Ch10 Forces Winds

Copyright 2015 by Roland Stull. Practical Meteorology: An Algebra-based Survey of Atmospheric Science.
10 Atmospheric Forces & Winds
Contents Winds power our wind turbines, push our sail-

boats, cool our houses, and dry our laundry. But
Winds and Weather Maps 290 winds can also be destructive in hurricanes,
Heights on Constant-Pressure Surfaces 290 thunderstorms, or mountain downslope wind-
Plotting Winds 291 storms. We design our bridges and skyscrapers to
Newtons 2nd Law 292 withstand wind gusts. Airplane flights are planned
Lagrangian 292 to compensate for headwinds and crosswinds.
Eulerian 293 Winds are driven by forces acting on air. But
Horizontal Forces 294 these forces can be altered by heat and moisture car-
Advection of Horizontal Momentum 294 ried by the air, resulting in a complex interplay we
Horizontal Pressure-Gradient Force 295 call weather. Newtons laws of motion describe how
Centrifugal Force 296 forces cause winds a topic called dynamics.
Coriolis Force 297
Many forces such as pressure-gradient, advection,
Turbulent-Drag Force 300
Summary of Forces 301
and frictional drag can act in all directions. Inertia
creates an apparent centrifugal force, caused when
Equations of Horizontal Motion 301
centripetal force (an imbalance of other forces)
Horizontal Winds 302 makes wind change direction. Local gravity acts
Geostrophic Wind 302 mostly in the vertical. But a local horizontal compo-
Gradient Wind 304
nent of gravity due to Earths non-spherical shape,
Atmospheric-Boundary-Layer Wind 307
ABL Gradient (ABLG) Wind 309
combined with the contribution to centrifugal force
Cyclostrophic Wind 311 due to Earths rotation, results in a net force called
Inertial Wind 312 Coriolis force.
Antitriptic Wind 312 These different forces are present in different
Summary of Horizontal Winds 313 amounts at different places and times, causing large
Horizontal Motion 314 variability in the winds. For example, Fig. 10.1 shows
Equations of Motion Again 314 changing wind speed and direction around a low-
Scales of Horizontal Motion 315 pressure center. In this chapter we explore forces,
Vertical Forces and Motion 315 winds, and the dynamics that link them.
Conservation of Air Mass 317
Continuity Equation 317
Incompressible Idealization 318
Boundary-Layer Pumping 318
Kinematics 320 OPSUI
Measuring Winds 321 1
L1B
Review 322

Homework Exercises 322
Broaden Knowledge & Comprehension 322
Apply 323
Evaluate & Analyze 325
Synthesize 327

TPVUI
XFTU FBTU
Practical Meteorology: An Algebra-based Survey
of Atmospheric Science by Roland Stull is licensed
under a Creative Commons Attribution-NonCom-
mercial-ShareAlike 4.0 International License. View this license at Figure 10.1
http://creativecommons.org/licenses/by-nc-sa/4.0/ . This work is Winds (arrows) around a low-pressure center (L) in the N.
available at http://www.eos.ubc.ca/books/Practical_Meteorology/ .
Hemisphere. Lines are isobars of sea-level pressure (P).
289
290 chapter 10 Atmospheric Forces & Winds
B

1
Winds and Weather Maps

[ L1B Heights on Constant-Pressure Surfaces
LN

Pressure-gradient force is the most important
[
force because it is the only one that can drive winds

LN in the horizontal. Other horizontal forces can alter
an existing wind, but cannot create a wind from
calm air. All the forces, including pressure-gradient
XFTU FBTU force, are explained in the next sections. However,
to understand the pressure gradient, we must first
understand pressure and its atmospheric variation.
We can create weather maps showing values
C
[
of the pressures measured at different horizontal
locations all at the same altitude, such as at mean-
L1B sea-level (MSL). Such a map is called a constant-
height map. However, one of the peculiarities of
meteorology is that we can also create maps on oth-
er surfaces, such as on a surface connecting points
UI
OPS
of equal pressure. This is called an isobaric map.

- Both types of maps are used extensively in meteo-
LN rology, so you should learn how they are related.
UI
TPV
In Cartesian coordinates (x, y, z), z is height above

XFTU FBTU some reference level, such as the ground or sea level.
Sometimes we use geopotential height H in place of
z, giving a coordinate set of (x, y, H) (see Chapter 1).
Can we use pressure as an alternative vertical
coordinate instead of z? The answer is yes, because
D
[LNTVSGBDF E
1L1BTVSGBDF pressure changes monotonically with altitude. The
OPSUI
OPSUI
word monotonic means that the value of the de-

1L1B [LN
pendent variable changes in only one direction (nev-
er decreases, or never increases) as the value of the

-
independent variable increases. Because P never in-
creases with increasing z, it is indeed monotonic, al-
lowing us to define pressure coordinates (x, y, P).
An isobaric surface is a conceptual curved sur-
TPVUI
TPVUI
face that connects points of equal pressure, such as

XFTU FBTU XFTU FBTU the shaded surface in Fig. 10.2b. The surface is higher
above sea level in high-pressure regions, and lower
in low-pressure regions. Hence the height contour
lines for an isobaric surface are good surrogates for
Figure 10.2 pressures lines (isobars) on a constant height map.
Sketch of the similarity of (c) pressures drawn on a constant Contours on an isobaric map are analogous to el-
height surface to (d) heights drawn on a constant pressure sur- evation contours on a topographic map; namely, the
face. At any one height such as z = 5 km (shown by the thin map itself is flat, but the contours indicate the height
dotted line in the vertical cross section of Fig. a), pressures at of the actual surface above sea level.
one location on the map might differ from pressure at other loca- High pressures on a constant height map corre-
tions. In this example, a pressure of 50 kPa is located midway spond to high heights of an isobaric map. Similarly,
between the east and west limits of the domain at 5 km altitude.
regions on a constant-height map that have tight
Pressure always decreases with increasing height z, as sketched
in the vertical cross section of (a). Thus, at the other locations
packing (close spacing) of isobars correspond to
on the cross section, the 50 kPa pressure will be found at higher regions on isobaric maps that have tight packing of
altitudes, as sketched by the thick dashed line. This thick dashed height contours, both of which are regions of strong
line and the corresponding thin dotted straight line are copied pressure gradients that can drive strong winds. This
into the 3-D view of the same scenario is sketched in Fig. b. one-to-one correspondence of both types of maps
L indicates the cyclone center, having low pressure and low (Figs. 10.2c & d) makes it easier for you to use them
heights. interchangeably.
R. Stull Practical Meteorology 291
Isobaric surfaces can intersect the ground, but

two different isobaric surfaces can never intersect INFO Why use isobaric maps?
because it is impossible to have two different pres- There are five reasons for using isobaric charts.
sures at the same point. Due to the smooth mono- 1) During the last century, the radiosonde (a
tonic decrease of pressure with height, isobaric sur- weather sensor hanging from a free helium balloon
faces cannot have folds or creases. that rises into the upper troposphere and lower strato-
We will use isobaric charts for most of the up- sphere) could not measure its geometric height, so in-
per-air weather maps in this book when describing stead it reported temperature and humidity as a func-
upper-air features (mostly for historical reasons; see tion of pressure. For this reason, upper-air charts
INFO box). Fig. 10.3 is a sample weather map show- (i.e., maps showing weather above the ground) tradi-
ing height contours of the 50 kPa isobaric surface. tionally have been drawn on isobaric maps.
2) Aircraft altimeters are really pressure gauges.
Aircraft assigned by air-traffic control to a specific al-
Plotting Winds titude above 18,000 feet MSL will actually fly along
Symbols on weather maps are like musical notes an isobaric surface. Many weather observations and
in a score they are a shorthand notation that con- forecasts are motivated by aviation needs.
cisely expresses information. For winds, the symbol 3) Air pressure is created by the weight of air mol-
is an arrow with feathers (or barbs and pennants). ecules. Thus, every point on an isobaric map has the
The tip of the arrow is plotted over the observation same mass of air molecules above it.
(weather-station) location, and the arrow shaft is 4) An advantage of using equations of motion in
aligned so that the arrow points toward where the pressure coordinates is that you do not need to con-
wind is going. The number and size of the feath- sider density, which is not routinely observed.
ers indicates the wind speed (Table 10-1, copied from 5) Numerical weather prediction models some-
Table 9-9). Fig. 10.3 illustrates wind barbs. times use pressure as the vertical coordinate.
Items (1) and (5) are less important these days, be-
cause modern radiosondes use GPS (Global Posi-
Table 10-1. Interpretation of wind barbs.
tioning System) to determine their (x, y, z) position.
Symbol Wind Speed Description So they report all meteorological variables (including
calm two concentric circles pressure) as a function of z. Also, some of the modern
weather forecast models do not use pressure as the
1 - 2 speed units shaft with no barbs vertical coordinate. Perhaps future weather analyses
5 speed units a half barb (half line) and numerical predictions will be shown on constant-
height maps.
10 speed units each full barb (full line)
50 speed units each pennant (triangle)
The total speed is the sum of all barbs and pennants.

8
For example, indicates a wind from the west at
speed 75 units. Arrow tip is at the observation location.
-/

CAUTION: Different organizations use different
8

speed units, such as knots, m s1, miles h1, km h1, etc.

Look for a legend to explain the units. When in doubt, /
/
assume knots the WMO standard. For unit conver-

)
sion, a good approximation is 1 m s1 2 knots.

8
Sample Application

/
Draw wind barb symbol for winds from the:

(a) northwest at 115 knots; (b) northeast at 30 knots.
)
Find the Answer
) [LN
8
(a) 115 knots = 2 pennants + 1 full barb + 1 half barb. /

(b) 30 knots = 3 full barbs L1B)FJHIUT LN
8JOET LOPUT
65$ 4FQ
B
C
Check: Consistent with Table 10-1.

Exposition: Feathers (barbs & pennants) should be on Figure 10.3
Winds (1 knot 0.5 m s 1) and heights (km) on the 50 kPa iso-
the side of the shaft that would be towards low pres-
baric surface. The relative maxima and minima are labeled as H
sure if the wind were geostrophic.
(high heights) and L (low heights). Table 10-1 explains winds.
INFO Newtons Laws of Motion

Isaac Newtons published his laws in Latin, the
Newtons 2nd Law
language of natural philosophy (science) at the
time (1687). Here is the translation from Newtons Lagrangian
Philosophi Naturalis Principia Mathematica (Math- For a Lagrangian framework (where the coordi-
ematical Principles of Natural Philosophy):
nate system follows the moving object), Newtons
Second Law of Motion is
Law I. Every body perseveres in its state of be-
ing at rest or of moving uniformly straight forward,
except inasmuch as it is compelled by impressed forc- F = m a (10.1)

es to change its state.
where F = ism a
aforce vector, m is mass of the ob-

Law II. Change in motion is proportional to the ject, = m a is the acceleration vector of the object.
F and
motive force impressed and takes place following the Namely, the object accelerates in the direction of the
straight line along which that force is impressed. applied force.
Acceleration is the velocity V change during a
Law III. To any action, there is always a con- short time interval t:
trary, equal reaction; in other words, the actions of
two bodies each upon the other are always equal and
opposite in direction. V
a = (10.2)
t
Corollary 1. A body under the joint action of
Plugging eq. (10.2) into (10.1) gives:
forces traverses the diagonal of a parallelogram in the
same time as it describes the sides under their sepa-
rate actions.
V
F = m (10.3a)
t

Recall that momentum is defined as m V . Thus, if
the objects mass is constant, you can rewrite New-
tons 2nd Law as Lagrangian momentum bud-
get:

Sample Application (m V )
If a 1200 kg car accelerates from 0 to 100 km h1 in 7 F = (10.3b)
s, heading north, then: (a) What is its average accelera-
t
tion? (b) What vector force caused this acceleration?
Namely, this equation allows you to forecast the rate
Find the Answer of change of the objects momentum.
If the object is a collection of air molecules mov-
Given: V initial = 0, V final = 100 km h1 = 27.8 m s1 ing together as an air parcel, then eq. (10.3a) allows
tinitial = 0, tfinal = 7 s. Direction is north.
you to forecast the movement of the air (i.e., the
m = 1200 kg.

wind). Often many forces act simultaneously on an

Find: m a = ? ms2 , (b) F = m
F = (a) ?N a air parcel, so we should rewrite eq. (10.3a) in terms

of the net force:
V
(a) Apply eq. (10.2): a =
t
= (27.8 0 m s1) / (7 0 s) = 3.97 ms2 to the V Fnet (10.4)
=
north t m

(b) Apply eq. (10.1): F = m a kg) ( 3.97 ms2 )
(1200
= 4766 N to the north = is
where F net m the
a vector sum of all applied forces, as
where 1 N = 1 kgms2 (see Appendix A). given by Newtons Corollary 1 (see the INFO box).

For situations where F net =/m
m=
a 0, eq. (10.4) tells us
Check: Physics and units are reasonable.
that the flow will maintain constant velocity due to
Exposition: My small car can accelerate from 0 to 100
km in 20 seconds, if I am lucky. Greater acceleration inertia. Namely, V /t = 0 implies that
consumes more fuel, so to save fuel and money, you
V = constant (not that V = 0).
should accelerate more slowly when you drive.
In Chapter 1 we defined the (U, V, W) wind com-

Sample Application
ponents in the (x, y, z) coordinate directions (positive
Initially still air is acted on by force Fy net/m = 5x10 4
toward the East, North, and up). Thus, we can split ms2 . Find the final wind speed after 30 minutes.
eq. (10.4) into separate scalar (i.e., non-vector) equa-
tions for each wind component: Find the Answer
Given: V(0) = 0, Fy net/m = 5x10 4 ms2 , t = 1800 s
U Fx net (10.5a) Find: V(t) = ? m s1. Assume: U = W = 0.
=
t m
Apply eq. (10.6b): V(t+t) = V(t) + t (Fy net/m)
V Fy net = 0 + (1800s)(5x10 4 ms2) = 0.9 m s 1.
= (10.5b)
t m
Check: Physics and units are reasonable.
W Fz net Exposition: This wind toward the north (i.e., from
= (10.5c)
t m 180) is slow. But continued forcing over more time
could make it faster.
where Fx net is the sum of the x-component of all the
applied forces, and similar for Fy net and Fz net .
From the definition of = final initial, you can
expand U/t to be [U(t+t) U(t)]/t. With similar A SCIENTIFIC PERSPECTIVE Creativity
expansions for V/t and W/t, eq. (10.5) becomes As a child at Woolsthorpe, his mothers farm in
Fx net England, Isaac Newton built clocks, sundials, and
U(t + t) = U(t) + t (10.6a) model windmills. He was an average student, but his
m schoolmaster thought Isaac had potential, and recom-
Fy net mended that he attend university.
V (t + t) = V (t) + t (10.6b) Isaac started Cambridge University in 1661. He
m was 18 years old, and needed to work at odd jobs to
pay for his schooling. Just before the plague hit in
Fz net 1665, he graduated with a B.A. But the plague was
W (t + t) = W (t) + t (10.6c)
m spreading quickly, and within 3 months had killed
10% of London residents. So Cambridge University
These are forecast equations for the wind, and are
was closed for 18 months, and all the students were
known as the equations of motion. The Numeri- sent home.
cal Weather Prediction (NWP) chapter shows how While isolated at his mothers farm, he continued
the equations of motion are combined with budget his scientific studies independently. This included
equations for heat, moisture, and mass to forecast much of the foundation work on the laws of motion,
the weather. including the co-invention of calculus and the expla-
nation of gravitational force. To test his laws of mo-
tion, he built his own telescope to study the motion of
Eulerian planets. But while trying to improve his telescope, he
While Newtons 2nd Law defines the fundamen- made significant advances in optics, and invented the
tal dynamics, we cannot use it very easily because reflecting telescope. He was 23 - 24 years old.
it requires a coordinate system that moves with the
air. Instead, we want to apply it to a fixed location It is often the young women and men who are most
(i.e., an Eulerian framework), such as over your creative in the sciences as well as the arts. Enhanc-
house. The only change needed is to include a new ing this creativity is the fact that these young people
term called advection along with the other forces, have not yet been overly swayed (perhaps misguided)
when computing the net force Fnet in each direction. in their thinking by the works of others. Thus, they
are free to experiment and make their own mistakes
All these forces are explained in the next section.
and discoveries.
But knowing the forces, we need additional infor- You have an opportunity to be creative. Be wary
mation to use eqs. (10.6) we need the initial winds of building on the works of others, because subcon-
[U(t), V(t), W(t)] to use for the first terms on the right sciously you will be steered in their same direction
side of eqs. (10.6). Hence, to make numerical weather of thought. Instead, I encourage you to be brave, and
forecasts, we must first observe the current weather explore novel, radical ideas.
and create an analysis of it. This corresponds to an This recommendation may seem paradoxical. You
initial-value problem in mathematics. are reading my book describing the meteorological
Average horizontal winds are often 100 times advances of others, yet I discourage you from reading
stronger than vertical winds, except in thunder- about such advances. You must decide on the best
storms and near mountains. We will focus on hori- balance for you.
zontal forces and winds first.
Horizontal Forces
Five forces contribute to net horizontal accelera-

tions that control horizontal winds: pressure-gra-
dient force (PG), advection (AD), centrifugal
force (CN), Coriolis force (CF), and turbulent
drag (TD):
(10.7a)
Fx net Fx AD Fx PG FxCN FxCF FxTD
= + + + +
m m m m m m
(10.7b)
Fy net Fy AD Fy PG Fy CN Fy CF FyTD
= + + + +
m m m m m m
Centrifugal force is an apparent force that allows us

to include inertial effects for winds that move in a
curved line. Coriolis force, explained in detail later,
includes the gravitational and compound centrifu-
gal forces on a non-spherical Earth. In the equations
above, force per unit mass has units of N kg1. These
units are equivalent to units of acceleration (ms2 ,
see Appendix A), which we will use here.
Advection of Horizontal Momentum

Advection is not a true force. Yet it can cause a
change of wind speed at a fixed location in Eulerian
coordinates, so we will treat it like a force here. The
wind moving past a point can carry specific mo-
mentum (i.e., momentum per unit mass). Recall
that momentum is defined as mass times velocity,
B
Z C
D
hence specific momentum equals the velocity (i.e.,
7 the wind) by definition. Thus, the wind can move
(advect) different winds to your fixed location.
Y This is illustrated in Fig. 10.4a. Consider a mass
NT of air (grey box) with slow U wind (5 m s1) in the
north and faster U wind (10 m s1) in the south.
7 6
Thus, U decreases toward the north, giving U/y
= negative. This whole air mass is advected toward
0 0
NT
0
NT
the north over a fixed weather station O by a south
wind (V = positive). At the later time sketched in Fig.
7 6 10.4b, a west wind of 5 m s1 is measured at O.
Even later, at the time of Fig. 10.4c, the west wind has
increased to 10 m s1 at the weather station. The rate
NT NT of increase of U at O is larger for faster advection
(V), and is larger if U/y is more negative.
6 Thus, U/t = V U/y for this example. The
advection term on the RHS causes an acceleration
of U wind on the LHS, and thus acts like a force per
NT
unit mass: U/t = Fx AD/m = V U/y .
You must always include advection when mo-
Figure 10.4 mentum-budget equations are written in Eulerian
Illustration of V advection of U wind. O is a fixed weather frameworks. This is similar to the advection terms
station. Grey box is an air mass containing a gradient of U in the moisture- and heat-budget Eulerian equations
wind. Initial state (a) and later states (b and c). that were in earlier chapters.
For advection, the horizontal force components

Sample Application
are
Minneapolis (MN, USA) is about 400 km north of
Fx AD U U U Des Moines (IA, USA). In Minneapolis the wind com-
= U V W (10.8a) ponents (U, V) are (6, 4) m s1 , while in Des Moines
m x y z
they are (2, 10) m s1. What is the value of the advective
Fy AD V V V force per mass?
= U V W (10.8b)
m x y z
Find the Answer
Recall that a gradient is defined as change across a Given: (U, V) = (6, 4) m s1 in Minneapolis,
distance, such as V/y. With no gradient, the wind (U, V) = (2, 10) m s1 in Des Moines
cannot cause accelerations. y = 400 km, x = is not relevant
Vertical advection of horizontal wind (WU/z Find: Fx AD/m =? ms2 , Fy AD/m =? ms2
in eq. 10.8a, and WV/z in eq. 10.8b) is often very
Use the definition of a gradient:
small outside of thunderstorms.
U/y = (6 2 m s1)/400,000 m = 1.0x10 5 s1
U/x = not relevant, U/z = not relevant,
Horizontal Pressure-Gradient Force V/y = (4 10 m s1)/400,000 m = 1.5x10 5 s1
In regions where the pressure changes with dis- V/x = not relevant, V/z = not relevant
Average U = (6 + 2 m s1)/2 = 4 m s1
tance (i.e., a pressure gradient), there is a force
Average V = (4 + 10 m s1)/2 = 7 m s1
from high to low pressure. On weather maps, this
force is at right angles to the height contours or iso- Use eq. (10.8a):
bars, directly from high heights or high pressures to Fx AD/m = (7m s1)(1.0x10 5 s1 )
low. Greater gradients (shown by a tighter packing = 7x10 5 ms2
of isobars; i.e., smaller spacing d between isobars Use eq. (10.8b):
on weather maps) cause greater pressure-gradient Fy AD/m = (7m s1)(1.5x10 5 s1 )
force (Fig. 10.5). Pressure-gradient force is indepen- = 1.05x10 4 ms2
dent of wind speed, and thus can act on winds of
any speed (including calm) and direction. Check: Physics and units are reasonable.
For pressure-gradient force, the horizontal com- Exposition: The slower U winds from Des Moines
ponents are: are being blown by positive V winds toward Minne-
apolis, causing the U wind speed to decrease at Min-
Fx PG 1 P neapolis. But the V winds are increasing there because
= (10.9a)
m x of the faster winds in Des Moines moving northward.
Fy PG 1 P (10.9b)
=
m y Sample Application
Minneapolis (MN, USA) is about 400 km north
where P is the pressure change across a distance of of Des Moines (IA, USA). In (Minneapolis , Des Moines)
either x or y, and is the density of air. the pressure is (101, 100) kPa. Find the pressure-gradi-
ent force per unit mass? Let = 1.1 kgm3.
Z L1B 1L1B
Find the Answer
11 ) Given: P =101 kPa @ x = 400 km (north of Des Moines).

P =100 kPa @ x = 0 km at Des Moines. = 1.1 kgm3.
Find: Fy PG/m = ? ms2

L1B '1( Apply eq. (10.9b):
1
Z Fy PG 1 (101, 000 100, 000)Pa
= 3

E m (1.1kgm ) ( 400, 000 0)m
Y = 2.27x10 3 ms2.
Hint, from Appendix A: 1 Pa = 1 kgm1s2.
-
Y Check: Physics and units are reasonable.
Exposition: The force is from high pressure in the
Figure 10.5 north to low pressure in the south. This direction is
The dark arrow shows the direction of pressure-gradient force
indicated by the negative sign of the answer; namely,
FPG from high (H) to low (L) pressure. This force is perpendicu-
the force points in the negative y direction.
lar to the isobars (solid curved lines).
If pressure increases toward one direction, then the

Sample Application
force is in the opposite direction (from high to low
If the height of the 50 kPa pressure surface decreas-
es by 10 m northward across a distance of 500 km, what P); hence, the negative sign in these terms.
is the pressure-gradient force? Pressure-gradient-force magnitude is
Find the Answer FPG 1 P (10.10)

=
Given: z = 10 m, y = 500 km, |g|= 9.8 ms2 . m d
Find: FPG/m = ? ms2
where d is the distance between isobars.
Use eqs. (10.11a & b): Eqs. (10.9) can be rewritten using the hydrostatic
Fx PG/m = 0 ms2 , because z/x = 0. Thus, FPG/m eq. (1.25) to give the pressure gradient components
= Fy PG/m. as a function of spacing between height contours on
Fy PG z m 10m an isobaric surface:
= g = 9.8 2
m y s 500, 000m
Fx PG z
FPG/m = 0.000196 ms 2 = g
m x (10.11a)
Check: Physics, units & sign are reasonable.
Exposition: For our example here, height decreases Fy PG z
= g
toward the north, thus a hypothetical ball would roll m y (10.11b)
downhill toward the north. A northward force is in
the positive y direction, which explains the positive for a gravitational acceleration magnitude of |g| =
sign of the answer. 9.8 ms2 . z is the height change in the x or y di-
rections; hence, it is the slope of the isobaric surface.
Extending this analogy of slope, if you conceptually
place a ball on the isobaric surface, it will roll down-
Table 10-2. To apply centrifugal force to separate Car- hill (which is the pressure-gradient force direction).
tesian coordinates, a (+/) sign factor s is required. The magnitude of pressure-gradient force is
For winds encircling a
FPG z (10.12)
Hemisphere Low Pressure High Pressure = g
m d
Center Center
Southern 1 +1 where d is distance between height contours.
Northern +1 1 The one force that makes winds blow in the hori-
zontal is pressure-gradient force. All the other forc-
es are a function of wind speed, hence they can only
change the speed or direction of a wind that already
exists. The only force that can start winds blowing
from zero (calm) is pressure-gradient force.
Sample Application
500 km east of a high-pressure center is a north
wind of 5 m s1. Assume N. Centrifugal Force
Hemisphere. What is the Inertia makes an air parcel try to move in a
centrifugal force? straight line. To get its path to turn requires a force in
'$/ a different direction. This force, which pulls toward
Find the Answer
Given: R = 5x105 m,
3
) the inside of the turn, is called centripetal force.
Centripetal force is the result of a net imbalance of
U = 0, V = 5 m s1 7 (i.e., the nonzero vector sum of) other forces.
Find: Fx CN/m = ? ms2.
For mathematical convenience, we can define
Apply eq. (10.13a). In Table 10-2 find s = 1. an apparent force, called centrifugal force, that is
opposite to centripetal force. Namely, it points out-
FxCN ( 5m/s) (5m/s)
= 1 = 5x10 5 ms 2. ward from the center of rotation. Centrifugal-force
m 5 105 components are:
Fx CN VM
Check: Physics and units OK. Agrees with sketch. = +s (10.13a)
Exposition: To maintain a turn around the high-pres- m R
sure center, other forces (the sum of which is the cen- Fy CN UM
tripetal force) are required to pull toward the center. = s (10.13b)
m R
where M = ( U2 + V2 )1/2 is wind speed (always posi- -PX-BUJUVEFT )JHI-BUJUVEFT
/PSUIFSO)FNJTQIFSF
tive), R is radius of curvature, and s is a sign factor
from Table 10-2 as determined by the hemisphere
(North or South) and synoptic pressure center (Low
E
E
E
JO
JO
JO
JO
X
X
or High).
X
'$' '$' '$' '$'
Centrifugal force magnitude is proportional to
wind speed squared:
'$' '$' '$'
4PVUIFSO)FN
'$'

FCN M2 (10.14)
=
E
E
m R
E
JO
JO
JO
JO
X
X
X
Coriolis Force
Figure 10.6
An object such as an air parcel that moves relative
Coriolis force (FCF) vs. latitude, wind-speed, and hemisphere.
to the Earth experiences a compound centrifugal
force based on the combined tangential velocities of
the Earths surface and the object. When combined Sample Application ()
with the non-vertical component of gravity, the re- a) Plot Coriolis parameter vs. latitude.
b) Find FCF/m at Paris, given a north wind of 15 m s1.
sult is called Coriolis force (see the INFO box on the
next page). This force points 90 to the right of the
Find the Answer:
wind direction in the Northern Hemisphere (Fig.
a) Given: = 48.874N at Paris.
10.6), and 90 to the left in the S. Hemisphere.
Find fc (s1) vs. () using eq. (10.16). For example:
The Earth rotates one full revolution (2 radians) fc = (1.458x10 4 s1)sin(48.874) = 1.1x10 4 s1.
during a sidereal day (i.e., relative to the fixed stars,

Psidereal is a bit less than 24 h, see Appendix B), giving
an angular rotation rate of -BUJUVEF

= 2 / Psidereal (10.15)
4 1
= 0.729 211 6 x 10 radians s
The units for are often abbreviated as s1. Using m

this rotation rate, define a Coriolis parameter as:
m
f c = 2 sin() (10.16)
m
m m m
where is latitude, and 2 = 1.458423x10 4 s1. Thus, $PSJPMJT1BSBNFUFS G D m T m
the Coriolis parameter depends only on latitude. Its

magnitude is roughly 1x10 4 s1 at mid-latitudes. b) Given: V = 15 m s1 . Find: FCF/m= ? m s2
The Coriolis force in the Northern Hemisphere Assume U = 0 because no info, thus Fy CF/m= 0.
is: Apply eq. (10.17a):
Fx CF Fx CF/m = (1.1x10 4 s1)(15 m s1) = 1.65x103 m s 2
= fc V (10.17a)
Exposition: This Coriolis force points to the west.
m
Fy CF
= fc U (10.17b)
m INFO Coriolis Force in 3-D
In the Southern Hemisphere the signs on the right Eqs. (10.17) give only the dominant components
side of eqs. (10.17) are opposite. Coriolis force is zero of Coriolis force. There are other smaller-magnitude
under calm conditions, and thus cannot create a Coriolis terms (labeled small below) that are usually
neglected. The full Coriolis force in 3-dimensions is:
wind. However, it can change the direction of an ex-
isting wind. Coriolis force cannot do work, because Fx CF
= fc V 2cos() W (10.17c)
it acts perpendicular to the objects motion. m
[small because often W<<V]
The magnitude of Coriolis force is:
Fy CF
= fc U (10.17d)
| FCF /m | 2 |sin()M| (10.18a) m
or
Fz CF
| FCF /m | | fc M | (10.18b) = 2cos() U (10.17e)
m
[small relative to other vertical forces]
INFO On Coriolis Force INFO On 7

Coriolis Force
Gaspar Gustave Coriolis explained a compound '$/7
(continuation) '()
centrifugal force on a rotating non-spherical planet '$/
such as Earth (Anders Persson: 1998, 2006, 2014). 3
'$/)
'( )
Basics
&BSUI '(7
On the rotating Earth an imbalance can occur be-
tween gravitational force and centrifugal force. G
For an object of mass m moving at tangential &RVBUPS
SPUBUJPO
BYJTPG
speed Mtan along a curved path having radius of cur-
vature R, centrifugal force was shown earlier in
this chapter to be FCN/m = (Mtan)2/R. In Fig 10.a the
object is represented by the black dot, and the center
of rotation is indicated by the X.
Figure 10.c. Horizontal & vertical force components.
.UBO Split the vectors of true gravity into local vertical
FGV and horizontal FGH components. Do the same
'$/N
for the centrifugal force (FCNV , FCNH) of Earths rota-
.UBO
3 tion (Fig. 10.c). Total centrifugal force FCN is parallel
3 to the equator (EQ). Thus, for an object at latitude ,
you can use trig to show FCNH FCNsin().
9
Objects at Rest with respect to Earths Surface

Figure 10.a. Looking down towards the north pole (NP), the
Basics of Earth turns counterclockwise with angular veloc-
centrifugal force (FCN). ity = 360/(sidereal day) (Fig. 10.d). Over a time
interval t, the amount of rotation is t. Any object
The Earth was mostly molten early in its forma- (black dot) at rest on the Earths surface moves with
tion. Although gravity tends to make the Earth the Earth at tangential speed Mtan = R (grey arrow),
spherical, centrifugal force associated with Earths ro- where R = Rocos() is the distance from the axis of
tation caused the Earth to bulge slightly at the equa- rotation. Ro = 6371 km is average Earth radius.
tor. Thus, Earths shape is an ellipsoid (Fig. 10.b). But because the object is at rest, its horizontal
The combination of gravity FG and centrifugal component of centrifugal force FCNH associated with
force FCN causes a net force that we feel as effective movement following the curved latitude (called a par-
gravity FEG. Objects fall in the direction of effective allel) is the same as that for the Earth, as plotted in
gravity, and it is how we define the local vertical (V) Fig. 10.c above. But this horizontal force is balanced
direction. Perpendicular to vertical is the local hori- by the horizontal component of gravity FGH, so the
zontal (H) direction, along the ellipsoidal surface. object feels no net hori-
An object initially at rest on this surface feels no net '$/) zontal force.
PS
horizontal force. [Note: Except at the poles and equa-
BU
V
tor, FG does not point exactly to Earths center, due to .UBO

&R
gravitational pull of the equatorial bulge.] '() 8p3

&BSUI 8pU
/PSUI1PMF 7
9 3
/1
/1
"OZ
'$/

EF
1
SB J
UV
BU
B
) MMFM
DPOTUM
Figure 10.d.
&BSUI XJUI '(
FYBHHFSBUFE '&(
Looking down on the
PCMBUFOFTT
North Pole (NP), for an
&RVBUPS &2
object at rest on Earths surface.

Objects Moving East or West relative to Earth
Suppose an object moves with velocity M due east
relative to the Earth. This velocity (thin white arrow
41
in Fig. 10.e) is relative to Earths velocity, giving the
4PVUI1PMF object a faster total velocity (grey arrow), causing
Figure 10.b. Earth cross section (exaggerated). greater centrifugal force and greater FCNH. But FGH is
(continues in next column) constant. (continues in next column)
INFO On Coriolis Force (continuation) INFO Coriolis Force (continuation)
'$/) Objects moving south have a Coriolis force to

'$'
. '$/) the right due to the larger radius of curvature. Re-
. gardless of the direction of motion in the Northern
'$' Hemisphere, Coriolis force acts 90 to the right of the
'() objects motion relative to the Earth. When viewing
8pU '() the Southern Hemisphere from below the south pole,
9 3 9 3
/1 the Earth rotates clockwise, causing a Coriolis force
&BSUI /1 that is 90 to the left of the relative motion vector.
&BSUI
Coriolis-force Magnitude Derivation
Figure 10.e. Figure 10.f. From Figs. 10.c & d, see that an object at rest (sub-
Eastward moving object. Westward moving object. script R) has
Horizontal force FCNH does NOT balance FGH. FGH = FCNH FCNHR (C1)
The thick white arrow (Fig. 10.e) shows that the force
difference FCF is to the right relative to the objects and Mtan rest = R (C2)
motion M. FCF is called Coriolis force.
The opposite imbalance of FCNH and FGH occurs From Fig. 10.e, Coriolis force for an eastward-mov-
for a westward-moving object (thin white arrow), ing object is defined as
because the object has slower net tangential veloc-
ity (grey arrow in Fig. 10.f). This imbalance, Coriolis FCF FCNH FGH
force FCF, is also to the right of the relative motion M.
Apply eq. (C1) to get
Northward-moving Objects FCF = FCNH FCNHR

When an object moves northward at relative speed or
M (thin white arrow in Fig. 10.g) while the Earth is ro- FCF = sin() [FCN FCNR] (from Fig. 10.c)
tating, the path traveled by the object (thick grey line)
has a small radius of curvature about point X that is Divide by mass m, and plug in the definition for cen-
displaced from the North Pole. The smaller radius trifugal force as velocity squared divided by radius:
R causes larger centrifugal force FCNH pointing out-
ward from X. FCF / m = sin() [ (Mtan)2/R (Mtan rest)2/R ]
Component FCNH-ns of centrifugal force balances
the unchanged horizontal gravitational force FGH. Use Mtan = Mtan rest + M, along with eq. (C2):
But there remains an unbalanced east-west compo-
nent of centrifugal force FCNH-ew which is defined as FCF / m = sin() [ (R+M)2/R (R)2/R ]
Coriolis force FCF. Again, it is to the right of the rela-
tive motion vector M of the object. FCF / m = sin() [(2M) + (M2/R)]
'$/) The first term is usually much larger than the last,
'$/)OT allowing the following approximation for Coriolis
force per mass:
FCF /m 2sin() M (10.18)
'$/)FX &BSUI
. Define a Coriolis parameter as fc 2sin() . Thus,
'$'
FCF /m fc M
MBU
JUVE
3
FMJOF QBSBMM
/1 9 HIGHER MATH Apparent Forces

$FOUFSPG
'() 3PUBUJPO In vector form, centrifugal force/mass for an object at
rest on Earth is ( r), and Coriolis force/mass
FM
is 2 V , where vector points along the Earths

Figure 10.g. Northward moving object. axis toward the north pole, r points from the Earths
center to the object, V is the objects velocity relative
(continues in next column)
to Earth, and is the vector cross product.
[ Turbulent-Drag Force
Surface elements such as pebbles, blades of grass,
GSFF
BUNPTQIFSF crops, trees, and buildings partially block the wind,
XJOE and disturb the air that flows around them. The
[J combined effect of these elements over an area of
BUNPTQIFSJD ground is to cause resistance to air flow, thereby
'5%
CPVOEBSZ slowing the wind. This resistance is called drag.
DF
UIFSNBM
UVSCVMFO
MBZFS "#-
At the bottom of the troposphere is a layer of air
roughly 0.3 to 3 km thick called the atmospheric
boundary layer (ABL). The ABL is named because
it is at the bottom boundary of the atmosphere. Tur-
. ( Y
bulence in the ABL mixes the very-slow near-surface
FEEJFTDBVTFECZ air with the faster air in the ABL, reducing the wind
TIFBSUVSCVMFODF
speed M throughout the entire ABL (Fig. 10.7).
Figure 10.7 The net result is a drag force that is normally only
Wind speed M (curved black line with white highlights) is slow- felt by air in the ABL. For ABL depth zi the drag is:
er than geostrophic G (vertical dashed line) because of turbulent
drag force FTD in the atmospheric boundary layer. Fx TD
U (10.19a)
= wT
m zi
Fy TD V (10.19b)
= wT
m zi
where wT is called a turbulent transport velocity.

The total magnitude of turbulent drag force is
Sample Application FTD M

What is the drag force per unit mass opposing a U = wT (10.20)
m zi
= 15 m s1 wind (with V = 0) for a: (a) statically neutral
ABL over a rough forest; & (b) statically unstable ABL
having convection with wB = 50 m s1, given zi = 1.5
and is always opposite to the wind direction.
km. For statically unstable ABLs with light winds,
where a warm underlying surface causes thermals
Find the Answer of warm buoyant air to rise (Fig. 10.7), this convec-
Given: U = M = 15 m s1, zi = 1500 m, tive turbulence transports drag information upward
CD = 2x10 2, wB = 50 m s1. at rate:
Find: Fx TD/m = ? ms2. wT = bD wB (10.22)
(a) Plugging eq. (10.21) into eq. (10.19a) gives: where dimensionless factor bD = 1.83x10 3. The
FxTD U (15m/s)2 buoyancy velocity scale, wB, is of order 10 to 50 m
= CD M = (0.02)
m zi 1500m s1, as is explained in the Heat Budget chapter.
For statically neutral conditions where strong
= 3x10 3 ms2.
winds M and wind shears (changes of wind di-
(b) Plugging eq. (10.22) into eq. (10.19a) gives: rection and/or speed with height) create eddies and
Fx TD mechanical turbulence near the ground (Fig. 10.7),
U
= bD wB the transport velocity is
m zi
(15m/s)
= (0.00183)(50m/s) wT = CD M (10.21)
1500m
= 9.15x10 4 ms2. where the drag coefficient CD is small (2x10 3
dimensionless) over smooth surfaces and is larger
Check: Physics and units are reasonable. (2x10 2) over rougher surfaces such as forests.
Exposition: Because the wind is positive (blowing In fair weather, turbulent-drag force is felt only
toward the east) it requires that the drag be negative in the ABL. However, thunderstorm turbulence
(pushing toward the west). Shear (mechanical) tur-
can mix slow near-surface air throughout the tro-
bulence and convective (thermal/buoyant) turbulence
can both cause drag by diluting the faster winds high-
posphere. Fast winds over mountains can create
er in the ABL with slower near-surface winds. mountain-wave drag felt in the whole atmosphere
(see the Regional Winds chapter).
Summary of Forces
Table 10-3. Summary of forces.

Name of Magnitude Horiz. (H) Remarks (item is in col-
Item Direction
Force (N kg1) or Vert. (V) umn 1; H & V in col. 5)
FG hydrostatic equilibrium
1 gravity down = g = 9.8 ms2 V
m when items 1 & 2V balance
pressure from high to low FPG z the only force that can drive
2 = g V&H
gradient pressure m d horizontal winds
90 to right (left)
FCF geostrophic wind when
Coriolis of wind in North-
= 2 sin() M
3 H* 2H and 3 balance (explained
(compound) ern (Southern) m later in horiz. wind section)
Hemisphere
FTD M atm. boundary-layer wind

turbulent = wT
4 opposite to wind H* when 2H, 3 and 4 balance (ex-
drag m zi
plained in horiz. wind section)
centripetal = opposite of
centrifugal away from center FCN M2
5 = H* centrifugal. Gradient wind
(apparent) of curvature m R when 2H, 3 and 5 balance
advection FAD U neither creates nor destroys

6 (any) = M V&H
(apparent) m d momentum; just moves it
*Horizontal is the direction we will focus on. However, Coriolis force has a small vertical component for zonal winds. Turbulent drag
can exist in the vertical for rising or sinking air, but has completely different form than the boundary-layer drag given above. Cen-
trifugal force can exist in the vertical for vortices with horizontal axes. Note: units N kg1 = ms 2.
Equations of Horizontal Motion
Combining the forces from eqs. (10.7, 10.8, 10.9, The terms on the right side of eqs. (10.23) can all
10.17, and 10.19) into Newtons Second Law of Mo- be of order 1x10 4 to 10x10 4 ms2 (which is equiva-
tion (eq. 10.5) gives simplified equations of horizon- lent to units of N kg1, see Appendix A for review).
tal motion: However, some of the terms can be neglected under
(10.23a) special conditions where the flow is less compli-
cated. For example, near-zero Coriolis force occurs
U U U U 1 P U
= U V W + fc V wT near the equator. Near-zero turbulent drag exists
t x y z x zi above the ABL. Near-zero pressure gradient is at
low- and high-pressure centers.
(10.23b) Other situations are more complicated, for which
additional terms should be added to the equations of
V V V V 1 P V horizontal motion. Within a few mm of the ground,
= U V W fc U wT
t x y z y zi molecular friction is large. Above mountains
during windy conditions, mountain-wave drag is
}
}
}
}
}
large. Above the ABL, cumulus clouds and thunder-

pressure turbulent
tendency advection
gradient
Coriolis
drag storms can create strong convective mixing.
For a few idealized situations where many terms
These are the forecast equations for wind. in the equations of motion are small, it is possible
For special conditions where steady winds to solve those equations for the horizontal wind
around a circle are anticipated, centrifugal force can speeds. These theoretical winds are presented in
be included. the next section. Later in this chapter, equations to
forecast vertical motion (W) will be presented.
Table 10-4. Names of idealized steady-state horizon-

tal winds, and the forces that govern them. Horizontal Winds
U 1 P U VM
0 = U + fc V wT + s When air accelerates to create wind, forces that
x x zi R
are a function of wind speed also change. As the
pressure turbulent centri-
Forces: gradient
Coriolis
drag fugal winds continue to accelerate under the combined ac-
tion of all the changing forces, feedbacks often oc-
Wind Name cur to eventually reach a final wind where the forces
Geostrophic balance. With a zero net force, there is zero accelera-
Gradient tion.
Atm.Bound. Layer Such a final, equilibrium, state is called steady
ABL Gradient state:
Cyclostrophic
U V
Inertial = 0, = 0 (10.24)
t t
Antitriptic
Caution: Steady state means no further change to

1L1B - the non-zero winds. Do not assume the winds are
zero.
'1( Under certain idealized conditions, some of the
forces in the equations of motion are small enough
L1B to be neglected. For these situations, theoretical
steady-state winds can be found based on only the
( remaining larger-magnitude forces. These theoreti-
cal winds are given special names, as listed in Table
L1B 10-4. These winds are examined next in more detail.
As we discuss each theoretical wind, we will learn
where we can expect these in the real atmosphere.
'$' Z
L1B
Geostrophic Wind
) Y For special conditions where the only forces are
Coriolis and pressure-gradient (Fig. 10.8), the result-
Figure 10.8 ing steady-state wind is called the geostrophic wind,
Idealized weather map for the Northern Hemisphere, showing
with components (Ug , Vg). For this special case, the
geostrophic wind (G, grey arrow) caused by a balance between
only terms remaining in, eqs. (10.23) are:
two forces (black arrows): pressure-gradient force (FPG) and
Coriolis force (FCF). P is pressure, with isobars plotted as thin 1 P
black lines. L and H are low and high-pressure regions. The 0 = + f c V (10.25a)
x
small sphere represents an air parcel.
1 P
0 = f c U (10.25b)
y
Sample Application
Find geostrophic wind components at a location
where = 1.2 kg m3 and fc = 1.1x10 4 s1. Pressure Define U Ug and V Vg in the equations above,
decreases by 2 kPa for each 800 km of distance north. and then solve for these wind components:
Find the Answer 1 P

Ug =
Given: fc =1.1x104 s1, P= 2 kPa, =1.2 kg m3, y=800 km. f c y (10.26a)
Find: (Ug , Vg) = ? m s1
But P/x = 0 implies Vg = 0. For Ug, use eq. (10.26a): 1 P (10.26b)

Vg = +
1 ( 2 kPa) f c x
Ug = =18.9 m s 1
(1.2 kg/m 3 )(1.1 10 4 s 1 ) (800km)
where fc = (1.4584x10 4 s1)sin(latitude) is the Coriolis
Check: Physics & units OK. Agrees with Fig. 10.10.
parameter, is air density, and P/x and P/y are
Exposition: As the pressure gradient accelerates air
the horizontal pressure gradients.
northward, Coriolis force turns it toward the east.
Z [E NLN

E
E

XJ XJ XJ XJ XJ XJ
OE OE OE OE OE OE

(
NT

MBUJUVEF
1L1B
Y
Figure 10.9
Isobars (black lines) that are more closely spaced (i.e., tightly
packed) cause stronger geostrophic winds (arrows), for N.
Hemisphere.
Real winds are nearly geostrophic at locations

where isobars or height contours are relatively
straight, for altitudes above the atmospheric bound-
ary layer. Geostrophic winds are fast where isobars 1EBUTFBMFWFM L1BLN
are packed closer together. The geostrophic wind

Figure 10.10
direction is parallel to the height contours or iso-
Variation of geostrophic wind speed (G) with horizontal pres-
bars. In the (N., S. ) hemisphere the wind direction sure gradient (P/d) at sea level. Top scale is height gradient
is such that low pressure is to the winds (left, right), of any isobaric surface.
see Fig. 10.9.
The magnitude G of the geostrophic wind is:
G = U g 2 + Vg 2 (10.27) Sample Application

Find the geostrophic wind for a height increase of
50 m per 200 km of distance toward the east. Assume,
If d is the distance between two isobars (in the
fc = 0.9x10 4 s1 .
direction of greatest pressure change; namely, per-
pendicular to the isobars), then the magnitude (Fig. Find the Answer
10.10) of the geostrophic wind is: Given: x = 200 km, z = 50 m, fc = 0.9x10 4 s1 .
Find: G = ? m s1
1 P (10.28)
G=
f c d No north-south height gradient, thus Ug = 0.
Apply eq. (10.29b) and set G = Vg :
Above sea level, weather maps are often on iso- g z 9.8m s -2 50m
baric surfaces (constant pressure charts), from which Vg = + = =27.2 m s 1
fc x 0.00009s -1 200, 000m
the geostrophic wind (Fig. 10.10) can be found from
the height gradient (change of height of the isobaric
surface with horizontal distance): Check: Physics & units OK. Agrees with Fig. 10.10.
Exposition: If height increases towards the east, then
g z (10.29a)
you can imagine that a ball placed on such a surface
Ug = would roll downhill toward the west, but would turn
f c y
to its right (toward the north) due to Coriolis force.
g z (10.29b)
Vg = +
f c x
INFO Approach to Geostrophy
where the Coriolis parameter is fc , and gravitational
acceleration is |g| = 9.8 ms2. The corresponding How does an air parcel, starting from rest, ap-
magnitude of geostrophic wind on an isobaric chart proach the final steady-state geostrophic wind speed
is: G sketched in Fig. 10.8?
Start with the equations of horizontal motion (10.23),
g z and ignore all terms except the tendency, pressure-
G= (10.29c) gradient force, and Coriolis force. Use the definition of
fc d
continues on next page
If the geopotential = |g|z is substituted in

INFO Appr. to Geostrophy (continuation) eqs. (10.29), the resulting geostrophic winds are:
geostrophic wind (eqs. 10.26) to write the resulting
simplified equations as: 1
Ug =
U / t = fc (Vg V ) fc y (10.30a)
V / t = fc (U g U )
1
Vg = (10.30b)
Next, rewrite these as forecast equations: fc x
U new = U old t fc (Vg Vold )
Vnew = Vold + t fc (U g U new ) Gradient Wind
If there is no turbulent drag, then winds tend to
Start with initial conditions (Uold, Vold) = (0, 0), and
then iteratively solve the equations on a spreadsheet
blow parallel to isobar lines or height-contour lines
to forecast the wind. even if those lines are curved. However, if the lines
For example, suppose P = 1 kPa, fc = 10 4 s1, x = curve around a low-pressure center (in either hemi-
500 km, = 1 kg m3, where we would anticipate the sphere), then the wind speeds are subgeostrophic
wind should approach (Ug, Vg) = (0, 20) m s1. The ac- (i.e., slower than the theoretical geostrophic wind
tual evolution of winds (U, V) and air parcel position speed). For lines curving around high-pressure
(X, Y) are shown in Figs. below. centers, wind speeds are supergeostrophic (faster
than theoretical geostrophic winds). These theoreti-
I
cal winds following curved isobars or height con-
I tours are known as gradient winds.
7 NT
I Gradient winds differ from geostrophic winds be-
cause Coriolis force FCF and pressure-gradient force
6H 7H

FPG do not balance, resulting in a non-zero net force
NT
Fnet. This net force is called centripetal force, and is

I what causes the wind to continually change direc-
I tion as it goes around a circle (Figs. 10.11 & 10.12). By
U
describing this change in direction as causing an ap-
I
parent force (centrifugal), we can find the equations
m that define a steady-state gradient wind:
6 NT
Surprisingly, the winds never reach geostrophic 1 P VM

0 = x + fc V + s R (10.31a)
equilibrium, but instead rotate around the geostrophic
wind. This is called an inertial oscillation, with
period of 2/fc. For our case, the period is 17.45 h.
Twice this period is called a pendulum day. 1 P UM
The net result in the figure below is that the 0 = f c U s (10.31b)
y R
wind indeed moves at the
geostrophic speed of 20 m I
}
}
}
s1 to the north ( 1250 km in

pressure
17.45 h), but along the way it I gradient
Coriolis centrifugal
staggers west and east with
an additional ageostrophic I Because the gradient wind is for flow around a
(non-geostrophic) part.
circle, we can frame the governing equations in ra-
I
Inertial oscillations are I

: LN
I dial coordinates, such as for flow around a low:
sometimes observed at night
I
in the atmospheric boundary
layer, but rarely higher in the 1 P M2
(NT
atmosphere. Why not? (1) IU = fc Mtan + tan (10.32)

R R
The ageostrophic component
of wind (wind from the East
I
in this example) moves air where R is radial distance from the center of the
mass, and changes the pres- I
I circle, fc is the Coriolis parameter, is air density,
sure gradient. (2) Friction P/R is the radial pressure gradient, and Mtan is
damps the oscillation toward m
the magnitude of the tangential velocity; namely,
a steady wind. 9 LN
the gradient wind.
-
L1B

L1B
)
B

L 1
1

1 L1B

B
L1

'$'

'1(
L
1 B
BU
M
DF
Q
BS I PG
Q BJ B

G BJ S SQ L1
P BSDF
L1 'OFU QBUI .UBO M 'OFU
B (
.UBO (
B
L '$' L1
1B '1(
Z
Z
Y ) Y -
Figure 10.11 Figure 10.12
Forces (dark arrows) that cause the gradient wind (solid grey Forces (dark arrows) that cause the gradient wind (solid grey
arrow, Mtan) to be slower than geostrophic (hollow grey arrow) arrow, Mtan) to be faster than geostrophic (hollow grey arrow)
when circling around a low-pressure center (called a cyclone in for an air parcel (grey sphere) circling around a high-pressure
the N. Hem.). The short white arrow with black outline shows center (called an anticyclone in the N. Hemisphere).
centripetal force (the imbalance between the other two forces).
Centripetal force pulls the air parcel (grey sphere) inward to
L1B
force the wind direction to change as needed for the wind to turn
Z
along a circular path.
.UBO
(
By re-arranging eq. (10.32) and plugging in the
definition for geostrophic wind speed G, you can get
an implicit solution for the gradient wind Mtan:
(
M 2 (10.33) .UBO
Mtan = G tan
fc R
(
In this equation, use the + sign for flow around high-
pressure centers, and the sign for flow around lows .UBO 1L1B
(Fig. 10.13).
Y
Figure 10.13
Comparison of gradient winds Mtan vs. geostrophic wind G for
flows around low (L) and high (H) pressures. N. Hemisphere.
Sample Application
What radius of curvature causes the gradient wind to equal the geostrophic wind?
Find the Answer

Given: Mtan = G Find: R = ? km
Use eq. (10.33), with Mtan = G: G = G G2/( fcR)

This is a valid equality G = G only when the last term in eq. (10.33) approaches zero; i.e., in the limit of R = .
Check: Eq. (10.33) still balances in this limit. Exposition: Infinite radius of curvature is a straight line, which (in
the absence of any other forces such as turbulent drag) is the condition for geostrophic wind.
Eq. (10.33) is a quadratic equation that has two so-

Sample Application
lutions. One solution is for the gradient wind Mtan
If the geostrophic wind around a low is 10 m s1,
then what is the gradient wind speed, given fc = 10 4 s1 around a cyclone (i.e., a low):
and a radius of curvature of 500 km? Also, what is the
curvature Rossby number? 4G
Mtan = 0.5 fc R 1 + 1 +
fc R (10.34a)

Find the Answer
Given: G = 10 m s1, R = 500 km, fc = 10 4 s1 The other solution is for flow around an anticy-
Find: Mtan = ? m s1, Roc = ? (dimensionless) clone (i.e., a high):
Use eq. (10.34a) 4G
Mtan = 0.5 fc R 1 1 (10.34b)
4 1 fc R
Mtan = 0.5 (10 s )(500000m)
4 (10m / s)
1 + 1 + To simplify the notation in the equations above,
4 1
(10 s )(500000m) let
G
= 8.54 m s 1 Roc = (10.35)
fc R
Use eq. (10.35):
(10m/s)
where we can identify (Roc) as a curvature Rossby
Roc = = 0.2 number because its length scale is the radius of cur-
(104 s 1 )(5 105 m)
vature (R). When Roc is small, the winds are roughly
Check: Physics & units are reasonable. geostrophic; namely, pressure gradient force nearly
Exposition: The small Rossby number indicates that balances Coriolis force. [CAUTION: In later chap-
the flow is in geostrophic balance. The gradient wind ters you will learn about a Rossby radius of defor-
is indeed slower than geostrophic around this low. mation, which is distinct from both Roc and R.]
For winds blowing around a low, the gradient
wind is:
G
Mtan = 2 Ro 1 + ( 1 + 4 Roc ) (10.36a)
1/2
c
and for winds around a high) the gradient wind is:
G
1 ( 1 4 Roc ) (10.36b)
1/2
Mtan =
2 Roc

1 ) where G is the geostrophic wind.

While the differences between solutions (10.36a
L1B
& b) appear subtle at first glance, these differences

have a significant impact on the range of winds that
are physically possible. Any value of Roc can yield
physically reasonable winds around a low-pressure
center (eq. 10.36a). But to maintain a positive argu-

- ment inside the square root of eq. (10.36b), only val-
ues of Roc 1/4 are allowed for a high.
3 LN
Thus, strong radial pressure gradients with small
radii of curvature, and strong tangential winds can
Figure 10.14 exist near low center. But only weak pressure gra-
Illustration of how mean sea-level pressure P can vary with dis- dients with large radii of curvature and light winds
tance R from a high-pressure (H) center. The anticyclone (i.e., are possible near high-pressure centers (Figs. 10.14
the high) has zero horizontal pressure gradient and calm winds and 10.15). To find the maximum allowable hori-
in its center, with weak pressure gradient (P/R) and gentle
zontal variations of height z or pressure P near an-
winds in a broad region around it. The cyclone (i.e., the low) can
ticyclones, use Roc = 1/4 in eq. (10.35) with G from
have steep pressure gradients and associated strong winds close
to the low center (L), with a pressure cusp right at the low cen- (10.29c) or (10.28):
ter. In reality (dotted line), turbulent mixing near the low center
smooths the cusp, allowing a small region of light winds at the ( )
z = zc fc 2 R 2 / ( 8 g ) (10.37a)
low center surrounded by stronger winds. Although this graph or
was constructed using eq. (10.37b), it approximates the pressure
variation along the cross section shown in the next figure.
c c (
P = P f 2 R 2 / 8
) (10.37b)
'BTU (
i4MPXFS
BSPVOE-PXTu
Canada .UBO
5VSCVMFOU%SBH
Europe ."#-(
8JOE
Atlantic 4QFFE
Ocean .UBO
5VSCVMFOU%SBH
i'BTUFS
."#-( BSPVOE)JHITu
Figure 10.15 4MPX (
Illustration of strong pressure gradients (closely-spaced isobars)
around the low-pressure center (L) over eastern Canada, and -PX )JHI
weak pressure gradients (isobars spaced further apart) around 1SFTTVSF$FOUFS
the high (H) over the NE Atlantic Ocean. NCEP reanalysis
of daily-average mean sea-level pressure (Pa) for 5 Feb 2013. Figure 10.16
Pressures in the low & high centers were 96.11 & 104.05 kPa. Relative magnitudes of different wind speeds around low- and
Pressure variation along the dotted line is similar to that plotted high-pressure centers. G = geostrophic wind, Mtan = gradient
in the previous figure. [Courtesy of the NOAA/NCEP Earth wind speed, M ABLG = atmospheric-boundary-layer gradient
Systems Research Laboratory. http://www.esrl.noaa.gov/psd/ wind speed. G is smaller in highs than in lows, because it is not
data/gridded/data.ncep.reanalysis.html ] physically possible to have strong pressure gradients to drive
strong steady-state winds at high centers.
where the center pressure in the high (anticyclone) is
Pc , or for an isobaric surface the center height is zc,
the Coriolis parameter is fc , |g| is gravitational ac-
celeration magnitude, is air density, and the radius
from the center of the high is R (see Fig. 10.14).
Figs. 10.14 and 10.15 show that pressure gradi-
ents, and thus the geostrophic wind, can be large
near low centers. However, pressure gradients, and
thus the geostrophic wind, must be small near high
centers. This difference in geostrophic wind speed
G between lows and highs is sketched in Fig. 10.16.
The slowdown of gradient wind Mtan (relative to
geostrophic) around lows, and the speedup of gra-
1L1B
-
dient wind (relative to geostrophic) around highs
is also plotted in Fig. 10.16. The net result is that
'1(
gradient winds, and even atmospheric boundary-
."#-
layer gradient winds M ABLG (described later in this
L1B
chapter), are usually stronger (in an absolute sense)
around lows than highs. For this reason, low-pres- B (
sure centers are often windy. '5%
L1B
Atmospheric-Boundary-Layer Wind
If you add turbulent drag to winds that would have '$' Z
been geostrophic, the result is a subgeostrophic
(slower-than-geostrophic) wind that crosses the iso- L1B
bars at angle () (Fig. 10.17). This condition is found
in the atmospheric boundary layer (ABL) where the ) Y
isobars are straight. The force balance at steady state

is: 1 P U Figure 10.17
0 = + f c V wT (10.38a) Balance of forces (black arrows) creating an atmospheric-bound-
x zi ary-layer wind (M ABL, solid grey arrow) that is slower than
geostrophic (G, hollow grey arrow). The grey sphere represents
1 P V an air parcel. Thin black lines are isobars. L and H are low and
0 = f c U wT (10.38b)
y zi high-pressure centers.
Namely, the only forces acting for this special case

Sample Application
are pressure gradient, Coriolis, and turbulent drag
For statically neutral conditions, find the winds in
the boundary layer given: zi = 1.5 km, Ug = 15 m s1, (Fig. 10.17).
Vg = 0, fc = 10 4 s1, and CD = 0.003. What is the cross- Replace U with UABL and V with VABL to indi-
isobar wind angle? cate these winds are in the ABL. Eqs. (10.38) can be
rearranged to solve for the ABL winds, but this solu-
Find the Answer tion is implicit (depends on itself):
Given: zi = 1.5 km, Ug = 15 m s1, Vg = 0, fc = 10 4 s1,
wT VABL
CD = 0.003. U ABL = U g (10.39a)
Find: VABL =? m s1, UABL =? m s1, M ABL =? m s1, fc zi
= ?
First: G =(Ug2 + Vg2)1/2 = 15 m s1. Now apply eq.(10.41) wT U ABL
VABL = Vg + (10.39b)
0.003
fc zi
a = = 0.02 s/m
(10 4 s 1 )(1500m) where (Ug, Vg) are geostrophic wind components, fc

Check: aG = (0.02 s m1)(15 m s1) =0.3 (is < 1. Good.) is Coriolis parameter, zi is ABL depth, and wT is the
UABL=[10.35(0.02s m1)(15m s1)](15m s1)13.4m s 1 turbulent transport velocity.
VABL=[10.5(0.02s m1)(15m s1)] You can iterate to solve eqs. (10.39). Namely, first
(0.02s m1)(15m s1)(15m s1) 3.8 m s 1 you guess a value for VABL to use in the right side
of the first eq. Solve eq. (10.39a) for UABL and use it
M ABL = U ABL 2 + VABL 2 = 13.42 + 3.82 = 13.9 m s 1
in the right side of eq. (10.39b), which you can solve
Isobars are parallel to the geostrophic wind. Thus, the for VABL. Plug this back into the right side of eq.
cross-isobar angle is: (10.39a) and repeat this procedure until the solution
=tan1(VABL/UABL) = tan1(3.8/13.4) = 15.8 . converges (stops changing very much). The mag-
nitude of the boundary-layer wind is:
Check: Physics & units are reasonable.
Exposition: Drag both slows the wind (13.4 m s1) M ABL = [U2ABL + V2ABL ]1/2 (10.40)
in the boundary layer below its geostrophic value (15
m s1) and turns it at a small angle (15.8) towards
low pressure. Given N. Hem. (because of the positive For a statically neutral ABL under windy condi-
Coriolis parameter), the ABL wind direction is 254.2. tions, then wT = CDM ABL, where CD is the drag coef-
ficient (eq. 10.21). For most altitudes in the neutral
ABL, an approximate but explicit solution is:
(10.41a)
Sample Application U ABL (1 0.35 a U g )U g (1 0.5 a Vg ) a Vg G
For statically unstable conditions, find winds in the
ABL given Vg = 0, Ug = 5 m s1, wB = 50 m s1, zi = 1.5 (10.41b)
km, bD = 1.83x10 3, and fc = 10 4 s1. What is the cross- VABL (1 0.5 a U g ) a G U g + (1 0.35 a Vg ) Vg
isobar wind angle?
where the parameter is a = CD/(fczi), G is the
Find the Answer geostrophic wind speed and a solution is possible
Given: (use convective boundary layer values above)
only if aG < 1. If this condition is not met, or if no
Find: M ABL =? m s1, VABL =? m s1, UABL =? m s1,
=?
reasonable solution can be found using eqs. (10.41),
then use the iterative approach described in the next
Apply eqs. (10.42): section, but with the centrifugal terms set to zero.
Eqs. (10.41) do not apply to the surface layer (bot-
(1.83 10 3 )(50m/s)
c1 = = 0.61 (dimensionless) tom 5 to 10% of the neutral boundary layer).
(10 4 s 1 )(1500m) If the ABL is statically unstable (e.g., sunny with
c2 = 1/[1+(0.61)2] = 0.729 (dimensionless) slow winds), use wT = bDwB (see eq. 10.22). Above
UABL = 0.729[(5m s1) 0 ] = 3.6 m s 1 the surface layer there is an exact solution that is ex-
VABL = 0.729[0 + (0.61)(5m s1)] = 2.2 m s 1 plicit:
Use eq. (10.40): U ABL = c2 [U g c1 Vg ] (10.42a)
M ABL = [U2ABL + V2ABL ]1/2 = 4.2 m s 1
=tan1(VABL/UABL) = tan1(2.2/3.6) = 31.4
VABL = c2 [Vg + c1 U g ] (10.42b)
Exposition: Again, drag slows the wind and causes bD wB 1
where c1 = , and c2 = .
it to cross the isobars toward low pressure. The ABL f c zi [1 + c12 ]
wind direction is 238.6.
The factors in c1 are given in the Forces section.
In summary, both wind-shear turbulence and

convective turbulence cause drag. Drag makes the
ABL wind slower than geostrophic (subgeostrophic),
and causes the wind to cross isobars at angle such
that it has a component point to low pressure.
ABL Gradient (ABLG) Wind -

L1B
For curved isobars in the atmospheric boundary
JSQBSDFM

1 L1B

layer (ABL), there is an imbalance of the following
forces: Coriolis, pressure-gradient, and drag. This
IPGB

imbalance is a centripetal force that makes ABL air

spiral outward from highs and inward toward lows '1(
L
QBU
1B
(Fig. 10.18). An example was shown in Fig. 10.1. ."#-(
If we devise a centrifugal force equal in mag-

nitude but opposite in direction to the centripetal 'OFU .UBO
L1 B
force, then the equations of motion can be written B (
for spiraling flow that is steady over any point on '5%
the Earths surface (i.e., NOT following the parcel):
'$'
L
1B
1 P U VM Z
0 = + f c V wT + s (10.43a)
x zi R
1 P V UM
Y
)
0 = f c U wT s (10.43b)
y zi R Figure 10.18
Imbalance of forces (black arrows) yield a net centripetal force
(Fnet) that causes the atmospheric-boundary-layer gradient
}
}
}
}
pressure turbulent wind (M ABLG, solid grey arrow) to be slower than both the
Coriolis centrifugal
gradient drag gradient wind (Mtan) and geostrophic wind (G). The resulting
air-parcel path crosses the isobars at a small angle toward low
pressure.
We can anticipate that the ABLG winds should
be slower than the corresponding gradient winds,
and should cross isobars toward lower pressure at
some small angle (see Fig. 10.19). Z
Lows are often overcast and windy, implying that
JTP
the atmospheric boundary layer is statically neutral. CB
S Y
For this situation, the transport velocity is given by:
T
wT = CD M = CD U 2 + V 2 (10.21 again)
Because this parameterization is nonlinear, it in-

creases the nonlinearity (and the difficulty to solve),
eqs. (10.43).
- 3
Highs often have mostly clear skies with light

winds, implying that the atmospheric boundary
-(
layer is statically unstable during sunny days, and . "#
statically stable at night. For daytime, the transport 7
velocity is given by:
6
wT = bD wB (10.22 again)
This parameterization for wB is simple, and does not

depend on wind speed. For statically stable condi- Figure 10.19
Tangential ABLG wind component (U) and radial ABLG wind
tions during fair-weather nighttime, steady state is
component (V) for the one vector highlighted as the thick black
unlikely, meaning that eqs. (10.43) do not apply. arrow. N. Hemisphere.
Nonlinear coupled equations (10.43) are difficult

Sample Application
to solve analytically. However, we can rewrite the
If G = 10 m s1 at R = 400 km from the center of a
N. Hem. cyclone, CD = 0.02, zi = 1 km, and fc = 10 4 s1, equations in a way that allows us to iterate numeri-
then find the ABLG wind speed and components. cally toward the answer (see the INFO box below for
instructions). The trick is to not assume steady state.
Find the Answer Namely, put the tendency terms (U/t , V/t) back
Given: (see the data above) in the left hand sides (LHS) of eqs. (10.43). But recall
Find: MBLG = ? m s1, UBLG = ? m s1, VBLG = ? m s1, that U/t = [U(t+t) U]/t, and similar for V.
For this iterative approach, first re-frame eqs.
Use a spreadsheet to iterate (as discussed in the INFO (10.43) in cylindrical coordinates, where (U, V) are
box) eqs. (10.44) & (1.1) with a time step of t = 1200 s. the (tangential, radial) components, respectively (see
Use U = V = 0 as a first guess. Fig. 10.19). Also, use G, the geostrophic wind defini-
tion of eq. (10.28), to quantify the pressure gradient.
G (m/s)= 10 zi (km)= 1
For a cyclone in the Northern Hemisphere (for
R (km)= 400 fc (s1)= 0.0001
CD = 0.02 t(s)= 1200
which s = +1 from Table 10-2), the atmospheric
boundary layer gradient wind eqs. (10.43) become:
Iteration U ABLG VABLG M ABLG U ABLG VABLG
Counter (m s1) (m s1) (m s1) (m s1) (m s1) M = ( U2 + V2 )1/2 (1.1 again)
0 0.00 0.00 0.00 0.000 1.200
1 0.00 1.20 1.20 0.148 1.165 (10.44a)
2 0.15 2.37 2.37 0.292 1.047 CD M U VM
3 0.44 3.41 3.44 0.408 0.861 U (t + t) = U + t fc V +s
zi R
4 0.85 4.27 4.36 0.480 0.640
5 1.33 4.91 5.09 0.502 0.420
(10.44b)
...
29 4.16 4.33 6.01 -0.003 0.001 C MV U M
30 4.16 4.33 6.01 -0.002 0.001 V (t + t) = V + t fc (G U ) D s
zi R
The evolution of the
iterative solution where (U, V) represent (tangential, radial) parts for

is plotted at right the wind vector south of the low center. These cou-
as it approaches pled equations are valid both night and day.
7 NTm
the final answer of

UABLG = 4.16 m s 1, (
#-
VABLG = 4.33 m s 1, ."
MABLG= 6.01 m s 1,
where (UABLG , INFO Find the Answer by Iteration

VABLG) are (tangen-
Equations (10.44) are difficult to solve analytically,
tial, radial) parts. 6 NTm
but you can iterate as an alternative way to solve for
the ABLG wind components. Here is the procedure.
Check: Physics & units are reasonable. You should
do the following what if experiments on the spread- (1) Make an initial guess for (U, V), such as (0, 0).
sheet to check the validity. I ran experiments using a (2) Use these (U, V) values in the right sides of
modified spreadsheet that relaxed the results using a eqs. (10.44) and (1.1), and solve for the new
weighted average of new and previous winds. values of [U(t+t), V(t+t)].
(a) As R approaches infinity and CD approaches (3) In preparation for the next iteration, let
zero, then M ABLG should approach the geostrophic U = U(t+t) , and V = V(t+t) .
wind G. I got UABLG = G = 10 m s1, VABLG = 0. (4) Repeat steps 2 and 3 using the new (U, V)
(b) For finite R and 0 drag, then M ABLG should on the right hand sides .
equal the gradient wind Mtan. I got UABLG = 8.28 (5) Keep iterating. Eventually, the [U(t+t), V(t+t)]
m s1, VABLG = 0. values stop changing (i.e., reach steady state),
(c) For finite drag but infinite R, then M ABLG should giving UABLG = U(t+t), and VABLG = V(t+t).
equal the atmospheric boundary layer wind M ABL. I
got UABLG = 3.91 m s1, VABLG = 4.87 m s1. Because Because of the repeated, tedious calculations, I recom-
this ABL solution is based on the full equations, it mend you use a spreadsheet (see the Sample Applica-
gives a better answer than eqs. (10.41). tion) or write your own computer program.
Exposition: If you take slightly larger time steps, the As you can see from the Sample Application, the
solution converges faster. But if t is too large, the it- solution spirals toward the final answer as a damped
eration method fails (i.e., blows up). inertial oscillation.
For daytime fair weather conditions in anticy-

clones, you could derive alternatives to eqs. (10.44)
that use convective parameterizations for atmo-
spheric boundary layer drag.
Because eqs. (10.44) include the tendency terms,
you can also use them for non-steady-state (time
varying) flow. One such case is nighttime during
fair weather (anticyclonic) conditions. Near sunset,
when vigorous convective turbulence dies, the drag
coefficient suddenly decreases, allowing the wind
to accelerate toward its geostrophic equilibrium
value. However, Coriolis force causes the winds to
turn away from that steady-state value, and forces
the winds into an inertial oscillation. See a pre-
vious INFO box titled Approach to Geostrophy for an -
L1B
Z
example of undamped inertial oscillations.
During a portion of this oscillation the winds can Y
1 1B

become faster than geostrophic (supergeostrophic),

L
'1(
leading to a low-altitude phenomenon called the
nocturnal jet. See the Atmospheric Boundary

L
Layer chapter for details. 'OFU
1B
M
DF
BS
SQ
Cyclostrophic Wind P G BJ

Winds in tornadoes are about 100 m s1, and in L1 QBUI .
B DT
waterspouts are about 50 m s1. As a tornado first
forms and tangential winds increase, centrifugal '$'
force increases much more rapidly than Coriolis
force. Centrifugal force quickly becomes the domi-
nant force that balances pressure-gradient force (Fig. Figure 10.20
10.20). Thus, a steady-state rotating wind is reached Around tornadoes, pressure gradient force FPG is so strong that
at much slower speeds than the gradient wind it greatly exceeds all other forces such as Coriolis force FCF.
speed. The net force (Fnet) pulls the air around the tight circle at the
If the tangential velocity around the vortex is cyclostrophic wind speed (Mcs).
steady, then the steady-state force balance is:
1 P VM (10.45a)
0 = + s
x R
Sample Application
0 = 1 P s U M (10.45b) A 10 m radius waterspout has a tangential velocity
y R of 45 m s1. What is the radial pressure gradient?
}
}
pressure
Find the Answer
gradient
centrifugal Given: Mcs = 45 m s1, R = 10 m.
Find: P/R = ? kPa m1.
You can use cylindrical coordinates to simplify
Assume cyclostrophic wind, and = 1 kg m3.
solution for the cyclostrophic (tangential) winds
Rearrange eq. (10.46):
Mcs around the vortex. The result is:
P 2 (1kg/m 3 )( 45m/s)2
= Mcs =
R P (10.46) R R 10m
Mcs =
R P/R = 202.5 kgm1s2 / m = 0.2 kPa m1.
where the velocity Mcs is at distance R from the vor- Check: Physics & units are reasonable.
tex center, and the radial pressure gradient in the Exposition: This is 2 kPa across the 10 m waterspout
vortex is P/R. radius, which is 1000 times greater than typical synop-
tic-scale pressure gradients on weather maps.
Z Recall from the Gradient Wind section that anticy-

clones cannot have strong pressure gradients, hence
3
winds around highs are too slow to be cyclostrophic.
Y Around cyclones (lows), cyclostrophic winds can
turn either counterclockwise or clockwise in either
hemisphere, because Coriolis force is not a factor.
'OFU'$'
QB
UI PG Inertial Wind
BJS
QBSD Steady-state inertial motion results from a bal-
FM
.J ance of Coriolis and centrifugal forces in the absence
of any pressure gradient:
Figure 10.21
Coriolis force (FCF, thick black arrow, behind the white arrow) Mi 2 (10.47)
0 = f c Mi +
on an air parcel (grey ball), creating an anticyclonic inertial R
wind Mi, grey arrow). R is radius of curvature. White arrow
is net force Fnet. where Mi is inertial wind speed, fc is the Coriolis pa-
rameter, and R is the radius of curvature. Since both
of these forces depend on wind speed, the inertial
Sample Application wind cannot start itself from zero. It can occur only
For an inertial ocean current of 5 m s 1, find the after some other force first causes the wind to blow,
radius of curvature and time period to complete one and then that other force disappears.
circuit. Assume a latitude where fc = 10 4 s1.
The inertial wind coasts around a circular path
of radius R,
Find the Answer
Given: Mi = 5 m s1, fc = 10 4 s1. Mi
Find: R = ? km, Period = ? h R= (10.48)
fc
Use eq. (10.48): R = (5 m s1) / (10 4 s1) = 50 km
Use Period = 2/fc = 62832 s = 17.45 h where the negative sign implies anticyclonic rotation
(Fig. 10.21). The time period needed for this inertial
Check: Units & magnitudes are reasonable. oscillation to complete one circuit is Period = 2/fc,
Exposition: The tracks of drifting buoys in the ocean which is half of a pendulum day (see Approach to
are often cycloidal, which is the superposition of a Geostrophy INFO Box earlier in this chapter).
circular inertial oscillation and a mean current that Although rarely observed in the atmosphere,
gradually translates (moves) the whole circle. inertial oscillations are frequently observed in the
ocean. This can occur where wind stress on the
ocean surface creates an ocean current, and then
- Z
after the wind dies the current coasts in an inertial
oscillation.
1L1B
.B Y
Antitriptic Wind
'1( A steady-state antitriptic wind Ma could result
L1B from a balance of pressure-gradient force and tur-
bulent drag:
(
0 = 1 P wT Ma (10.49)
L1B d zi
'5% where P is the pressure change across a distance

d perpendicular to the isobars, wT is the turbulent
) transport velocity, and zi is the atmospheric bound-
ary-layer depth.
Figure 10.22 This theoretical wind blows perpendicular to the
Balance of forces (F, black arrows) that create the antitriptic isobars (Fig. 10.22), directly from high to low pres-
wind Ma (grey arrow). G is the theoretical geostrophic wind. sure:
FTD is turbulent drag, and FPG is pressure-gradient force.
zi fc G Sample Application
Ma = (10.50)
In a 1 km thick convective boundary layer at a loca-
wT
tion where fc = 10 4 s1, the geostrophic wind is 5 m s1.
The turbulent transport velocity is 0.02 m s1. Find the
For free-convective boundary layers, wT = bDwB is antitriptic wind speed.
not a function of wind speed, so Ma is proportional
to G. However, for windy forced-convection bound- Find the Answer
ary layers, wT = CDMa, so solving for Ma shows it to Given: G = 5 m s1, zi = 1000 m, fc = 10 4 s1,
be proportional to the square root of G. wT = 0.02 m s1
This wind would be found in the atmospheric Find: Ma = ? m s1
boundary layer, and would occur as an along-valley
component of long gap winds (see the Regional Use eq. (10.50):
Ma = (1000m)(10 4 s1)(5m s1) / (0.02 m s1)
Winds chapter). It is also sometimes thought to be
= 25 m s 1
relevant for thunderstorm cold-air outflow and for
steady sea breezes. However, in most other situa-
Check: Magnitude is too large. Units reasonable.
tions, Coriolis force should not be neglected; thus,
Exposition: Eq. (10.50) can give winds of Ma > G for
the atmospheric boundary-layer wind and BL Gra- many convective conditions, for which case Coriolis
dient winds are much better representations of na- force would be expected to be large enough that it
ture than the antitriptic wind. should not be neglected. Thus, antitriptic winds are
unphysical. However, for forced-convective bound-
ary layers where drag is proportional to wind speed
Summary of Horizontal Winds squared, reasonable solutions are possible.
Table 10-5 summarizes the idealized horizontal
winds that were discussed earlier in this chapter.
On real weather maps such as Fig. 10.23, isobars
or height contours have complex shapes. In some
regions the height contours are straight (suggesting
that actual winds should nearly equal geostrophic
or boundary-layer winds), while in other regions the
height contours are curved (suggesting gradient or
boundary-layer gradient winds). Also, as air parcels INFO The Rossby Number
move between straight and curved regions, they are
sometimes not quite in equilibrium. Nonetheless, The Rossby number (Ro) is a dimensionless ratio
when studying weather maps you can quickly esti- defined by
mate the winds using the summary table. M M
Ro = or Ro =
fc L fc R
where M is wind speed, fc is the Coriolis parameter,
L is a characteristic length scale, and R is radius of

- curvature.
In the equations of motion, suppose that advection
- terms such as UU/x are order of magnitude M2/L,
and Coriolis terms are of order fcM. Then the Rossby

$BOBEB "UMBOUJD0DFBO number is like the ratio of advection to Coriolis terms:
(M2/L) / (fcM) = M/(fcL) = Ro. Or, we could consider
&VSPQF the Rossby number as the ratio of centrifugal (order of
[LN
M2/R) to Coriolis terms, yielding M/(fcR) = Ro.
Use the Rossby number as follows. If Ro < 1, then
) Coriolis force is a dominant force, and the flow tends
to become geostrophic (or gradient, for curved flow).
If Ro > 1, then the flow tends not to be geostrophic.
Figure 10.23 For example, a midlatitude cyclone (low-pressure
One-day average geopotential heights z (thick lines in km, thin system) has approximately M = 10 m s1, fc = 10 4 s1,
lines in m) on the 20 kPa isobaric surface for 5 Feb 2013. Close and R = 1000 km, which gives Ro = 0.1 . Hence, mid-
spacing (tight packing) of the height contours indicate faster latitude cyclones tend to adjust toward geostrophic
winds. This upper-level chart is for the same day and location balance, because Ro < 1. In contrast, a tornado has
(Atlantic Ocean) as the mean-sea-level pressure chart in Fig. roughly M = 50 m s1, fc = 10 4 s1, and R = 50 m,
10.15. [Courtesy of NOAA/NCEP Earth System Research Lab. which gives Ro = 10,000, which is so much greater
http://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanaly- than one that geostrophic balance is not relevant.
sis.html ]
Table 10-5. Summary of horizontal winds**.

Name of
Item Forces Direction Magnitude Where Observed
Wind
faster where isobars
parallel to straight are closer together. aloft in regions
pressure-gradient,
1 geostrophic isobars with Low pres- g z where isobars are
Coriolis
sure to the winds left* G= nearly straight
fc d
similar to geostrophic
slower than
wind, but following
pressure-gradient, geostrophic around aloft in regions
curved isobars. Clock-
2 gradient Coriolis, Lows, faster than where isobars are
wise* around Highs,
centrifugal geostrophic around curved
counterclockwise*
Highs
around Lows.
similar to geostrophic near the ground
atmospheric pressure-gradient, slower than
wind, but crosses in regions where
3 boundary Coriolis, geostrophic (i.e.,
isobars at small angle isobars are nearly
layer drag subgeostrophic)
toward Low pressure straight
atmospheric pressure-gradient, similar to gradient
near the ground in
boundary- Coriolis, wind, but crosses slower than gradient
4 regions where iso-
layer drag, isobars at small angle wind speed
bars are curved
gradient centrifugal toward Low pressure
either clockwise or tornadoes, water-
stronger for lower
pressure-gradient, counterclockwise spouts (& sometimes
5 cyclostrophic pressure in the vortex
centrifugal around strong vortices in the eye-wall of
center
of small diameter hurricanes)
coasts at constant
Coriolis, anticyclonic circular ocean-surface
6 inertial speed equal to its
centrifugal rotation currents
initial speed
* For Northern Hemisphere. Direction is opposite in Southern Hemisphere. ** Antitriptic winds are unphysical; not listed here.
Horizontal Motion
Equations of Motion Again

The geostrophic wind can be used as a surrogate For winds turning around a circle, you can add a
for the pressure-gradient force, based on the defini- term for centrifugal force, which is an artifice to ac-
tions in eqs. (10.26). Thus, the equations of hori- count for the continual changing of wind direction
zontal motion (10.23) become: caused by an imbalance of the other forces (where
the imbalance is the centripetal force).
(10.51a) The difference between the actual and
geostrophic winds is the ageostrophic wind (Uag,
U
t
= U
U
x
V
U
y
W
U
z
(
+ fc V Vg wT
U
zi
) Vag). The term in eqs. (10.51) containing these differ-
ences indicates the geostrophic departure.
(10.51b) Uag = U Ug (10.52a)

V
t
= U
V
x
V
V
y
W
V
z
(
fc U U g wT
V
zi
) Vag = V Vg (10.52b)
}
}
}
}
}
pressure turbulent
tendency advection Coriolis
gradient drag
Scales of Horizontal Motion 5BCMF)PSJ[POUBMTDBMFTPGNPUJPOJOUIFUSPQPTQIFSF

A wide range of horizontal scales of motion (Ta-
)PSJ[POUBM4DBMF
ble 10-6) are superimposed in the atmosphere: from
4J[F %FTJHOBUJPO /BNF
large global-scale circulations through extra-tropi-
LN
cal cyclones, thunderstorms, and down to swirls of NBDSPB QMBOFUBSZTDBMF
LN
turbulence. NBDSPC TZOPQUJDTDBMF
LN
The troposphere is roughly 10 km thick, and this NFTPB
LN
constrains the vertical scale of most weather phe- NFTPC NFTPTDBMF
LN
nomena. Thus, phenomena of large horizontal scale NFTPH
LN
will have a constrained vertical scale, causing them NJDSPB CPVOEBSZMBZFSUVSCVMFODF
N
NJDSPTDBMF
to be similar to a pancake. However, phenomena NJDSPC TVSGBDFMBZFSUVSCVMFODF
N
with smaller horizontal scale can have aspect ratios NJDSPH
N JOFSUJBMTVCSBOHFUVSCVMFODF
(width/height) of about one; namely, their character-
NN
istics are isotropic. NJDSPE
NN GJOFTDBMFUVSCVMFODF
Larger-scale meteorological phenomena tend to
NN
exist for longer durations than smaller-scale ones. WJTDPVT EJTTJQBUJPOTVCSBOHF
N
Fig. 10.24 shows that time scales and horizontal NFBOGSFFQBUICFUXFFONPMFD
N NPMFDVMBS
length scales of many meteorological phenomena NPMFDVMFTJ[FT

nearly follow a straight line on a log-log plot. This
/PUF%JTBHSFFNFOUBNPOHEJGGFSFOUPSHBOJ[BUJPOT
implies that 4ZOPQUJD".4LN8.0LN
/o = (/o)b (10.53) .FTPTDBMF".4LN8.0LN
.JDSPTDBMF".4LN8.0DNLN
where o 10 3 h, o 10 3 km, and b 7/8. XIFSF".4"NFSJDBO.FUFPSPMPHJDBM4PDJFUZ
BOE8.08PSME.FUFPSPMPHJDBM0SHBOJ[BUJPO
In the next several chapters, we cover weather
phenomena from largest to smallest horiz. scales:
Chapter 11 General Circulation (planetary)
Chapter 12 Fronts & Airmasses (synoptic)
Chapter 13 Extratropical Cyclones (synoptic)
EFDBEF
Chapter 14 Thunderstorm Fundam. (meso ) HMPCBM
DJSDVMBUJPO
Chapter 15 Thunderstorm Hazards (meso ) DMJNBUF
ZFBS
Chapter 16 Tropical Cyclones (meso & ) WBSJBUJPOT
Chapter 17 Regional Winds (meso & ) NPO

Chapter 18 Atm. Boundary Layers (microscale) NPOUI TPPO
3PTTCZ
Although hurricanes are larger than thunderstorms, XFFL XBWFT
5JNF4DBMF I
we cover thunderstorms first because they are the GSPOUT DZDMPOFT
building blocks of hurricanes. Similarly, midlati- EBZ IVSSJDBOF

UIVOEFSTUPSNT
MPDBM
tude cyclones often contain fronts, so fronts are cov- XJOET .$4
ered before extratropical cyclones.

IPVS
UPSOBEP
UIFSNBMT &BSUI$JSDVNGFSFODF
UVSCVMFODF
QSPEVDUJPO
Vertical Forces and Motion
NJOVUF
UVSCVMFODF
DBTDBEF
Forces acting in the vertical can cause or change
vertical velocities, according to Newtons Second
Law. In an Eulerian framework, the vertical com- )PSJ[POUBM4DBMF LN
ponent of the equations of motion is:

(10.54) Figure 10.24
Typical time and spatial scales of meteorological phenomena.
W W W W 1 P Fz TD
= U V W g
t x y z z m
}
}
}
}
pressure gra- turb.

tendency advection
gradient vity drag
S LHNm
where the vertical acceleration given in the left side
of the equation is determined by the sum of all

forces/mass acting in the vertical, as given on the
right. For Cartesian directions (x, y, z) the velocity
components are (U, V, W). Also in this equation are
air density (), pressure (P), vertical turbulent-drag
force (Fz TD), mass (m), and time (t). Magnitude of
gravitational acceleration is |g| = 9.8 ms2. Coriolis
TUSBUPTQIFSF
[ force is negligible in the vertical (see the INFO box
LN
on Coriolis Force in 3-D, earlier in this chapter), and is
not included in the equation above.
Recall from Chapter 1 that our atmosphere has
an extremely large pressure gradient in the verti-
cal, which is almost completely balanced by gravity
USPQPTQIFSF (Fig. 10.25). Also, there is a large density gradient
in the vertical. We can define these large terms as a
mean background state or a reference state of
1 L1B
the atmosphere. Use the overbar over variables to
indicate their average background state. Define this
Figure 10.25 background state such that it is exactly in hydro-
Background state, showing change of mean atmospheric pres- static balance (see Chapter 1):
P = P +and
sure = +with
P ' mean density height z, based on a standard
atmosphere from Chapter 1. P (10.55)
= g
z
However, small deviations in density and pres-

sure from the background state can drive important
non-hydrostatic vertical motions, such as in ther-
mals and thunderstorms. To discern these effects,
we must first remove the background state from the
full vertical equation of motion. From eq. (10.54), the
gravity and pressure-gradient terms are:
1 P
g (10.56)
Sample Application z
Suppose your neighborhood has a background en-
vironmental temperature of 20C, but at your particu- But total density can be divided into background
lar location the temperature is 26C with a 4 m s1 west =( )+ and = +( ) components: = + . Do
deviation
wind and no vertical velocity. Just 3 km west is an 5 m the same for pressure: P = P + P ' . Thus, eq. (10.56)
s1 updraft. Find the vertical acceleration. can be expanded as:
Find the Answer
Given: Te = 273+20 = 293 K, = 6C, U= 4 m s1 1 P P
z z g g (10.57)
W/x= (5 m s1 0) / (3,000m 0)
Find: W/t = ? ms2
( )
+
Assume: Because W = 0, there is zero drag. Because
the air is dry: Tv = T. Given no V info, assume zero. The first and third terms in square brackets in eq.
(10.57) cancel out, due to hydrostatic balance (eq.
Apply eq. (10.59): W W p e 10.55) of the background state.
= U + g
t x Te In the atmosphere, density perturbations = +( ) are
W/t = (4 m s1)(5ms1/3,000m)+(6/293)(9.8ms2) usually much smaller than mean density. Thus densi-
= 0.0067 + 0.20 = 0.21 ms 2 ty perturbations can be neglected everywhere except
in the gravity term, where g / ( + ) ( / ) g .
Check: Physics & units are reasonable. This is called the Boussinesq approximation.
Exposition: Buoyancy dominated over advection for Recall from the chapters 1 and 5 that you can use
this example. Although drag was zero initially be-
virtual temperature (Tv) with the ideal gas law in
cause of zero initial vertical velocity, we must include
place of air density (but changing the sign because
the drag term once the updraft forms.
low virtual temperatures imply high densities):
v p v e INFO Etvs Effect

g = vg = g = g ' (10.58)
Tv Tv e When you move along a path at constant distance
R above Earths center, gravitational acceleration ap-
where subscripts p & e indicate the air parcel and the pears to change slightly due to your motion. The
environment surrounding the parcel, and where g measured gravity |gobs| = |g| ar , where:
is called the reduced gravity. The virtual potential
ar = 2cos()U + (U2 + V2)/R
temperature v can be in either Celsius or Kelvin, but
units of Kelvin must be used for Tv and Tve. The first term is the vertical component of Coriolis
Combining eqs. (10.54) & (10.58) yields: force (eq. 10.17e in the INFO box on p.297), and the last
term is centrifugal force as you follow the curvature
W W W W of the Earth. Thus, you feel lighter traveling east and
= U V W heavier traveling west. This is the Etvs effect.
t x y z
(10.59)
1 P v p v e Fz TD
+ g
z Tve m 88
Terms from this equation will be used in the Region-

al Winds chapter and in the Thunderstorm chapters Y
to explain strong vertical velocities. Z
When an air parcel rises or sinks it experiences
resistance (turbulent drag, Fz TD) per unit mass m as 77 [
it tries to move through the surrounding air. This
is a completely different effect than air drag against
6 66
the Earths surface, and is not described by the same 7
drag equations. The nature of Fz TD is considered in
the chapter on Air Pollution Dispersion, as it affects [
the rise of smoke-stack plumes. Fz TD = 0 if the air Z
parcel and environment move at the same speed. 8 Figure 10.26
Y
Air-mass flows to and from
a fixed Eulerian volume.
Conservation of Air Mass

Sample Application
Hurricane-force winds of 60 m s1 blow into an
Due to random jostling, air molecules tend to
north-facing entrance of a 20 m long pedestrian tun-
distribute themselves uniformly within any volume. nel. The door at the other end of the tunnel is closed.
Namely, the air tends to maintain its continuity. The initial air density in the tunnel is 1.2 kg m3. Find
Any additional air molecules entering the volume the rate of air density increase in the tunnel.
that are not balanced by air molecules leaving (Fig.
10.26) will cause the air density (, mass of air mol- Find the Answer
ecules in the volume) to increase, as described below Given: VN. entrance = 60 m s1, VS. entrance = 0 m s1,
by the continuity equation. = 1.2 kg m3, y = 20 m,
Find: /t = ? kgm3s1 initially.
Continuity Equation Use eq. (10.60), with U = W = 0 because the other walls,
For a fixed Eulerian volume, the mass budget roof, and floor prevent winds in those directions:
equation (i.e., the continuity equation) is: V VS.entr . kg ( 60 0)m/s
(10.60) = N .entr . = 1.2 3
t ytunnel m (20)m
U V W /t = +3.6 kgm3s1 .
= U V W + +
t x y z x y z
The terms in curly braces { } describe advection. Exposition: As air density increases, so will air pres-
sure. This pressure might be sufficient to blow open
With a bit of calculus one can rewrite this equation
the other door at the south end of the pedestrian tun-
as:
nel, allowing the density to decrease as air escapes.
( U ) ( V ) ( W ) vertical velocities increase with height where there

= + + (10.61) is horizontal convergence:
t x y z
W
= D (10.64)
where U, V, and W are the wind components in the x, z
y, and z directions, respectively, and t is time.
Boundary-Layer Pumping
When you calculate wind gradients, be sure to
Consider an extratropical cyclone, where the
take the wind and space differences in the same di-
boundary-layer gradient wind spirals in toward the
rection. For example: U/x = (U2U1)/(x2x1).
low-pressure center. Those spiraling winds consist
of a tangential component following the isobars as
Incompressible Idealization they encircle the low center, and a radial component
Mean air density changes markedly with alti- having inflow velocity Vin (Fig. 10.27).
tude, as was sketched in Fig. 10.25. However, at any But volume inflow (2Rz Vin) through the sides
one altitude the density changes only slightly due of the cylindrical volume of radius R and height
to local changes in humidity and temperature. For z must be balanced by net volume outflow (R2
non-tornadic, non-thunderstorm conditions where W/z) through the top and bottom. Equating these
Fig. 10.25 is valid, we can make a reasonable sim- incompressible flows gives:
plifying idealization that density is constant (
2Vin W
0) at any one altitude. Namely, air behaves as if it is = (10.65a)
incompressible. R z
If we make this idealization, then the advection Thus, for horizontal inflow everywhere (positive
terms of eq. (10.60) are zero, and the time-tendency Vin), one finds that W must also be positive.
term is zero. The net result is volume conserva- If a cylinder of air is at the ground where W = 0 at
tion, where volume outflow equals volume inflow: the cylinder bottom, then W at the cylinder top is:
W = (2 Vin z) / R (10.65b)
U V W
+ + = 0 (10.62)
x y z Namely, extratropical cyclones have rising air, which
causes clouds and rain due to adiabatic cooling. This
Fig. 10.26 illustrates such incompressible con- forcing of a broad updraft regions by horizontal-
tinuity. Can you detect an error in this figure? It wind drag around a cyclone is known as bound-
shows more air leaving the volume in each coor- ary-layer pumping or Ekman pumping.
dinate direction than is entering impossible for For atmospheric boundary-layer gradient (ABLG)
incompressible flow. A correct figure would have winds around anticyclones (highs), the opposite oc-
changed arrow lengths, to indicate net inflow in one curs: horizontal outflow and a broad region of de-
or two directions, balanced by net outflow in the scending air (subsidence). The subsidence causes
other direction(s). adiabatic warming, which evaporates any clouds
As will be explained in the last section of this and creates fair weather.
chapter, divergence is where more air leaves a vol-
ume than enters (corresponding to positive terms in
eq. 10.62). Convergence is where more air enters
than leaves (corresponding to negative terms in eq.
10.62). Thus, volume (mass) conservation of incom- 88
pressible flow requires one or two terms in eq. (10.62)
to be negative (i.e., convergence), and the remaining
term(s) to be positive (i.e., divergence) so that their
sum equals zero. 3
7JO
Horizontal divergence (D) is defined as 7JO
[
U V (10.63)
D= +
x y 7JO
7JO
Negative values of D correspond to convergence. 8
Plugging this definition into eq. (10.62) shows that
Figure 10.27
Volume conservation for an idealized cylindrical extratropical
cyclone.
Recall from the ABLG wind section that an ana-

Sample Application
lytical solution could not be found for VABLG (which
Atmospheric boundary-layer gradient winds have
is the needed Vin for eq. 10.65). Instead, we can ap- an inflow component of 3 m s1 at radius 500 km from
proximate Vin VABL for which an analytical solu- the center of a midlatitude cyclone. What is the updraft
tion exists. But VABL is always larger than VABLG for speed?
flow around cyclones, so we must be aware that our
analytical answer will always give winds that are Find the Answer
slightly faster than occur around lows in nature. Given: R = 500 km, Vin = 3 m s1. Let z = zi = 1 km
To solve for VABL, we need to make an assump- Find: W = ? m s1
tion about the static stability of the atmospheric
boundary layer. Because cyclones generally have Apply eq. (10.65b):
overcast skies and strong winds, we can safely as- z 1km
W = 2 Vin = 2 (3m/s)
sume neutral stability. In this case, eq. (10.41b) gives R 500km
the cross-isobaric inflow velocity.
= 0.012 m s 1
Use VABL for Vin in eq. (10.65b) and solve for W
(which we will call WABL the vertical velocity at Check: Physics & units are reasonable.
the atmospheric boundary-layer top, as sketched in Exposition: This updraft speed of 1.2 cm/s, while
Fig. 10.28): slow, can cause significant lifting over many hours.
2 Nonetheless, this updraft is slow enough to enable us
2 b CD G (10.66)
WABL = to approximate the atmosphere as being hydrostatic.
fc R
with geostrophic wind G, radius of curvature R,

Coriolis parameter fc, and drag coefficient CD for
statically neutral boundary conditions. For flow
over land, CD 0.005 .
8"#-
Eq. (10.41b) can be used to find b = { 1 0.5[CDG/
(fczi)] } for an atmospheric boundary layer of thick- 7"#-( 7"#-(
ness zi. If you dont know the actual atmospheric
boundary-layer depth, then a crude approxima-
- 7"#-( [J
7"#-( (
tion for cyclones (not valid for anticyclones) is :
7"#-(
7"#-( 7"#-(
G
zi (10.67) [ Z
N BV Y
Figure 10.28
In this approximation, you must use a Brunt-Visl
Application of atmospheric boundary-layer gradient winds
frequency NBV that is valid for the statically stable VABLG to estimate the vertical velocity WABL due to atmospher-
air in the troposphere above the top of the statically ic boundary-layer pumping around a low-pressure center (L).
neutral atmospheric boundary layer. For this spe- G = geostrophic wind.
cial approximation: b = { 1 0.5[CDNBV/fc]}. A re-
quired condition for a physically realistic solution is
[CDNBV/fc] < 1.
You can interpret eq. (10.66) as follows. Stronger
pressure gradients (which cause larger geostrophic
wind G), larger drag coefficients, and smaller radii
of curvature cause greater atmospheric boundary-
layer pumping WABL.
Although the equations above allow a complete
approximate solution, we can rewrite them in terms
of a geostrophic relative vorticity:
2 G
g = (10.68)
R
which indicates air rotation. Vorticity is introduced
later in this chapter, and is covered in greater detail
in the General Circulation chapter.
Eq. (10.66) can be modified to use geostrophic

Sample Application
vorticity. The resulting Ekman pumping at the at-
At 500 km from the center of a midlatitude cyclone
at latitude where fc = 0.0001 s1, the pressure gradient mospheric boundary layer top in a midlatitude cy-
can drive a 15 m s1 geostrophic wind. Assume a stan- clone is:
dard atmosphere static stability above the top of the
G C N
atmospheric boundary layer (ABL), and a drag coeffi- WABL = CD g 1 0.5 D BV (10.69)
cient of 0.004 at the bottom. Find the Ekman pumping fc fc
updraft speed out of the atmospheric boundary-layer
top. Also, what are the geostrophic relative vorticity, The first four factors on the right side imply that
the depth of the ABL, and the internal Rossby defor- larger drag coefficients (i.e., rougher terrain with
mation radius? more trees or buildings) and stronger pressure gra-
dients (as indicated by larger geostrophic wind)
Find the Answer driving winds around smaller radii of curvature
Given: fc = 0.0001 s1, R = 5x105 m, G = 15 m s1,
(i.e., larger geostrophic vorticity) at lower latitudes
NBV = 0.0113 s1 (from a previous Sample Appli-
cation using the standard atmos.), CD = 0.004, (i.e., smaller fc) create stronger updrafts. Also, stron-
Find: WABL = ? m s1, g = ? s1, zi = ? m, R = ? km ger static stabilities (i.e., larger Brunt-Visl fre-
For depth of the troposphere, assume zT = 11 km. quency NBV) in the troposphere above atmospheric
boundary-layer top reduce updraft speed by oppos-
Apply eq. (10.67): ing vertical motion.
zi G/NBV = (15 m s1)/(0.0113 s1) = 1327 m One can write an internal Rossby deforma-
tion radius based on the eq. (10.67) approximation
Apply eq. (10.68): for depth zi of the atmospheric boundary layer:
2 (15m/s)
g = = 6x10 5 s1 G zT
5 105 m R (10.70)
f c zi
Apply eq. (10.70):
(15m/s) 11km where tropospheric depth is zT. Internal and exter-
R = 1243 km
(0.0001s 1 ) 1.327 km nal Rossby deformation radii are described further
We need to check to ensure that [CDNBV/fc] < 1. in the General Circulation and Fronts & Airmasses
[0.004(0.0113s1)/(0.0001s1)] = 0.452 < 1. chapters, respectively.
Thus, we can expect our approximate solution should The Rossby deformation radius can be used to
work for this case. write yet another expression for Ekman pumping
vertical velocity out of the top of the atmospheric
Apply eq. (10.71): WABL = boundary layer:
(1.327 km)
0.004 (1.243 106 m)(6 10 5 s 1 ) zi
(11km) WABL = CD R g 1 0.5 CD R (10.71)
1243km zT zT
1 0.5 (0.004)
11km
= (0.036 m s ) [0.774] = 0.028 m s 1
1
Check: Physics and units are reasonable. Kinematics

Exposition: The updraft speed 2.8 cm s1 is slow, but
over many hours can cause significant lifting. As the
Kinematics is the study of patterns of motion,
rising air cools adiabatically, clouds form and latent
heat is released due to condensation. Hence, clouds without regard to the forces that cause them. We
and bad weather are often associated with midlatitude will focus on horizontal divergence, vorticity, and
cyclones. deformation. All have units of s1 .
We have already encountered horizontal diver-
gence, D, the spreading of air:
U V (10.72)
D= +
x y
Figure 10.29a shows an example of pure divergence.

Its sign is positive for divergence, and negative for
convergence (when the wind arrows point toward
a common point).
Vorticity describes the rotation of air (Fig. B

%JWFSHFODF
10.29b). The relative vorticity, r , about a locally ver-
tical axis is given by:
V U (10.73)
r =
x y
Z
The sign is positive for counterclockwise rotation
(i.e., cyclonic rotation in the N. Hemisphere), and
negative for clockwise rotation. Vorticity is dis-
Y
cussed in greater detail in the General Circulation
chapter. Neither divergence nor vorticity vary with
rotation of the axes they are rotationally in-
variant.
C
7PSUJDJUZ
Two types of deformation are stretching defor-
mation and shearing deformation (Figs. 10.29c & d).
Stretching deformation, F1, is given by:
U V (10.74)
F1 =
x y Z
The axis along which air is being stretched (Fig.

10.29c) is called the axis of dilation (x axis in this Y
example), while the axis along which air is com-
pressed is called the axis of contraction (y axis in
this example).
Shearing deformation, F2, is given by: D
4USFUDIJOH%FGPSNBUJPO E
4IFBSJOH%FGPSNBUJPO
V U (10.75)
F2 = +
x y
As you can see in Fig. 10.29d, shearing deformation

is just a rotated version of stretching deformation.
The total deformation, F, is:
Z
F12 2 1/2 (10.76)
F= + F2
Y
Deformation often occurs along fronts. Most real
flows exhibit combinations of divergence, vorticity,
and deformation. Figure 10.29
Kinematic flow-field definitions. Black arrows represent wind
velocity.
Measuring Winds
For weather stations at the Earths surface, wind

direction can be measured with a wind vane
mounted on a vertical axel. Fixed vanes and other
shapes can be used to measure wind speed, by us-
ing strain gauges to measure the minute deforma-
tions of the object when the wind hits it.
The generic name for a wind-speed measuring
device is an anemometer. A cup anemometer
has conic- or hemispheric-shaped cups mounted on
spokes that rotate about a vertical axel. A propellor
anemometer has a propellor mounted on a hori- If all the forces vector-sum to zero, then there
zontal axel that is attached to a wind vane so it al- is no net force and winds blow at constant speed.
ways points into the wind. For these anemometers, Theoretical winds based on only a small number
the rotation speed of the axel can be calibrated as a of forces are given special names. The geostrophic
wind speed. wind occurs when pressure-gradient and Coriolis
Other ways to measure wind speed include a hot- forces balance, causing a wind that blows parallel
wire or hot-film anemometer, where a fine metal to straight isobars. For curved isobars around lows
wire is heated electrically, and the power needed to and highs, the imbalance between these two forces
maintain the hot temperature against the cooling turns the wind in a circle, with the result called the
effect of the wind is a measure of wind speed. A gradient wind. Similar winds can exist in the atmo-
pitot tube that points into the wind measures the spheric boundary layer, where turbulent drag of the
dynamic pressure as the moving air stagnates in a air against the Earths surface slows the wind and
dead-end tube. By comparing this dynamic pres- causes it to turn slightly to cross the isobars toward
sure with the static pressure measured by a differ- low pressure.
ent sensor, the pressure difference can be related to Waterspouts and tornadoes can have such strong
wind speed. winds that pressure-gradient force is balanced by
Sonic anemometers send pulses of sound back centrifugal force, with the resulting wind speed
and forth across a short open path between two op- known as the cyclostrophic wind. In oceans, cur-
posing transmitters and receivers of sound. The rents can inertially flow in a circle.
speed of sound depends on both temperature and The two most important force balances at mid-
wind speed, so this sensor can measure both. Trac- latitudes are hydrostatic balance in the vertical, and
ers such as smoke, humidity fluctuations, or clouds geostrophic balance in the horizontal.
can be tracked photogramatically from the ground Conservation of air mass gives the continuity
or from remote sensors such as laser radars (lidars) equation, for which an incompressible approxima-
or satellites, and the wind speed then estimated tion can be used in most places except in thunder-
from the change of position of the tracer between storms. Mechanisms that cause motion in one di-
successive images. rection (horizontal or vertical) will also indirectly
Measurements of wind vs. height can be made cause motions in the other direction as the air tries
with rawinsonde balloons (using a GPS receiver to maintain continuity, resulting in a circulation.
in the sonde payload to track horizontal drift of the Kinematics is the word that describes the be-
balloons with time), dropsondes (like rawinsondes, havior and effect of winds (such as given by the con-
only descending by parachute after being dropped tinuity equation) without regard to the forces that
from aircraft), pilot balloons (carrying no payload, cause them. The word dynamics describes how
but being tracked instead from the ground using forces cause winds (as given by Newtons 2nd law).
radar or theodolites), wind profilers, Doppler
weather radar (see the Satellites & Radar chapter),
and via anemometers mounted on aircraft.
Homework Exercises
Some of these questions are inspired by exercises

Review in Stull, 2000: Meteorology for Sci. & Engr. 2nd Ed.,
Brooks/Cole, 528 pp.
According to Newtons second law, winds are
driven by forces. The pressure-gradient creates a
Broaden Knowledge & Comprehension
force, even in initially calm (windless) conditions.
For all the exercises in this section, collect informa-
This force points from high to low pressure on a con-
tion off the internet. Dont forget to cite the web sites
stant altitude chart (such as at sea-level), or points
you use.
from high to low heights on an isobaric chart (such
as the 50 kPa chart). Pressure-gradient force is the
B1. a. Find a weather map showing todays sea-level
main force that drives the winds.
pressure isobars near your location. Calculate pres-
Other forces exist only when there is already a
sure-gradient force (N) based on your latitude and
wind. One example is turbulent drag against the
the isobar spacing (km/kPa).
ground, which pushes opposite to the atmospheric
b. Repeat this for a few days, and plot the pres-
boundary-layer wind direction. Another example is
sure gradient vs. time.
Coriolis force, which is related to centrifugal force of
winds relative to a rotating Earth.
B2. Get 50 kPa height contour maps (i.e., 500 hPa the altitude of the storm (ie., a map for any pressure
heights) over any portion of the Northern Hemi- level between 85 to 60 kPa). At the eye-wall location,
sphere. In 2 locations at different latitudes hav- use the height-gradient to calculate the cyclostrophic
ing straight isobars, compute the geostrophic wind wind speed. Compare this with the observed hur-
speed. In 2 locations of curved isobars, compute ricane winds at that same approximate location, and
the gradient wind speed. How do these theoreti- discuss any differences.
cal winds compare with wind observations near the
same locations? B10. Get a 500 hPa (= 50 kPa) geopotential height
contour map that is near or over the equator. Com-
B3. Similar exercise B2, but for 2 locations in the pute the theoretical geostrophic wind speed based
Southern Hemisphere. on the height gradients at 2 locations on that map
where there are also observed upper-air wind
B4. a. Using your results from exercise B2 or B3, plot speeds. Explain why these theoretical wind speeds
the geostrophic wind speed vs. latitude and pres- disagree with observed winds.
sure gradient on a copy of Fig. 10.10. Discuss the
agreement or disagreement of your results vs. the
Apply
lines plotted in that figure.
A1. Plot the wind symbol for winds with the follow-
b. Using your results from exercise B4 or B5,
ing directions and speeds:
show that gradient winds are indeed faster than
a. N at 5 kt b. NE at 35 kt c. E at 65 kt
geostrophic around high-pressure centers, and
d. SE at 12 kt e. S at 48 kt f. SW at 105 kt
slower around low-pressure centers.
g. W at 27 kt h. NW at 50 kt i. N at 125 kt
B5. Discuss surprising insights regarding Isacc
A2. How fast does an 80 kg person accelerate when
Newtons discoveries on forces and motion.
pulled with the force given below in Newtons?
a. 1 b. 2 c. 5 d. 10 e. 20 f. 50
B6. Get a map of sea-level pressure, including isobar
g. 100 h. 200 i. 500 j. 1000 k. 2000
lines, for a location or date where there are strong
low and high-pressure centers adjacent to each
A3. Suppose the following force per mass is applied
other. On a printed copy of this map, use a straight
on an object. Find its speed 2 minutes after starting
edge to draw a line connecting the low and high
from rest.
centers, and extend the line further beyond each
a. 5 N kg1 b. 10 ms2
center. Arbitrarily define the high center as location
c. 15 N kg1 d. 20 ms2
x = 0. Then, along your straight line, add distance
e. 25 N kg1 f. 30 ms2
tic marks appropriate for the map scale you are us-
g. 35 N kg1 h. 40 ms2
ing. For isobars crossing your line, create a table
i. 45 N kg1 j. 50 ms2
that lists each pressure P and its distance x from the
high. Then plot P vs. x and discuss how it compares
A4. Find the advective force per unit mass given
with Fig. 10.14. Discuss the shape of your curve in
the following wind components (m s1) and hori-
the low- and high-pressure regions.
zontal distances (km):
a. U=10, U=5, x=3
B7. Which animations best illustrate Coriolis Force?
b. U=6, U=10, x=5
c. U=8, V=20, x=10
B8. a. Get a map of sea-level pressure isobars that
d. U=4, V=10, x=2
also shows observed wind directions. Discuss why
e. V=3, U=10, y=10
the observed winds have a direction that crosses the
f. V=5, U=10, y=4
isobars, and calculate a typical crossing angle..
g. V=7, V=2, y=50
b. For regions where those isobars curve around
h. V=9, V=10, y=6
cyclones or anticyclones, confirm that winds spiral
into lows and out of highs.
A5. Town A is 500 km west of town B. The pressure
c. For air spiraling in toward a cyclone, estimate
at town A is given below, and the pressure at town B
the average inflow radial velocity component, and
is 100.1 kPa. Calculate the pressure-gradient force/
calculate WBL based on incompressible continuity.
mass in between these two towns.
a. 98.6 b. 98.8 c. 99.0 d. 99.2 e. 99.4
B9. For a typhoon or hurricane, get a current or past
f. 99.6 g. 99.8 h. 100.0 i. 100.2 j. 100.4
weather map showing height-contours for any one
k. 100.6 l. 100.8 m. 101.0 n. 101.2 o. 101.4
isobaric level corresponding to an altitude about 1/3
A6. Suppose that U = 8 m s1 and V = 3 m s1, and a. pressure gradient b. Coriolis

latitude = 45 Calculate centrifugal-force compo- c. centrifugal d. drag
nents around a:
a. 500 km radius low in the N. hemisphere A12. Given the pressure gradient magnitude
b. 900 km radius high in the N. hemisphere (kPa/1000 km) below, find geostrophic wind speed
c. 400 km radius low in the S. hemisphere for a location having fc = 1.1x10 4 s1 and = 0.8 kg
d. 500 km radius high in the S. hemisphere m3.
a. 1 b. 2 c. 3 d. 4 e. 5
A7. What is the value of fc (Coriolis parameter) at: f. 6 g. 7 h. 8 i. 9 j. 10
a. Shanghai k. 11 m. 12 n. 13 o. 14 p. 15
b. Istanbul
c. Karachi A13. Suppose the height gradient on an isobaric sur-
d. Mumbai face is given below in units of (m km1). Calculate
e. Moscow the geostrophic wind at 55N latitude.
f. Beijing a. 0.1 b. 0.2 c. 0.3 d. 0.4 e. 0.5
g. So Paulo f. 0.6 g. 0.7 h. 0.8 i. 0.9 j. 1.0
h. Tianjin k. 1.1 m. 1.2 n. 1.3 o. 1.4 p. 1.5
i. Guangzhou
j. Delhi A14. At the radius (km) given below from a low-
k. Seoul pressure center, find the gradient wind speed
l. Shenzhen given a geostrophic wind of 8 m s1 and given fc =
m. Jakarta 1.1x10 4 s1.
n. Tokyo a. 500 b. 600 c. 700 d. 800 e. 900
o. Mexico City f. 1000 g. 1200 h. 1500 i. 2000 j. 2500
p. Kinshasa
q. Bangalore A15. Suppose the geostrophic winds are Ug = 3 m
r. New York City s1 with Vg = 8 m s1 for a statically-neutral bound-
s. Tehran ary layer of depth zi = 1500 m, where fc = 1.1x10 4 s1.
t. (a city specified by your instructor) For drag coefficients given below, what is the atmos.
boundary-layer wind speed, and at what angle does
A8. What is the magnitude and direction of Coriolis this wind cross the geostrophic wind vector?
force/mass in Los Angeles, USA, given: a. 0.002 b. 0.004 c. 0.006 d. 0.008 e. 0.010
U (m s1 ) V (m s1) f. 0.012 g. 0.014 h. 0.016 i. 0.018 j. 0.019
a. 5 0
b. 5 5 A16. For a statically unstable atmos. boundary layer
c. 5 5 with other characteristics similar to those in exercise
d. 0 5 A15, what is the atmos. boundary-layer wind speed,
e. 0 5 at what angle does this wind cross the geostrophic
f. 5 0 wind vector, given wB (m s1) below?
g. 5 5 a. 75 b. 100 c. 50 d. 200 e. 150
h. 5 5 f. 225 g. 125 h. 250 i. 175 j. 275
A9. Same wind components as exercise A8, but find A17(). Review the Sample Application in the At-
the magnitude and direction of turbulent drag force/ mospheric Boundary Layer Gradient Wind section.
mass in a statically neutral atmospheric boundary Re-do that calculation for MABLG with a different
layer over an extensive forested region. parameter as given below:
a. zi = 1 km b. CD = 0.003 c. G = 8 m s1
A10. Same wind components as exercise A8, but find d. fc = 1.2x10 4 s1 e. R = 2000 km
the magnitude and direction of turbulent drag force/ f. G = 15 m s1 g. zi = 1.5 km h. CD = 0.005
mass in a statically unstable atmospheric boundary i. R = 1500 km j. fc = 1.5x10 4 s1
layer with a 50 m/s buoyant velocity scale. Hint: Assume all other parameters are unchanged.
A11. Draw a northwest wind of 5 m s1 in the S. A18. Find the cyclostrophic wind at radius (m) given
Hemisphere on a graph, and show the directions below, for a radial pressure gradient = 0.5 kPa m1:
of forces acting on it. Assume it is in the boundary a. 10 b. 12 c. 14 d. 16 e. 18
layer. f. 20 g. 22 h. 24 i. 26 j. 28 k 30
E2. Suppose that the initial winds are unknown.

A19. For an inertial wind, find the radius of curva- Can a forecast still be made using eqs. (10.6)? Ex-
ture (km) and the time period (h) needed to com- plain your reasoning.
plete one circuit, given fc = 10 4 s1 and an initial
wind speed (m s1) of: E3. Considering eq. (10.7), suppose there are no forc-
a. 1 b. 2 c. 3 d. 4 e. 6 f. 7 g. 8 h. 9 es acting. Based on eq. (10.5), what can you antici-
i. 10 j. 11 k. 12 m. 13 n. 14 o. 15 pate about the wind speed.
A20. Find the antitriptic wind for the conditions of E4. We know that winds can advect temperature
exercise A15. and humidity, but how does it work when winds
advect winds? Hint, consider eqs. (10.8).
A21. Below is given an average inward radial wind
component (m s1) in the atm. boundary layer at ra- E5. For an Eulerian system, advection describes the
dius 300 km from the center of a cyclone. What is influence of air that is blown into a fixed volume. If
the average updraft speed out of the atm. boundary- that is true, then explain why the advection terms
layer top, for a boundary layer that is 1.2 km thick? in eq. (10.8) is a function of the wind gradient (e.g.,
a. 2 b. 1.5 c. 1.2 d. 1.0 U/x) instead of just the upwind value?
e. 0.5 f. 1 g. 2.5 h. 3 i. 0.8 j. 0.2
E6. Isobar packing refers to how close the isobars
A22. Above an atmospheric boundary layer, assume are, when plotted on a weather map such as Fig.
the tropospheric temperature profile is T/z = 0. 10.5. Explain why such packing is proportional to
For a midlatitude cyclone, estimate the atm. bound- the pressure gradient.
ary-layer thickness given a near-surface geostrophic
wind speed (m s1) of: E7. Pressure gradient has a direction. It points to-
a. 5 b. 10 c. 15 d. 20 e. 25 f. 30 ward low pressure for the Northern Hemisphere.
g. 35 h. 40 i. 3 j. 8 k. 2 l. 1 For the Southern Hemisphere, does it point toward
high pressure? Why?
A23(). For atm. boundary-layer pumping, plot
a graph of updraft velocity vs. geostrophic wind E8. To help you interpret Fig. 10.5, consider each hor-
speed assuming an atm. boundary layer of depth izontal component of the pressure gradient. For an
0.8 km, a drag coefficient 0.005 . Do this only for arbitrary direction of isobars, use eqs. (10.9) to dem-
wind speeds within the valid range for the atmos. onstrate that the vector sum of the components of
boundary-layer pumping eq. Given a standard at- pressure-gradient do indeed point away from high
mospheric lapse rate at 30 latitude with radius of pressure, and that the net direction is perpendicular
curvature (km) of: to the direction of the isobars.
a. 750 b. 1500 c. 2500 d. 3500 e. 4500
f. 900 g. 1200 h. 2000 i. 3750 j. 5000 E9. For centrifugal force, combine eqs. (10.13) to
show that the net force points outward, perpendicu-
A24. At 55N, suppose the troposphere is 10 km lar to the direction of the curved flow. Also show
thick, and has a 10 m s1 geostrophic wind speed. that the magnitude of that net vector is a function of
Find the internal Rossby deformation radius for an tangential velocity squared.
atmospheric boundary layer of thickness (km):
a. 0.2 b. 0.4 c. 0.6 d. 0.8 e. 1.0 E10. Why does fc = 0 at the equator for an air parcel
f. 1.2 g. 1.5 h. 1.75 i. 2.0 j. 2.5 that is stationary with respect to the Earths surface,
even though that air parcel has a large tangential ve-
A25. Given U/x = V/x = (5 m s1) / (500 km), locity associated with the rotation of the Earth?
find the divergence, vorticity, and total deformation
for (U/y , V/y) in units of (m s1)/(500 km) as E11. Verify that the net Coriolis force is perpendicu-
given below: lar to the wind direction (and to its right in the N.
a. (5, 5) b. (5, 0) c. (0, 5) d. (0, 0) e. (0, 5) Hemisphere), given the individual components de-
f. (5, 0) g. (5, 5) h. (5, 5) i. (5, 5) scribed by eqs. (10.17).
E12. For the subset of eqs. (10.1 - 10.17) defined by

Evaluate & Analyze your instructor, rewrite them for flow in the South-
E1. Discuss the relationship between eqs. (1.24) and
ern Hemisphere.
(10.1).
E13. Verify that the net drag force opposes the wind wind against the more exact iterative solutions to
by utilizing the drag components of eqs. (10.19). the implicit form in eq. (10.39).
Also, confirm that drag-force magnitude for stati-
cally neutral conditions is a function of wind-speed E26. No explicit solution exists for the neutral atmo-
squared. spheric boundary layer winds, but one exists for the
statically unstable ABL? Why is that?
E14. How does the magnitude of the turbulent-
transport velocity vary with static stability, such as E27. Plug eqs. (10.42) into eqs (10.38) or (10.39) to con-
between statically unstable (convective) and stati- firm that the solution is valid.
cally neutral (windy) situations?
E28(). a. Create your own spreadsheet that gives
E15. Show how the geostrophic wind components the same answer for ABLG winds as in the Sample
can be combined to relate geostrophic wind speed Application in the ABLG-wind section.
to pressure-gradient magnitude, and to relate b. Do what if experiments with your spread-
geostrophic wind direction to pressure-gradient di- sheet to show that the full equation can give the gra-
rection. dient wind, geostrophic wind, and boundary-layer
wind for conditions that are valid for those situa-
E16. How would eqs. (10.26) for geostrophic wind be tions.
different in the Southern Hemisphere? c. Compare the results from (b) against the re-
spective analytical solutions (which you must com-
E17. Using eqs. (10.26) as a starting point, show your pute yourself).
derivation for eqs. (10.29).
E29. Photocopy Fig. 10.13, and enhance the copy
E18. Why are actual winds finite near the equator by drawing additional vectors for the atmospheric
even though the geostrophic wind is infinite there? boundary-layer wind and the ABLG wind. Make
(Hint, consider Fig. 10.10). these vectors be the appropriate length and direc-
tion relative to the geostrophic and gradient winds
E19. Plug eq. (10.33) back into eqs. (10.31) to confirm that are already plotted.
that the solution is valid.
E30. Plug the cyclostrophic-wind equation into eq.
E20. Plug eqs. (10.34) back into eq. (10.33) to confirm (10.45) to confirm that the solution is valid for its spe-
that the solution is valid. cial case.
E21. Given the pressure variation shown in Fig. 10.14. E31. Find an equation for cyclostrophic wind based
Create a mean-sea-level pressure weather map with on heights on an isobaric surface. [Hint: Consider
isobars around high- and low-pressure centers such eqs. (10.26) and (10.29).]
that the isobar packing matches the pressure gradi-
ent in that figure. E32. What aspects of the Approach to Geostrophy INFO
Box are relevant to the inertial wind? Discuss.
E22. Fig. 10.14 suggests that any pressure gradient is
theoretically possible adjacent to a low-pressure cen- E33. a. Do your own derivation for eq. (10.66) based
ter, from which we can further infer that any wind on geometry and mass continuity (total inflow = to-
speed is theoretically possible. For the real atmo- tal outflow).
sphere, what might limit the pressure gradient and b. Drag normally slows winds. Then why does
the wind speed around a low-pressure center? the updraft velocity increase in eq. (10.66) as drag
coefficient increases?
E23. Given the geopotential heights in Fig. 10.3, c. Factor b varies negatively with increasing drag
calculate the theoretical values for gradient and/or coefficient in eq. (10.66). Based on this, would you
geostrophic wind at a few locations. How do the ac- change your argument for part (b) above?
tual winds compare with these theoretical values?
E34. Look at each term within eq. (10.69) to justi-
E24. Eq. 10.39 is an implicit solution. Why do we fy the physical interpretations presented after that
say it is implicit? equation.
E25. Determine the accuracy of explicit eqs. (10.41)

by comparing their approximate solutions for ABL
E35. Consider eq. (10.70). For the internal Rossby S8. Suppose that wind speed M = cF/m, where c =
deformation radius, discuss its physical interpreta- a constant, m = mass, and F is force. Describe the
tion in light of eq. (10.71). resulting climate and weather.
E36. What type of wind would be possible if the S9. What if Earths axis of rotation was pointing di-
only forces were turbulent-drag and Coriolis. Dis- rectly to the sun. Describe the resulting climate and
cuss. weather.
E37. Derive equations for Ekman pumping around S10. What if there was no limit to the strength of
anticyclones. Physically interpret your resulting pressure gradients in highs. Describe the resulting
equations. climate, winds and weather.
E38. Rewrite the total deformation as a function of S11. What if both the ground and the tropopause
divergence and vorticity. Discuss. were rigid surfaces against which winds experience
turbulent drag. Describe the resulting climate and
weather.
Synthesize
S1. For zonal (east-west) winds, there is also a verti-
S12. If the Earth rotated half as fast as it currently
cal component of Coriolis force. Using your own di-
does, describe the resulting climate and weather.
agrams similar to those in the INFO box on Coriolis
Force, show why it can form. Estimate its magni-
S13. If the Earth had no rotation about its axis, de-
tude, and compare the magnitude of this force to
scribe the resulting climate and weather.
other typical forces in the vertical. Show why a ver-
tical component of Coriolis force does not exist for
S14. Consider the Coriolis-force INFO box. Create
meridional (north-south) winds.
an equation for Coriolis-force magnitude for winds
that move:
S2. On Planet Cockeyed, turbulent drag acts at right
a. westward b. southward
angles to the wind direction. Would there be any-
thing different about winds near lows and highs on
S15. What if a cyclostrophic-like wind also felt drag
Cockeyed compared to Earth?
near the ground? This describes conditions at the
bottom of tornadoes. Write the equations of mo-
S3. The time duration of many weather phenomena
tion for this situation, and solve them for the tan-
are related to their spatial scales, as shown by eq.
gential and radial wind components. Check that
(10.53) and Fig. 10.24. Why do most weather phe-
your results are reasonable compared with the pure
nomena lie near the same diagonal line on a log-log
cyclostrophic winds. How would the resulting
plot? Why are there not additional phenomena that
winds affect the total circulation in a tornado? As
fill out the relatively empty upper and lower tri-
discoverer of these winds, name them after your-
angles in the figure? Can the distribution of time
self.
and space scales in Fig. 10.24 be used to some ad-
vantage?
S16. What if F = ca, where c = a constant not equal
to mass, a = acceleration, and F is force. Describe the
S4. What if atmospheric boundary-layer drag were
resulting dynamics of objects such as air parcels.
constant (i.e., not a function of wind speed). De-
scribe the resulting climate and weather.
S17. What if pressure-gradient force acted parallel
to isobars. Would there be anything different about
S5. Suppose Coriolis force didnt exist. Describe the
our climate, winds, and weather maps?
resulting climate and weather.
S18. For a free-slip Earth surface (no drag), describe
S6. Incompressibility seems like an extreme simpli-
the resulting climate and weather.
fication, yet it works fairly well? Why? Consider
what happens in the atmosphere in response to
S19. Anders Persson discussed issues related to
small changes in density.
Coriolis force and how we understand it (see Weath-
er, 2000.) Based on your interpretation of his paper,
S7. The real Earth has locations where Coriolis force
can Coriolis force alter kinetic energy and momen-
is zero. Where are those locations, and what does
tum of air parcels, even though it is only an apparent
the wind do there?
force? Hint, consider whether Newtons laws would
be violated if your view these motions and forces

from a fixed (non-rotating) framework.
S20. If the Earth was a flat disk spinning about the

same axis as our real Earth, describe the resulting
climate and weather.
S21. Wind shear often creates turbulence, and tur-

bulence mixes air, thereby reducing wind shear.
Considering the shear at the ABL top in Fig. 10.7,
why can it exist without mixing itself out?
S22. Suppose there was not centrifugal or centrip-

etal force for winds blowing around lows or highs.
Describe the resulting climate, winds and weather.
S23. Suppose advection of the wind by the wind

were impossible. Describe the resulting climate and
weather.

Ch10 Forces Winds

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch10 Forces Winds

Uploaded by

Copyright:

Available Formats

Copyright 2015 by Roland Stull. Practical Meteorology: An Algebra-based Survey of Atmospheric Science.

10 Atmospheric Forces & Winds

Contents Winds power our wind turbines, push our sail-

of equal pressure. This is called an isobaric map.

In Cartesian coordinates (x, y, z), z is height above

word monotonic means that the value of the de-

face that connects points of equal pressure, such as

Isobaric surfaces can intersect the ground, but

speed units, such as knots, m s1, miles h1, km h1, etc.

assume knots the WMO standard. For unit conver-

Sample Application  

Draw wind barb symbol for winds from the:

(a) 115 knots = 2 pennants + 1 full barb + 1 half barb.  /

Check: Consistent with Table 10-1.

INFO Newtons Laws of Motion

In Chapter 1 we defined the (U, V, W) wind com-

Five forces contribute to net horizontal accelera-

Centrifugal force is an apparent force that allows us

Advection of Horizontal Momentum

For advection, the horizontal force components

1 1 ) Given: P =101 kPa @ x = 400 km (north of Des Moines).

If pressure increases toward one direction, then the

Find the Answer FPG 1 P (10.10)

where M = ( U2 + V2 )1/2 is wind speed (always posi- -PX-BUJUVEFT )JHI-BUJUVEFT

The units for are often abbreviated as s1. Using m

the Coriolis parameter depends only on latitude. Its

INFO On Coriolis Force INFO On 7

Objects at Rest with respect to Earths Surface

tor, FG does not point exactly to Earths center, due to .UBO

gravitational pull of the equatorial bulge.] '() 8p3

INFO On Coriolis Force (continuation) INFO Coriolis Force (continuation)

'$/) Objects moving south have a Coriolis force to

Northward-moving Objects FCF = FCNH FCNHR

/1 9 HIGHER MATH Apparent Forces

is 2 V , where vector points along the Earths

where wT is called a turbulent transport velocity.

Sample Application FTD M

Table 10-3. Summary of forces.

FTD M atm. boundary-layer wind

advection FAD U neither creates nor destroys

Equations of Horizontal Motion

large. Above the ABL, cumulus clouds and thunder-

Table 10-4. Names of idealized steady-state horizon-

Caution: Steady state means no further change to

Find the Answer 1 P

But P/x = 0 implies Vg = 0. For Ug, use eq. (10.26a): 1 P (10.26b)

Real winds are nearly geostrophic at locations

are packed closer together. The geostrophic wind

G = U g 2 + Vg 2 (10.27) Sample Application

If the geopotential = |g|z is substituted in

Fnet. This net force is called centripetal force, and is

Surprisingly, the winds never reach geostrophic 1 P VM

s1 to the north ( 1250 km in

Inertial oscillations are I

atmosphere. Why not? (1) IU = fc Mtan + tan (10.32)

Find the Answer

Use eq. (10.33), with Mtan = G: G = G G2/( fcR)

Eq. (10.33) is a quadratic equation that has two so-

and for winds around a high) the gradient wind is:

1  ) where G is the geostrophic wind.

& b) appear subtle at first glance, these differences

isobars are straight. The force balance at steady state

Namely, the only forces acting for this special case

In summary, both wind-shear turbulence and

Sample Application

(a) 115 knots = 2 pennants + 1 full barb + 1 half barb. /

11 ) Given: P =101 kPa @ x = 400 km (north of Des Moines).

The units for are often abbreviated as s1. Using m

tor, FG does not point exactly to Earths center, due to .UBO

Inertial oscillations are I

atmosphere. Why not? (1) IU = fc Mtan + tan (10.32)

1 ) where G is the geostrophic wind.

ABL Gradient (ABLG) Wind -

the final answer of

Scales of Horizontal Motion 5BCMF)PSJ[POUBMTDBMFTPGNPUJPOJOUIFUSPQPTQIFSF

Chapter 17 Regional Winds (meso & ) NPO

building blocks of hurricanes. Similarly, midlati- EBZ IVSSJDBOF