Abstract

Remarkable progress has been achieved in understanding the vorticity source responsible for tornadogenesis. Nevertheless, the answer to this question remains elusive, particularly after introducing surface friction in realistic tornado simulations. In this study, a simulation using the Weather Research and Forecasting (WRF) Model is conducted based on the F3 supercell tornado that hit Tsukuba City, Japan, on 6 May 2012. The simulation uses triply nested domains, and the tornado is successfully reproduced in the innermost domain with 50-m horizontal grid spacing. The circulation analyses reveal that the frictional term is the dominant vorticity source responsible for the vortices at both the pretornadic and tornadogenesis times. The detailed vorticity source analyses of the air parcels show that the vorticity of the tornado at the genesis time mainly originates from the frictionally generated crosswise vorticity near the ground. The crosswise vorticity is directly tilted (or first exchanged into streamwise vorticity and then tilted) into vertical vorticity when the air parcels enter the tornado. A rear-flank downdraft (RFD) surge from the south and west sides of a primary low-level mesocyclone (LMC) may trigger tornadogenesis by increasing the convergence near the ground. The RFD surge is not necessarily associated with the baroclinically generated vorticity. In this study, the baroclinity is weak across the hook echo, which may cause a lack of baroclinically generated vorticity in the RFD surge.

1. Introduction

To understand the mechanism of supercell tornadogenesis, identifying the responsible vorticity sources is a vitally important issue. Despite several decades of study on this issue, a complete understanding remains ambiguous. Previous studies included field observations and idealized numerical simulations. It was suggested that the baroclinic effects are prominent in generating the vorticity of tornadoes (e.g., Davies-Jones and Brooks 1993; Adlerman et al. 1999; Straka et al. 2007). The baroclinic effects are due to the horizontal buoyancy gradient associated with the downdrafts in a storm. Horizontal vorticity is generated by baroclinity when airflow approaches a tornado along the rear-flank downdraft (RFD; Straka et al. 2007; Markowski et al. 2008), along the forward-flank gust front (FFGF; Wicker and Wilhelmson 1995; Markowski et al. 2012; Markowski and Richardson 2014), or at the periphery of downdrafts aloft (e.g., Dahl et al. 2014).

Although idealized simulations have traditionally neglected surface drag, several new studies in the past several years have included parameterized drag. Roberts et al. (2016) suggested that surface drag generates crosswise vorticity, and the crosswise vorticity is consequently converted into vertical vorticity in a tornado. Markowski (2016) showed that the contributions of environmental and storm-generated (baroclinic or frictional) vorticity to the circulation of a near-surface mesocyclone depend on different environmental wind profiles and supercell stages.

Realistic tornado simulations are becoming more frequent in recent years. Mashiko et al. (2009) reproduced a minisupercell tornado in a rainband of Typhoon Shanshan using the Japan Meteorological Agency Nonhydrostatic Model (JMANHM). Environmental streamwise vorticity was suggested to be the main component of the vorticity sources responsible for tornadogenesis. The environmental vorticity is associated with the strong low-level veering shear in the typhoon. Schenkman et al. (2014) utilized the Advanced Regional Prediction System (ARPS) to examine a supercell tornado that occurred in the 2003 Oklahoma City storm. The vorticity sources of four pretornadic vortices that led to two tornadoes were investigated. The crosswise vorticity generated by surface drag was found to be responsible for all the vortices. The crosswise vorticity was exchanged into streamwise vorticity and consequently tilted into vertical vorticity. Mashiko (2016b, hereafter M16) simulated the 2012 Tsukuba tornado and proposed that frictionally generated circulation is the dominant source of the pretornadic vortex, while baroclinically generated circulation is prominent for the tornado at genesis time.

In addition to the controversy regarding the vorticity sources of tornadic vortices in different studies, recent studies indicate that even for a given tornado, the vorticity sources can be different with different model settings. Yokota et al. (2018, hereafter Y18) simulated the same 2012 Tsukuba tornado by ensemble forecasts and analyzed the circulation sources of the vortices at the times of tornadogenesis and maximum intensity. The authors revealed that circulation generation is mainly (not always) due to the frictional terms. The baroclinic terms can also be dominant at times.

The RFD surge (also named the internal outflow surge or secondary RFD) behind the primary rear-flank gust front (RFGF) was found to be related to tornadogenesis in some recent studies. These studies suggested that the RFD surge enhances convergence, which leads to tornadogenesis (e.g., Kosiba et al. 2013; Mashiko et al. 2009). However, the generation of vorticity associated with the RFD surge is still under investigation. Marquis et al. (2012) suggested that the secondary RFD may baroclinically generate vorticity and contribute to the tornado at the maintenance stage by vortex tilting. M16 suggested that the baroclinically generated vorticity on the tip of the hook echo is transported into the tornado via the RFD surge. Schenkman et al. (2014) suggested that the frictionally generated vorticity is prominent in the RFD surge.

Previous studies have also found that the RFD surge can induce cyclic mesocyclogenesis (Adlerman et al. 1999; Adlerman 2003; Beck et al. 2006). The developing RFD wraps around the primary mesocyclone and causes the gust front to surge outward. This surging gust front initiates a new updraft on the downshear flank of the primary mesocyclone. Then, a new mesocyclone is formed in the new updraft.

The Weather Research and Forecasting (WRF) Model (Skamarock et al. 2008) is a widely used meteorological model. However, the application of this model in resolving tornadoes has been rarely demonstrated. Most studies have been limited by coarse grid resolutions. Xu et al. (2015) simulated a bow echo system with 0.8-km horizontal grid spacing. The significant contribution of surface friction to the generation of low-level mesovortices was emphasized. Marquis et al. (2014, 2016) investigated the 5 June 2009 Goshen County tornadic supercell. Vorticity budget analyses were conducted on several sets of air parcels from a near-surface mesocyclone.1 The authors suggested that the baroclinity within the forward-flank region is the main source of vorticity of the near-surface mesocyclone throughout its lifetime. A more recent study (Flournoy and Coniglio 2019) simulated a tornadic mesovortex in a quasi-linear convective system with 333-m horizontal grid spacing. Three separate airstreams entered the mesovortex and contributed to the vertical vorticity in different ways. In the southerly inflow, the tilting of crosswise vorticity is predominant, while in the northerly inflow and a rear inflow jet, the tilting of streamwise vorticity is predominant. Sun et al. (2019) simulated an EF4 supercell tornado in Funing, China, and discussed the sensitivity of flow features in the tornado to grid resolutions. The authors recommended that horizontal grid spacing of 50 m or less is needed for studies of tornado dynamics.

As mentioned above, the discussion of the vorticity sources of a tornado remains controversial particularly after introducing frictional effects. The main purpose of this study is to understand the mechanism of tornadogenesis by considering the sources responsible for the vorticity. To achieve a detailed understanding of the tornadogenesis process, a realistic simulation by WRF with large-eddy simulation implementation (WRFLES; Moeng et al. 2007) is conducted on the 2012 Tsukuba tornado. The finest horizontal grid spacing is 50 m, and the tornado is explicitly resolved. The vorticity generation mechanism of the flow entering the tornado is explored by backward trajectory with a particular focus on an RFD surge. Y18 indicated the uncertainty in the tornadogenesis mechanism for a given tornado. Thus, the generated tornadoes in M16 and Y18 and the present study can all be regarded as solutions of the Tsukuba tornado. The atmospheric condition resolved in the present study is physically possible and favorable for tornadogenesis. In this context, the analysis of tornadogenesis mechanism in the present study is expected to achieve meaningful results.

Section 2 introduces the experimental design. Section 3 describes the general characteristics of the mesocyclone and the explicitly resolved tornado. Section 4 introduces the methods of circulation and vorticity analyses. Section 5 discusses the circulation sources of near-ground vortices at the pretornadic and tornadogenesis stages. Vorticity budgets along the trajectories of air parcels are investigated in section 6. A discussion and conclusions are presented in section 7.

2. Experimental design

The present study focuses on the Tsukuba tornado, which is an F3 tornado that occurred on the afternoon of 6 May 2012 (Japan Meteorological Agency 2012). The WRF Model (version 3.7.1) simulation has triply one-way nested domains (Fig. 1a). The outermost domain (WRFLES1km) covers most of the area of eastern Japan using 300 × 300 horizontal grid points with 1-km spacing. This domain aims to resolve the environment of the storm and runs from 0000 to 0600 UTC. Nested within WRFLES1km, the second domain (WRFLES250m) starts from 0200 UTC with 250-m horizontal grid spacing. The number of horizontal grid points of WRFLES250m is 801 × 701, and the domain covers most of the Kanto Plain, where the convective storm developed. Starting from 0230 UTC, the innermost domain (WRFLES50m) with 50-m horizontal grid spacing is nested within WRFLES250m. The number of horizontal grid points of WRFLES50m is 1501 × 1501, and the tornado is explicitly resolved (Fig. 1b). WRFLES1km and WRFLES250m have 62 vertical levels with spacings that stretch from approximately 40 m near the ground to approximately 640 m at the model top with a constant pressure of 60 hPa. WRFLES50m has 81 vertical levels in which the spacing near the ground is 10 m, and there are 24 levels below the 1-km height. A fifth-order horizontal advection scheme and a third-order vertical advection scheme are used. A third-order time-split Runge–Kutta scheme is used to advance the computation in time.

Fig. 1.

The topography map of the triply nested simulation domains. (a) The topography map is plotted on the entire WRFLES1km domain. Locations of WRFLES250m and WRFLES50m are shown with black and green rectangles, respectively. The locations of Tokyo City, Tsukuba City, and Kanto Plain are also shown. (b) The topography of WRFLES50m with observed and simulated (WRFLES50m) Tsukuba tornado paths.

Fig. 1.

The topography map of the triply nested simulation domains. (a) The topography map is plotted on the entire WRFLES1km domain. Locations of WRFLES250m and WRFLES50m are shown with black and green rectangles, respectively. The locations of Tokyo City, Tsukuba City, and Kanto Plain are also shown. (b) The topography of WRFLES50m with observed and simulated (WRFLES50m) Tsukuba tornado paths.

The model uses the Morrison et al. two-moment bulk microphysics scheme (Morrison et al. 2009). Cumulus parameterization is turned off. Radiation is handled through the RRTMG scheme (Iacono et al. 2008). The five-layer thermal diffusion land surface scheme and the revised MM5 Monin–Obukhov surface layer scheme (Jiménez et al. 2012) are also included. The Monin–Obukhov similarity theory (MOST) is based on steady and well-mixed flows. However, the outflow of a thunderstorm is normally unsteady, which would hinder the reliability of the surface layer scheme. Markowski et al. (2019) suggested that the MOST tends to underpredict the surface layer wind shear in the storm environment, especially in the outflow of the storm.

In recent years, the “gray zone” problem (Wyngaard 2004) has been encountered in downscaling simulations using mesoscale models. When the grid size is at an intermediate scale between the LES upper limit (slightly less than the integral turbulence scale) and the mesoscale lower limit (no resolved turbulence), neither the planetary boundary layer (PBL) schemes nor the LES models are appropriate. For the simulation of a supercell storm, Fiori et al. (2010) suggested that even for 1-km horizontal grid spacing, the LES closure produces better results than the PBL schemes and “gives acceptable representation of storm-scale supercell structure and evolution”. We also reach similar conclusions in sensitivity tests (not shown). Moreover, considering that the simulation period is daytime, and the atmospheric condition is unstable in the storm environment, it is reasonable to expect deep mixing. Thus, the present study uses an LES model. The LES model is based on the 1.5-order turbulent kinetic energy (TKE) closure scheme (Deardorff 1972). A drawback of the LES model should be noted here. Markowski and Bryan (2016) showed that when simulating a homogeneous, quasi-steady boundary layer using LES, unrealistically large wind shear would develop if the turbulence is not sufficiently resolved (especially in a laminar flow). The environmental inflow of a storm is likely error prone due to this phenomenon. In the present simulation, turbulence has partly developed in the environmental inflow, which may alleviate this problem. Another problem with the LES model is that the 1.5-order TKE model may overestimate the wind shear on the lowest model levels because the resolution of these grids is inadequate to resolve the turbulent eddies near the ground (Mason and Thomson 1992). Note that the overestimation of wind shear caused by the two problems may be partly offset by the underestimation caused by the MOST scheme.

The topography and land cover of WRFLES1km are interpolated from the U.S. Geological Survey (USGS) dataset, the resolution of which is 30 arc-s. For WRFLES250m and WRFLES50m, a 100-m meshed dataset provided by the Geospatial Information Authority of Japan is used. The JMA mesoscale analysis data are utilized for the initial and boundary conditions of WRFLES1km. These data are updated every 3 h and have a 5-km horizontal resolution. Data assimilation is not used in the present study.

3. Overview of the mesocyclone and tornado in the WRFLES simulation

Figure 2 shows the supercell generated in WRFLES250m. A midlevel mesocyclone (MMC; defined at the 3-km height) is associated with the hook echo of the supercell. The MMC has a diameter of approximately 3 km and a central pressure deficit of approximately 8 hPa. The storm translational velocity is (17, 6) m s−1 (zonal and meridional components), which is defined by the average moving velocity of the MMC center from 0258 to 0308 UTC. The MMC center is defined at the location of maximum vertical vorticity. Although the 250-m horizontal grid spacing simulations can hardly explicitly resolve small structures, such as tornadoes, near-ground rotation is still generated at 0259 UTC. At 200 m AGL, this rotation has a vertical vorticity greater than 0.1 s−1, reaching the mature stage from 0302 to 0308 UTC with a vertical vorticity greater than 0.2 s−1. The rotation dissipates at 0315 UTC when it approaches the mountainous area.

Fig. 2.

(a) Distribution of the mixing ratio of hydrometeors at the 3-km height at 0307 UTC in WRFLES250m. The purple line shows the path of the observed tornado. The area of (b) is enclosed by the red rectangle. (b) Distribution of the vertical vorticity at the 3-km height. Arrows represent the storm-relative horizontal wind vectors. Black lines in two panels denote the isobars at 3-hPa intervals.

Fig. 2.

(a) Distribution of the mixing ratio of hydrometeors at the 3-km height at 0307 UTC in WRFLES250m. The purple line shows the path of the observed tornado. The area of (b) is enclosed by the red rectangle. (b) Distribution of the vertical vorticity at the 3-km height. Arrows represent the storm-relative horizontal wind vectors. Black lines in two panels denote the isobars at 3-hPa intervals.

In WRFLES50m, a strong rotation (tornado) is successfully generated at 0253 UTC, which is 6 min earlier than the generation of the near-ground rotation in WRFLES250m. However, the two rotations are still regarded as physically consistent because the near-ground vorticity development processes are morphologically similar. Compared with the actual tornado, the simulated tornadogenesis occurs about half an hour earlier and the simulated path deviates by approximately 12 km north (Fig. 1b). Other studies of this tornado also show earlier tornadogenesis and north deviation of the path (Mashiko 2016a; Seko et al. 2015). The hydrometeor distribution of the supercell (Fig. 3a) shows a similar pattern to the observations (Yamauchi et al. 2013) and M16 (see his Fig. 2). A low pressure region is located within the hook echo of the supercell and extends to the tornado center. The central pressure deficit of the tornado near the ground is approximately 25 hPa (Fig. 3b).

Fig. 3.

(a) Mixing ratio of hydrometeors at the 1-km height at 0254:30 UTC in WRFLES50m. Arrows denote storm-relative horizontal wind vectors. The zoomed-in view of the area enclosed by the purple rectangle is superimposed on (a) adding isobars. The area of (b) is also enclosed by the purple rectangle. (b) Distribution of the vertical vorticity at 5 m AGL. Arrows represent ground-relative horizontal wind vectors. Black contours denote isobars at 4-hPa intervals.

Fig. 3.

(a) Mixing ratio of hydrometeors at the 1-km height at 0254:30 UTC in WRFLES50m. Arrows denote storm-relative horizontal wind vectors. The zoomed-in view of the area enclosed by the purple rectangle is superimposed on (a) adding isobars. The area of (b) is also enclosed by the purple rectangle. (b) Distribution of the vertical vorticity at 5 m AGL. Arrows represent ground-relative horizontal wind vectors. Black contours denote isobars at 4-hPa intervals.

The time series of the maximum vertical vorticity and maximum horizontal wind speed at the lowest scalar model level (approximately 5 m AGL; ζz=5m, VHz=5m, respectively) are plotted in Fig. 4. ζz=5m and VHz=5m are relatively small during the first 20 min of WRFLES50m. ζz=5m increases quickly from 0252 UTC and exceeds 1.2 s−1 at 0254:30 UTC, while VHz=5m increases until 0255 UTC and exceeds 60 m s−1. The period between 0253 and 0255 UTC is indicated as the tornadogenesis and intensification stages. The minimum sea level pressure drops by 25 hPa over this period. From 0255 to 0310 UTC, ζz=5m and VHz=5m fluctuates between 0.45–1.1 s−1 and 43–63 m s−1, respectively. Starting at approximately 0310 UTC, the decay stage can be indicated by the decreasing tendencies of ζz=5m and VHz=5m.

Fig. 4.

Time series of the maximum horizontal wind velocity (black line) and maximum vertical vorticity (blue line) at the lowest scalar points (approximately 5 m AGL) of the entire WRFLES50m domain. Red line denotes the minimum sea level pressure. The time series are plotted at 30-s intervals.

Fig. 4.

Time series of the maximum horizontal wind velocity (black line) and maximum vertical vorticity (blue line) at the lowest scalar points (approximately 5 m AGL) of the entire WRFLES50m domain. Red line denotes the minimum sea level pressure. The time series are plotted at 30-s intervals.

4. Methodologies: Circulation and vorticity analyses

To quantitatively investigate the vorticity sources, circulation and vorticity analyses are applied on material circuits (attached to air parcels) and air parcels from the near-ground vortices, respectively. The locations of the air parcels are calculated by backward trajectories.

The backward trajectories are calculated using velocity data of WRFLES50m saved every 0.5 s (two model time steps). A fourth-order Runge–Kutta (RK4) method is used with a time step of 0.5 s. The velocity at the air parcel location is obtained by tricubic and linear interpolation in space and time, respectively. Dahl et al. (2012) pointed out that even when the model output saved every model time step are used, trajectories that are traced backward from tornadoes could still be spurious. In the present study, both the model and backward trajectory utilize smaller time steps than most recent studies of the same model grid size (Schenkman et al. 2014; M16; Y18). To check the sensitivity to the saving frequency of model output, the circulation analysis of a tornadic vortex is also tested based on model output saved every four model time steps (not shown). In most periods of the analysis, the difference in the circulation budget is slight between the case of every four model time steps and the default (every two model time steps) case. Nonetheless, the authors still note the caveat that the backward trajectory calculations may result in spurious trajectories.

Brioude et al. (2012) mentioned the uncertainties in interpolating the wind velocity field from the Eulerian model to the Lagrangian model. The authors suggested avoiding using the Cartesian vertical velocity when the terrain is complex. In this study, the vertical velocity is represented by the instantaneous mass-weighted sigma dot vertical velocity (σ˙). The σ˙ is defined on the vertical coordinate of the WRF Model and is directly used in the advection terms of the governing equations (Skamarock et al. 2008). Thus, using σ˙ is less error prone than using the Cartesian vertical velocity. Meanwhile, the terrain of the regions related to backward trajectory analyses is quite flat (see the Kanto Plain in Fig. 1). Thus, the analyses in this study should not be subject to large error induced by complex terrain.

When the air parcel is below the lowest scalar model level, there is uncertainty in specifying the location, velocity and scalar values of the parcel. Guchte and Dahl (2018) explained the unphysical tendency of the scalar variables on the trajectory below the lowest scalar model level in a simulation using free-slip bottom conditions. For the semislip conditions, as in the present study, the flow is still poorly defined below the lowest scalar model level, and it may lead to errors in the trajectory analyses. Previous semislip studies assumed a logarithmic wind profile (Mashiko et al. 2009; M16), while in the present study, the horizontal velocity is determined by the second-order polynomial extrapolation from the upper levels. Using the extrapolation, the interpolated vorticity and the integrated vorticity (see section 6) show better consistency than those calculated with the logarithmic profile. Y18 also mentioned that using the linear extrapolation results in more consistent circulations than using the logarithmic profile. The authors suggested that the stratification in the tornado was not neutral, and the logarithmic profile might not be suitable for tornadic conditions.

Here ωs, ωc, ζ represent the three components of vorticity under the seminatural coordinates (s, c, z),where s, c, z are horizontal streamwise, horizontal crosswise, and vertical coordinates, respectively. In the present study, the streamwise (and crosswise) vorticity is defined with respect to the ground-relative wind unless otherwise specified. The vorticity equations of an air parcel are given as (Mashiko et al. 2009; Schenkman et al. 2014; Roberts et al. 2016; Y18)

 
DωsDt=ωs(VHψc+wz)+ωcVHc+ζVHz+1ρ2(ρcpzρzpc)+1ρ(FzcFczFsψz)+ωcDψDt,
(1)
 
DωcDt=ωc(wz+VHs)+ζVHψz+ωsVHψs+1ρ2(ρzpsρspz)+1ρ(FszFzsFcψz)ωsDψDt,
(2)
 
DζDt=ζ(VHs+VHψc)+ωsws+ωcwc+1ρ2(ρspcρcps)+1ρ(FcsFsc+Fsψs+Fcψc),
(3)

where ψ = tan−1(υ/u) represents the direction of the horizontal ground-relative wind VH = (u, υ); w, p, ρ represent vertical velocity, pressure, and density, respectively; and Fs, Fc, Fz represent three components of the diffusion terms, which consist of subgrid-scale (SGS) mixing and implicit terms. The SGS mixing terms are the spatial derivative of momentum stress. The momentum stress consists of the surface shear stress (parameterized in the surface layer scheme) and the internal mixing stress (the momentum stress above the lowest level, calculated by the LES model). The implicit terms are from the upwind advection schemes. At the lowest two levels, the vertical advection uses central difference; thus, the implicit terms are zero. For the air parcels analyzed in the present study, the implicit term is much smaller than the SGS mixing term. This result is likely because most air parcels are located at low levels. The first three terms on the right-hand side (rhs) of the equations are the stretching and tilting terms, which embody the effects of vortex deformation with fluid. The fourth and fifth terms on the rhs are the baroclinic (solenoidal) and frictional terms. The last terms on the rhs of Eqs. (1) and (2) represent the conversion between crosswise vorticity and streamwise vorticity as the parcel changes moving direction.

In calculating the terms of the vorticity equations, errors may be induced by the large velocity gradient in a highly converged flow (Dahl et al. 2012). Circulation analysis can be used to alleviate this problem by distributing the material circuit surrounding the vortex. The circulation C is defined as

 
C(t)=υdl,
(4)

in which υ = (u, υ, w), and l is the line vector along the circuit.

The time change of the circulation can be derived as

 
DDtC(t)=dpρ+Fdl,
(5)

in which F is the frictional force that is physically the same as (Fs, Fc, Fz) in Eqs. (1)(3). The first and second terms on the rhs are the baroclinic and frictional terms.

The circuit is traced back by the backward trajectories of many air parcels along the circuit. To exclude ambiguity, the initial times of the circuits mentioned in this study refer to the times when the backward trajectory calculations start. Initially, air parcels are distributed along the circuit at 1-m intervals. During the backward integration, if the distance between two adjacent air parcels exceeds 50 m, an additional parcel is added between the two.

The source terms in the circulation and vorticity equations [rhs of Eqs. (1)(3) and (5)] are interpolated from the nearby model grid points to air parcel locations. If an air parcel falls below the lowest scalar level, it is uncertain to value the source terms, which may lead to errors in the analysis. In the present study, the source terms at the lowest scalar level are used (i.e., constant value). Calculation of the source terms using second-order extrapolation from the upper levels has also been tested. The sum of the time-integrated source terms on the rhs of Eq. (5) is more consistent with the C(t) in Eq. (4) if using the constant value (not shown).

5. Circulation analysis

a. Tornadogenesis process

Figure 5 shows the evolution of the vertical vorticity, horizontal velocity, and vertical velocity at 200 m AGL from 0250 UTC to tornadogenesis (0253 UTC). The FFGF and RFGF are indicated by the regions of horizontal velocity convergence. The FFGF is regarded as the left-flank convergence boundary (LFCB) described by Beck and Weiss (2013) because its orientation and location are different from those of a traditional FFGF (Lemon and Doswell 1979). The updraft region is aligned with the FFGF and RFGF, which is especially strong near the region where the FFGF intersects with the RFGF (hereafter, intersection zone). At 0250 UTC, two vortices exist (Fig. 5a-1). One vortex is located near the intersection zone. This vortex is regarded as the pretornadic vortex owing to the consistency with the tornado (Figs. 5a-2–5a-4). Another vortex is approximately 2 km to the southwest of the pretornadic vortex. This vortex is connected with a primary low-level mesocyclone (LMC, defined below the 1-km height);2 thus, it is named the primary near-ground mesocyclone (NMC). Several local vertical vorticity extrema exist in the primary NMC. The vorticity extrema also evolve, rotating around each other (see online supplement). The RFD region is located behind (on the south and west sides of) the RFGF. In the RFD region, flow of locally intensified horizontal momentum is evident (Fig. 5b-1). This flow is from a locally intensified downdraft, which corresponds to an RFD surge. [The RFD surge reaches the ground around 0250 UTC; thus, the downdraft is less evident in the panels of vertical velocity (Fig. 5c-1)]. Horizontal convergence is enhanced at the front edge of the RFD surge. At 0250 UTC, the RFD surge is located on the south side of the primary LMC and reaches the ground to the south of the intersection zone. Then, the RFD surge heads quickly northeastward, approaching the intersection zone (0251 UTC; Figs. 5b-2,c-2). When the RFD surge reaches the RFGF (0252 UTC; Figs. 5b-3,c-3), the updraft on the left front edge of the RFD surge is merged with the updraft in the pretornadic vortex, leading to the intensification of vertical vorticity (tornadogenesis at 0253 UTC; Figs. 5a-4,b-4,c-4). This indicates that the RFD surge may play a role in triggering tornadogenesis by increasing the convergence near the ground. The phenomenon that an RFD surge causes a tornado to the east of a primary mesocyclone is similar to cyclic mesocyclogenesis (Adlerman et al. 1999; Adlerman 2003; Beck et al. 2006).

Fig. 5.

The (a1)–(a4) vertical vorticity, (b1)–(b4) horizontal velocity, (c1)–(c4) and vertical velocity distributions on the horizontal plane near the ground are shown at 1-min intervals from 0250 to 0253 UTC. The horizontal planes in (a) are at 200 m AGL. The horizontal planes in (b) and (c) are at 50 m AGL because of the relatively low location of the RFD surge. Thin dashed contour lines represent isobars at 1-hPa intervals. Arrows denote the horizontal ground-relative wind vectors. In (b), horizontal convergence larger than 0.04 s−1 is shaded by cyan color except that the convergence at the front edge of the RFD surge is shaded by blue color. The front edge is estimated from the blue-color convergence area and indicated by thick dashed lines in (c). The FFGF and RFGF are indicated by thick solid lines. In (a), the pretornadic vortex, primary near-ground mesocyclone (NMC), and tornado are pointed out by thick arrows.

Fig. 5.

The (a1)–(a4) vertical vorticity, (b1)–(b4) horizontal velocity, (c1)–(c4) and vertical velocity distributions on the horizontal plane near the ground are shown at 1-min intervals from 0250 to 0253 UTC. The horizontal planes in (a) are at 200 m AGL. The horizontal planes in (b) and (c) are at 50 m AGL because of the relatively low location of the RFD surge. Thin dashed contour lines represent isobars at 1-hPa intervals. Arrows denote the horizontal ground-relative wind vectors. In (b), horizontal convergence larger than 0.04 s−1 is shaded by cyan color except that the convergence at the front edge of the RFD surge is shaded by blue color. The front edge is estimated from the blue-color convergence area and indicated by thick dashed lines in (c). The FFGF and RFGF are indicated by thick solid lines. In (a), the pretornadic vortex, primary near-ground mesocyclone (NMC), and tornado are pointed out by thick arrows.

To be more intuitive, the evolution of the primary mesocyclone and tornadogenesis is shown by 3D rendering in Fig. 6. Before tornadogenesis (Fig. 6a) the primary LMC and NMC are connected to an MMC (named the primary MMC). A new LMC is formed at approximately 0251 UTC (Fig. 6b) above the pretornadic vortex. During the intensification of the new LMC, the pretornadic vortex is rapidly intensified and connected to the new LMC (Fig. 6c). Eventually, the pretornadic vortex becomes the tornado after 0252 UTC (Fig. 6d). The new LMC is named the tornadic LMC, hereafter.

Fig. 6.

The 3D development of vertical vorticity (the blue, light red, and dark red isosurfaces, respectively, represent the 0.1, 0.2, and 0.3 s−1 vertical vorticity isosurfaces) from 0249:42 to 0252:12 UTC. The horizontal plane shows the horizontal velocity convergence (negative divergence) at 5 m AGL. The midlevel mesocyclones (MMC), low-level mesocyclones (LMCs), near-ground mesocyclone (NMC), pretornadic vortex, and tornado are pointed out by dashed circles. Vortex lines that pass through the local vorticity extrema in the primary NMC and tornado at 5 m AGL are plotted using black and red lines, respectively. The coordinate on the horizontal axis (“east”) refers to the distance to the southern boundary of WRFLES50m.

Fig. 6.

The 3D development of vertical vorticity (the blue, light red, and dark red isosurfaces, respectively, represent the 0.1, 0.2, and 0.3 s−1 vertical vorticity isosurfaces) from 0249:42 to 0252:12 UTC. The horizontal plane shows the horizontal velocity convergence (negative divergence) at 5 m AGL. The midlevel mesocyclones (MMC), low-level mesocyclones (LMCs), near-ground mesocyclone (NMC), pretornadic vortex, and tornado are pointed out by dashed circles. Vortex lines that pass through the local vorticity extrema in the primary NMC and tornado at 5 m AGL are plotted using black and red lines, respectively. The coordinate on the horizontal axis (“east”) refers to the distance to the southern boundary of WRFLES50m.

To illustrate the changes in the flow entering the near-ground vortex at approximately the onset of tornadogenesis, circuits initialized from the pretornadic and tornadic vortices are traced backward for 10 min (Fig. 7). The spatial coverages of the circuits after backward integration could qualitatively represent the distributions of air parcels that were initially enclosed by the circuits. The circuits are initiated as rings3 surrounding the cores of the vortices at 200 m AGL (e.g., Fig. 8a). Four initial times are selected from 0250 to 0253 UTC. For the circuit initiated at 0250 UTC, most part of the circuit is traced back near the surface in the eastern inflow.4 For the circuits initiated after 0251 UTC, segment of the circuits are traced back to the FFGF region. The circuit segment in the FFGF region is gradually traced back to approximately the 300-m height, which means the segment has experienced a downdraft. Considering that the FFGF lies between the cold pool and the warm updraft, the horizontal vorticity could be generated by baroclinity. When an air parcel traveling along the FFGF experiences a downdraft, it leads to a “slippage” between the velocity and vorticity vectors. Thus, the horizontal vorticity is tilted into vertical vorticity during the descent, and then, the vertical vorticity can be intensified by stretching (Davies-Jones and Brooks 1993; Markowski and Richardson 2014; Rotunno et al. 2017). For the circuits initialized after 0252 UTC, large segments of the circuits are traced back to the RFD region on the south and west sides of the primary LMC/NMC. It is reiterated that the RFD surge may play a role in triggering tornadogenesis. Note that the segments of circuits become very contorted. The rotational flow and the local vorticity extrema in the primary LMC/NMC may cause this contortion (see online supplement of the detailed evolution in a zoom-in view). As suggested by Dahl et al. (2012), a strongly curved, confluent flow may induce errors in the backward trajectory calculations because the gradients of the velocity field could be poorly resolved. A caveat is noted here about the possible errors in backtracking the circuits.

Fig. 7.

Locations of circuits traced backward for 10 min. The circuits are initiated as circles surrounding the cores of the vortices at 200 m AGL at 0250, 0251, 0252, and 0253 UTC. Gray-filled contour represents the mixing ratio of hydrometeors at the 1-km height. Dashed lines show the potential temperature at 200 m AGL. The colors on the circuits denote height AGL.

Fig. 7.

Locations of circuits traced backward for 10 min. The circuits are initiated as circles surrounding the cores of the vortices at 200 m AGL at 0250, 0251, 0252, and 0253 UTC. Gray-filled contour represents the mixing ratio of hydrometeors at the 1-km height. Dashed lines show the potential temperature at 200 m AGL. The colors on the circuits denote height AGL.

Fig. 8.

(a) The initial position of a material circuit (cyan line) surrounding the vortex core at 200 m AGL at 0250 UTC. Filled color contours refer to the vertical vorticity. Also present are black lines (isobars at 1-hPa intervals) and purple lines (solid: positive vertical velocity incremented by 2 m s−1; dashed: negative vertical velocity incremented by 1 m s−1). (b) Time series of the circulation, integrated (Int.) circulation, baroclinic (Bar.) term, and frictional (Fri.) term on the circuit from 0250 UTC (time = 20 min) to 0230 UTC (time = 0 min).

Fig. 8.

(a) The initial position of a material circuit (cyan line) surrounding the vortex core at 200 m AGL at 0250 UTC. Filled color contours refer to the vertical vorticity. Also present are black lines (isobars at 1-hPa intervals) and purple lines (solid: positive vertical velocity incremented by 2 m s−1; dashed: negative vertical velocity incremented by 1 m s−1). (b) Time series of the circulation, integrated (Int.) circulation, baroclinic (Bar.) term, and frictional (Fri.) term on the circuit from 0250 UTC (time = 20 min) to 0230 UTC (time = 0 min).

To investigate the change in the circulation sources before and just after tornadogenesis, the circulation sources of the near-ground vortices at the pretornadic (0250 UTC) and tornadogenesis (0253 UTC) times are discussed in the following subsections.

b. Circulation analysis of the pretornadic vortex

The circuit initialized surrounding the pretornadic vortex is shown in Fig. 8. The circulation calculated by Eq. (4) is compared with the circulation integrated from the terms on the rhs of Eq. (5) (named the integrated circulation). The histories of the two circulations show good agreement, which verifies the reliability of the circulation calculation. The magnitude of the frictional term is much larger than that of the baroclinic term. The positive frictional term lasts from approximately 0230 to 0234 UTC. After 0234 UTC, the frictional term is negative, resulting in a decrease in circulation. The decrease in circulation just before a circuit arrives at a vortex is also noticed in previous observational (e.g., Markowski et al. 2012) and numerical (e.g., M16; Schenkman et al. 2012) studies. The circuit is located near the ground (mostly lower than 200 m AGL; see Fig. 7a), implying that the frictional force reduces circulation when the flow converges near the ground.

c. Circulation analysis of the vortex at tornadogenesis time

The circulation analysis of the vortex at tornadogenesis time is shown in Fig. 9. The frictional term is also the dominant source. The frictional term is positive throughout most of the period. The negative frictional term also occurs during the last 1.5 min before the circuit arrives at the tornado. The baroclinic term is negligible compared to the frictional term. This result apparently differs from M16, who emphasized the contribution of baroclinity.

Fig. 9.

As in Fig. 8, but (a) time is at 0253 UTC; (b) time is from 0253 UTC (time = 23 min) to 0240 UTC (time = 10 min).

Fig. 9.

As in Fig. 8, but (a) time is at 0253 UTC; (b) time is from 0253 UTC (time = 23 min) to 0240 UTC (time = 10 min).

According to the caveat mentioned in section 5a, the complex flow in the primary LMC/NMC may induce error in the trajectory analysis. In addition, it could be supposed that if the flow is more complex, the error in trajectory calculation are more sensitive to the saving frequencies of model output. Evidence for the error is shown in the sensitivity tests. Prior to min 15 in Fig. 9, the period in which large circuit segments are contorted by the primary LMC/NMC, discernable differences in the circulation and the frictional term appear between the default case (model output saved every two time steps) and the case of every four time steps (not shown). The errors in trajectory calculation may also be the reason for the intensive oscillation of the frictional term prior to min 15.

The configurations of the circuit traced back in time are shown in Figs. 10a–10e. The aforementioned circuit segment on the south and west sides of the primary LMC is present before 0249 UTC. A primary hook echo is associated with the primary LMC. M16 found that the hook echo was accompanied by large baroclinity, which was the vorticity source of the tornado. To investigate the weak baroclinity in the present study, the buoyancy around the primary hook echo at 0249 UTC is plotted (Fig. 10f). The buoyancy term is estimated from the Boussinesq version of the vertical momentum equation:

 
dwdt=1ρ¯pz1ρ¯ρg,
(6)

where g is the gravitational acceleration, ρ′ is the density deviation from ρ¯, p′ is the perturbation pressure, which is the deviation from the base-state pressure, and ρ¯ is the average density over the 7.5 km by 7.5 km horizontal region around the hook echo. The first and second terms on the rhs are the vertical perturbation pressure gradient force (VPPGF) and buoyancy terms, respectively.

Fig. 10.

History of the horizontal projections of circuit traced back from 0253 to 0245 UTC. Height of the circuit is shown by different colors. In (a)–(e), the filled color contours represent the distribution of hydrometeors at the 500-m height. The dashed red and blue circles indicate the primary mesocyclone (P) and the mesocyclone associated with the tornado (T), respectively. The area of (f) is enclosed by the magenta rectangle. In (f), the buoyancy term in Eq. (6) is plotted; the blue dashed line denotes the mixing ratio of hydrometeors equal to 0.05 g kg−1, and vectors denote the storm-relative horizontal wind.

Fig. 10.

History of the horizontal projections of circuit traced back from 0253 to 0245 UTC. Height of the circuit is shown by different colors. In (a)–(e), the filled color contours represent the distribution of hydrometeors at the 500-m height. The dashed red and blue circles indicate the primary mesocyclone (P) and the mesocyclone associated with the tornado (T), respectively. The area of (f) is enclosed by the magenta rectangle. In (f), the buoyancy term in Eq. (6) is plotted; the blue dashed line denotes the mixing ratio of hydrometeors equal to 0.05 g kg−1, and vectors denote the storm-relative horizontal wind.

The buoyancy term difference across the hook echo is approximately 0.06 m s−2, which is smaller than that in M16 (approximately 0.2 m s−2, see his Fig. 11b). This difference between the two studies implies that compared with M16, the present study has weaker baroclinity at the hook echo and a warmer cold pool. The microphysics schemes may cause the difference. M16 utilized a single-moment scheme in which a fixed intercept parameter (8 × 106 m−4) was used to estimate the raindrop size distribution. This value is typically used in a number of previous studies (e.g., Schenkman et al. 2014), but there is still uncertainty regarding this parameter which is adjusted in some other studies (e.g., Dahl et al. 2014; Roberts et al. 2016). However, under realistic conditions, the intercept parameter of raindrops varies over a wide range. Multimoment schemes diagnose both mixing ratios and number concentrations; thus, the intercept number does not need to be predefined. Dawson et al. (2015, 2016) also suggested that multimoment microphysics schemes show advantages over single-moment schemes in predicting tornado track and intensity, because single-moment schemes tend to produce cold pools that are too strong. The two-moment microphysics scheme used in the present study supposedly resolves a more realistic condition.

Moreover, because the circuit segment is on the south and west sides of the hook echo rather than across the tip of the hook echo, there is no baroclinic vortex line penetrating the circuit, resulting in the small baroclinic term in circulation budget.5 This configuration of the circuit segment near the hook echo may result from the relationship between the primary LMC/hook echo and the tornado. The primary LMC/hook echo provides the downdraft6 (the RFD surge), which enters the tornado at genesis time. However, it is the tornadic LMC (associated with a tornadic hook echo) that provides the updraft for tornadogenesis. In M16, the circuit connects the downdraft and updraft regions associated with the same hook echo, allowing the baroclinity to be maintained. In the present study, the baroclinity is less likely to be maintained.

The discussion above indicates that the baroclinity at the hook echo is not necessarily a major vorticity source responsible for tornadogenesis. Here, we reiterated that frictional term is the major vorticity source. The fractions of the air parcels along the circuit in different height bins are shown in Fig. 11a. After being traced back for 3 min, most of the air parcels are below 100 m AGL. The fraction of the air parcels near the ground increases when the circuit is traced backward to an earlier time. Physically, this implies that a large proportion of the air in the tornado is drawn from the near-ground levels. The distribution of frictional term (Fig. 11b) is similar to that of the fraction of air parcels. The most contribution from friction occurs near the ground, especially in the first layer (below 10 m AGL). This result implies that surface drag has the dominant contribution to the frictionally generated circulation.

Fig. 11.

(a) The fraction of air parcels and (b) the frictional term in different height bins over the integration period for the circuit initialized at 0253 UTC. The abscissa is the time defined the same as Fig. 10. The ordinate is height. The bins are 10 s along the abscissa and 10 m along the ordinate. In (a), the shading represents the fraction of all air parcels. The total number of air parcels actually varies with time. In (b), the frictional term is the frictional term on the rhs of Eq. (5).

Fig. 11.

(a) The fraction of air parcels and (b) the frictional term in different height bins over the integration period for the circuit initialized at 0253 UTC. The abscissa is the time defined the same as Fig. 10. The ordinate is height. The bins are 10 s along the abscissa and 10 m along the ordinate. In (a), the shading represents the fraction of all air parcels. The total number of air parcels actually varies with time. In (b), the frictional term is the frictional term on the rhs of Eq. (5).

6. Vorticity budget analyses of air parcels

To explore the large frictionally generated vorticity in detail, the vorticity budgets of 140 air parcels in the tornado at genesis time (0253 UTC) are analyzed (Fig. 12a). The air parcels are initialized at 200 m AGL and are located at the centers of the 140 grids enclosed by the circuit analyzed in section 5c. The trajectories of the air parcels are calculated backward for 10 min (Fig. 12b). Based on where the air parcels originate, the air parcels can be categorized into three classes: (class-A) air parcels descending with the RFD surge, (class-B) air parcels from the north, and (class-C) air parcels from the low level in the eastern inflow.7

Fig. 12.

Backward trajectories of 140 air parcels from the tornadic vortex at 0253 UTC. (a) Initial locations of the air parcels at 0253 UTC. Dashed lines are isobars at 3-hPa intervals. The cyan circle denotes the same circuit as that in Fig. 9a. (b) Lines show the backward trajectories which are plotted in the storm-relative frame for 10 min. Filled gray contours show the potential temperature at 200 m AGL. Arrows refer to the storm-relative horizontal wind vectors. Thick points in (a) represent three representative air parcels. The colors of red and black represent air parcels in class-A and class-C, respectively. The colors of green and purple represent the air parcels in the first group and the second group of class-B, respectively. (c),(d) The partitions of circulation (area integral of vertical vorticity) and number of air parcels in (a) from each class.

Fig. 12.

Backward trajectories of 140 air parcels from the tornadic vortex at 0253 UTC. (a) Initial locations of the air parcels at 0253 UTC. Dashed lines are isobars at 3-hPa intervals. The cyan circle denotes the same circuit as that in Fig. 9a. (b) Lines show the backward trajectories which are plotted in the storm-relative frame for 10 min. Filled gray contours show the potential temperature at 200 m AGL. Arrows refer to the storm-relative horizontal wind vectors. Thick points in (a) represent three representative air parcels. The colors of red and black represent air parcels in class-A and class-C, respectively. The colors of green and purple represent the air parcels in the first group and the second group of class-B, respectively. (c),(d) The partitions of circulation (area integral of vertical vorticity) and number of air parcels in (a) from each class.

All three classes of the air parcels have contributions to the total circulation (the area integral of the vertical vorticity) of the tornado (Fig. 12c), while class-C contributes the most (60.6%). Class-C also occupies 44.3% of the total number of air parcels (Fig. 12d). Class-A and class-B only contribute 10.4% and 29% of the total circulation, respectively. This finding indicates that most vorticity sources should be attributed to the eastern inflow (class-C) and may also indicate that the main function of the RFD surge (class-A) is to enhance the near-ground convergence rather than to provide vorticity. It should be noted that these indications are only based on the tornado at genesis time. The conclusions could be different in other periods of the tornado (Marquis et al. 2016).

For trajectories in each class, histories of the vorticity terms show qualitatively similar tendencies. The vorticity calculated by interpolation from the vorticity on nearby model grid points is compared with the vorticity calculated by integrating the terms on the rhs of the vorticity equations. The former vorticity is named the interpolated vorticity. The latter is named the integrated vorticity. All of the trajectories show good agreement between the two vorticities (look ahead to Figs. 13c, 14c, and 16c for examples). This result indicates that the calculated trajectories and vorticity terms are reliable. In each class, the air parcel that has the best agreement is selected as a representative to be discussed in detail.

Fig. 13.

(a) Backward trajectories of the air parcels descending with the RFD surge (class-A) from 0253 to 0243 UTC. The colors of the trajectories denote height. The representative trajectory is plotted with thick line; horizontal vorticity vectors are plotted along the representative trajectory every 0.25 min (time is shown omitting “min”) using brown arrows. The horizontal plane shows the vertical vorticity at 35 m AGL at 0252:45 UTC when vertical tilting is prominent. Horizontal velocity vectors are presented by thin black arrows. (b)–(f) The history of the representative air parcel is shown. On the time axes, 0 represents 0253 UTC. (b) The trajectory height and (c) the interpolated vorticity (solid) and integrated vorticity (dashed) on vertical (black), streamwise (red), and crosswise (blue) components. (d)–(f) The time integrations of each term on the rhs of Eqs. (1)(3) from 0243 to 0253 UTC, starting from zero at 0243 UTC. In (d)–(f), red, black, orange, and green lines show stretching, tilting, frictional, and solenoidal (baroclinic) terms, respectively. In (d), the tilting of the streamwise (crosswise) vorticity into the vertical is shown by the green (red) dashed line. In (e) and (f), the exchange terms are shown by black dashed lines. All the streamwise and crosswise terms are defined in ground-relative frame.

Fig. 13.

(a) Backward trajectories of the air parcels descending with the RFD surge (class-A) from 0253 to 0243 UTC. The colors of the trajectories denote height. The representative trajectory is plotted with thick line; horizontal vorticity vectors are plotted along the representative trajectory every 0.25 min (time is shown omitting “min”) using brown arrows. The horizontal plane shows the vertical vorticity at 35 m AGL at 0252:45 UTC when vertical tilting is prominent. Horizontal velocity vectors are presented by thin black arrows. (b)–(f) The history of the representative air parcel is shown. On the time axes, 0 represents 0253 UTC. (b) The trajectory height and (c) the interpolated vorticity (solid) and integrated vorticity (dashed) on vertical (black), streamwise (red), and crosswise (blue) components. (d)–(f) The time integrations of each term on the rhs of Eqs. (1)(3) from 0243 to 0253 UTC, starting from zero at 0243 UTC. In (d)–(f), red, black, orange, and green lines show stretching, tilting, frictional, and solenoidal (baroclinic) terms, respectively. In (d), the tilting of the streamwise (crosswise) vorticity into the vertical is shown by the green (red) dashed line. In (e) and (f), the exchange terms are shown by black dashed lines. All the streamwise and crosswise terms are defined in ground-relative frame.

Fig. 14.

As in Fig. 13, but the air parcels are from north (class-B); the plane in (a) is at 120 m AGL; in (b), purple lines denote the times plotted in Fig. 15.

Fig. 14.

As in Fig. 13, but the air parcels are from north (class-B); the plane in (a) is at 120 m AGL; in (b), purple lines denote the times plotted in Fig. 15.

a. Class-A: Air parcel descending with the RFD surge

The representative air parcel of class-A (Fig. 13) starts descending from approximately 200 m AGL at 4 min before tornadogenesis (T − 4 min, hereafter). From T − 2 min to approximately T − 0.25 min, the air parcel is near the ground and moves quickly northeastward toward the tornado (also described in section 5a). During this period, crosswise vorticity increases mainly via frictional term. Streamwise vorticity increases mainly via the exchange from the crosswise vorticity. As mentioned previously, the streamwise and crosswise directions are defined with respect to the ground-relative wind. Just before arriving at the tornado (from T − 0.25 min), the air parcel moves northward in the east of the updraft region. Therefore, the crosswise and streamwise vorticity are tilted into the vertical (Fig. 13a). After the air parcel enters the tornado, the increase in vertical vorticity is dominated by stretching.

Baroclinic terms are small in the horizontal vorticity budgets, which is consistent with the weak baroclinity discussed in section 5c. The increase in the horizontal vorticity is mainly caused by the friction near the ground. This result is similar to Schenkman et al. (2014, see their Fig. 12).

b. Class-B: Air parcel from the north

The air parcels in class-B can be further divided into two groups. In the first group the air parcels originate from the cold pool to the west of the FFGF (green color in Fig. 12), while in the second group the air parcels are along the FFGF (purple color in Fig. 12). A distinct feature of the first group is that the air parcels experience several “internal” boundaries to the west of the FFGF. The representative air parcel is from the first group (Fig. 14). Before arriving at the FFGF, the air parcel transverses three “internal” boundaries (Fig. 15; as mentioned in section 5a, the FFGF shows consistency with the LFCB, thus labeled as “LFCB”). Updrafts are associated with the “internal” boundaries and there are downdrafts between the boundaries, which cause the air parcel to oscillate vertically.

Fig. 15.

The history of the representative air parcel (green dot) in Fig. 14 (class-B). (a1)–(a4) Horizontal convergence, and (b1)–(b4) vertical velocity at 120 m AGL. Three “internal” boundaries (“IB-1” to “IB-3”) and LFCB are pointed out by curved red arrows. Four times are plotted when the air parcel is at the “internal” boundaries and LFCB.

Fig. 15.

The history of the representative air parcel (green dot) in Fig. 14 (class-B). (a1)–(a4) Horizontal convergence, and (b1)–(b4) vertical velocity at 120 m AGL. Three “internal” boundaries (“IB-1” to “IB-3”) and LFCB are pointed out by curved red arrows. Four times are plotted when the air parcel is at the “internal” boundaries and LFCB.

From T − 0.5 to T − 0 min, the tilting of crosswise vorticity is predominant for the increase in the vertical vorticity. The crosswise vorticity is generated mainly owing to frictional and tilting terms from T − 1.5 to T − 1 min when the air parcel is at its nadir. This observation indicates that the frictional term may result from surface drag. Schenkman et al. (2014) generally found that large frictional generation was present when the parcels were closer to 50 m AGL and lower, while herein, the nadir of the air parcel is approximately 60 m AGL thus the frictional term is relatively small. For several other air parcels in the first group that have nadirs lower than 60 m AGL, the frictional terms are more significant (not shown).

To investigate the vertical influential range of the surface drag, the history of the vertical profile of the crosswise frictional term8 at the location of the representative air parcel is checked (not shown). It is found that from T − 1.5 to T − 1 min, the frictional term is enhanced between 50 and 150 m AGL. This enhancement is probably caused by the strong updraft in the FFGF where the air parcel is located (Figs. 15a-4,b-4). The strong updraft may extend the vertical range of the frictional term. At the levels above the lowest one, more active internal mixing stress may be induced by the large velocity gradient associated with the updraft and the gust front, and consequently results in the large frictional term.9 Similar (but weaker) enhancement is also found on the “internal” boundaries, which is consistent with the (weaker) updraft and convergence on these boundaries.

For all the air parcels in class-B, the baroclinic term is quite small, which may result from the weak density gradient across the FFGF. The potential virtual temperature difference between the cold pool and the warm inflow is approximately 4 K in this study (not shown). In M16, this difference is much larger, which is approximately 15 K (W. Mashiko 2018, personal communication). The result of the present study is different from those in several previous studies that emphasized the baroclinic effects (Wicker and Wilhelmson 1995; Markowski and Richardson 2014; Dahl 2015). In the second group, the air parcels originate from lower levels. The large crosswise vorticity is generated by friction. The baroclinic terms are also insignificant. The histories of the air parcels are similar to those of the air parcels in class-C except for the different moving directions; thus, a detailed description is omitted.

c. Class-C: Air parcel from the low level in the eastern inflow

The ground-relative inflow of the “eastern” parcels is actually from the south. Thus, the eastern inflow is also named southerly flow in this subsection. From T − 9 to T − 2 min, the crosswise vorticity of the representative air parcel (Fig. 16) increases owing to frictional term. The air parcel is near the ground during this period. Figure 17a depicts the history of the vertical profile of the crosswise frictional term at the location of the air parcel, showing that the discernible magnitude of the term is mainly below 50 m AGL. This implies that the frictional term is mainly induced by surface drag. From T − 0.5 to T − 0 min, the air parcel enters the tornado from the east side of the updraft. Thus, the crosswise vorticity is tilted into the vertical because the updraft region is located on the left side of the air parcel motion.

Fig. 16.

As in Fig. 13, but the air parcels are from the low level in the eastern inflow (class-C); (a1) shows all the air parcels and the horizontal plane at 0243 UTC; (a2) shows only the representative air parcel and the horizontal plane at 0252:45 UTC.

Fig. 16.

As in Fig. 13, but the air parcels are from the low level in the eastern inflow (class-C); (a1) shows all the air parcels and the horizontal plane at 0243 UTC; (a2) shows only the representative air parcel and the horizontal plane at 0252:45 UTC.

Fig. 17.

The time–height sections of (a) the crosswise frictional term and (b) the crosswise vorticity. The profile is at the location of the representative air parcel in the eastern inflow (class-C; Fig. 16). The time is the same as that in Fig. 16. The black lines represent the height of the air parcel.

Fig. 17.

The time–height sections of (a) the crosswise frictional term and (b) the crosswise vorticity. The profile is at the location of the representative air parcel in the eastern inflow (class-C; Fig. 16). The time is the same as that in Fig. 16. The black lines represent the height of the air parcel.

The environmental crosswise vorticity is also quite large (see T − 10 min in Fig. 16c). Because the environmental inflow (southerly inflow) is almost steady (see the eastern part in Fig. 16a-1), it is thought that the environmental inflow is similar to the flow in a typical boundary layer. Additionally, there is no comparable crosswise vorticity above 100 m AGL in the environmental inflow (see T − 10 min in Fig. 17b). Thus, this result indicates that the environmental crosswise vorticity (vertical wind shear) is also caused by surface drag.

Although the vertical grid is fine near the ground (10-m spacing), several air parcels in class-C are still traced back below the lowest scalar model level. Higher vertical resolution near the ground is necessary in future research (Rotunno et al. 2017).

7. Discussion and conclusions

a. Discussion

As mentioned previously, for the representative air parcels, the tilting of the crosswise vorticity into the vertical is predominant. However, significant upward tilting of streamwise vorticity is also found in several other parcels in each class. The streamwise vorticity is attained via exchange from crosswise vorticity when the moving directions of the air parcels change. This behavior is consistent with the second mechanism demonstrated by Roberts et al. (2016), and it is more prominent for the air parcels that have experienced large changes in moving directions (e.g., the black-marked air parcels in the south of the vortex in Fig. 12a, which was also noted in Markowski 2016). Furthermore, after the representative air parcels enter the tornado and swirl, the streamwise vorticity would increase and its tilting into vertical vorticity would be more significant.

Davies-Jones (1984) demonstrated that to form the updraft rotation (helical vortex) in supercell storms, the environmental horizontal vorticity should have a streamwise (in storm-relative frame) component (see his Figs. 7,8). This concern may raise some doubt about whether the tilting of crosswise vorticity in the present study can truly turn into the helical vortex of the tornado. Actually, as mentioned previously, the “crosswise” in the present study is ground-relative.10 We checked the vorticity in storm-relative frame and found that the tilting of storm-relative streamwise vorticity is prominent (Fig. 18). Figures 18a–18d show the ground-relative and storm-relative flow and the vortex lines passing the representative air parcels during (T − 10 s) the tilting of horizontal vorticity. The horizontal vorticity vectors are roughly parallel (perpendicular) to storm-relative (ground-relative) horizontal velocity vectors (especially evident in the eastern inflow region). The storm-relative streamwise vorticity is turned into the vertical (Figs. 18b,d–g) and the helical vortex is formed. Thus, the tilting of horizontal vorticity in this study is consistent with Davies-Jones (1984).

Fig. 18.

(a)–(d) The vortex lines passing the representative air parcels at 0252:50 UTC (T − 10 s). A, B, and C represent the three representative air parcels, respectively. “Start” and “end” denote the start and end of the vortex lines. The colors of the lines show the height AGL. The black points denote the representative air parcels. The planes are at 35 m AGL in (a)–(b) and 170 m AGL in (c)–(d). The black and green arrows denote the horizontal velocity and horizontal vorticity vectors, respectively. In (a) and (c), the shading is the vertical tilting of ground-relative crosswise vorticity (gr. ωc) and the black arrows are ground-relative velocity vectors. In (b) and (d), the shading is the vertical tilting of storm-relative streamwise vorticity (sr. ωs) and the black arrows are storm-relative velocity vectors. The black contours are the vertical vorticity larger than 0.1 s−1 at 0.3 s−1 intervals. (e)–(g) As in Figs. 13d, 14d, and 16d, respectively, but the streamwise and crosswise directions are in the storm-relative frame and only the last one minute is plotted.

Fig. 18.

(a)–(d) The vortex lines passing the representative air parcels at 0252:50 UTC (T − 10 s). A, B, and C represent the three representative air parcels, respectively. “Start” and “end” denote the start and end of the vortex lines. The colors of the lines show the height AGL. The black points denote the representative air parcels. The planes are at 35 m AGL in (a)–(b) and 170 m AGL in (c)–(d). The black and green arrows denote the horizontal velocity and horizontal vorticity vectors, respectively. In (a) and (c), the shading is the vertical tilting of ground-relative crosswise vorticity (gr. ωc) and the black arrows are ground-relative velocity vectors. In (b) and (d), the shading is the vertical tilting of storm-relative streamwise vorticity (sr. ωs) and the black arrows are storm-relative velocity vectors. The black contours are the vertical vorticity larger than 0.1 s−1 at 0.3 s−1 intervals. (e)–(g) As in Figs. 13d, 14d, and 16d, respectively, but the streamwise and crosswise directions are in the storm-relative frame and only the last one minute is plotted.

b. Conclusions

In this study, the 6 May 2012 Tsukuba tornado was successfully reproduced utilizing WRFLES. The tornado was explicitly resolved with 50-m horizontal grid spacing. The vorticity sources responsible for the tornadogenesis were investigated by circulation and vorticity analyses. The following conclusions can be drawn.

Circulation budgets for the circuits initialized around the pretornadic and tornadic vortices showed that the magnitude of the frictional term was consistently much larger than the magnitude of the baroclinic term. In the case of the pretornadic vortex several minutes before tornadogenesis, friction actually reduced the circulation when the circuit approached the vortex. For the vortex at the time of tornadogenesis, friction contributed significant positive circulation. Considering the controversy of the vorticity sources of tornadoes, the present study emphasized again the dominant contribution of frictionally generated vorticity as proposed by Schenkman et al. (2014). The air parcels entering the tornado at genesis time could be categorized into three classes: (class-A) air parcels descending with the RFD surge, (class-B) air parcels from the north, and (class-C) air parcels from the low level in the eastern inflow. For each class, the surface drag was the main vorticity source of the tornado. The frictionally generated crosswise vorticity was directly tilted (or first exchanged into the streamwise vorticity and then tilted) into the vertical. Among the three classes, class-C contributed the most to the circulation of the tornado.

An RFD surge from the downdraft region on the south and west sides of the primary LMC/hook echo might play a role in triggering a tornado by increasing the convergence near the ground. The baroclinity across the hook echo was weaker than that in M16, which might lead to the lack of baroclinity in the vorticity sources.

Caveats should be mentioned that the MOST surface layer scheme might be questionable in the outflow; the LES might suffer from the overestimated wind shear due to underresolved turbulence and inadequate grid resolution near the surface. Another caveat was that as the air parcels were allowed to descend below the lowest scaler model level, the backward trajectories were prone to errors. The errors consisted of the uncertainty in both the location of air parcels and the extrapolated relevant fields. Moreover, the backward trajectory calculations might still need a longer time to reveal all the possible vorticity sources, especially considering that after being calculated backward for 20 min, the circulation of the circuit from the pretornadic vortex was even larger than the initial circulation. The circulation of the circuit from the tornado almost reached zero after a 13-min backward calculation. However, the circulation source terms in a prior time might still be uncertain. This uncertainty can be indicated from a circuit studied in Markowski (2016). The baroclinic term was not prominent until the circuit was calculated backward for 15 min, and the frictional term showed a remarkable negative value after a 25-min backward calculation (see his Fig. 23).

Acknowledgments

Dr. Hidenori Kawai contributed a considerable amount of help in the simulations of this study. The authors appreciate the valuable advice about the numerical configurations and trajectory calculations from Dr. Wataru Mashiko. The authors also acknowledge Prof. Hiroshi Niino, and Drs. Hiromu Seko and Kazuo Saito for the helpful discussions about the WRF simulation of tornadoes. We also thank Christopher Weiss and the anonymous reviewers for their constructive comments and suggestions on the original manuscript. The simulations were performed using the supercomputer TSUBAME 3.0 at the Tokyo Institute of Technology, financially assisted by the MEXT FLAGSHIP project: “Advancement of meteorological and global environment predictions utilizing observational ‘Big Data’” of the social and scientific priority issues (Theme 4). The authors also appreciate the support from the China Scholarship Council (CSC) for overseas doctoral study.

REFERENCES

REFERENCES
Adlerman
,
E. J.
,
2003
:
Numerical simulations of cyclic storm behaviors: Mesocyclogenesis and tornadogenesis. Ph.D. dissertation, University of Oklahoma, 219 pp
.
Adlerman
,
E. J.
,
K. K.
Droegemeier
, and
R.
Davies-Jones
,
1999
:
A numerical simulation of cyclic mesocyclogenesis
.
J. Atmos. Sci.
,
56
,
2045
2069
, https://doi.org/10.1175/1520-0469(1999)056<2045:ANSOCM>2.0.CO;2.
Beck
,
J. R.
, and
C.
Weiss
,
2013
:
An assessment of low-level baroclinity and vorticity within a simulated supercell
.
Mon. Wea. Rev.
,
141
,
649
669
, https://doi.org/10.1175/MWR-D-11-00115.1.
Beck
,
J. R.
,
J. L.
Schroeder
, and
J. M.
Wurman
,
2006
:
High-resolution, dual-Doppler analyses of the 29 May 2001 Kress, Texas, cyclic supercell
.
Mon. Wea. Rev.
,
134
,
3125
3148
, https://doi.org/10.1175/MWR3246.1.
Brioude
,
J.
,
W. M.
Angevine
,
S. A.
McKeen
, and
E. Y.
Hsie
,
2012
:
Numerical uncertainty at mesoscale in a Lagrangian model in complex terrain
.
Geosci. Model Dev.
,
5
,
1127
1136
, https://doi.org/10.5194/GMD-5-1127-2012.
Dahl
,
J. M.
,
2015
:
Near-ground rotation in simulated supercells: On the robustness of the baroclinic mechanism
.
Mon. Wea. Rev.
,
143
,
4929
4942
, https://doi.org/10.1175/MWR-D-15-0115.1.
Dahl
,
J. M.
,
M. D.
Parker
, and
L. J.
Wicker
,
2012
:
Uncertainties in trajectory calculations within near-surface mesocyclones of simulated supercells
.
Mon. Wea. Rev.
,
140
,
2959
2966
, https://doi.org/10.1175/MWR-D-12-00131.1.
Dahl
,
J. M.
,
M. D.
Parker
, and
L. J.
Wicker
,
2014
:
Imported and storm-generated near-ground vertical vorticity in a simulated supercell
.
J. Atmos. Sci.
,
71
,
3027
3051
, https://doi.org/10.1175/JAS-D-13-0123.1.
Davies-Jones
,
R.
,
1984
:
Streamwise vorticity: The origin of updraft rotation in supercell storms
.
J. Atmos. Sci.
,
41
,
2991
3006
, https://doi.org/10.1175/1520-0469(1984)041<2991:SVTOOU>2.0.CO;2.
Davies-Jones
,
R.
, and
H.
Brooks
,
1993
:
Mesocyclogenesis from a theoretical perspective. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 105–114
.
Dawson
,
D. T.
,
M.
Xue
,
J. A.
Milbrandt
, and
A.
Shapiro
,
2015
:
Sensitivity of real-data simulations of the 3 May 1999 Oklahoma City tornadic supercell and associated tornadoes to multimoment microphysics. Part I: Storm- and tornado-scale numerical forecasts
.
Mon. Wea. Rev.
,
143
,
2241
2265
, https://doi.org/10.1175/MWR-D-14-00279.1.
Dawson
,
D. T.
,
M.
Xue
,
A.
Shapiro
,
A.
Milbrandt
, and
A. D.
Schenkman
,
2016
:
Sensitivity of real-data simulations of the 3 May 1999 Oklahoma City tornadic supercell and associated tornadoes to multimoment microphysics. Part II: Analysis of buoyancy and dynamic pressure forces in simulated tornado-like vortices
.
J. Atmos. Sci.
,
73
,
1039
1061
, https://doi.org/10.1175/JAS-D-15-0114.1.
Deardorff
,
J. W.
,
1972
:
Numerical investigation of neutral and unstable planetary boundary layers
.
J. Atmos. Sci.
,
29
,
91
115
, https://doi.org/10.1175/1520-0469(1972)029<0091:NIONAU>2.0.CO;2.
Fiori
,
E.
,
A.
Parodi
, and
F.
Siccardi
,
2010
:
Turbulence closure parameterization and grid spacing effects in simulated supercell storms
.
J. Atmos. Sci.
,
67
,
3870
3890
, https://doi.org/10.1175/2010JAS3359.1.
Flournoy
,
M. D.
, and
M. C.
Coniglio
,
2019
:
Origins of vorticity in a simulated tornadic mesovortex observed during PECAN on 6 July 2015
.
Mon. Wea. Rev.
,
147
,
107
134
, https://doi.org/10.1175/MWR-D-18-0221.1.
Guchte
,
A. V.
, and
J. M.
Dahl
,
2018
:
Sensitivities of parcel trajectories beneath the lowest scalar model level of a Lorenz vertical grid
.
Mon. Wea. Rev.
,
146
,
1427
1435
, https://doi.org/10.1175/MWR-D-17-0190.1.
Iacono
,
M. J.
,
J. S.
Delamere
,
E. J.
Mlawer
,
M. W.
Shepard
,
S. A.
Clough
, and
W. D.
Collins
,
2008
:
Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models
.
J. Geophys. Res.
,
113
,
D13103
, https://doi.org/10.1029/2008JD009944.
Japan Meteorological Agency
,
2012
:
Tornadoes generated on 6 May 2012 (in Japanese). Meteorological Research Institute, Tokyo District Meteorological Observatory, and Sendai District Meteorological Observatory, 14 pp., http://www.jma.go.jp/jma/menu/tatsumaki-portal/tyousa-houkoku.pdf
.
Jiménez
,
P. A.
,
J.
Dudhia
,
J. F.
González-Rouco
,
J.
Navarro
,
J. P.
Montávez
, and
E.
García-Bustamante
,
2012
:
A revised scheme for the WRF surface layer formulation
.
Mon. Wea. Rev.
,
140
,
898
918
, https://doi.org/10.1175/MWR-D-11-00056.1.
Kosiba
,
K.
,
J.
Wurman
,
Y.
Richardson
,
P.
Markowski
,
P.
Robinson
, and
J.
Marquis
,
2013
:
Genesis of the Goshen County, Wyoming, Tornado on 5 June 2009 during VORTEX2
.
Mon. Wea. Rev.
,
141
,
1157
1181
, https://doi.org/10.1175/MWR-D-12-00056.1.
Lemon
,
L. R.
, and
C. A.
Doswell
,
1979
:
Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis
.
Mon. Wea. Rev.
,
107
,
1184
1197
, https://doi.org/10.1175/1520-0493(1979)107<1184:STEAMS>2.0.CO;2.
Markowski
,
P. M.
,
2016
:
An idealized numerical simulation investigation of the effects of surface drag on the development of near-surface vertical vorticity in supercell thunderstorms
.
J. Atmos. Sci.
,
73
,
4349
4385
, https://doi.org/10.1175/JAS-D-16-0150.1.
Markowski
,
P. M.
, and
Y. P.
Richardson
,
2014
:
The influence of environmental low-level shear and cold pools on tornadogenesis: Insights from idealized simulations
.
J. Atmos. Sci.
,
71
,
243
275
, https://doi.org/10.1175/JAS-D-13-0159.1.
Markowski
,
P. M.
, and
G. H.
Bryan
,
2016
:
LES of laminar flow in the PBL: A potential problem for convective storm simulations
.
Mon. Wea. Rev.
,
144
,
1841
1850
, https://doi.org/10.1175/MWR-D-15-0439.1.
Markowski
,
P. M.
,
Y. P.
Richardson
,
E.
Rasmussen
,
J.
Straka
,
R.
Davies-Jones
, and
R. J.
Trapp
,
2008
:
Vortex lines within low-level mesocyclones obtained from pseudo-dual-Doppler radar observations
.
Mon. Wea. Rev.
,
136
,
3513
3535
, https://doi.org/10.1175/2008MWR2315.1.
Markowski
,
P. M.
, and et al
,
2012
:
The pretornadic phase of the Goshen County, Wyoming, supercell of 5 June 2009 intercepted by VORTEX2. Part II: Intensification of low-level rotation
.
Mon. Wea. Rev.
,
140
,
2916
2938
, https://doi.org/10.1175/MWR-D-11-00337.1.
Markowski
,
P. M.
,
N. T.
Lis
,
D. D.
Turner
,
T. R.
Lee
, and
M. S.
Buban
,
2019
:
Observations of near-surface vertical wind profiles and vertical momentum fluxes from VORTEX-SE 2017: Comparisons to Monin–Obukhov similarity theory
.
Mon. Wea. Rev.
,
147
,
3811
3824
, https://doi.org/10.1175/MWR-D-19-0091.1.
Marquis
,
J. N.
,
Y. P.
Richardson
,
P.
Markowski
,
D.
Dowell
, and
J.
Wurman
,
2012
:
Tornado maintenance investigated with high-resolution dual-Doppler and EnKF analysis
.
Mon. Wea. Rev.
,
140
,
3
27
, https://doi.org/10.1175/MWR-D-11-00025.1.
Marquis
,
J. N.
,
Y. P.
Richardson
,
P.
Markowski
,
D.
Dowell
,
K.
Kosiba
,
P.
Robinson
, and
G.
Romine
,
2014
:
An investigation of the Goshen County, Wyoming, tornadic supercell of 5 June 2009 using EnKF assimilation of mobile mesonet and radar observations collected during VORTEX2. Part I: Experiment design and verification of the EnKF analyses
.
Mon. Wea. Rev.
,
142
,
530
554
, https://doi.org/10.1175/MWR-D-13-00007.1.
Marquis
,
J. N.
,
Y. P.
Richardson
,
P.
Markowski
,
J.
Wurman
,
K.
Kosiba
, and
P.
Robinson
,
2016
:
An investigation of the Goshen County, Wyoming, tornadic supercell of 5 June 2009 using EnKF assimilation of mobile mesonet and radar observations collected during VORTEX2. Part II: Mesocyclone-scale processes affecting tornado formation, maintenance, and decay
.
Mon. Wea. Rev.
,
144
,
3441
3463
, https://doi.org/10.1175/MWR-D-15-0411.1.
Mashiko
,
W.
,
2016a
:
A numerical study of the 6 May 2012 Tsukuba City supercell tornado. Part I: Vorticity sources of low-level and midlevel mesocyclones
.
Mon. Wea. Rev.
,
144
,
1069
1092
, https://doi.org/10.1175/MWR-D-15-0123.1.
Mashiko
,
W.
,
2016b
:
A numerical study of the 6 May 2012 Tsukuba City supercell tornado. Part II: Mechanisms of tornadogenesis
.
Mon. Wea. Rev.
,
144
,
3077
3098
, https://doi.org/10.1175/MWR-D-15-0122.1.
Mashiko
,
W.
,
H.
Niino
, and
T.
Kato
,
2009
:
Numerical simulation of tornadogenesis in an outer-rainband minisupercell of Typhoon Shanshan on 17 September 2006
.
Mon. Wea. Rev.
,
137
,
4238
4260
, https://doi.org/10.1175/2009MWR2959.1.
Mason
,
P. J.
, and
D. J.
Thomson
,
1992
:
Stochastic backscatter in large-eddy simulations of boundary layers
.
J. Fluid Mech.
,
242
,
51
78
, https://doi.org/10.1017/S0022112092002271.
Moeng
,
C.-H.
,
J.
Dudhia
,
J.
Klemp
, and
P.
Sullivan
,
2007
:
Examining two-way grid nesting for large eddy simulation of PBL using the WRF Model
.
Mon. Wea. Rev.
,
135
,
2295
2311
, https://doi.org/10.1175/MWR3406.1.
Morrison
,
H.
,
G.
Thompson
, and
V.
Tatarskii
,
2009
:
Impact of cloud microphysics on the development of trailing stratiform precipitation in a simulated squall line: Comparison of one- and two-moment schemes
.
Mon. Wea. Rev.
,
137
,
991
1007
, https://doi.org/10.1175/2008MWR2556.1.
Naylor
,
J.
, and
M. S.
Gilmore
,
2014
:
Vorticity evolution leading to tornadogenesis and tornadogenesis failure in simulated supercells
.
J. Atmos. Sci.
,
71
,
1201
1217
, https://doi.org/10.1175/JAS-D-13-0219.1.
Roberts
,
B.
, and
M.
Xue
,
2017
:
The role of surface drag in mesocyclone intensification leading to tornadogenesis within an idealized supercell simulation
.
J. Atmos. Sci.
,
74
,
3055
3077
, https://doi.org/10.1175/JAS-D-16-0364.1.
Roberts
,
B.
,
M.
Xue
,
A. D.
Schenkman
, and
D. T.
Dawson
II
,
2016
:
The role of surface drag in tornadogenesis within an idealized supercell simulation
.
J. Atmos. Sci.
,
73
,
3371
3395
, https://doi.org/10.1175/JAS-D-15-0332.1.
Rotunno
,
R.
,
P. M.
Markowski
, and
G. H.
Bryan
,
2017
:
“Near ground” vertical vorticity in supercell thunderstorm models
.
J. Atmos. Sci.
,
74
,
1757
1766
, https://doi.org/10.1175/JAS-D-16-0288.1.
Schenkman
,
A. D.
,
M.
Xue
, and
A.
Shapiro
,
2012
:
Tornadogenesis in a simulated mesovortex within a mesoscale convective system
.
J. Atmos. Sci.
,
69
,
3372
3390
, https://doi.org/10.1175/JAS-D-12-038.1.
Schenkman
,
A. D.
,
M.
Xue
, and
M.
Hu
,
2014
:
Tornadogenesis in a high-resolution simulation of the 8 May 2003 Oklahoma City supercell
.
J. Atmos. Sci.
,
71
,
130
154
, https://doi.org/10.1175/JAS-D-13-073.1.
Seko
,
H.
,
M.
Kunii
,
S.
Yokota
,
T.
Tsuyuki
, and
T.
Miyoshi
,
2015
:
Ensemble experiments using a nested LETKF system to reproduce intense vortices associated with tornadoes of 6 May 2012 in Japan
.
Prog. Earth Planet. Sci.
,
2
,
42
, https://doi.org/10.1186/s40645-015-0072-3.
Skamarock
,
W. C.
, and et al
,
2008
: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., https://doi.org/10.5065/D68S4MVH.
Straka
,
J. M.
,
E. N.
Rasmussen
,
R. P.
Davies-Jones
, and
P. M.
Markowski
,
2007
:
An observational and idealized numerical examination of low-level counter-rotating vortices in the rear flank of supercells
.
Electron. J. Severe Storms Meteor.
,
2
(
8
), http://www.ejssm.org/ojs/index.php/ejssm/article/viewArticle/32.
Sun
,
Z.
,
M.
Xue
,
K.
Zhu
, and
B.
Zhou
,
2019
:
Prediction of an EF4 supercell tornado in Funing, China: Resolution dependency of simulated tornadoes and their structures
.
Atmos. Res.
,
229
,
175
189
, https://doi.org/10.1016/j.atmosres.2019.06.019.
Wicker
,
L. J.
, and
R. B.
Wilhelmson
,
1995
:
Simulation and analysis of tornado development and decay within a three-dimensional supercell thunderstorm
.
J. Atmos. Sci.
,
52
,
2675
2703
, https://doi.org/10.1175/1520-0469(1995)052<2675:SAAOTD>2.0.CO;2.
Wyngaard
,
J. C.
,
2004
:
Toward numerical modeling in the “terra incognita.”
J. Atmos. Sci.
,
61
,
1816
1826
, https://doi.org/10.1175/1520-0469(2004)061<1816:TNMITT>2.0.CO;2.
Xu
,
X.
,
M.
Xue
, and
Y.
Wang
,
2015
:
The genesis of mesovortices within a real-data simulation of a bow echo system
.
J. Atmos. Sci.
,
72
,
1963
1986
, https://doi.org/10.1175/JAS-D-14-0209.1.
Yamauchi
,
H.
,
H.
Niino
,
O.
Suzuki
,
Y.
Shoji
,
E.
Sato
,
A.
Adachi
, and
W.
Mashiko
,
2013
:
Vertical structure of the Tsukuba F3 tornado on 6 May 2012 as revealed by a polarimetric radar. 36th Conf. on Radar Meteorology, Breckenridge, CO, Amer. Meteor. Soc., 320, https://ams.confex.com/ams/36Radar/webprogram/Paper228827.html
.
Yokota
,
S.
,
H.
Niino
,
H.
Seko
,
M.
Kunii
, and
H.
Yamauchi
,
2018
:
Important factors for tornadogenesis as revealed by high-resolution ensemble forecasts of the Tsukuba supercell tornado of 6 May 2012 in Japan
.
Mon. Wea. Rev.
,
146
,
1109
1132
, https://doi.org/10.1175/MWR-D-17-0254.1.
For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Footnotes

1

The horizontal grid spacing in their simulation was 500 m. Thus, the tornado was not explicitly resolved. The near-surface rotation that presented characteristics of a tornado is referred to as near-surface mesocyclone.

2

In WRFLES250m, the LMC is defined where vertical vorticity (ζ) at the 1-km height is larger than 0.05 s−1. ζ is concentrated at the intersection between FFGF and RFGF; thus, the LMC is easy to identify. In WRFLES50m, at the corresponding location there are several local vorticity extrema.

3

Circuits located along vorticity isopleths have also been tested. The resulting circulation budgets and locations are qualitatively similar to those of circuits initiated as rings.

4

In the storm-relative frame, the inflow direction is westward and is thus named the eastern inflow.

5

The circuit segment in M16 is across the tip of the hook echo. As a result, the baroclinic term is large.

6

The downdraft is mainly forced by the negative VPPGF term on the rear side of the primary LMC (not shown).

7

About five air parcels in class-C (trajectories from the southmost in Fig. 12b) are traced back to a relatively higher location in weak convection from the south. These air parcels also have a similar history as the other air parcels in class-C except for some pre-existing vertical vorticity.

In Dahl et al. (2012) the air parcels that are traced back to the eastern inflow are regarded as erroneous parcels because the storm outflow forms a boundary preventing the inflow from entering the tornado near the ground. The authors utilized free-slip bottom condition. On the other hand, several recent studies did show air parcels coming from the eastern inflow (e.g., Naylor and Gilmore 2014; Schenkman et al. 2014; Marquis et al. 2016; M16; Roberts and Xue 2017). Most of these studies used nonfree-slip bottom condition. This may indicate that the nonfree-slip condition alters storm outflow features, enabling the eastern inflow to enter the low level of the tornado. Despite this, the possible error in trajectory is noted here.

8

Crosswise frictional term means the fifth (the frictional) term on the rhs of the crosswise vorticity equation [Eq. (2)].

9

This explanation may lead to a thought that internal mixing is also a vorticity source. However, Markowski (2016) found that in free-slip simulation, internal mixing actually decreases the circulation. Thus, we consider that internal mixing should not be regarded as a vorticity source. Surface drag is the vorticity source, and the effect of surface drag is extended upward via internal mixing.

10

The horizontal vorticity induced by surface drag is generally perpendicular to ground-relative wind. To clearly show this vorticity, the “crosswise” is thus defined ground-relative. The vertical titling of ground-relative crosswise vorticity is also evident (Figs. 18a,c).