Available online at www.sciencedirect.

com
Procedia
Engineering
Procedia Engineering 00 (2011)17000–000
Procedia Engineering (2011) 141 – 150
www.elsevier.com/locate/procedia

The 2nd International Symposium on Aircraft Airworthiness (ISAA 2011)

3D Engineering Model of Downburst Evolution in

Thunderstorm
TANG Chua *, HONG Guanxina
a
School of Aeronautics Science and Engineering, Beihang University, Beijing, P.R.Chia

Abstract

Downburst in thunderstorm mainly refers to the outflow on frontier and the strong wind shear caused by the
downburst. In this paper, the thunderstorm evolution is divided into three stages as growing, mature and dissipating
stages according to its development. Based on meteorological data, trends of relevant parameters are presented. Then
with the 3D ring vortices method, the classical fluid mechanics theory is applied through adding linear source model
to simulate the distribution of velocity in the wind shear field and its evolution over time. This modeling method
could effectively simulate the evolution of downburst, and is suitable for engineering application.

© 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Airworthiness
Technologies Research Center NLAA, and Beijing Key Laboratory on Safety of Integrated Aircraft and
Propulsion Systems, China Open access under CC BY-NC-ND license.

Keyword: Thunderstorm, Downburst, Evolution, Vortex rings model, Linear source model

1. Introduction

A variety of meteorological phenomena can cause downburst. Generally, strong convection current in
certain conditions will cause downburst. Downburst is easy to form, determined by the characteristics of
thunderstorm. Downburst affects flight safety, and for an aircraft flying through downburst, the wind
changes from head wind into tail wind in short time, which will cause deviations of the aircraft from the

* Corresponding author. Tel.: +86-010-82338971.

E-mail address: tangchu@ase.buaa.edu.cn.

1877-7058 © 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2011.10.017
2142 TANG
TANG Chu,Chu and HONG
HONG Guanxin
Guanxin/ / Procedia
Procedia Engineering
Engineering 17000–000
00 (2011) (2011) 141 – 150

original flight path and difficulty for to pilot to control it , especially in take-off and landing operations.
Severe downburst even causes flight accidents. In order to explore the dynamics of aircraft in the
downburst wind field, it is necessary to build up a model to reflect the real fluid characteristics of
downburst.
Three kinds of methods have gained favor in past to simulate downburst. The first one is to simulate
the fluid field characteristics by disposing fundamental solutions of potential flow theory and selecting its
positions and numbers. The second method is to derive atmospheric dynamic equations with basic fluid
dynamics equations and thermodynamic laws and then solve for the velocity distributions of the wind
field and its change with time. The third one is to establish a database with data from real measurements,
such as JAWS and Doppler radars.
The first method is simple and easy to use, and models with this method include the doublet sheet
model [1], the vortex section model [2], and the vortex ring model [3,4], etc., all of which are 3D models
that can simulate downburst field characteristics at a certain time. However, the assumption of frozen
field is generally used, which means velocity is only the function of position, so it cannot continuously
describe the evolution stage of the downburst field. The second method is based on atmospheric dynamics,
describing wind field characteristic completely, such as wind speed, rainfall, temperature, pressure and
radar reflectivity factor. It shows the evolution of downburst clearly, but is not suitable for engineering
research because of its complexity and huge amount of calculation. The third models, though close to
reality, are restricted by measuring methods and costs, with limited usable data.
Therefore, it is necessary to develop a simple method to describe the evolution of downburst from
beginning to end that facilitates engineering applications. With current engineering simulation methods
and simulation results of atmospheric dynamics equations, this paper presents a 3D evolution model of
downburst by introducing time varying ring vortices and linear source models.

2. Modeling Method

2.1. The Basic character and Evolution Downburst in Thunderstorm

In different zones, two kinds of wind shear can be formed by downburst in thunderstorm as the
downdraft under thunderstorm and the outflow after it reaches ground [5].
Thunderstorm evolution can be divided into three stages as growing, mature and dissipating phase. For
different scales, evolution times also differ, from a few to dozens of minutes.
Figure 1 shows the life cycle of a typical downburst. From (T-5) to T is the growing phase. In this
phase, the main form is downdraft, and then the outflow appears gradually as time passes. From T to
(T+10) is the mature phase. During this phase, outflow velocity value reaches the maximum and the
horizontal velocity becomes the largest at (T+10). After (T+10) is the dissipating phase. The radar detects
horizontal wind shear at (T-2) and gives the alarm till the end of the life cycle.

Fig. 1. the life cycle of downburst [6]

 TANG
TANGChu andHONG
Chu, HONGGuanxin/
GuanxinProcedia
/ Procedia Engineering
Engineering 0017 (2011)
(2011) 141 – 150
000–000 3 143

Mark R. Hjelmfelt generalized the rule of downburst evolution after studying the Doppler radar data of
26 cases in JAWS. In the beginning, the descending downburst showed no observable wind shear. When
downdraft reached the surface, the Doppler radar first found horizontal wind shear and only a short time
before surface-level divergence was seen. As the downdraft reached the ground surface, the outflow
began to take form. The wind shear increased to intensity and began to grow. The horizontal vortex
circulation developed and began to move away from the thunderstorm core. After reaching maximum
intensity, about half of the downdraft continued to expand. The rest weakened without further growth.
Finally, some outflows grew beyond downdraft scale and became larger-scale outflows before dissipating
[6].
The biggest threat for flight through a downburst is the strong horizontal wind shear from several
dozens to a hundred meters above ground. Data shows that the horizontal shear value of 10-2s-1 is the
danger limit for commercial aircraft [7]. The danger will increase when the maximum vertical velocity
height is lower.

2.2. Downburst Calculation Model

As mentioned above, this paper uses ring vortex and linear source models with time varying positions
and intensities to simulate the characteristics of downburst wind field. It is assumed that, the ground
boundary layer can be ignored due to its small thickness; the wind field is filled with incompressible and
inviscid ideal gas; the effects of rain and temperature changes can be ignored.
Both ring vortex and linear source models are axisymmetric, therefore when studying the induced
speed field, only the field on one side of the symmetric axis need to be considered. Select cylindrical
coordinate system to describe the model, where r is the radial coordinate, and h is the axial coordinate.
The wind field dimension is 2R×H (R is the maximum radius, H is the maximum height), and P(r, h) is
an arbitrary point inside the wind field. Zero height means the ground. The ring vortex has strength of Γ,
with its center at (rv, hv) and a radius of rv. The unit linear source model has a strength of Q with its two
ends at (0, 0) and (0, hv).
2.2.1. The Ring Vortex Model
In order to find the speed components at a point caused by the ring vortex, it is required to calculate
the Stokes’ stream function, Ψ. The Stokes’ stream function due to a ring vortex circulation Γ is given by
(see Fig 2) [8]，
Γ
ψ=
− ( R1 + R2 )  F1 ( λ ) − E1 ( λ ) (1)
2π
Where,
r2 + r1
λ=
r2 − r1
R1 and R2 are the farthest and the nearest distance from P to the ring vortex.
=R1 max ( (r − r ) + (h − h ) ,
v
2
v
2 2
)
( r + rv ) + ( h − hv )
2

min ( ( r − r ) + ( h − h ) , (r + r ) + (h − h ) )
2 2 2 2
=
R2 v v v v

F1 ( λ ) and E1(λ) are elliptic integrals. In the calculation of F1(λ)–E1(λ), the approximation method
employed by Ivan [3] was used, quoted as Equation (2). For 0≤λ2≤1.
4144 TANG
TANG Chu,Chu and HONG
HONG Guanxin
Guanxin/ / Procedia
Procedia Engineering
Engineering 17000–000
00 (2011) (2011) 141 – 150

0.788λ 2
[ F1 (λ ) − E1 (λ )] = (2)
0.25 + 0.75 1 − λ 2
The radial and vertical speeds (vrr and vhr respectively) are calculated with,

 1 ∂ψ
vrr = r ∂h
 (3)
v = − 1 ∂ψ

hr
r ∂r

There are areas within the velocity field where Equation (1) and Equation (3) cannot be used. The
resulting division by zero can be avoided by setting vrr= 0 along the vertical axis. The corresponding
vertical velocity is then calculated from the equation for the velocity induced along the axis of a ring
vortex.
vrr = 0

 −
3
 Γ   h 2  2 (4)
=vhr 1 +   
 2rv   2rv  

The velocity field is calculated by Equation (3) and (4).

h
The Ring Vortex Model
(-rv, hv) (rv, hv)
(0, hv)

R1 R2

α P(r,h)
dh R

hs
r
The Linear Source Model

Fig. 2. The ring vortex and linear source model diagram

2.2.2. The Linear Source Model

The linear source model is the continuous distributions of point source between the center of the ring
vortex and the ground. The model is used to simulate outflow of the downburst. dh is an element of the
linear source (see Fig 2), and the induced velocity at P by the element dh is
 TANG
TANGChu andHONG
Chu, HONGGuanxin/
GuanxinProcedia
/ Procedia Engineering
Engineering 0017 (2011)
(2011) 141 – 150
000–000 5 145

Qdh
dv =
4π R 2
Where,

r 2 + ( h − hs )
2
R=

The induced velocity components are

 r Qdh
dvrs = R 4π R 2

dv = ( h − hs ) Qdh
 hs R 4π R 2

Integrating above equations leads to the induced velocity components of the linear source model at
P(r,h) as in Equation (5),

 hv r Qdh
vrs = ∫0 R 4π R 2
 (5)
v = hv ( h − hs ) Qdh
 hs ∫0 R 4π r

As shown in Figure 2, a variable α is introduce to help solve Equation (5), and α has such relations
with r, R, h, hs as

 r
sin α = R
 (6)
cos α = h − hs
 R

Velocity components from Equation (5) and (6) are

 Qr
=vrs 4π ( cos α 2 − cos α1 )

v = Q ( sin α − sin α )
 hs 4π r 2 1

Where,
r h − hs
sin α1
= cos α1
,=
r +h
2 2
r 2 + h2
r h − hs
sin α 2
= cos α 2
,=
r 2 + ( h − hv ) r 2 + ( h − hv )
2 2

2.2.3. Ground Boundary Conditions

The velocity field to be simulated is that of the model close to the ground. This implies zero vertical
6146 TANG
TANG Chu,Chu and HONG
HONG Guanxin
Guanxin/ / Procedia
Procedia Engineering
Engineering 17000–000
00 (2011) (2011) 141 – 150

velocity (vh= 0) along the ground surface plane. In order to produce proper flow boundary conditions,
mirror models reflected in the ground plane is introduced. The circulation strength of the real ring vortex
is Γ, so that of its image must be -Γ. Also, the center of the image ring vortex is (rv, -hv), and the
strength of the image linear source is ―Q. As above, the two ends of the image linear source model is at
(0, 0) and (0, -hv), and the velocity components in the image models are vr΄ and vh΄.
2.2.4. Superposed model
To simulate the field of downburst, the ring vortex model will be superposed on linear source model.
And, the velocity field is the superposition of speeds of the real and the image models. Furthermore, the
thunderstorm is moving with a translation speed of V0, so the total vector is the summation of the
translation speed and the wind speed generated by the model, as showed in Equation (7),
V = Vr + Vr′ + Vs + Vs′ + V0 (7)
Where, Vr and Vs are the velocity vectors of ring vortex and linear source respectively in the real
models and Vr’ and Vs’ in the image models.
2.2.5. Downburst evolution
Evolution is the time history of velocity field. Through analysis, the results of atmospheric dynamics
showed that the parameters include strength, radius, height of the ring vortex, and strength of the linear
source.
The evolution model is divided into three stages. See Table 1.

Table 1. Phases and main form of downburst evolution

Phase Time Main form

Growing 0-t1 Downdraft
Mature t1-t2 Outflow appears and reaches the maximum speed
Dissipating t2-t3 Velocity decays till the cycle ends

In the two-dimensional rectilinear vortex model, Swolinshky proposed the rule that the induced
velocity and the radius follow the exponential rule along with the time change [9]. Therefore, introduce
time-varying parameters (Γ, rv, hv, Q), such as in Equation (8) ，
Γ (t ) =
aΓ ebΓ t
rv ( t ) = ar ebr t
 a ebh t (0 < t ≤ t1 )
hv ( t ) =  h (8)
 H (t1 ) (t > t2 )

a e Q
b t
(0 < t ≤ t2 )
Q(t ) =  Q
 Q ( t2 ) ( t > t2 )
Where, aГ, ar, ah, bГ, br, bh are corresponding parameters determined by the field.

3. Simulation Results and Analysis

3.1. Comparison of Simulation results

As an example, the downburst on June 30, 1982 in Denver is simulated with this method. Measured
values and the simulation results of the mature phase are compared. See Table 2.
In this paper, the whole evolution process of this downburst is simulated. Table 2 lists the simulation
 TANG Chu
TANG andHONG
Chu, HONGGuanxin/
Guanxin Procedia
/ ProcediaEngineering
Engineering0017(2011)
(2011)000–000
141 – 150 7 147

results, atmospheric dynamics simulation results and the measured results of the mature phase. The
results show that, the simulation results with this method are close to those obtained with atmospheric
dynamics. Yet the diameter and velocity vectors are smaller than measured values.

Table 2. Simulation values and measured values

Simulation value of this paper Simulation value in literature [10] Measured value [11]
umax(m/s) 18.7 18.4 17
Δu 37.4 36.81 33
vmin(m/s) -9.60 -10.6 -12~-15
Diameter of umax(m) 3000 3300 3600
Height of vmin(m) 800 740-1000 750-1000

3.2. Analysis of Simulation Results

As the horizontal scale of an isolated thunderstorm is usually a dozen kilometers, define the simulation
region as 16km×4km, and the horizontal and vertical grid spacing is 0.2km. The coefficients of time-
varying parameters (Γ, rv, hv, Q) are shown in Table 3.
From the growing to the mature phase, trends with time of the vortex ring strength, radius, height and
the linear source strength are calculated and the results are shown in fig 3. In the dissipating phase,
diameter of the ring vortex is beyond the horizontal scale of thunderstorm and when rv>10km, the
evolution ends.
This model simulates velocity field of downburst from the growing to the dissipating stage. At 7.5min,
downdraft reached the ground. Before 7.5min was the growing phase, from 7.5min to 13min was the
mature phase, and after 13min was the dissipating phase. The vertical velocity was 9.08m/s at 7.5min, and
then the outflow appeared, and the horizontal velocity reached its maximum value. After 13min, velocity
was gradually decreasing. At 12min, the radius of the outflow center was more than 2km, and the outflow
region was expanding.

Table 3. Time-varying parameters

Γ rv hv Q
a 232 341 15488 5711
b 2/π 1/2π -1/3 4/5π

Fig 4 shows graphics of the maximum horizontal velocities and the maximum vertical velocities from
4min to 20min in the field. The horizontal velocity is warning to 7.7m/s, ualarm [12]. T is the time the
alarm lasted. In the growing phase, both horizontal and vertical velocities increased over time, however,
the two values are not significant. At the mature phase, downdraft reached the ground and formed the
outflow, indicating that the horizontal velocity increased rapidly. Finally, in the dissipating phase,
horizontal and vertical velocities were both decaying.
According to the results, the maximum horizontal velocities from 7.5min to12min are larger than the
danger limit of the horizontal shear; the risk peak is even 1.29×10-2s-1 at 9min. After 7.5min, the vertical
velocity reached a peak value, more than 9m/s, and the height was less than 1km, which will increase the
risk of flight.
148
8 TANG
TANG Chu
Chu, and HONG
HONG Guanxin
Guanxin/ / Procedia
Procedia Engineering
Engineering 17 (2011)
00 (2011) 000–000141 – 150

2.0x105
16

14

1.5x105 12

10
Γ

1.0x105

hv / km
8

5.0x104 4

0.0 0
0 4 8 12 16 20 0 4 8 12 16 20
t / min t / min

(a) (b)
9 1.6x105

1.2x105
6
rv / km

Q 8.0x104

4.0x104

0 0.0
0 4 8 12 16 20 0 4 8 12 16 20
t / min t / min

(c) (d)

Fig. 3. (a) Γ as a function of time; (b) time history of hv; (c) time history of rv; (d) time history of Q

20

umax
vmax
15 ualarm
Vmax / m· s-1

10

5
T

0
4 6 8 10 12 14 16 18 20
t / min

Fig. 4. time history of the maximum horizon/ vertical velocity

Wind field simulation of each phase is shown in fig 5. Velocity values from the smallest to the largest
are represented with changing colors with blue for the smallest to red for the largest. It can be seen from
the figure that, in the growing phase, the altitude of downburst was higher (more than 2km), the outflow
did not appear near the ground, and the downdraft was the main factor affecting flight. In the mature
phase, outflow appeared near the ground. Flying through a downburst region, an aircraft encounters
outflow and downdraft (at altitude less than 1km), and they all affect flight safety. It is extremely
dangerous to fly through the wind field during this phase. That is the reason why most models only
 TANG
TANGChu andHONG
Chu, HONGGuanxin/
GuanxinProcedia
/ Procedia Engineering
Engineering 0017 (2011)
(2011) 141 – 150
000–000 9 149

simulate the wind field of this phase. In the dissipating phase, outflow became larger. The velocity change
of outflow front will also affect flight safety.
4000
4000

2000
3000

h/m
y/m

0 2000

1000
-2000

0
-4000
-4000 -2000 0 2000 4000
-4000 -2000 0 2000 4000
x/m x/m

a) Growing phase
4000
4000

2000
3000
h/m
y/m

0 2000

-2000
1000

0
-4000 -4000 -2000 0 2000 4000
-4000 -2000 0 2000 4000
x/m x/m

b) Mature phase
8000

5000

6000
h/m

4000
y/m

2000
-5000

0
-5000 0 5000
-5000 0 5000
x/m x/m

c) Dissipating phase

Fig. 5. Flow lines of horizontal section and velocity of longitudinal profile of each phase
150
10 TANG
TANG Chu,Chu and HONG
HONG Guanxin
Guanxin/ / Procedia
Procedia Engineering
Engineering 17000–000
00 (2011) (2011) 141 – 150

4. Conclusions

This paper presents an engineering model to simulate downburst evolution through introducing time-
varying parameters of ring vortex and linear source models, and the results are compared with measured
values. Following conclusions are obtained:
1) Based on the ring vortex model of downburst, through superposing the linear source model, the
new model can describe downburst evolution well. The method is simple and easy for calculation.
2) By introducing the linear source model, the new model could describe outflow when the
downdraft reaches the ground.
3) Parameters affecting the downburst evolution include the strength, radius and altitude of the ring
vortex and the strength of the linear source. These parameters follow the exponential rule with
time.
4) During mature and dissipating phases, the horizontal shear value is larger and it can even exceed
the danger limit, thus threaten flight safety.
5) The model is able to simulate the whole evolution process of downburst thus can be used in the
research of flight characteristics through downburst.
This paper simulated the downburst under an isolated axisymmetric thunderstorm without translation
speed. The model describes the structure of downburst flow with some meteorological parameters
neglected. In addition, thunder clouds often accompany severe isolated thunderstorm, so the interactions
between them need further study.

References

[1] Zhu Shangxiang, Etkin B. Model of the Wind Field in a downburst. Journal of Aircraft, 1985;22(7):595-601
[2] Liu Gang, Wang Xingren, Jia Rongzhen. Engineering Simulation of Variable Wind Field in Synthetic Natural Environment.
Journal of System Simulation, 2006,18(2):297-300
[3] Ivan M. A Ring-Vortex Downburst Model for Flight Simulators. Journal of Aircraft, 1986;23 (3): 232-236
[4] Gao Zhenxing, Gu Hongbin, Liu Hui. Real-time Simulation of Large Aircraft Flying Through Micoburst. Chinese Journal of
Aeronautics, 2009(22):459-466
[5] Zhu Shangxiang. Hydrokinetics Model of Microburst. Flight Dynamics, 1985;59-72
[6] Mark R. Hjelmfelt. Structure and Life Cycle of Microburst Outflows Observed in Colorado. Journal of applied meteorology,
1987;(27): 900-927
[7] Fred H. Proctor. Numerical Simulation of an Isolated Microburst. Part I: Dynamics and Structure. Journal of the atmospheric
sciences,1988;45 (21): 3137-3160
[8] Paul Aaron Robinson. The Modeling of Turbulence and Downbursts for Flight Simulators. Toronto: University of Toronto,
1991
[9] Xiao Yelun, Jin Changjiang. Flight Principle in Atmospheric Disturbance. Beijing: National Defence Industry Press, 1993
[10] Kong Fanyou, Huang Meiyuan, Xu Huanying. Numerical Simulations of the Microburst. Sciatica Atmospheric Sinica ,
1994;18 (1):11-21
[11] Sun Qingmin, Jin Changjiang. Numerical Simulation of Microburst Case. Flight Dynamics,1995;13 (4):29-32
[12] Chen Huizhi. Aviation Flight and Safety under Dangerous Weather Condition. Southwest Jiaotong University. 2003