Fundamentals for Atmospheric Instability

1. Buoyancy and Static Stability

a). Buoyancy
When we discuss scale analysis for large-scale motions in the Tropics,
we assumed the motion is in quasi-hydrostatic balance. However, this
assumption is not valid for small scale motions, such as in convective
clouds. To describe the motion of cumulus convection, the full vertical
momentum equation needs to be considered. In the absence of friction, the
vertical component of the momentum equation is
Dw
1 p

g
Dt
z

(2.1)

Assuming a basic state that is in motionless and thus horizontally uniform,

we then can write the pressure and density in the form
p ( x, y , z , t ) p 0 ( z ) p ' ( x, y , z , t )

( x, y , z , t ) 0 ( z ) ' ( x, y , z , t )

(2.2)

where p0 and 0 are unperturbed pressure and density, and p and are
deviations from the basic state values of p0 and 0. For an atmosphere at rest,
p and would be zero. Without any perturbations, the atmosphere is
generally in hydrostatic, that means
1 p 0
g
0 z

(2.3)

Assuming that / <<1 so that (0+)-10-1(1-/0), we can find that

Dw
1 p
1 p'

g
B
Dt
z
0 z

(2.4)

where B is the so-called buoyancy and is defined as

B g

'
0

(2.5)

For moist air, the state equation is

p Rd Tv

(2.6)
1

where Tv=T(1+0.608qv), the virtual temperature and qv is the mixing ratio of

water vapor. When the air is saturated and contains cloud particles
hydrometeors, the total density becomes
a (1 q H )

(2.7)

where qH is the total mixing ratio of all hydrometeors.

With all above factors included, the final buoyancy for a cloudy air
can be written as
T'
p'
B g

0.608q v q H
T0 p 0

(2.8)

We can see that positive temperature perturbation, negative pressure

perturbation, and any moisture in the air contribute to positive buoyancy,
while hydrometeors in the cloud reduce the buoyancy.
b). Static stability
Static stability is a measure of the vertical acceleration of an air parcel
displaced adiabatically in the vertical from its equilibrium level. The static
stability is determined by the stratification of temperature of the
environment. A critical assumption for the static stability is the small vertical
displacement of a parcel without disturbing its environment, and thus p is
assumed zero in (2.4) and (2.8).
i) Unsaturated air
Further assuming the dry motion, we can obtain
Dw
T'
'
g
g
Dt
T0
0

(2.9)

Since the air parcel is assumed to be displaced dry adiabatically in the

vertical, it follows
( z0 z ) ( z0 ),

and

T ( z 0 z ) T ( z 0 ) d z ,

(2.10)

where d=g/Cp=9.8K/km is the adiabatic lapse rate. z is the small vertical

displacement of the parcel from its equilibrium level where we have
( z 0 ) 0 ( z 0 ) and

T(z0)=T0(z0). For the environment, we can have

0
z , and T0 ( z 0 z ) T0 ( z 0 ) z , (2.11)
z

0 ( z0 z ) 0 ( z0 )

where =-T0/z is the lapse rate of the environment. Substituting (2.10) and
(2.11) into (2.9) and noting that T=T-T0, =-0, and w=Dz/Dt, we can get
ln 0
D2
g
(z )
( d )z g
z N 2z ,
2
Dt
T0
z

(2.12)

The solution of (2.12) is either exponential (when N2>0) or oscillatory

(when N2<0). Since for the oscillatory solution, the frequency equals exactly
to N, N is thus usually called buoyancy frequency, or often referred to as
Brunt-Visl frequency.
N2

ln 0
g
(d ) g
T0
z

(2.13)

In (2.12), we refer
N 2 0,

, Statically stable
N 0,
d , Statically neutral
(2.14)
2
N 0, d , Statically unstable
Physically, when N2>0, if a parcel of unsaturated air is raised slightly
2

upward adiabatically, its temperature is lower than the environmental

temperature, the parcel tends to return to its original (Fig. 2.1.a) (indeed,
because of the inertial, the air parcel will oscillate around its original level,
namely a buoyancy oscillation). However, when N2<0, the air parcel will be
warmer than its environment and thus will tend to further accelerate away
from its original level (Figure 2.1.b), namely statically unstable.

Figure 2.1. Conditions for (a) statically stable and (b) statically unstable stratifications
for unsaturated air (from Wallace and Hobb 1977).

ii). Saturated air

When the air is saturated, the lapse rate d in (2.12) should be
replaced by the so-called pseudoadiabatic lapse rate. In this case a saturated
parcel is assumed to undergo pseudoadiabatic ascent. Following the
saturated parcel, the lapse rate can be written as (See Appendix D for detail
in Holton 1992).
s

1 Lc q s /( Rd T )
dT
d
dz
1 0.622 L2c q s /(C p Rd T 2 ) ,

(2.15)

s is the pseudoadiabatic lapse rate. It explains the rate of change of

temperature with respect to height for a saturated parcel undergoing
pseudoadiabatic ascent. Namely, when a saturated air parcel ascends, it will
remain saturated and the excess water vapor will be condensed and fallen
out immediately from the parcel while the latent heat will warm the parcel.
As a result, s is always less than d because of the latent heat release in the
parcel. Observed values of s range from ~4 K km-1 in warm humid air
masses in the lower troposphere to ~6-7 K km-1 in the mid-troposphere. The
static stability of the saturated air thus is similar to the unsaturated case as
shown schematically in Figure 2.1 but the d needs to be replaced by s.

2. Conditional and Convective Instability

a). Conditional instability
Static stability discussed above applied to either unsaturated or
saturated. In case the lapse rate lies between dry adiabatic and
pseudoadiabatic values, namely,

s d ,

the atmosphere is stably

stratified with respect to dry adiabatic displacements but unstable with

respect to pseudoadiabatic displacements. Such a situation is referred to as
conditional instability, namely, the instability is conditional to saturation of
air parcel.

Figure 2.2. Conditions for conditional instability. s and d are the saturated and dry
adiabatic lapse rates and is the actual lapse rate of temperature of the air (From
Wallace and Hobb 1977) .

As illustrated in Figure 2.2, a parcel lifted from its equilibrium level at

O cools dry adiabatically until it reaches its lifting condensation level (LCL)
at A. At this level, it is colder than its environmental temperature. Further
lifting produces cooling at the moist adiabatic lapse rate, so the temperature
of the air parcel follows the moist adiabatic lapse rate ABC. If the air parcel
is sufficiently moist and is forced to continue to rise, it will reach a level at
B above which it becomes buoyant relative to its surroundings. Once above
5

level B, the parcel can be freely accelerate upward even in the absence of
further lifting. For this reason, the level B is referred to as the level of free
convection (LFC).
b). Convective instability
Convective instability is different from the static stability and
conditional instability in that it is an instability associated with the lifting of
a layer of air instead of a parcel. It is related to not only the temperature
stratification but also the vertical stratification of water vapor. In the profile
shown in Figure 2.3, the mixing ratio decreases rapidly with height within
the inversion layer AB which marks the top of a moist layer.

Figure 2.3. Conditions for convective instability. T and Td are the temperature and dew
point of the air.

Now suppose that this layer is lifted. The air parcel at A will reach its
LCL almost immediately, and beyond LCL it will cool moist adiabatically.
On the other hand, an air parcel starting at point B will cool dry adiabatically
through a deep layer before it reaches its LCL. Therefore, as the inversion
layer is lifted, the top part cools much more rapidly than the bottom part and
the lapse rate of the layer quickly becomes destabilized. Sufficient lifting

may cause the layer to become conditionally unstable, even if the entire
sounding is absolutely stable to begin with. The instability resulting from
this layer lifting is referred as to convective instability or potential
instability. The criterion for convective instability is

d e
0
dz

in the layer,

where e is the equivalent potential temperature and is defined as

Lq
e exp c s
C T
p

(2.16)

Equivalent potential temperature is the potential temperature that a

parcel of air would have if all its moisture were condensed and the resultant
latent heat used to warm the parcel. The temperature of an air parcel can be
brought to its equivalent potential value by raising the parcel from its
original level until all water vapor in the parcel has condensed and fallen out
and then compressing the parcel adiabatically to a pressure of 1000 hPa.
Because the condensed water is assumed to fall out, the temperature increase
during the compression will be at the dry adiabatic rate and the parcel will
arrive back at its original level with a temperature that is higher than its
original temperature. Thus, the process is irreversible and called a
pseudoadiabatic process.
Note that the expression (2.16) may also be used to compute e for an
unsaturated parcel provided that the temperature T used in the formula is the
temperature that the parcel would have if expanded adiabatically to
saturation, namely the temperature at LCL (TLCL) and the saturated mixing
ratio is replaced by the actual mixing ratio of the initial state. Thus,
equivalent potential is conserved for a parcel during both dry adiabatic and
pseudoadiabatic ascents.
3. Convective Available Potential Energy (CAPE) and Convective
Inhibition (CIN)

a). Convective available potential energy (CAPE)

Development of convective storms depends on the presence of
environmental conditions favorable for the occurrence of deep convection. A
useful measure of the susceptibility of a given temperature and moisture
profile to the occurrence of deep convection is the convective available
potential energy (CAPE). CAPE provides a measure of the maximum
possible kinetic energy that a statically unstable parcel can acquire,
assuming that the parcel ascends without mixing with the environment and
adjusts instantaneously to the local environmental pressure. The vertical
momentum equation for such a parcel is (2.4), which can be rewritten
following the vertical motion of the parcel as
Dw Dz Dw
Dw

w
B
Dt
Dt Dz
Dz

(2.17)

where B is the buoyancy given by

Bg

env parcel
parcel

T parcel Tenv

(2.18)

Tenv

where subscript env designates the parameter of the environment while

parcel that of the parcel. Integrating (2.18) vertically from the LFC, zLFC to
the level of neutral buoyancy, zLNB following the motion of the parcel, we can
get the so-called CAPE as
CAPE

2
z LNB T parcel Tenv
wmax
g
z LFC
2
Tenv

dz

(2.19)

Therefore, CAPE is the maximum kinetic energy per unit mass that a
buoyant parcel could obtain by ascending from a state of rest at the level of
free convection to the level of neutral buoyancy near the tropopause. The
unit of CAPE is thus m2 s-2 or J kg-1.
In a typical tropical oceanic sounding, parcel temperature excesses of
1-2 K may occur over a depth of 10-12 km. A typical value of CAPE is then
about 500 m2 s-2. In severe storm conditions in the Midwest of North
America, however, parcel temperature excesses can be 7-10 K and CAPE

can be 2000-3000 m2 s-2. Observed updrafts in the latter case (up to 50 m/s)
are much stronger than in the former case (5-10 m/s). The small value of
CAPE in the mean tropical environment is the major reason that updraft
velocities in tropical cumulonimbus are observed to be much smaller than
those in midlatitude thunderstorms.

Figure 2.5. buoyancy profiles for a lifted parcel of mixed air from the lowest 50 hPa in a
COARE mean sounding. Heavy line is for undiluted ascent with condensed water content
truncated at 3 g km-1, representing precipitation processes, and latent heat of
sublimation added for water condensed above the 0oC (550 hPa) level. Light line above
this level is without sublimation. Light lines below are for continuously entraining
parcels with mixing rates of (from right to left) 15%, 30%, 45% of parcel mass per 100hPa ascent. Dashed line is for pseudoadiabatic ascent (no condensate loading allowed).
(from Mapes 2000).

Note that the CAPE is an overestimate of the actual kinetic energy

realized by a nonentraining parcel, as the negative buoyancy contribution of
liquid water reduces the effective buoyancy, especially in the tropics. In
several applications, some modifications have been made to (2.19) to
include the effect of water vapor and condensed water loading on buoyancy
and the effect of ice phase on the pseudoadiabatic ascent (Williams and
Renno 1993; Doswell and Rasmussen 1994). Note that in some practical
applications, some just use the temperature difference as the buoyancy,
9

which thus has a unit of K or oC (e.g., Xu and Emanuel 1989). There have
been also some discussions on the effect of irreversible or reversible parcel
assumptions. These complications can be referred to some references
provided. One example is given in Figure 2.5, which shows buoyancy
profiles for a lifted parcel of mixed air from the lowest 50 hPa in a COARE
mean sounding. What we see here is that the buoyancy cab be largely
affected by assumptions and both water loading and ice cloud microphysical
processes. Further the effect by entrainment of environmental are is
tremendous and is the topic to be discussed in our next chapter.
b). Convective inhibition (CIN)
When we discuss the CAPE, we only took into account the positive
buoyancy and did not consider how large forcing is required to lift the air
parcel to the LFC. This is measured by another parameter called convective
inhibition (CIN), which measures the energy required to lift a parcel through
the stable layer to its LFC. CIN thus equals the negative work done by the
mean atmospheric boundary layer as the undiluted parcel rises through the
stable layer to its LFC.
CIN

z LFC
T T parcel
w 2 0 max
g env
dz
z0
2
Tenv

(2.20)

where z0 is the level near the surface where the air parcel starts to be
negatively buoyant. Similar to (2.19), CIN is the maximum kinetic energy
initially required for a parcel to penetrate the stable layer and reach the LFC.
It is a good indication of how strong large-scale forcing is required to allow
the deep convection to occur. In Figure 2.6 we can see the negatively
buoyancy in the lowest layer of the atmosphere. Another example is given is
Figure 2.7 in which large CIN is below CAPE in the left panel.

10

Figure 2.6. Lower portion of skew-T-logp diagram illustrating (a) elevated environment
(no surface-based CAPE), and (b) environment with positive surface-based CAPE above
CIN. Thermodynamic profiles of dewpoint and temperature are shown as thick solid
line/curves. Important features are labeled with arrows. Curves with thick round dots
show surface parcel ascent. Curve with thick square dots in (a) is elevated lifted parcel
ascent. Hatching in (b) denotes negative area of surface-based CIN. The virtual
temperature correction is not included for simplicity. (from Davies 2004).

References
Wallace, J.M., and P.V. Hobb, 1977: Atmospheric Science-An Introductory Survey,
Academic Press. (Chapter 2.7 The concept of static stability).
Holton, J.R., 1992: An Introduction to Dynamic Meteorology, Fourth Edition, Elsevier
Academic Press. (Chapter 9.5 Cumulus Convection).
Doswell III, C.A., and E.N. Rasmussen, 1994: The effect of neglecting the virtual
temperature correction on CAPE calculations. Wea. Forecasting, 9, 625-629.
Williams, E., and N. Renno, 1993: An analysis of the conditional instability of the
tropical atmosphere. Mon. Wea. Rev., 121, 21-36.
Xu, K.-M., and K.A. Emanuel, 1989: Is the troposphere conditionally unstable? Mon.
Wea. Rev., 117, 1471-1479.
Mapes, B.E., 2000: Convective inhibition, subgrid-scale triggering energy, and
stratifiform instability in a toy tropical wave model. J. Atmos. Sci., 57, 15151535.
Critical thinking: (1) to explain why the buoyancy profiles are so different in Figure 2.6;
(2) do you expect strong convection for soundings in Figure 2.7 and explain why.

11