1. Introduction
Tropical cyclones are an annual threat to Australia each year from November to April, often with devastating consequences. A tropical cyclone (TC) in the Australian region (90°–160°E in the Southern Hemisphere) is defined by the Australian Bureau of Meteorology as “a nonfrontal, synoptic-scale cyclone that has developed over tropical waters, with a 10-min average wind speed, V ≥ 63 km h−1 [17 m s−1] near the center of the organized wind circulation” (Dare and Davidson 2004). A “severe” TC is one that produces sustained winds of at least 118 km h−1 (33 m s−1), with gusts of 170–225 km h−1 (47–63 m s−1). An average of 12.5 TCs occur in the Australian region each year, with about 5 making landfall.
The far northeastern region of Australia (see location map Fig. 1a) is significant because the complex terrain in close proximity to the coastline acts to enhance the destructive potential of landfalling TCs. Between one and two TCs make landfall on the northeast Australian coast each year (Dare and Davidson 2004). While less prone to TCs than more sparsely populated northwestern Australia, TCs that impact the far northeast coast pose an extreme threat to large cities such as Cairns and Townsville, Australia, smaller coastal communities, and the agricultural industry of the region.
Severe TC Larry crossed the northeastern Australian coast near the town of Innisfail, Australia (Fig. 1b) on the morning of 20 March 2006, causing significant damage to infrastructure and crops, with an estimated cost of about $500 million Australian dollars. Analyses of damage to the region suggested that the local topography played a significant role in modifying Larry’s wind field, as evidenced by varying degrees of damage across relatively small distances. A preliminary damage investigation by engineers at James Cook University, Townsville, Australia, found that wind speed up over topographic ridges resulted in a significant increase in damage to buildings, whereas those structures sheltered by topography suffered much less damage. The town of Babinda, for example, to the northwest of Innisfail (Fig. 1b), sustained disproportionate damage relative to its distance from Larry’s eye by the funneling of westerly downslope winds. Westerly downslope winds also caused moderate damage in and around the Cairns region (Fig. 1b).
Cairns is often affected by these damaging westerly winds when TCs make landfall south of the region. Callaghan (2003) noted at least three historical TC events that resulted in substantial westerly wind damage in Cairns, the most extreme of which was TC Agnes in 1956. TC Agnes crossed the coast near Townsville, about 280 km south of Cairns, with a central pressure of 960 hPa. Despite its modest intensity at landfall, Agnes’s very large circulation resulted in extensive westerly wind damage in the Cairns region, with wind speeds varying from nearly calm to 43 m s−1 over very short time periods (Whittingham 1964). The damage caused by TC Agnes at Cairns was comparable to the damage near where the eye crossed the coast at Townsville. In 1986 TC Winifred, though smaller than Agnes in size, crossed the coast just south of Innisfail and caused scattered westerly wind damage to the Cairns region.
Previous studies of the influence of orography on TC structure and intensity have focused mainly on idealized numerical simulations (e.g., Chang 1982; Bender et al. 1985), or on TCs that interact with the Central Mountain Range (CMR) of Taiwan (e.g., Bender et al. 1987; Chang et al. 1993; Yeh and Elsberry 1993a, b; Lin et al. 1999, 2002, 2005, 2006; Wu and Kuo 1999; Wu 2001; Wu et al. 2002). The Sierra Madre Mountains of Mexico have also been shown to influence TCs that approach from the Gulf of Mexico (e.g., Zehnder 1993; Farfan and Zehnder 2001). Other mountainous regions affected by TCs include Luzon in the northern Philippines (Brand and Blelloch 1973) and the Caribbean Islands of Cuba, Hispaniola, and Puerto Rico (Bender et al. 1987).
The above studies have provided considerable understanding of how orography affects the track and structure of TCs. While the importance of orographic influences on landfalling TCs over northeastern Queensland has long been recognized (e.g., Whittingham 1964; Harper 1999; Callaghan 2003), to our knowledge there have been no comprehensive high-resolution numerical simulations that have explored the influence of orography on Australian TCs.
Observations of TC Larry (more information available online at http://www.bom.gov.au/weather/qld/cyclone/tc_larry; Henderson et al. 2006) motivated this high-resolution numerical simulation. Our main objective is to understand how the complex terrain of the northeastern Australian region affects TC track, winds, and precipitation, and to assess how different the impact of TC Larry would have been if the region was flat.
Section 2 is a summary of the major observational aspects of TC Larry, including its track and its characteristics before, during, and after landfall. Section 3 describes the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5) model configuration and the proposed numerical experiments. The results are given in section 4, and section 5 summarizes the findings and suggests possible future work.
2. Observations of TC Larry
TC Larry developed from a tropical low in the Coral Sea (see location in Fig. 1a), initially located about 1500 km east of Cairns at 0600 UTC 17 March 2006. The low was upgraded to an Australian category 1 TC 12 h later (1800 UTC 17 March 2006) and continued to steadily intensify as it moved westward, reaching the Australian severe TC classification during the morning of 18 March 2006. Ten hours prior to landfall (1110 UTC 19 March 2006) a wind gust of 59 m s−1 was recorded at Flinders Reef, located about 260 km east of Innisfail. An analysis of high-resolution microwave satellite imagery suggests that Larry may have briefly reached Australian category 5 intensity (i.e., gusts of greater than 78 m s−1) before making landfall as a category 4 storm (with gusts 63–78 m s−1) near the town of Innisfail (Fig. 1b). Radar scans from Willis Island showed a persistent convective asymmetry in TC Larry well before landfall, with the deepest convection located on the northern side the eyewall (P. Otto, Bureau of Meteorology 2008, personal communication).
The highest official observed wind gust near the time of landfall (2020–2120 UTC 19 March 2006) was 50 m s−1 recorded by the South Johnstone Automatic Weather Station (AWS), located about 11 km southwest of Innisfail (Fig. 1b) and 14 km inland. However, an analysis of damage to the regions most affected suggested that peak wind gusts were considerably higher, being estimated at 66 m s−1, which is consistent with an Australian category 4 TC. The highest unofficial measured wind gust was 82 m s−1 at Bellenden Ker Tower near the peak of Mt. Bellenden Ker (see Fig. 1b). The tower elevation is approximately 1450 m and is 30 km northwest of Innisfail. Cairns Airport reported a maximum wind gust of 30 m s−1 at 2201 UTC.
Heavy rainfall, with 3-h totals up to 139 mm, produced extensive flooding in coastal rivers. As TC Larry moved inland toward the low-lying area south of the Gulf of Carpentaria very heavy rain continued, with the highest recorded total of 436 mm at Gereta Station (about 700 km inland) in the 24 h preceding 2300 UTC 21 March 2006.
Larry was a small TC with destructive wind gusts of greater than 47 m s−1 extending no more than 50 km from the center. It had an unusually fast forward speed of about 8 m s−1 as it tracked to the west-northwest after landfall. Tropical Cyclone Larry decayed fairly rapidly as it moved inland, though it maintained category 1 strength or higher for several hundred kilometers. It was downgraded to a tropical low at 1500 UTC 20 March 2006, 18 h after landfall and about 500 km inland from the coast.
3. Model configuration and experimental design
Version 3.7 of MM5 was used for all simulations. A detailed description of the properties of MM5 can be found in Grell et al. (1994).
The MM5 simulations were computed on a quadruply nested, two-way interactive grid, as shown in Fig. 2. The four domains have dimensions and grid spacings of 93 × 100, 27 km (D1); 220 × 210, 9 km (D2); 445 × 286, 3 km (D3); and 385 × 268, 1 km (D4). In total, 46 vertical half-sigma levels were used with higher resolution in the boundary layer and upper troposphere to more accurately resolve the inflow and outflow layers of the TC. The model top is set at 50 hPa. High-resolution terrain data with horizontal resolution of 900 m was used in D4 to resolve the interaction of the TC wind field with topography. The complex nature of this topography in the region where Larry made landfall is shown in Fig. 1b. The steepest elevated regions of Mount Bartle Frere and Mount Bellenden Ker, at 1622 and 1593 m, respectively, were reproduced well in D4 with the highest-resolution model terrain rising to 1600 and 1484 m.
All simulations were initialized at 1200 UTC 17 March and run for 72 h until the storm decayed roughly to the Australian category 1 strength (i.e., gusts to 35 m s−1) after landfall. The initial and lateral boundary conditions were obtained from the National Centers for Environmental Prediction (NCEP) Final Analysis (FNL) dataset, with 1 × 1 degree horizontal grid spacing and 26 mandatory pressure levels from 1000 to 10 hPa. In addition, a bogus vortex was inserted into the initial state, following the TC bogussing scheme of Low-Nam and Davis (2001). The model failed to generate a TC when no bogus vortex was used, owing to the relatively coarse resolution of the FNL dataset.
PBL processes were parameterized using the Eta Model Mellor–Yamada 1.5-order local closure scheme (Janjic 1990, 1994) that includes a prognostic equation for turbulence kinetic energy (TKE). Betts–Miller cumulus parameterization (Betts 1986; Betts and Miller 1986) was used for D1 and D2 only, while cloud-microphysical processes were explicitly resolved in all domains using the Reisner mixed-phase scheme (Reisner et al. 1998). Both shortwave and longwave radiation were computed using the rapid radiation transfer model (RRTM; Mlawer et al. 1997).
To study the effect of orography on TC Larry, a sensitivity experiment (NOTOPOG) was carried out in which the terrain elevation was set to zero. Special care was taken to ensure that the sensitivity experiment had no memory of the original terrain data by allowing the model to run for a sufficiently long time (∼60 h) prior to TC landfall. By 48 h the surface and upper-level fields over land showed no sign of the orography that was removed at initialization.
4. Results
a. TC track and intensity
The simulated TC track for the realistic terrain experiment (hereafter CTRL) is in very good agreement with the observed track of TC Larry, as shown in Fig. 3. After an initial 24-h spinup period, the simulated TC developed a more westerly track, following the estimated best-track positions very closely from 1200 UTC 18 March to 1200 UTC 20 March 2006 when Larry was officially downgraded, 12 h after landfall. Varying the size and strength of the initial vortex by specifying different values of the radius of maximum wind (RMW) and the maximum tangential wind speed led to relatively large changes in simulated intensity, though only minor variations in track. Stronger initial bogus vortices generally resulted in more southerly storm tracks relative to Larry’s best track. Of more importance for the current study for the purpose of studying the effect of the local terrain on the TC’s wind field, are the time and location of landfall. Both of these agree very well with observation (Figs. 3 and 4). The simulated TC with topography crosses the coast about 2 h after the observed time of Larry’s landfall, with the southern half of the eye passing directly over Innisfail. The TC with no topography shares the same track as the CTRL TC until t = 36 h, after which it develops a more southerly component of motion that results in a landfall location about 80 km to the south of the CTRL TC and the observed landfall location of Larry (Fig. 3). Until about t = 30 h, the low-level and upper-level wind field patterns for the CTRL TC and NOTOPOG TCs are in close agreement, however by t = 36 h some notable differences begin to emerge. Analyses of the NOTOPOG TC’s wind field indicate increased mid-to-upper-level flow in the rear-left quadrant relative to the CTRL TC, resulting in greater advection of potential vorticity (toward the south) and a concomitant southerly component of the TC’s motion. While difficult to establish a direct link between the removal of orography and TC track differences prior to landfall, the relationship between orography and storm motion after landfall is more obvious. As the CTRL TC passes over the coastal range, the easterly flow on the left side of the storm descends and warms while westerly flow on the right side of the storm rises and cools (not shown). Therefore, storm-relative cyclonic vorticity is increased (decreased) on the left (right) side of the storm by stretching (compression) of vortex tubes. Hence, the vortex moves toward the southwest like the observed storm (Fig. 3), while the NOTOPOG TC maintains its westerly motion. In the Northern Hemisphere this same combination of adiabatic warming and vortex tube stretching in downslope flow results in TC movement toward the northwest (e.g., Lin et al. 2002, p. 2934).
The removal of orography results in a more intense TC relative to the CTRL TC. Figure 4 shows the estimated best-track intensity versus the simulated intensities for the CTRL and NOTOPOG simulations, respectively. After the 24-h spinup period, intensification rates for the CTRL and NOTOPOG TCs follow TC Larry’s rate of intensification closely. The central pressures in D2 are about 5–7 hPa higher than D3, however there is almost no difference between the pressures in D3 and D4, indicating that the TC eye is well resolved for grid spacing less than about 3 km. Rapid intensification from t = 36 h to t = 45 h is captured well by the model. By t = 51 h the best track and CTRL central pressures converge and remain close until landfall. Note that a substantial data-void period exists in the estimated best-track intensity between 1800 UTC 19 March and 0000 UTC 20 March 2006 (t = 54 and t = 60 in Fig. 4) including the time at which Larry crossed the coast near Innisfail around 2100 UTC 19 March 2006. The lowest official central pressure near landfall was 957 hPa at South Johnstone automatic weather station, which was near the southern edge of the eyewall, so the actual minimum pressure probably was lower. Starting from model initialization at 1712 UTC 17 March 2006, the intensification rate of the simulated TCs is the same until t = 30 h, after which the NOTOPOG TC deepens faster than the CTRL TC. The NOTOPOG TC attains a minimum central pressure in D3 of 911 hPa about 4 h prior to landfall; 18 hPa lower than the minimum central pressure (929 hPa) for the CTRL TC. Both TCs exhibit weakening in the 4 h leading up to landfall.
The time at which the intensities of the CTRL and NOTOPOG TCs begin to diverge (t = 30 h) is the same as when the TC tracks separate, suggesting differences in storm structure and/or the ambient environment. The sea surface temperature is the same for both simulations so should not cause differences in intensity. Rather, the difference in intensity appears to be the result of a distinct increase in the deep-layer (850–200 hPa) near-core vertical wind shear experienced by the CTRL TC from about t = 27 h to t = 36 h. This increase causes the CTRL vortex to tilt in the downshear direction (i.e., toward the southeast) such that the upper-level circulation is displaced about 30 km from the low-level circulation. The NOTOPOG vortex shows much less vertical tilt in accordance with the somewhat weaker deep-layer vertical shear. These results agree with previous studies that have investigated the effects of vertical wind shear on TC intensity (e.g., DeMaria 1996; Frank and Ritchie 2001; Braun et al. 2006).
b. TC structure during landfall
In addition to the structure and intensity differences prior to landfall discussed above, distinct differences also exist between the CTRL and NOTOPOG TC’s finescale characteristics during landfall. The CTRL TC makes landfall around 2300 UTC 19 March 2006 with a central pressure of 935 hPa (D4). The center of the eye crosses the coast 3 km north of Innisfail (Fig. 5). The diameter of the eye, as shown by the 1.8-km simulated radar reflectivity in Fig. 5c, is about 22 km; only slightly smaller than Larry’s actual eye, estimated to be about 25 km in diameter during landfall (Henderson et al. 2006).
In terms of finescale inner-core structure, we find that the equivalent potential temperature (θe) at 1 km within the eye is fragmented into three separate maxima, with values approaching 375 K concentrated around sections the eyewall (Fig. 6). Maxima in cyclonic vertical vorticity of up to −1.8 × 10−2 s−1 are found within the tight θe gradient along the southern and western sides of the eyewall (Fig. 6). This distribution of cyclonic vorticity reveals the distinct elliptical shape of the eye that is consistent with radar observations during Larry’s landfall (P. Otto, Bureau of Meteorology 2008, personal communication). While not an integral part of this paper, we note also that the CTRL simulation produces polygonal-shaped eyewalls in the hours leading up to landfall that match very closely with Larry’s polygonal eyewall structure observed by radar (Fig. 7). In Fig. 7 both the radar and the model eye walls appear to be pentagonal. However, at other times the eyewall takes other polygonal shapes (not shown). Previous studies have shown that polygonal eyewalls are the result of a breakdown of strong vorticity within the eyewall into discrete pools and that these play an import role in the inner-core dynamics of TCs (e.g., Schubert et al. 1999; Kossin and Schubert 2001; Wang 2002a, b).
Throughout this study, the term “surface wind” denotes the 10-m wind, which is calculated by the model based on boundary layer and land surface characteristics. For the CTRL TC, a region of maximum surface wind speed is evident in the rear-left quadrant of the circulation just offshore over the water, with velocities ranging between 50 and 58 m s−1. A second area of very strong surface winds, approaching 52 m s−1, is evident in the core region to the north of the eye over the ocean. The surface wind speed over land is significantly reduced owing to the increased surface roughness, and shows much more spatial variability as a result of the steep and complex terrain (Fig. 5a). The strongest winds over land (∼50 m s−1) are restricted to a small coastal area south of the eye exposed to the onshore flow on the left side of the storm. This onshore flow diminishes quickly with distance from the coast, slowing from 50 to 34 m s−1 over a distance of 3 km. Analyses of the tangential wind speed over land reveals two distinct maxima. The first is collocated with the eyewall with wind speeds up to 34 m s−1. The second region farther to the west is stronger, with speeds approaching 44 m s−1, despite a greater radial distance (∼25 km) from the eye. This second band of winds follows the contours of elevated terrain to the west, suggesting the TC’s tangential flow is accelerated by the orography, at least locally. No such strong, secondary inland wind maximum is evident in the NOTOPOG TC’s wind field during landfall (Fig. 8a). In addition, there are several smaller, localized, pockets of strong winds located on the upwind sides of Mt. Bartle Frere and Mt. Bellenden Ker (Fig. 5a; see location map Fig. 1b). Surface winds of up to 38 m s−1 are evident on their southern slopes where the tangential flow is partially blocked so it is forced to go around rather than over the mountains. The surface winds on the sheltered lee sides of these mountains are significantly lower (∼8 m s−1).
An east–west cross section of the meridional winds through the center of the TC during landfall reveals only a very slight asymmetry in the tangential winds, with a maximum southerly wind of 73 m s−1 about 600 m above ground level (AGL) and 21 km to the west of the TC center (Fig. 5b). On the eastern side of the TC, a similar northerly wind maximum of 77 m s−1 is present about 500 m MSL and 20 km east of the eye.
Relative to the CTRL TC, the NOTOPOG TC is more intense at landfall with a central pressure of 928 hPa (D4); however, it contains only slightly stronger surface winds with maximum values approaching 60 m s−1 (Fig. 8a). The finescale inner-core structure is characterized by a strong warm core with a maximum equivalent potential temperature of 380 K near the center of the eye (5 K higher than the 375 K for the CTRL TC) decreasing sharply with radial distance out from the center (Fig. 6b). The stronger, more symmetric warm core of the NOTOPOG TC relative to the CTRL TC is consistent with its lower central pressure at landfall.
The maximum surface winds are located in the eastern half of the circulation over water, collocated with a maximum in low-level cyclonic vorticity and a very strong gradient of equivalent potential temperature in the eyewall (Figs. 8a and 6b). An analysis of the sea level pressure field and the surface wind vectors reveals substantial cross-isobaric flow over land where the surface wind speed is generally 15–20 m s−1 less than over the ocean owing to an increase in surface roughness. The east–west cross section of the meridional wind field through the center of the TC reveals two distinct low-level jets at a radius of about 10 km from the axis of minimum winds in the eye. These jets occur at a height of about 250 m MSL. Marked asymmetry is present with a maximum of 82 m s−1 in the northerly flow east of the eye and 67 m s−1 in the southerly flow west of the eye over land (Fig. 8b). Several other, weaker, low-level wind maxima are evident to the west of the eye in Fig. 8b, extending as far as 50 km inland.
In addition to the differences in wind structure and intensity, the CTRL and NOTOPOG TCs show significant differences in their simulated reflectivities at landfall. The NOTOPOG TC is characterized by a region of relatively high reflectivity in the eastern half of the circulation offshore, with 55 dBZ up to 5 km MSL (Fig. 8c). In contrast, the CTRL TC (Fig. 5c) is more asymmetrical, with similar high values of reflectivity, approaching 55 dBZ, in the western half of the circulation over land. This region of high reflectivity results from a combination of strong frictional convergence upstream (nearer to the coast) and enhanced vertical motion as onshore flow is forced upward by the orography (not shown). A band of enhanced reflectivity caused by the melting of precipitation near the freezing level (∼5 km) is present in both simulations (Figs. 5d and 8d). The upward bulge of the freezing level within the eye signifies warming caused by strong subsidence. This feature is more pronounced in the NOTOPOG TC during landfall and is consistent with both the stronger warm core and lower central pressure relative to the CTRL TC.
c. Boundary layer turbulence
The production of turbulence within the TC boundary layer is linked closely to maximum wind gusts at the surface that may cause significant damage. Figure 9 shows the turbulence kinetic energy for the lowest model sigma level (∼19 m) during landfall for both CTRL (Figs. 9a,b) and NOTOPOG simulations (Fig. 9c). For the CTRL simulation, a marked difference in TKE is evident between land and water surfaces, with values up 180 J kg−1 over land compared with less than 80 J kg−1 over water. This difference is largely the result of enhanced shear production of TKE over land where the increased surface friction acts to reduce the surface wind speed while the flow aloft maintains its velocity. Indeed, the spatial distribution of the vertical shear in the lowest 100 m is in very close agreement with the spatial distribution of TKE. This increase in the boundary layer vertical wind shear, caused by flow transition from smooth to rough surfaces, is well documented in the literature (e.g., Powell 1982, 1987). There are several TKE maxima evident in Fig. 9a. The first of these is located on the stretch of coast exposed to the very strong onshore flow south of the TC eye. The TKE there is maximized only in the immediate vicinity of the coast and is consistent with very large shear in the lowest 100 m up to 30 m s−1. A second region of large TKE is evident in a narrow band about 20 km west of the eye associated with strong vertical wind shear. Finally, several local maxima of TKE are evident over the windward slopes of Mt. Bartle Frere and Mt. Bellenden Ker, both of which are located outside the eyewall at 2300 UTC. The elevated terrain acts to speed up the winds in the boundary layer and increase the production of shear-induced turbulence locally. For instance, the 50-m wind speed over the eastern slopes of Mt. Bartle Frere (elevation ∼1200 m) is 56 m s−1 whereas the surface wind is only 32 m s−1, resulting in extremely high shear in the lowest 50 m. In contrast, over the sheltered northern slopes of the mountain (elevation ∼960 m) the 50-m wind is only 16 m s−1. Similar speedup/sheltering effects that coincide with distinct maxima and minima of TKE are also noted over and around Mt. Bellenden Ker farther to the north.
The NOTOPOG simulation also reveals distinct regions of locally high TKE values during TC landfall, measuring up to 180 J kg−1 where the maximum tangential winds to the south of the eye experience a sudden increase in surface roughness (Fig. 9c). The overall structure is characterized by a ring of high TKE around the inner core of the circulation, particularly over land, and several smaller yet distinct bands of lesser TKE at larger radii. A noteworthy aspect of the TKE distribution in the NOTOPOG simulation is the apparent banding structure over land. These bands contain locally high values of TKE and are even more distinctive as the TC moves farther inland, particularly when the core of the circulation is entirely over land (not shown). The physical mechanisms behind these finescale bands, including their role in the exchange of heat and momentum within the TC boundary layer, are beyond the scope of the current study.
d. Influence of orography on TC winds
As introduced in the previous section, the interaction between the complex terrain and the TC wind field in the CTRL simulation results in substantial variability of low-level wind speed and direction. The terrain acts to alter the surface wind field before, during, and after landfall, as shown by the streamlines in Fig. 10. Blocking on the windward side of mountains, funneling through gaps, and leeside sheltering are all evident depending on angle of approach of the wind as well as the amount of forcing. The strongest forcing is found near the center of the vortex where the tangential flow is maximized, as should be expected. It is also apparent from Fig. 10 that some regions initially sheltered from the wind by orography are later exposed to locally high wind speeds as the TC moves westward and the flow changes direction.
As the TC approaches Mt. Bellenden Ker (Fig. 1b) from the southwest, westerly winds impinging on the western slope of the mountain range acquire enough momentum and depth to pass freely over the top of the ridge and down the other side. Figure 11 is an east–west cross section of the near-zonal winds over Mt. Bellenden Ker 20 min after landfall (2320 UTC). A maximum wind speed of 68 m s−1 is evident over the top of the ridge where the flow is locally accelerated. An observed westerly wind gust of 82 m s−1 was recorded during the passage of Larry at roughly the same location, suggesting good agreement between the simulated winds over elevated terrain and observations, despite the smoother model terrain compared with actual orography. Significant downslope winds are also evident on the lee side of the range where the near-surface westerly wind speeds range between 50 and 60 m s−1. Consequently, the mountains play an important role by facilitating the transport of pockets of higher-momentum flow down toward the surface. The Cairns region, located about 35 km to the north of Mount Bellenden Ker, also experiences these downslope winds but to a lesser degree because of the smaller mountains and larger distance from the TC eye. The simulated westerly winds in the Cairns region range from about 25 to 30 m s−1 as the TC passes to the south (not shown).
e. Downslope winds in the Port Douglas region
The morphology of the north–south-oriented mountain range situated about 13 km west of the popular tourist town of Port Douglas (Fig. 1b), with its gentle windward rise and steep leeward slope, is well-suited for generating severe downslope winds (e.g., Smith 1977; Lilly and Klemp 1979; Hoinka 1985). Indeed, CTRL successfully captures such an event. At 2300 UTC, the u-w component of the wind ranges from 4–8 m s−1 on the windward side of the range to 16–20 m s−1 on the leeward side (Fig. 12a). As the flow in the lowest 3 km approaches the mountain ridge from the west, it accelerates and is forced downward toward the surface on the leeward side. By 0030 UTC the near-surface wind speed reaches its maximum sustained value of approximately 24 m s−1 (Fig. 12b). The associated mountain wave, as depicted by the large vertical displacement of lines of constant potential temperature, is also amplified during this time. Vertical velocities up to 4 m s−1 at 2000 m AGL are noted on the leeward side of the range. The vertically propagating gravity wave is trapped by a critical layer at 10 km AGL where the zonal wind goes to zero (not shown). Critical layers have been shown to play an important role in the amplification of mountain waves and subsequent intensification of severe downslope windstorms, both in numerical and analytical studies (see e.g., Clark and Peltier 1984; Smith 1985). While the downslope winds generated here are not severe, a slightly stronger TC at landfall with stronger westerly winds at the crest of the mountain produces greater downslope winds of up to 36 m s−1 (not shown). The results suggest that a severe downslope windstorm is possible if a TC makes landfall to the south of the mountain range such that moderate-to-strong westerly flow impinges on the mountain crest.
f. Wind gusts
Wind gusts, as opposed to 1- or 10-min-averaged sustained winds, last only a few seconds but produce the majority of structural damage. The wind gust estimate (WGE) method of Brasseur (2001) is used here to estimate maximum surface gusts given certain characteristics of the model boundary layer. The assumptions of WGE are (i) wind gusts at the surface result from deflection of parcels flowing in the (whole) boundary layer, and (ii) the deflection process is accomplished by large turbulent eddies that transport air parcels to the surface. Analyses of these parameterized wind gusts for both the CTRL and NOTOPOG simulations reveal that the spatial distribution of wind gusts is far from homogeneous, even in the absence of topography, and that the estimated maximum gust does not decrease simply with distance from the center of the TC, but instead depends on both the shape and height of the underlying terrain, and the internal characteristics of the TC boundary layer.
For CTRL during landfall a region of maximum surface gusts with speeds ranging from 80 to 90 m s−1 is evident to the south of the eye over land and coincides with the area of maximum TKE shown in Fig. 9a. The maximum gust is roughly twice the speed of the surface wind for the same time and location. An analysis of wind speed profiles in the lowest 2 km to the south and west of the eye (as shown in Fig. 14) reveals accordance between the maximum wind within the PBL and the maximum estimated gust at the surface. The profile labeled A in Fig. 13, west of the eye, is characterized by a particularly strong jet of 83 m s−1 at 800 m AGL (Fig. 14a), suggesting a potential for these locally very strong winds to be transported to the surface by turbulent eddies and/or convectively driven downdrafts. Moreover, the wind increases from 26 m s−1 at the surface to 76 m s−1 at 500 m AGL resulting in favorable environmental shear for tornadogenesis (e.g., Novlan and Gray 1974; McCaul 1991). Although no tornadoes were reported, damage patterns were consistent with tornadic winds (more information available online at http://www.ga.gov.au/image_cache/GA8468.pdf). Wong and Chan (2007) showed that supergradient tangential winds near the top of the PBL are primarily the result of the vertical advection of the radial wind, in agreement with Kepert and Wang (2001). Inspection of the parameterized wind gusts at landfall reveals two distinct maxima: one south of the TC eye in the onshore flow and the other west of the eye and over land. The asymmetrical distribution of wind gusts persists as the TC moves inland (Fig. 13b), with the strongest gusts confined to the southern side of the circulation over an elevated plateau where tangential flow and translation speed are maximized. These wind gusts are about 10 m s−1 stronger than the gusts produced in the NOTOPOG TC for the same storm-relative location (i.e., south of the eye) and time after landfall, despite the lower central pressure of the NOTOPOG TC. Although the strongest gusts are in the southern half of the circulation after landfall, gusts up to 40 m s−1, capable of causing damage, extend 70 km northeast of the eye to Cairns and south along the coast.
In contrast to the CTRL TC, the parameterized wind gusts for the NOTOPOG TC during landfall are strongest in the eastern half of the circulation over the ocean and range from 70 to 80 m s−1 (Fig. 13c). These maximum gusts are about 1.6 times greater than the surface wind speed for the same time and location. As the TC moves inland, the region of maximum gusts moves cyclonically (clockwise) around the vortex to the southern half of the circulation 2 h after landfall (Fig. 13d). This transition in the location of maximum surface gusts is not surprising given the westward motion of the TC, resulting in an increase of the tangential winds south of the eye due to the translation speed of the vortex. Gusts of up to 50 m s−1 extend at least twice as far south of the eye as they do north (Fig. 13d).
g. Rainfall
Significant differences in the accumulated rainfall are noted between the CTRL and NOTOPOG run, both before and after landfall, as shown in Figs. 15 and 16. In the 12 h prior to landfall the heaviest precipitation occurs to the north of the storm track in the deep convection surrounding the TC eye. This bias is evident in both simulations, though it is particularly pronounced in the CTRL simulation. It is also in agreement with the radar-observed convective asymmetries. Both the CTRL and NOTOPOG simulations are characterized by an almost continuous swath of accumulated rainfall of greater than 200 mm to the north of the TC track, with embedded isolated amounts of more than 300 mm. The apparent asymmetry in the accumulated rainfall is likely due to the deep-layer ambient environmental shear. An analysis of the 850–200-hPa deep-layer shear reveals that the TCs in each simulation are embedded in generally weak (<5 m s−1) easterly shear in the 12 h leading up to landfall. Chen et al. (2006) show that for TCs in the Southern Hemisphere, enhanced precipitation is favored to the right of the deep-layer environmental shear vector, in agreement with the northward bias of heavy precipitation in Fig. 15. For CTRL, analyses of the column-integrated cloud liquid water content indicates maximum values occur generally in the front-left quadrant of the vortex, upstream of the heavy precipitation in the front-right/rear-right quadrants. It is likely that strong cyclonic advection around the eye plays an important role in the final distribution of precipitation. The orography in CTRL has a significant influence on the location and amount of rainfall before, during and after landfall. In the 3 h prior to landfall (2000–2300 UTC) the heaviest precipitation occurs to the north of the TC track with a distinct swath of more than 200 mm (Fig. 16a). Isolated amounts of between 75 and 100 mm are evident over the windward slopes of the coastal range where strong tangential flow is forced upward resulting in increased vertical motion (not shown). From 2300 UTC 19 March to 0200 UTC 20 March 2006, maximum accumulated rainfall values of up to 225 mm are evident over the highest orography to the north and south of the eye (Fig. 16b). The maximum on the southern side of the circulation is absent in the preceding 3-h period (Fig. 16a) highlighting the importance of orography in determining the location and amount of postlandfall precipitation. The highest 3-h observed rainfall total to 2300 UTC 19 March 2006 was 139 mm at The Boulders on Babinda Creek, at the base of Mt. Bellenden Ker. The modeled 3-h accumulated precipitation is 150 mm, in close agreement with the observed total. By 0500 UTC 20 March 2006, with the TC well inland, the accumulated rainfall amounts decrease substantially, ranging generally from 75 to 100 mm (Fig. 16c). Local rainfall amounts exceeding 100 mm occur over the mountains south of Innisfail, far from the TC core, and are due to strong onshore winds.
The amount and pattern of accumulated rainfall in NOTOPOG differs substantially from the CTRL simulation. In the 3 h prior to landfall, the heaviest precipitation occurs south of the eye with amounts exceeding 200 mm (Fig. 16d). This southward bias, in contrast to CTRL, is due partly to the enhanced low-level convergence caused by the shape of the coastline, including Hinchinbrook Island. The rainfall amounts in the 3 h after landfall (up to 175 mm; Fig. 16e) are less than in the CTRL simulation (225 mm), despite the NOTOPOG TC’s greater intensity. As the TC moves inland the precipitation increases relative to the CTRL TC with a large region of 75–100 mm and an embedded band of more than 100 mm (Fig. 16f). An analysis of the total precipitable water field suggests that the absence of steep coastal terrain allows the NOTOPOG TC to retain its moisture content farther inland compared to the CTRL TC.
h. Decay after landfall
The overall TC decay rate from landfall to 12 h after landfall is relatively insensitive to the underlying terrain, though there are marked differences in the trends of maximum surface wind speed from hour to hour (Fig. 17). The CTRL TC makes landfall with a central pressure of 935 hPa; 12 h after landfall the central pressure is 991 hPa—an increase of 56 hPa. The CTRL TC’s central pressure after landfall is in good agreement with Larry’s best-track central pressure (Fig. 4). The NOTOPOG TC makes landfall with a central pressure of 928 hPa that increases by 55 hPa to 983 hPa 12 h later. Thus, the simulated TCs decay at a rate relative to their respective central pressures at landfall, consistent with several previous studies (e.g., Schwerdt et al. 1979; Ho et al. 1987; Kaplan and DeMaria 1995).
For both CTRL and NOTOPOG simulations, the largest decrease in surface wind speed occurs in the first hour or two following landfall, as expected owing to the sudden increase in surface roughness over land. The sudden decrease in maximum surface wind is particularly pronounced in CTRL, where a reduction of 30 m s−1 occurs in the first hour after landfall.
Large differences are evident in the TCs respective decay rates, from hour to hour (Fig. 17). The influence of orography on the decay of the CTRL TC is indicated by the nonsteady nature of the decrease in maximum surface wind speed as a function of time, compared to the relatively smooth downward decay rate for the NOTOPOG TC. As the CTRL TC moves westward over the Atherton Tablelands the maximum surface wind speed actually increases from 32 to 35 m s−1 over a 2-h period. These results suggest that simple TC decay models (e.g., Kaplan and DeMaria 1995) should be used with caution when being applied to TCs that make landfall in regions of complex terrain.
5. Discussion and conclusions
The MM5 mesoscale model has been used to investigate the effects of complex terrain on TC Larry that made landfall on the northeastern Australian coast as a severe category 4 TC. The simulated track, speed of movement, strength, size, and decay rate of the simulated storm (CTRL) matched the observations closely, and reproduced many of the features of TC Larry over land. To compare with CTRL a contrasting simulation (NOTOPOG) was carried out in which the orography was completely removed, but other surface properties were retained. Both experiments were initialized using a bogus vortex to produce a TC with intensity close to the observed intensity of TC Larry as it approached landfall. Tropical cyclone genesis occurred only with a bogus initial vortex, owing to relatively coarse model input data.
Previous studies on the orographic influence of TCs have focused mainly on the effects of mountain ranges on TC motion. Examples of mountain ranges that affect TC motion include the Central Mountain Range of Taiwan and the Sierra Madre of Mexico (e.g., Chang 1982; Zehnder 1993; Yeh and Elsberry 1993a, b; Zehnder and Reeder 1997; Lin et al. 2002, 2005, 2006). The problem of orographically enhanced mesoscale precipitation associated with landfalling TCs has been the focus of a number of studies (e.g., Geerts et al. 2000; Wu 2001; Lin et al. 2002; Wu et al. 2002). Here, the focus is not only on the primary variables such as wind, pressure, and precipitation, but also on how complex terrain acts to modify the TC boundary layer. This is an important issue because the damaging winds associated with landfalling TCs are intrinsically related to characteristics of the underlying surface. The present study explores these complex interactions using high-resolution terrain data (∼900 m) and two fine nest domains (D3 and D4) with storm-scale grid spacing (3 and 1 km, respectively).
It was found that complex terrain has a significant impact on the wind and rainfall distribution of a landfalling TC. In the CTRL simulation, the orography is found to affect the TC track, intensity, and structure well prior to landfall by modifying the three-dimensional environmental wind upstream. Differences in the deep-layer vertical wind shear close to the TC core had a large impact on intensity, track and rainfall distribution prior to landfall. As the NOTOPOG TC approached landfall, a change in the near-core deep-layer shear from circularly asymmetric to circularly symmetric occurred, resulting in a transition of the spatial distribution of heaviest precipitation.
The direct influences of orography on the three-dimensional wind field of the CTRL TC are pronounced. Locally high surface wind speeds, ranging from 36 to 40 m s−1, occur over Mount Bellenden Ker and Mount Bartle Frere. The 50-m winds above the peaks were higher, reaching 68 m s−1 over Mount Bellenden Ker as the TC passed to its south. On the sheltered lee sides, wind speeds were much lower (∼8 m s−1). Analyses of the winds above the surface for CTRL during landfall revealed several wind speed maxima within the boundary layer, the strongest being 86 m s−1 at 631 m. These boundary layer jets produced strong low-level vertical wind shear (>30 m s−1); a favorable environment for tornadogenesis associated with landfalling TCs. However, the issue of how these boundary layer wind maxima are transported to the surface, resulting in wind gusts, remains an elusive area in TC research. The boundary layer parameterization schemes used in mesoscale models are too crude to accurately reproduce the small-scale exchanges of heat and momentum that result in wind gusts within the TC boundary layer, particularly over complex terrain.
Strong downslope winds were produced in the CTRL simulation when westerly flow from the outer TC circulation impinged on the north–south-oriented mountain range west of Port Douglas. The shape of this range is well-suited for generating severe downslope winds, with its steep leeside slope and gentle windward rise. During the most amplified phase of the mountain wave, the near-surface wind speed at the base of the range approached 24 m s−1. Even stronger surface winds of up to 36 m s−1 were observed in a separate simulation carried out in which westerly winds near the crest of the mountain increased (not shown). This numerical result confirms previous observational work on damaging westerly winds associated with landfalling TCs in the region (e.g., Callaghan 2003).
Rainfall amounts and patterns associated with TC Larry were reproduced well by the CTRL simulation, with 3-h totals in excess of 200 mm over the steep coastal orography. In contrast, the 3-h rainfall totals for the NOTOPOG TC were lower immediately following landfall, but increased relative to CTRL as the system moved farther inland.
Parameterized maximum surface wind gusts were spatially consistent with regions of locally high wind speeds in the TC boundary layer. The distribution and intensity of these gusts appear to be controlled primarily by vertical mixing associated with strong boundary layer vertical wind shear that is modified by the shape and height of the underlying terrain. Locally high values of TKE occurred over ridges where acceleration of the TC’s tangential winds increased vertical wind shear. However, given the simplistic assumptions of the WGE scheme the maximum surface gusts are probably overestimated.
Finally, an analysis of the postlandfall boundary layer wind field for the NOTOPOG simulation revealed finescale banding features that appear related to the increased surface roughness over land and the resultant strong low-level vertical wind shear. Similar boundary layer features have recently been documented in theoretical and observational studies (e.g., Foster 2005; Morrison et al. 2005; Romine and Wilhelmson 2006) and underline the need for further investigation regarding their possible connection with surface wind gusts.
Acknowledgments
This research was supported by funding from the Insurance Australia Group, Sydney, Australia. We thank Dr. David Stensrud (NOAA/National Severe Storms Laboratory) for valuable discussions as well as Dr. Bruce Buckley, Jeff Callaghan, and Peter Otto of the Australian Bureau of Meteorology for providing details of the observed characteristics of TC Larry. We also wish to thank the three anonymous reviewers for many beneficial suggestions, which helped to improve the manuscript. Finally, we acknowledge Greg Pearson and Norm Henry of the New Zealand Meteorological Service for providing the code to compute wind gusts in section 4f.
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.
Bender, M. A., R. E. Tuleya, and Y. Kurihara, 1985: A numerical study of the effect of a mountain-range on a landfalling tropical cyclone. Mon. Wea. Rev., 113 , 567–582.
Betts, A. K., 1986: A new convective adjustment scheme.1. Observational and theoretical basis. Quart. J. Roy. Meteor. Soc., 112 , 677–691.
Betts, A. K., and M. J. Miller, 1986: A new convective adjustment scheme 2. Single column tests using GATE wave, BOMEX, ATEX and arctic air-mass data sets. Quart. J. Roy. Meteor. Soc., 112 , 693–709.
Brand, S., and J. W. Blelloch, 1973: Changes in the characteristics of typhoons crossing the Philippines. J. Appl. Meteor., 12 , 104–109.
Brasseur, O., 2001: Development and application of a physical approach to estimating wind gusts. Mon. Wea. Rev., 129 , 5–25.
Braun, S. A., M. T. Montgomery, and Z. Pu, 2006: High-resolution simulation of Hurricane Bonnie (1998). Part I: The organization of eyewall vertical motion. J. Atmos. Sci., 63 , 19–42.
Callaghan, J., 2003: The tropical cyclone risk in Cairns. Nat. Hazards, 30 , 129–153.
Chang, C. P., T. C. Yeh, and J. M. Chen, 1993: Effects of terrain on the surface-structure of typhoons over Taiwan. Mon. Wea. Rev., 121 , 734–752.
Chang, S. W. J., 1982: The orographic effects induced by an island mountain-range on propagating tropical cyclones. Mon. Wea. Rev., 110 , 1255–1270.
Chen, S. S., J. A. Knaff, and F. D. Marks, 2006: Effects of vertical wind shear and storm motion on tropical cyclone rainfall asymmetries deduced from TRMM. Mon. Wea. Rev., 134 , 3190–3208.
Clark, T. L., and W. R. Peltier, 1984: Critical-level reflection and the resonant growth of nonlinear mountain waves. J. Atmos. Sci., 41 , 3122–3134.
Dare, R. A., and N. E. Davidson, 2004: Characteristics of tropical cyclones in the Australian region. Mon. Wea. Rev., 132 , 3049–3065.
DeMaria, M., 1996: The effect of vertical shear on tropical cyclone intensity change. J. Atmos. Sci., 53 , 2076–2088.
Farfan, L. M., and J. A. Zehnder, 2001: An analysis of the landfall of Hurricane Nora (1997). Mon. Wea. Rev., 129 , 2073–2088.
Foster, R. C., 2005: Why rolls are prevalent in the hurricane boundary layer. J. Atmos. Sci., 62 , 2647–2661.
Frank, W. M., and E. A. Ritchie, 2001: Effects of vertical wind shear on the intensity and structure of numerically simulated hurricanes. Mon. Wea. Rev., 129 , 2249–2269.
Geerts, B., G. M. Heymsfield, L. Tian, J. B. Halverson, A. Guillory, and M. I. Mejia, 2000: Hurricane Georges’s landfall in the Dominican Republic: Detailed airborne Doppler radar imagery. Bull. Amer. Meteor. Soc., 81 , 999–1018.
Grell, G. A., J. Dudhia, and D. R. Stauffer, 1994: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 122 pp.
Harper, B. A., 1999: Numerical modelling of extreme tropical cyclone winds. J. Wind Eng. Ind. Aerodyn., 83 , 35–47.
Henderson, D. J., J. D. Ginger, C. Leitch, G. N. Boughton, and D. Falck, 2006: Tropical Cyclone Larry: Damage to buildings in the Innisfail area. Rep. TR51, Cyclone Testing Station, James Cook University, Queensland, Australia, 96 pp.
Ho, F. P., J. C. Su, K. L. Hanevich, R. J. Smith, and F. P. Richards, 1987: Hurricane climatology for the Atlantic and Gulf coasts of the United States. NOAA Tech. Rep. NWS 38, NOAA/ NWS, Silver Spring, MD, 195 pp.
Hoinka, K. P., 1985: A comparison of numerical simulations of hydrostatic flow over mountains with observations. Mon. Wea. Rev., 113 , 719–735.
Janjic, Z. I., 1990: The step-mountain coordinate—Physical package. Mon. Wea. Rev., 118 , 1429–1443.
Janjic, Z. I., 1994: The step-mountain ETA coordinate model—Further developments of the convection, viscous sublayer, and turbulence closure schemes. Mon. Wea. Rev., 122 , 927–945.
Kaplan, J., and M. DeMaria, 1995: A simple empirical-model for predicting the decay of tropical cyclone winds after landfall. J. Appl. Meteor., 34 , 2499–2512.
Kepert, J., and Y. Wang, 2001: The dynamics of boundary layer jets within the tropical cyclone core. Part II: Nonlinear enhancement. J. Atmos. Sci., 58 , 2485–2501.
Kossin, J. P., and W. H. Schubert, 2001: Mesovortices, polygonal flow patterns, and rapid pressure falls in hurricane-like vortices. J. Atmos. Sci., 58 , 2196–2209.
Lilly, D. K., and J. B. Klemp, 1979: The effects of terrain shape on nonlinear hydrostatic mountain waves. J. Fluid Mech., 95 , 241–261.
Lin, Y. L., J. G. Han, D. W. Hamilton, and C. Y. Huang, 1999: Orographic influence on a drifting cyclone. J. Atmos. Sci., 56 , 534–562.
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.
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.
Lin, Y. L., N. C. Witcraft, and Y. H. Kuo, 2006: Dynamics of track deflection associated with the passage of tropical cyclones over a mesoscale mountain. Mon. Wea. Rev., 134 , 3509–3538.
Low-Nam, S., and C. Davis, 2001: Development of a tropical cyclone bogussing scheme for the MM5 system. Preprints. 11th PSU–NCAR Mesoscale Model Users’ Workshop, Boulder, CO, PSU–NCAR, 130–134.
McCaul, E. W., 1991: Buoyancy and shear characteristics of hurricane–tornado environments. Mon. Wea. Rev., 119 , 1954–1978.
Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough, 1997: Radiative transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave. J. Geophys. Res., 102 , 16663–16682.
Morrison, I., S. Businger, F. Marks, P. Dodge, and J. A. Businger, 2005: An observational case for the prevalence of roll vortices in the hurricane boundary layer. J. Atmos. Sci., 62 , 2662–2673.
Novlan, D. J., and W. M. Gray, 1974: Hurricane-spawned tornadoes. Mon. Wea. Rev., 102 , 476–488.
Powell, M. D., 1982: The transition of the hurricane Frederic boundary-layer wind-field from the open Gulf of Mexico to landfall. Mon. Wea. Rev., 110 , 1912–1932.
Powell, M. D., 1987: Changes in the low-level kinematic and thermodynamic structure of Hurricane Alicia (1983) at landfall. Mon. Wea. Rev., 115 , 75–99.
Reisner, J., R. M. Rasmussen, and R. T. Bruintjes, 1998: Explicit forecasting of supercooled liquid water in winter storms using the MM5 mesoscale model. Quart. J. Roy. Meteor. Soc., 124 , 1071–1107.
Romine, G. S., and R. B. Wilhelmson, 2006: Finescale spiral band features within a numerical simulation of Hurricane Opal (1995). Mon. Wea. Rev., 134 , 1121–1139.
Schubert, W. H., M. T. Montgomery, R. K. Taft, T. A. Guinn, S. R. Fulton, J. P. Kossin, and J. P. Edwards, 1999: Polygonal eyewalls, asymmetric eye contraction, and potential vorticity mixing in hurricanes. J. Atmos. Sci., 56 , 1197–1223.
Schwerdt, R. W., F. P. Ho, and R. R. Watkins, 1979: Meteorological criteria for standard project hurricane and probable maximum hurricane windfields in the Gulf and East Coasts of the United States. NOAA Tech. Rep. NWS 23, NOAA/NWS, Silver Spring, MD, 317 pp.
Smith, R. B., 1977: Steepening of hydrostatic mountain waves. J. Atmos. Sci., 34 , 1634–1654.
Smith, R. B., 1985: On severe downslope winds. J. Atmos. Sci., 42 , 2597–2603.
Wang, Y., 2002a: Vortex Rossby waves in a numerically simulated tropical cyclone. Part I: Overall structure, potential vorticity, and kinetic energy budgets. J. Atmos. Sci., 59 , 1213–1238.
Wang, Y., 2002b: Vortex Rossby waves in a numerically simulated tropical cyclone. Part II: The role in tropical cyclone structure and intensity changes. J. Atmos. Sci., 59 , 1239–1262.
Whittingham, H. E., 1964: Extreme wind gusts in Australia. Bureau of Meteorology Bulletin 46, 133 pp.
Wong, M. L. M., and J. C. L. Chan, 2007: Modeling the effects of land–sea roughness contrast on tropical cyclone winds. J. Atmos. Sci., 64 , 3249–3264.
Wu, C. C., 2001: Numerical simulation of Typhoon Gladys (1994) and its interaction with Taiwan terrain using the GFDL hurricane model. Mon. Wea. Rev., 129 , 1533–1549.
Wu, C. C., and Y. H. Kuo, 1999: Typhoons affecting Taiwan: Current understanding and future challenges. Bull. Amer. Meteor. Soc., 80 , 67–80.
Wu, C. C., T. H. Yen, Y. H. Kuo, and W. Wang, 2002: Rainfall simulation associated with Typhoon Herb (1996) near Taiwan. Part I: The topographic effect. Wea. Forecasting, 17 , 1001–1015.
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.
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.
Zehnder, J. A., 1993: The influence of large-scale topography on barotropic vortex motion. J. Atmos. Sci., 50 , 2519–2532.
Zehnder, J. A., and M. J. Reeder, 1997: A numerical study of barotropic vortex motion near a large-scale mountain range with application to the motion of tropical cyclones approaching the Sierra Madre. Meteor. Atmos. Phys., 64 , 1–19.
(a) Location map showing the far northeastern region of Australia, delineated by the black rectangle. (b) Map showing the region of complex terrain affected by TC Larry, including the town of Innisfail where the eye crossed during landfall. Terrain contours are every 100 m. The tallest peaks of Mount Bartle Frere and Bellenden Ker are marked with the initials BF and BK, respectively.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Location map showing the far northeastern region of Australia, delineated by the black rectangle. (b) Map showing the region of complex terrain affected by TC Larry, including the town of Innisfail where the eye crossed during landfall. Terrain contours are every 100 m. The tallest peaks of Mount Bartle Frere and Bellenden Ker are marked with the initials BF and BK, respectively.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Location map showing the far northeastern region of Australia, delineated by the black rectangle. (b) Map showing the region of complex terrain affected by TC Larry, including the town of Innisfail where the eye crossed during landfall. Terrain contours are every 100 m. The tallest peaks of Mount Bartle Frere and Bellenden Ker are marked with the initials BF and BK, respectively.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The quadruply nested grid configuration used for the simulations. The grid spacing for each domain is indicated next to the domain number.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The quadruply nested grid configuration used for the simulations. The grid spacing for each domain is indicated next to the domain number.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The quadruply nested grid configuration used for the simulations. The grid spacing for each domain is indicated next to the domain number.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Simulated storm tracks for the CTRL (white triangles) and NOTOPOG (white circles) experiments as well the best track of TC Larry (black squares). The white symbols indicate the position of central pressures every 3 h from 1712 UTC 17 Mar to 2012 UTC 20 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Simulated storm tracks for the CTRL (white triangles) and NOTOPOG (white circles) experiments as well the best track of TC Larry (black squares). The white symbols indicate the position of central pressures every 3 h from 1712 UTC 17 Mar to 2012 UTC 20 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Simulated storm tracks for the CTRL (white triangles) and NOTOPOG (white circles) experiments as well the best track of TC Larry (black squares). The white symbols indicate the position of central pressures every 3 h from 1712 UTC 17 Mar to 2012 UTC 20 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of best-track central pressure for TC Larry (black squares) and simulated central pressure for the CTRL (solid line with triangles) and NOTOPOG simulations (dashed line with circles) using data from D2. The vertical lines indicate the time of landfall for each simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of best-track central pressure for TC Larry (black squares) and simulated central pressure for the CTRL (solid line with triangles) and NOTOPOG simulations (dashed line with circles) using data from D2. The vertical lines indicate the time of landfall for each simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of best-track central pressure for TC Larry (black squares) and simulated central pressure for the CTRL (solid line with triangles) and NOTOPOG simulations (dashed line with circles) using data from D2. The vertical lines indicate the time of landfall for each simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface wind magnitude (gray shading with contour interval of 8 m s−1) and velocity vectors during landfall for the CTRL simulation. The terrain height is given by solid black contours with an interval of 200 m. (b) Corresponding east–west vertical cross section through the center of storm [marked by the black line in (a)] showing the υ component of the wind with contour interval of 5 m s−1. Solid (dashed) white contours indicate southerly (northerly) winds. (c) Simulated radar reflectivity at 1.8 km with contour interval of 5 dBZ. (d) East–west vertical cross section of the radar reflectivity corresponding to the position of the solid black line in (c). The freezing line is plotted in white.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface wind magnitude (gray shading with contour interval of 8 m s−1) and velocity vectors during landfall for the CTRL simulation. The terrain height is given by solid black contours with an interval of 200 m. (b) Corresponding east–west vertical cross section through the center of storm [marked by the black line in (a)] showing the υ component of the wind with contour interval of 5 m s−1. Solid (dashed) white contours indicate southerly (northerly) winds. (c) Simulated radar reflectivity at 1.8 km with contour interval of 5 dBZ. (d) East–west vertical cross section of the radar reflectivity corresponding to the position of the solid black line in (c). The freezing line is plotted in white.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface wind magnitude (gray shading with contour interval of 8 m s−1) and velocity vectors during landfall for the CTRL simulation. The terrain height is given by solid black contours with an interval of 200 m. (b) Corresponding east–west vertical cross section through the center of storm [marked by the black line in (a)] showing the υ component of the wind with contour interval of 5 m s−1. Solid (dashed) white contours indicate southerly (northerly) winds. (c) Simulated radar reflectivity at 1.8 km with contour interval of 5 dBZ. (d) East–west vertical cross section of the radar reflectivity corresponding to the position of the solid black line in (c). The freezing line is plotted in white.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Horizontal distribution of equivalent potential temperature (black contours with contour interval of 2 K) and cyclonic vertical vorticity starting from −1000 × 10−5 s−1 [gray shading with contour interval of (a) 100 and (b) 150] at z = 1 km for (a) CTRL simulation at landfall and (b) NOTOPOG simulation at landfall. Terrain elevation contours are designated by the black dashed lines.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Horizontal distribution of equivalent potential temperature (black contours with contour interval of 2 K) and cyclonic vertical vorticity starting from −1000 × 10−5 s−1 [gray shading with contour interval of (a) 100 and (b) 150] at z = 1 km for (a) CTRL simulation at landfall and (b) NOTOPOG simulation at landfall. Terrain elevation contours are designated by the black dashed lines.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Horizontal distribution of equivalent potential temperature (black contours with contour interval of 2 K) and cyclonic vertical vorticity starting from −1000 × 10−5 s−1 [gray shading with contour interval of (a) 100 and (b) 150] at z = 1 km for (a) CTRL simulation at landfall and (b) NOTOPOG simulation at landfall. Terrain elevation contours are designated by the black dashed lines.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Observed reflectivity of TC Larry at 1850 UTC 19 Mar 2006 showing a distinct pentagonal eyewall and (b) simulated reflectivity for the CTRL simulation (D4) at z = 2 km showing pentagonal eyewall structure at 2109 UTC 19 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Observed reflectivity of TC Larry at 1850 UTC 19 Mar 2006 showing a distinct pentagonal eyewall and (b) simulated reflectivity for the CTRL simulation (D4) at z = 2 km showing pentagonal eyewall structure at 2109 UTC 19 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Observed reflectivity of TC Larry at 1850 UTC 19 Mar 2006 showing a distinct pentagonal eyewall and (b) simulated reflectivity for the CTRL simulation (D4) at z = 2 km showing pentagonal eyewall structure at 2109 UTC 19 Mar 2006.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
As in Fig. 5, but for the NOTOPOG simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
As in Fig. 5, but for the NOTOPOG simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
As in Fig. 5, but for the NOTOPOG simulation.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Distribution of TKE (J kg−1) calculated at the lowest model sigma level (∼19 m) at the time of landfall for CTRL simulation. Terrain elevation contours are in black with interval of 200 m. (b) Distribution of TKE (gray shading) and the magnitude of the vertical shear in the lowest 100 m (black contours) at the time of landfall for the CTRL TC. (c) As in (b), but for the NOTOPOG TC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Distribution of TKE (J kg−1) calculated at the lowest model sigma level (∼19 m) at the time of landfall for CTRL simulation. Terrain elevation contours are in black with interval of 200 m. (b) Distribution of TKE (gray shading) and the magnitude of the vertical shear in the lowest 100 m (black contours) at the time of landfall for the CTRL TC. (c) As in (b), but for the NOTOPOG TC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Distribution of TKE (J kg−1) calculated at the lowest model sigma level (∼19 m) at the time of landfall for CTRL simulation. Terrain elevation contours are in black with interval of 200 m. (b) Distribution of TKE (gray shading) and the magnitude of the vertical shear in the lowest 100 m (black contours) at the time of landfall for the CTRL TC. (c) As in (b), but for the NOTOPOG TC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Surface streamlines and terrain height (shaded with a 250-m contour interval) for the CTRL simulation for times (a) 2100, (b) 2300, (c) 0100, and (d) 0300 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Surface streamlines and terrain height (shaded with a 250-m contour interval) for the CTRL simulation for times (a) 2100, (b) 2300, (c) 0100, and (d) 0300 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Surface streamlines and terrain height (shaded with a 250-m contour interval) for the CTRL simulation for times (a) 2100, (b) 2300, (c) 0100, and (d) 0300 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface streamlines and terrain height (gray shading with an interval of 200 m) at 2320 UTC 19 Mar 2006. (b) Vertical cross section of u–w wind component (gray shading with contour interval of 10 m s−1), circulation vectors (white vectors) and potential temperature (white contours every 3 K) at 2320 UTC 19 Mar 2006. The orientation of the cross section relative to the center of the TC is shown by the solid white line in (a).
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface streamlines and terrain height (gray shading with an interval of 200 m) at 2320 UTC 19 Mar 2006. (b) Vertical cross section of u–w wind component (gray shading with contour interval of 10 m s−1), circulation vectors (white vectors) and potential temperature (white contours every 3 K) at 2320 UTC 19 Mar 2006. The orientation of the cross section relative to the center of the TC is shown by the solid white line in (a).
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
(a) Surface streamlines and terrain height (gray shading with an interval of 200 m) at 2320 UTC 19 Mar 2006. (b) Vertical cross section of u–w wind component (gray shading with contour interval of 10 m s−1), circulation vectors (white vectors) and potential temperature (white contours every 3 K) at 2320 UTC 19 Mar 2006. The orientation of the cross section relative to the center of the TC is shown by the solid white line in (a).
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
East–west vertical cross section of u–w wind component (gray shading with contour interval of 4 m s−1), potential temperature (white contours every 3 K), and circulation vectors (white vectors) over the mountain range west of Port Douglas (see location map Fig. 1b) at time (a) 2300 and (b) 0030 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
East–west vertical cross section of u–w wind component (gray shading with contour interval of 4 m s−1), potential temperature (white contours every 3 K), and circulation vectors (white vectors) over the mountain range west of Port Douglas (see location map Fig. 1b) at time (a) 2300 and (b) 0030 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
East–west vertical cross section of u–w wind component (gray shading with contour interval of 4 m s−1), potential temperature (white contours every 3 K), and circulation vectors (white vectors) over the mountain range west of Port Douglas (see location map Fig. 1b) at time (a) 2300 and (b) 0030 UTC.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Parameterized maximum surface wind gusts (gray shading with contour interval of 10 m s−1) for the CTRL simulation at time (a) 2300 (landfall) and (b) 0100 UTC. The maximum surface wind gusts for the NOTOPOG simulation at time (c) 0009 (landfall) and (d) 0209 UTC. The letters A and B in (a) and (c) show the location of the wind profiles in Fig. 14.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Parameterized maximum surface wind gusts (gray shading with contour interval of 10 m s−1) for the CTRL simulation at time (a) 2300 (landfall) and (b) 0100 UTC. The maximum surface wind gusts for the NOTOPOG simulation at time (c) 0009 (landfall) and (d) 0209 UTC. The letters A and B in (a) and (c) show the location of the wind profiles in Fig. 14.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Parameterized maximum surface wind gusts (gray shading with contour interval of 10 m s−1) for the CTRL simulation at time (a) 2300 (landfall) and (b) 0100 UTC. The maximum surface wind gusts for the NOTOPOG simulation at time (c) 0009 (landfall) and (d) 0209 UTC. The letters A and B in (a) and (c) show the location of the wind profiles in Fig. 14.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Vertical profiles of wind speed in the lowest 2000 m corresponding to locations A and B as shown in Figs. 13a,c, for the CTRL (solid line with triangles) and NOTOPOG (dashed line with circles) simulations.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Vertical profiles of wind speed in the lowest 2000 m corresponding to locations A and B as shown in Figs. 13a,c, for the CTRL (solid line with triangles) and NOTOPOG (dashed line with circles) simulations.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Vertical profiles of wind speed in the lowest 2000 m corresponding to locations A and B as shown in Figs. 13a,c, for the CTRL (solid line with triangles) and NOTOPOG (dashed line with circles) simulations.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 12-h accumulated explicit precipitation in D3 prior to landfall (gray shading with a 50-mm contour interval) for (a) the CTRL simulation and (b) NOTOPOG simulation. The 100-mm contour is outlined in black and the 200- and 300-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 12-h accumulated explicit precipitation in D3 prior to landfall (gray shading with a 50-mm contour interval) for (a) the CTRL simulation and (b) NOTOPOG simulation. The 100-mm contour is outlined in black and the 200- and 300-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 12-h accumulated explicit precipitation in D3 prior to landfall (gray shading with a 50-mm contour interval) for (a) the CTRL simulation and (b) NOTOPOG simulation. The 100-mm contour is outlined in black and the 200- and 300-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 3-hourly accumulated precipitation in D4 (gray shading with a 25-mm contour interval) for the CTRL TC at (a) 2300, (b) 0200, and (c) 0500 UTC, and for the NOTOPOG TC at (d) 0009, (e) 0309, and (f) 0609 UTC. The 50- and 100-mm contours are outlined in black. The 150- and 200-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 3-hourly accumulated precipitation in D4 (gray shading with a 25-mm contour interval) for the CTRL TC at (a) 2300, (b) 0200, and (c) 0500 UTC, and for the NOTOPOG TC at (d) 0009, (e) 0309, and (f) 0609 UTC. The 50- and 100-mm contours are outlined in black. The 150- and 200-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
The 3-hourly accumulated precipitation in D4 (gray shading with a 25-mm contour interval) for the CTRL TC at (a) 2300, (b) 0200, and (c) 0500 UTC, and for the NOTOPOG TC at (d) 0009, (e) 0309, and (f) 0609 UTC. The 50- and 100-mm contours are outlined in black. The 150- and 200-mm contours are outlined in white. Terrain elevation is shown by dark gray contours with a contour interval of 100 m.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of maximum surface wind speed for the CTRL simulation (solid line with black triangles) and NOTOPOG simulation (dashed line with white circles) plotted every hour from the time landfall (0 h) to 12 h after landfall.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of maximum surface wind speed for the CTRL simulation (solid line with black triangles) and NOTOPOG simulation (dashed line with white circles) plotted every hour from the time landfall (0 h) to 12 h after landfall.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1
Time series of maximum surface wind speed for the CTRL simulation (solid line with black triangles) and NOTOPOG simulation (dashed line with white circles) plotted every hour from the time landfall (0 h) to 12 h after landfall.
Citation: Monthly Weather Review 136, 11; 10.1175/2008MWR2429.1