analysis emphasizes relative differences among ERA5 and the WRF experiments rather than requiring the vector direction to coincide with the instantaneous TC motion at every time.
To investigate the dynamical processes associated with track deflection, the horizontal-advection potential-vorticity tendency (PVTh) was calculated and averaged over 500, 600, and 700 hPa according to Equation (1):
P V T h = − u ∂ P V ∂ x + v ∂ P V ∂ y ,
where PV denotes potential vorticity, and u and v are the horizontal components of wind velocity. Positive PVTh indicates a local tendency for PV to increase, and its azimuthal distribution is used here to characterize the orientation of the dominant horizontal-advection asymmetry relative to the TC center. For the PVTh diagnostic, the WRF PV and horizontal wind fields at 500, 600, and 700 hPa were first vertically averaged and then interpolated onto a 0.25° regular latitude–longitude grid comparable to ERA5. Before the horizontal PV gradients were calculated, the layer-mean PV field was lightly smoothed with a Gaussian filter using σ = 0.8 grid points. After PVTh was calculated, the resulting field was further smoothed with σ = 1.0 grid point to suppress residual grid-scale noise. The same smoothing parameters were applied to the ERA5 diagnostic. The analysis therefore emphasizes the spatial orientation of dominant positive PVTh anomalies rather than grid-scale extrema or exact pointwise agreement in magnitude, and interprets that orientation together with the observed and simulated motion vectors.
Difference fields CTRL−Te0P and CTRL−HFX0 were calculated to compare the responses to terrain and sensible heating. CTRL−Te0P represents the combined environmental-field difference between simulations with realistic and removed terrain, whereas CTRL−HFX0 compares simulations with and without land sensible heat flux. As the simulated tracks diverge, these difference fields include not only direct environmental adjustment to the perturbed process but also contributions from differences in TC position, intensity, and structure. They are therefore interpreted together with the track, annular environmental-flow, streamline, and PVTh diagnostics rather than as pure responses to a single forcing.
All post-processing was performed in Python 3.12. WRF output was read with netCDF4 and wrf-python; numerical extraction, interpolation, area weighting, and statistical calculations used NumPy and pandas; Gaussian filtering used SciPy; and figures were produced with Matplotlib 3.10.0 and Cartopy 0.25.0. Intermediate annular-mean, directional, and error diagnostics were exported to machine-readable CSV or statistical tables where appropriate before plotting. These procedures were applied consistently when compiling the data shown in the Results figures.

3. Results

3.1. Track Responses and Mid-Tropospheric Circulation Changes

Figure 2 compares the CMA best track with the simulated tracks. CTRL lies markedly north of the CMA track, indicating that the control simulation did not accurately reproduce Yagi’s motion near Luzon. Te0P shifts westward and is generally closer to the CMA track, whereas Te2P exhibits the strongest northward bias. HFX0 differs from CTRL, but the difference is substantially smaller than those in the terrain-sensitivity experiments.
These results show that the northward CTRL track bias is closely related to the model sensitivity to Luzon topography. The closer agreement of Te0P with the CMA track does not imply that a no-terrain state is more realistic; rather, it indicates that the terrain-related northward track response in CTRL may be too strong under the present model configuration. The further northward displacement in Te2P confirms that stronger terrain forcing can amplify this response in the present case.
Figure 3 shows the evolution of the three-level mean environmental-flow proxy within the 300–550 km annulus. ERA5 exhibits a pronounced northward component at all analyzed times. Differences among the experiments are relatively small initially but increase after 1200 UTC 2 September. Te2P develops a stronger resultant flow and northward component, reaching approximately 6.1 and 5.4 m s−1, respectively, at 0000 UTC 3 September. Te0P maintains a weaker environmental flow and northward component. CTRL and HFX0 remain close, indicating that suppressing land sensible heat flux does not substantially alter the surrounding mid-tropospheric flow. The zonal component varies non-monotonically, showing that the terrain response is stage-dependent; its clearest signal is the strengthening of the northward component and resultant flow during the key interaction period, rather than a persistent weakening of the westward component throughout the simulation.
Figure 4 presents the three-level mean wind field at 1200 UTC 2 September. All datasets and experiments contain a pronounced cyclonic circulation, but the circulation center, strong-wind-band location, and streamline curvature differ. In ERA5, the midlevel circulation center lies southwest of the CMA best-track center, indicating that the reanalysis midlevel vortex center does not fully coincide with the low-level composite position represented by the best track. This offset may reflect incomplete vertical alignment during early development and the smoothing of the inner-core circulation in ERA5; however, the cause is not the focus of the present analysis and should be interpreted cautiously. The WRF midlevel circulations are approximately centered on their respective simulated centers. Te2P exhibits more tightly packed streamlines and a more concentrated strong-wind band to the east and northeast of the center. Te0P has a comparatively diffuse circulation, whereas CTRL and HFX0 are broadly similar.
These results indicate that Luzon terrain does not directly pull the TC toward a fixed direction. Instead, it changes the spatial organization of the surrounding mid-tropospheric circulation, the position of the strong-wind band, and streamline curvature, thereby establishing different environmental backgrounds for the track response. Together with Figure 3, the results show that enhanced terrain strengthens the northward component and resultant environmental flow during the key stage, while the zonal response remains stage-dependent.

3.2. PVTh Responses and Comparison of Terrain-Dynamic and Sensible-Heat Effects

Figure 5 shows the three-level mean horizontal-advection PVTh at the same time. In ERA5, positive PVTh anomalies are located mainly north to northwest of the CMA center, indicating a dynamical asymmetry favorable for northward or northwestward propagation. Positive PVTh in CTRL is also concentrated north to northwest of the center but is more localized. In Te2P, positive PVTh is more pronounced north and northeast of the center, consistent in orientation with its stronger northward track response. The PVTh configurations in Te0P and HFX0 differ from CTRL to varying degrees, indicating that both terrain height and land-surface thermal processes can modify horizontal PV advection, although the terrain response is more evident.
The PVTh analysis suggests one possible pathway by which terrain affects the track: modification of the mid-tropospheric horizontal wind changes the asymmetric distribution of horizontal PV advection. Concentration of positive PVTh north or northwest of the TC is generally favorable for propagation in those directions. This relationship is not one-to-one, however, and must be interpreted together with the environmental flow and actual motion.
Figure 6 compares the diagnosed PVTh orientation with the TC motion direction. The ERA5 PVTh-orientation vector contains a clear northward component, whereas the CMA motion is more northwestward, indicating an overall relationship but not exact correspondence. The PVTh-orientation error and motion-direction error in CTRL are 25.5° and 25.2°, respectively. CTRL therefore reproduces part of the northward PVTh characteristic in ERA5, although its motion remains too far north.
Among the sensitivity experiments, Te0P has the motion direction closest to the CMA track, with an error of only 2.4°, whereas Te2P has the largest motion-direction error, 49.1°. Removing terrain therefore weakens the modeled northward track response, while enhanced terrain amplifies the error. However, Te0P has the largest PVTh-orientation error relative to ERA5, 82.0°. HFX0 has a PVTh-orientation error of only 3.7° but a motion-direction error of 33.0°.
The closer Te0P track therefore does not imply that its complete mid-tropospheric dynamical structure is closest to ERA5. Instead, it may reflect compensating suppression of the overly strong terrain-related northward response in CTRL. PVTh is thus an important diagnostic linking changes in the environmental flow to the track response, but it is not a unique predictor of TC motion.
Figure 7 shows the mid-tropospheric field differences between CTRL and Te0P. At 0600 and 1200 UTC 2 September, the 500 hPa geopotential-height difference and the three-level mean wind-speed difference are concentrated near Luzon and within the outer TC circulation. Both the extent and magnitude of the differences increase markedly by 1200 UTC, indicating that terrain-height changes substantially modify the surrounding mid-tropospheric environment. Because the TC centers in the two experiments have already separated, strong differences near the centers include contributions from environmental adjustment as well as differences in vortex position and structure; they should not be interpreted entirely as direct terrain forcing.
In combination with the track, annular environmental flow, streamline, and PVTh results, the CTRL−Te0P differences indicate that Luzon terrain can reorganize the mid-tropospheric flow through blocking, flow splitting, and terrain–vortex interaction, and thereby modify the asymmetry of horizontal PV advection. The terrain-dynamic response is therefore the main factor separating the CTRL and Te0P tracks.
Figure 8 shows the corresponding CTRL−HFX0 differences. Compared with Figure 7, the three-level mean wind-speed differences are smaller in magnitude and more localized, mainly around the TC inner region, rainbands, and land areas. The 500 hPa geopotential-height difference is also primarily local and does not exhibit an adjustment over the same spatial extent as in the terrain experiment. Land sensible heat flux can therefore affect local thermal structure, boundary-layer convergence, and the near-center wind field, but its overall effect on the surrounding mid-tropospheric environmental flow and PVTh configuration is weaker than that of terrain-height changes.
Thus, in this case, land sensible heat flux primarily provides a local and secondary modulation rather than the dominant cause of the northward CTRL track bias.
Taken together, the track, environmental-flow, mid-tropospheric circulation, PVTh, and difference-field diagnostics are summarized conceptually in Figure 9. Terrain-height changes substantially reorganize the surrounding mid-tropospheric circulation and, during the key interaction period, strengthen the northward environmental-flow component while modifying the spatial configuration of positive PVTh anomalies, thereby favoring a stronger northward track response. In contrast, the relatively small CTRL–HFX0 track and circulation differences indicate that land sensible heating mainly provides more localized, secondary adjustment. Figure 9 is a conceptual synthesis of multiple diagnostics rather than a strict one-variable-to-track causal mapping.

4. Discussion

These findings imply that the influence of Luzon’s complex terrain cannot be reduced to the fixed deflection relationships obtained in some idealized Taiwan experiments. In the present case, enhanced terrain amplifies the northward rather than southward track response. This contrast indicates that terrain height alone does not determine the sign of deflection; the response depends on storm–terrain geometry, the background environmental flow, and the spatial configuration of the island terrain and coastline. Over Luzon, blocking, flow splitting, and flow-around effects first modify the lower-tropospheric circulation and can extend their influence upward through pressure adjustment, mass continuity, vertical momentum transport, and vortex-structure adjustment. The diagnosed 500–700 hPa environmental-flow and PVTh changes should therefore be interpreted as the integrated mid-tropospheric response to terrain–vortex interaction rather than as a direct mechanical forcing of the TC toward a fixed direction. An observational study of TCs crossing Mindanao and the Visayas also found that track-deflection behavior differed between island groups and between weaker and stronger storms [46]. Although that study did not isolate a causal terrain mechanism, its results support the cautious interpretation that terrain-related track responses may depend on island configuration and storm intensity.
From a model-error perspective, the closer Te0P track does not imply that a no-terrain atmosphere is more realistic. ERA5 already exhibits a clear northward environmental-flow component and positive PVTh anomalies on the northern side of the cyclone, indicating that the real environment contained a dynamical background favorable for poleward turning. The main CTRL problem is therefore not the creation of an entirely spurious northward signal, but a likely over-strong terrain-related circulation adjustment and track response under the present model configuration. Removing terrain suppresses this response and shifts Te0P westward, reducing the geometric track and motion-direction errors. However, Te0P has the smallest motion-direction error while its PVTh orientation is not the closest to ERA5. Its track improvement is therefore best interpreted as dynamical compensation rather than as a uniformly improved dynamical state. This result also cautions against judging the realism of an individual physical process solely from whether a sensitivity experiment produces a smaller track error.
The PVTh results further show that horizontal PV advection is a useful link between changes in the mid-tropospheric flow and the track response, but it is not a unique predictor of TC motion. The present PVTh diagnostic isolates only the horizontal-advection contribution and does not separately quantify diabatic PV generation associated with asymmetric latent heating. Because Yagi was still developing during the analyzed period, asymmetric convective heating, rainband reorganization, boundary-layer convergence, and vertical vortex alignment may also affect displacement or reorganization of the low-level center. An auxiliary circulation-center check at 850, 700, and 500 hPa did not show a systematic correspondence between vertical alignment and the inter-experiment track differences during the key deflection period. Vertical-structure evolution may therefore modulate individual early-stage center displacements, but it does not appear to be the primary explanation for the improved Te0P track.
Consistent with this limitation, Wang et al. [47] showed that terrain blocking produced an asymmetric rainfall pattern during Typhoon Fanapi (2010) and that the associated asymmetric latent heating caused a temporary reduction in translation speed. Their PV-tendency diagnosis explicitly included horizontal PV advection, vertical PV advection, and diabatic heating, confirming that the horizontal-advection PVTh term examined here represents only one component of the complete PV tendency. Compared with terrain-height sensitivity, the track response to land sensible heat flux is substantially weaker in this case. The similarity between CTRL and HFX0 and the more localized CTRL−HFX0 differences indicate that land sensible heating mainly perturbs boundary-layer thermodynamics and nearby circulation. Such perturbations may indirectly affect the developing vortex through changes in stability, turbulent mixing, low-level convergence, and convective asymmetry, but these pathways and the complete diabatic PV budget were not quantified separately here. The present evidence therefore supports a local and secondary sensitivity to land sensible heating, rather than a claim that CTRL contains a systematic sensible-heat bias or that sensible heating is the primary cause of the northward track error.
Several limitations remain. The ERA5 midlevel circulation center is offset from the CMA best-track center, so annular environmental-flow estimates centered on the CMA position may be affected by center mismatch. In addition, each WRF experiment uses its own simulated center for the annular average, so inter-experiment differences contain both genuine environmental changes and sampling differences caused by track separation. The CTRL−Te0P and CTRL−HFX0 difference fields also combine center displacement, intensity evolution, and structural deformation and therefore cannot be interpreted as pure responses to a single forcing. Finally, the conclusions are based on one early-stage TC case. Their generality should be tested with storms of different intensities, approach angles, terrain configurations, and environmental-flow regimes. A useful future direction is to convert high-resolution physical descriptors—such as terrain height, slope and orientation, storm–terrain geometry, land–sea contrast, blocking and flow-splitting indices, and terrain-induced circulation or PVTh asymmetries—into physically interpretable predictors for deep-learning track frameworks. Such a physics-informed approach could test whether the terrain-dynamic signals identified here improve track prediction near complex islands while retaining dynamical interpretability.

5. Conclusions

Using WRF control and sensitivity experiments together with the CMA best track and ERA5 reanalysis, we examined how Luzon topography and land sensible heat flux affected the simulated early-stage track of Typhoon Yagi (2024), with particular emphasis on the likely source of the northward CTRL track bias. The main conclusions are as follows.
  • Luzon terrain dynamics are the dominant source of the track differences among the sensitivity experiments. CTRL lies markedly north of the CMA best track; Te0P shifts westward and closer to observations, whereas Te2P further amplifies the northward bias. HFX0 remains much closer to CTRL. Together with the northward dynamical background present in ERA5, these results suggest that the CTRL bias is more consistent with a likely over-strong terrain-related circulation adjustment and northward track response under the present model configuration than with an entirely spurious northward signal.
  • Terrain-height changes affect the track mainly through stage-dependent reorganization of the surrounding mid-tropospheric circulation and PVTh asymmetry. During the key interaction period, Te2P develops a stronger 500–700 hPa northward environmental-flow component and resultant speed, whereas Te0P remains weaker; the zonal component does not show a persistent monotonic response. Te0P nevertheless has the smallest motion-direction error while its PVTh orientation differs substantially from ERA5, indicating dynamical compensation and confirming that PVTh orientation must be interpreted together with environmental flow and actual TC motion.
  • Land sensible heat flux provides a weaker, more localized modulation of the early-stage track than terrain dynamics. HFX0 remains close to CTRL in both track and surrounding mid-tropospheric flow, while CTRL−HFX0 differences are more localized than CTRL−Te0P differences. The sensible-heat effect is therefore most consistent with secondary adjustment of boundary-layer thermodynamics and nearby circulation rather than the primary cause of the CTRL northward bias.
The principal advantage of this study is the separation of terrain-dynamic sensitivity from land-surface thermal sensitivity for an organizing TC, while evaluating the resulting track response jointly with environmental-flow and PVTh diagnostics. The results show that geometric track improvement in a sensitivity experiment does not necessarily imply a uniformly more realistic dynamical structure, an important consideration for physical attribution of track errors near complex islands.
The conclusions are based on a single case and remain subject to uncertainties associated with center mismatch, track-dependent sampling, and vortex-position and structural differences in the difference fields. Their applicability should therefore be tested with additional storms spanning different intensities, approach tracks, terrain configurations, and environmental-flow regimes.

Author Contributions

Conceptualization, Y.Z. and W.L.; methodology, Y.Z.; software, Y.Z.; validation, Y.Z.; formal analysis, Y.Z.; investigation, W.L.; data curation, W.L.; writing—original draft preparation, Y.Z. and W.L.; writing—review and editing, Y.Z.; visualization, W.L.; supervision, W.L.; funding acquisition, W.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research is funded by the Guangdong Major Project of Basic and Applied Basic Research (Grant 2020B0301030004) and the Guangdong Basic and Applied Basic Research Foundation (grant numbers 2023A1515240045; 2025A1515510010).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The CMA tropical cyclone best-track dataset is available from the Shanghai Typhoon Institute of the China Meteorological Administration at (https://tcdata.typhoon.org.cn/ (accessed on 16 July 2026). ERA5 pressure-level reanalysis data are available from the Copernicus Climate Data Store at (https://cds.climate.copernicus.eu/datasets/reanalysis-era5-pressure-levels (accessed on 16 July 2026). The NCEP GFS 0.25° Global Forecast Grids Historical Archive is available from the NSF NCAR Geoscience Data Exchange at (https://gdex.ucar.edu/datasets/d084001/ (accessed on 16 July 2026).

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Yu, H.; Chen, G.; Zhou, C.; Wong, W.-K.; Yang, M.; Xu, Y.; Chen, P.; Wan, R.; Hu, X. Are we reaching the limit of tropical cyclone track predictability in the Western North Pacific? Bull. Am. Meteorol. Soc. 2022, 103, E410–E428. [Google Scholar] [CrossRef] [Scilit]
  2. Conroy, A.; Titley, H.; Rivett, R.; Feng, X.; Methven, J.; Hodges, K.; Brammer, A.; Burton, A.; Chakraborty, P.; Chen, G.; et al. Track forecast: Operational capability and new techniques—Summary from the Tenth International Workshop on Tropical Cyclones (IWTC-10). Trop. Cyclone Res. Rev. 2023, 12, 64–80. [Google Scholar] [CrossRef] [Scilit]
  3. Tang, C.K.; Chan, J.C.L.; Yamaguchi, M. Large tropical cyclone track forecast errors of global numerical weather prediction models in the Western North Pacific basin. Trop. Cyclone Res. Rev. 2021, 10, 151–169. [Google Scholar] [CrossRef] [Scilit]
  4. Ye, Y.; Zhou, F. Reasons for different predictability of tropical cyclone tracks in the Western North Pacific and Atlantic Oceans. Bull. Am. Meteorol. Soc. 2024, 105, E2263–E2282. [Google Scholar] [CrossRef] [Scilit]
  5. Chan, J.C.L. The physics of tropical cyclone motion. Annu. Rev. Fluid Mech. 2005, 37, 99–128. [Google Scholar] [CrossRef] [Scilit]
  6. Ashcroft, J.; Schwendike, J.; Griffiths, S.D.; Ross, A.N.; Short, C.J. The impact of weak environmental steering flow on tropical cyclone track predictability. Q. J. R. Meteorol. Soc. 2021, 147, 4122–4142. [Google Scholar] [CrossRef] [Scilit]
  7. Akter, N.; Tsuboki, K. Recurvature and movement processes of tropical cyclones over the Bay of Bengal. Q. J. R. Meteorol. Soc. 2021, 147, 3681–3702. [Google Scholar] [CrossRef] [Scilit]
  8. Li, Y.; Heming, J.; Torn, R.D.; Lai, S.; Xu, Y.; Chen, X. Unusual tracks: Statistical characteristics, controlling factors, and model prediction. Trop. Cyclone Res. Rev. 2023, 12, 309–322. [Google Scholar] [CrossRef] [Scilit]
  9. Li, H.; Yan, Z.; Peng, M.; Ge, X.; Wang, Z. Unusual tropical cyclone tracks under the influence of an upper-tropospheric cold low. Mon. Weather Rev. 2024, 152, 39–58. [Google Scholar] [CrossRef] [Scilit]
  10. Chang, S.W.-J. The orographic effects induced by an island mountain range on propagating tropical cyclones. Mon. Weather Rev. 1982, 110, 1255–1270. [Google Scholar] [CrossRef] [Scilit]
  11. Bender, M.A.; Tuleya, R.E.; Kurihara, Y. A numerical study of the effect of island terrain on tropical cyclones. Mon. Weather Rev. 1987, 115, 130–155. [Google Scholar] [CrossRef] [Scilit]
  12. Yeh, T.-C.; Elsberry, R.L. Interaction of typhoons with the Taiwan orography. Part I: Upstream track deflections. Mon. Weather Rev. 1993, 121, 3193–3212. [Google Scholar] [CrossRef] [Scilit]
  13. Yeh, T.-C.; Elsberry, R.L. Interaction of typhoons with the Taiwan orography. Part II: Continuous and discontinuous tracks across the island. Mon. Weather Rev. 1993, 121, 3213–3233. [Google Scholar] [CrossRef] [Scilit][Green Version]
  14. Lin, Y.-L.; Chen, S.-Y.; Hill, C.M.; Huang, C.-Y. Control parameters for the influence of a mesoscale mountain range on cyclone track continuity and deflection. J. Atmos. Sci. 2005, 62, 1849–1866. [Google Scholar] [CrossRef] [Scilit]
  15. Huang, Y.-H.; Wu, C.-C.; Wang, Y. The influence of island topography on typhoon track deflection. Mon. Weather Rev. 2011, 139, 1708–1727. [Google Scholar] [CrossRef] [Scilit]
  16. Tang, C.K.; Chan, J.C.L. Idealized simulations of the effect of Taiwan and Philippines topographies on tropical cyclone tracks. Q. J. R. Meteorol. Soc. 2014, 140, 1578–1589. [Google Scholar] [CrossRef] [Scilit]
  17. Huang, K.-C.; Wu, C.-C. The impact of idealized terrain on upstream tropical cyclone track. J. Atmos. Sci. 2018, 75, 3887–3910. [Google Scholar] [CrossRef] [Scilit]
  18. Jian, G.-J.; Wu, C.-C. A numerical study of the track deflection of Supertyphoon Haitang (2005) prior to its landfall in Taiwan. Mon. Weather Rev. 2008, 136, 598–615. [Google Scholar] [CrossRef] [Scilit]
  19. Huang, C.-Y.; Sha, S.-H.; Kuo, H.-C. A modeling study of Typhoon Lekima (2019) with the topographic influence of Taiwan. Mon. Weather Rev. 2022, 150, 1993–2011. [Google Scholar] [CrossRef] [Scilit]
  20. Huang, C.-Y.; Hsu, Y.-C.; Kuo, H.-C. A numerical study for Tropical Cyclone Atsani (2020) past offshore of southern Taiwan under topographic influences. Atmosphere 2022, 13, 618. [Google Scholar] [CrossRef] [Scilit]
  21. Wu, Y.-C.; Yang, M.-J.; Rogers, R.F. Examining terrain effects on the evolution of precipitation and vorticity of Typhoon Fanapi (2010) after departing the Central Mountain Range of Taiwan. Mon. Weather Rev. 2022, 150, 1517–1540. [Google Scholar] [CrossRef] [Scilit]
  22. Luo, C.; Shang, S.; Xie, Y.; He, Z.; Wei, G.; Zhang, F.; Wang, L.; Li, X. Effects of terrain near Taiwan Island on typhoons with different tracks and typhoon waves. Water 2023, 15, 3661. [Google Scholar] [CrossRef] [Scilit]
  23. Chi, Y.-S.; Huang, C.-Y.; Skamarock, W.C. Track deflection of Typhoon Chanthu (2021) near Taiwan as investigated using a high-resolution global model. Mon. Weather Rev. 2024, 152, 2169–2188. [Google Scholar] [CrossRef] [Scilit]
  24. Chen, H.-C. An innovative dynamic model for predicting typhoon track deflections over complex terrain. Atmosphere 2024, 15, 1372. [Google Scholar] [CrossRef] [Scilit]
  25. Vu, D.-H.; Huang, C.-Y.; Nguyen, T.-C. Numerical investigation of track and intensity evolution of Typhoon Doksuri (2023). Atmosphere 2024, 15, 1105. [Google Scholar] [CrossRef] [Scilit]
  26. Wu, L.; Wang, B. A potential vorticity tendency diagnostic approach for tropical cyclone motion. Mon. Weather Rev. 2000, 128, 1899–1911. [Google Scholar] [CrossRef] [Scilit]
  27. Chan, J.C.L.; Ko, F.M.F.; Lei, Y.M. Relationship between potential vorticity tendency and tropical cyclone motion. J. Atmos. Sci. 2002, 59, 1317–1336. [Google Scholar] [CrossRef] [Scilit]
  28. Xie, T.; Wu, L.; Yu, J. Application of potential vorticity tendency diagnosis method to high-resolution simulation of tropical cyclones. Front. Earth Sci. 2022, 10, 994647. [Google Scholar] [CrossRef] [Scilit]
  29. Wang, C.-C.; Chen, Y.-H.; Li, M.-C.; Kuo, H.-C.; Tsuboki, K. On the separation of upper- and low-level centres of Tropical Storm Kong-Rey (2013) near Taiwan in association with asymmetric latent heating. Q. J. R. Meteorol. Soc. 2021, 147, 1135–1149. [Google Scholar] [CrossRef] [Scilit]
  30. Wong, M.L.M.; Chan, J.C.L. Tropical cyclone motion in response to land surface friction. J. Atmos. Sci. 2006, 63, 1324–1337. [Google Scholar] [CrossRef] [Scilit]
  31. Wu, K.; Wang, C.; Wu, L.; Zhao, H.; Cao, J. Slowdown in landfalling tropical cyclone motion in South China. Geophys. Res. Lett. 2022, 49, e2022GL100428. [Google Scholar] [CrossRef] [Scilit]
  32. Li, Y.; Chan, J.C.L.; Li, X.; Feng, W.; Zhang, Y. Challenges in forecasting Super Typhoon Yagi (2024). Trop. Cyclone Res. Rev. 2025, 14, 317–322. [Google Scholar] [CrossRef] [Scilit]
  33. He, J.Y.; Lau, T.K.; Chan, Y.W.; Cheung, P.; Lam, C.C.; Choy, C.W.; Chan, P.W. An observational analysis of Super Typhoon Yagi (2024) over the South China Sea. Meteorol. Atmos. Phys. 2025, 137, 21. [Google Scholar] [CrossRef] [Scilit]
  34. Ying, M.; Zhang, W.; Yu, H.; Lu, X.; Feng, J.; Fan, Y.; Zhu, Y.; Chen, D. An overview of the China Meteorological Administration tropical cyclone database. J. Atmos. Ocean. Technol. 2014, 31, 287–301. [Google Scholar] [CrossRef] [Scilit]
  35. Lu, X.; Yu, H.; Ying, M.; Zhao, B.; Zhang, S.; Lin, L.; Bai, L.; Wan, R. Western North Pacific tropical cyclone database created by the China Meteorological Administration. Adv. Atmos. Sci. 2021, 38, 690–699. [Google Scholar] [CrossRef] [Scilit]
  36. Hersbach, H.; Bell, B.; Berrisford, P.; Hirahara, S.; Horányi, A.; Muñoz-Sabater, J.; Nicolas, J.; Peubey, C.; Radu, R.; Schepers, D.; et al. The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 2020, 146, 1999–2049. [Google Scholar] [CrossRef] [Scilit]
  37. National Centers for Environmental Prediction; National Weather Service; NOAA; U.S. Department of Commerce. NCEP GFS 0.25 Degree Global Forecast Grids Historical Archive; NSF National Center for Atmospheric Research: Boulder, CO, USA, 2015. [CrossRef] [Scilit]
  38. Skamarock, W.C.; Klemp, J.B.; Dudhia, J.; Gill, D.O.; Liu, Z.; Berner, J.; Wang, W.; Powers, J.G.; Duda, M.G.; Barker, D.M.; et al. A Description of the Advanced Research WRF Model Version 4; NCAR Technical Note NCAR/TN-556+STR; National Center for Atmospheric Research: Boulder, CO, USA, 2021. [Google Scholar] [CrossRef]
  39. Hong, S.-Y.; Lim, J.-O.J. The WRF Single-Moment 6-Class Microphysics Scheme (WSM6). J. Korean Meteorol. Soc. 2006, 42, 129–151. [Google Scholar]
  40. Iacono, M.J.; Delamere, J.S.; Mlawer, E.J.; Shephard, M.W.; Clough, S.A.; Collins, W.D. Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res. Atmos. 2008, 113, D13103. [Google Scholar] [CrossRef] [Scilit]
  41. Kain, J.S. The Kain–Fritsch convective parameterization: An update. J. Appl. Meteorol. 2004, 43, 170–181. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  42. Hong, S.-Y.; Noh, Y.; Dudhia, J. A new vertical diffusion package with an explicit treatment of entrainment processes. Mon. Weather Rev. 2006, 134, 2318–2341. [Google Scholar] [CrossRef] [Scilit]
  43. Jiménez, P.A.; Dudhia, J.; González-Rouco, J.F.; Navarro, J.; Montávez, J.P.; García-Bustamante, E. A revised scheme for the WRF surface layer formulation. Mon. Weather Rev. 2012, 140, 898–918. [Google Scholar] [CrossRef] [Scilit]
  44. Chen, F.; Dudhia, J. Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Weather Rev. 2001, 129, 569–585. [Google Scholar] [CrossRef] [Scilit]
  45. Ek, M.B.; Mitchell, K.E.; Lin, Y.; Rogers, E.; Grunmann, P.; Koren, V.; Gayno, G.; Tarpley, J.D. Implementation of Noah land surface model advances in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res. Atmos. 2003, 108, 8851. [Google Scholar] [CrossRef] [Scilit]
  46. Petilla, C.E.R.; Tonga, L.P.S.; Olaguera, L.M.P.; Matsumoto, J. Changes in intensity and tracks of tropical cyclones crossing the central and southern Philippines from 1979 to 2020: An observational study. Prog. Earth Planet. Sci. 2023, 10, 32. [Google Scholar] [CrossRef] [Scilit]
  47. Wang, C.-C.; Chen, Y.-H.; Kuo, H.-C.; Huang, S.-Y. Sensitivity of typhoon track to asymmetric latent heating/rainfall induced by Taiwan topography: A numerical study of Typhoon Fanapi (2010). J. Geophys. Res. Atmos. 2013, 118, 3292–3308. [Google Scholar] [CrossRef] [Scilit]
Atmosphere 17 00888 g001
Figure 1. Configuration of the three nested WRF domains.
Figure 1. Configuration of the three nested WRF domains.
Atmosphere 17 00888 g001
Atmosphere 17 00888 g002
Figure 2. CMA best track and simulated tracks from 1800 UTC 1 September to 0000 UTC 3 September 2024. The red line denotes the CMA best track; black, blue, green, and purple lines denote CTRL, Te0P, Te2P, and HFX0, respectively. Markers indicate TC-center positions at different times, and shading indicates terrain elevation over Luzon.
Figure 2. CMA best track and simulated tracks from 1800 UTC 1 September to 0000 UTC 3 September 2024. The red line denotes the CMA best track; black, blue, green, and purple lines denote CTRL, Te0P, Te2P, and HFX0, respectively. Markers indicate TC-center positions at different times, and shading indicates terrain elevation over Luzon.
Atmosphere 17 00888 g002
Atmosphere 17 00888 g003
Figure 3. Time evolution of the three-level mean environmental-flow proxy at 500, 600, and 700 hPa within a 300–550 km annulus around the TC center from 0000 UTC 2 September to 0000 UTC 3 September 2024: (a) Vector direction; (b) resultant speed; and (c) zonal component u and meridional component v. Direction angles of 0, 90, 180, and 270° indicate eastward, northward, westward, and southward vectors, respectively. Solid and dashed lines in panel (c) indicate u and v, respectively. “Real” denotes ERA5. Red, black, blue, green, and purple lines denote ERA5 (“Real”), CTRL, Te0P, Te2P, and HFX0, respectively.
Figure 3. Time evolution of the three-level mean environmental-flow proxy at 500, 600, and 700 hPa within a 300–550 km annulus around the TC center from 0000 UTC 2 September to 0000 UTC 3 September 2024: (a) Vector direction; (b) resultant speed; and (c) zonal component u and meridional component v. Direction angles of 0, 90, 180, and 270° indicate eastward, northward, westward, and southward vectors, respectively. Solid and dashed lines in panel (c) indicate u and v, respectively. “Real” denotes ERA5. Red, black, blue, green, and purple lines denote ERA5 (“Real”), CTRL, Te0P, Te2P, and HFX0, respectively.
Atmosphere 17 00888 g003
Atmosphere 17 00888 g004
Figure 4. Three-level mean horizontal wind fields at 500, 600, and 700 hPa at 1200 UTC 2 September 2024: (a) ERA5 (“Real”); (b) CTRL; (c) Te0P; (d) Te2P; and (e) HFX0. Shading and streamlines indicate the three-level mean wind speed and horizontal flow, respectively. Thick black contours denote 500 hPa geopotential height. The star in the ERA5 panel marks the CMA best-track center, whereas stars in the WRF panels mark the respective simulated centers.
Figure 4. Three-level mean horizontal wind fields at 500, 600, and 700 hPa at 1200 UTC 2 September 2024: (a) ERA5 (“Real”); (b) CTRL; (c) Te0P; (d) Te2P; and (e) HFX0. Shading and streamlines indicate the three-level mean wind speed and horizontal flow, respectively. Thick black contours denote 500 hPa geopotential height. The star in the ERA5 panel marks the CMA best-track center, whereas stars in the WRF panels mark the respective simulated centers.
Atmosphere 17 00888 g004
Atmosphere 17 00888 g005
Figure 5. Three-level mean horizontal-advection potential-vorticity tendency (PVTh) at 500, 600, and 700 hPa at 1200 UTC 2 September 2024: (a) ERA5 (“Real”); (b) CTRL; (c) Te0P; (d) Te2P; and (e) HFX0. Red and blue shading indicate positive and negative PVTh, respectively. Stars mark the CMA best-track center or the respective simulated centers, and gray circles denote the azimuthal diagnostic region. Gray arrows indicate the three-level mean horizontal wind vectors, and the reference arrow denotes 30 m s−1.
Figure 5. Three-level mean horizontal-advection potential-vorticity tendency (PVTh) at 500, 600, and 700 hPa at 1200 UTC 2 September 2024: (a) ERA5 (“Real”); (b) CTRL; (c) Te0P; (d) Te2P; and (e) HFX0. Red and blue shading indicate positive and negative PVTh, respectively. Stars mark the CMA best-track center or the respective simulated centers, and gray circles denote the azimuthal diagnostic region. Gray arrows indicate the three-level mean horizontal wind vectors, and the reference arrow denotes 30 m s−1.
Atmosphere 17 00888 g005
Atmosphere 17 00888 g006
Figure 6. Comparison of PVTh orientation and TC motion direction at 1200 UTC 2 September 2024: (a) Normalized PVTh-orientation and motion-direction vectors for ERA5 and the WRF experiments; (b) PVTh-orientation error of each WRF experiment relative to ERA5; and (c) motion-direction error of each WRF experiment relative to the CMA best track. Vector lengths in panel (a) are normalized and indicate direction only.
Figure 6. Comparison of PVTh orientation and TC motion direction at 1200 UTC 2 September 2024: (a) Normalized PVTh-orientation and motion-direction vectors for ERA5 and the WRF experiments; (b) PVTh-orientation error of each WRF experiment relative to ERA5; and (c) motion-direction error of each WRF experiment relative to the CMA best track. Vector lengths in panel (a) are normalized and indicate direction only.
Atmosphere 17 00888 g006
Atmosphere 17 00888 g007
Figure 7. Mid-tropospheric field differences between CTRL and Te0P. Panels (a,c) show the 500 hPa geopotential-height difference (CTRL−Te0P) at 0600 and 1200 UTC 2 September 2024, respectively; panels (b,d) show the corresponding three-level mean wind-speed difference at 500, 600, and 700 hPa. Stars and squares denote the CTRL and Te0P TC centers, respectively.
Figure 7. Mid-tropospheric field differences between CTRL and Te0P. Panels (a,c) show the 500 hPa geopotential-height difference (CTRL−Te0P) at 0600 and 1200 UTC 2 September 2024, respectively; panels (b,d) show the corresponding three-level mean wind-speed difference at 500, 600, and 700 hPa. Stars and squares denote the CTRL and Te0P TC centers, respectively.
Atmosphere 17 00888 g007
Atmosphere 17 00888 g008
Figure 8. Mid-tropospheric field differences between CTRL and HFX0. Panels (a,c) show the 500 hPa geopotential-height difference (CTRL−HFX0) at 0600 and 1200 UTC 2 September 2024, respectively; panels (b,d) show the corresponding three-level mean wind-speed difference at 500, 600, and 700 hPa. Stars and squares denote the CTRL and HFX0 TC centers, respectively.
Figure 8. Mid-tropospheric field differences between CTRL and HFX0. Panels (a,c) show the 500 hPa geopotential-height difference (CTRL−HFX0) at 0600 and 1200 UTC 2 September 2024, respectively; panels (b,d) show the corresponding three-level mean wind-speed difference at 500, 600, and 700 hPa. Stars and squares denote the CTRL and HFX0 TC centers, respectively.
Atmosphere 17 00888 g008
Atmosphere 17 00888 g009
Figure 9. Conceptual schematic of the relative roles of terrain dynamics and land sensible heating as an early-stage TC approaches Luzon: (a) terrain-dynamic effects on the surrounding mid-tropospheric circulation, PVTh, and track response; and (b) local and secondary modulation by land sensible heat flux. The red shaded area indicates the region favorable for positive PVTh anomalies, and dashed lines indicate the track responses of the sensitivity experiments.
Figure 9. Conceptual schematic of the relative roles of terrain dynamics and land sensible heating as an early-stage TC approaches Luzon: (a) terrain-dynamic effects on the surrounding mid-tropospheric circulation, PVTh, and track response; and (b) local and secondary modulation by land sensible heat flux. The red shaded area indicates the region favorable for positive PVTh anomalies, and dashed lines indicate the track responses of the sensitivity experiments.
Atmosphere 17 00888 g009
Table 1. Design of the control and sensitivity experiments.
Table 1. Design of the control and sensitivity experiments.
ExperimentTerrain ConfigurationLand Sensible Heat Flux
CTRLRealistic terrainRetained
Te0PTerrain elevation in the Luzon target region set to 0 m; land properties retainedRetained
Te2PTerrain elevation in the target region set to twice that in CTRLRetained
HFX0Realistic terrainSet to 0 W m−2 over land
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhu, Y.; Li, W. Effects of Luzon Topography on Track Deflection During the Early Development of Typhoon Yagi (2024). Atmosphere 2026, 17, 888. https://doi.org/10.3390/atmos17090888

AMA Style

Zhu Y, Li W. Effects of Luzon Topography on Track Deflection During the Early Development of Typhoon Yagi (2024). Atmosphere. 2026; 17(9):888. https://doi.org/10.3390/atmos17090888

Chicago/Turabian Style

Zhu, Yuan, and Weibiao Li. 2026. "Effects of Luzon Topography on Track Deflection During the Early Development of Typhoon Yagi (2024)" Atmosphere 17, no. 9: 888. https://doi.org/10.3390/atmos17090888

APA Style

Zhu, Y., & Li, W. (2026). Effects of Luzon Topography on Track Deflection During the Early Development of Typhoon Yagi (2024). Atmosphere, 17(9), 888. https://doi.org/10.3390/atmos17090888

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop