Skip to content

Thermodynamics of Dry and Moist Air and the Clausius–Clapeyron Equation

CSIR NET Part C meteorology: potential temperature, humidity variables, Clausius–Clapeyron, moist adiabats, equivalent potential temperature, tephigrams.

  • 18 min read
  • Updated
On this page · 14 sections
  1. Prerequisites
  2. The Hook: Why a Warmer World Is a Wetter World
  3. Dry Air as an Ideal Gas
  4. Water Vapour: The Humidity Variables
  5. The Clausius–Clapeyron Equation
  6. Moist Processes
  7. Thermodynamic Diagrams
  8. Memory Tricks
  9. Don’t Confuse
  10. Common Mistakes
  11. Practice
  12. Quick Revision Table
  13. Key Takeaways
  14. What’s Next

Prerequisites

The Hook: Why a Warmer World Is a Wetter World

Dry Air as an Ideal Gas

Equation of state and constants

$$p = \rho R_d T,\qquad R_d = 287\ \text{J kg}^{-1}\text{K}^{-1}$$
ConstantValue
\( c_p \) (dry air)1004 J kg\(^{-1}\) K\(^{-1}\)
\( c_v \) (dry air)717 J kg\(^{-1}\) K\(^{-1}\)
\( c_p - c_v \)\( R_d \)
\( \kappa = R_d/c_p \)0.286 (about 2/7, as for a diatomic ideal gas)
\( \gamma = c_p/c_v \)1.40

First law in meteorological form

Per unit mass, with specific volume \( \alpha = 1/\rho \):

$$\delta q = c_v\,dT + p\,d\alpha = c_p\,dT - \alpha\,dp$$

The second form is the useful one, because pressure is what a rising parcel follows.

Dry adiabatic processes and potential temperature

For an adiabatic process (\( \delta q = 0 \)): \( c_p\,dT = \alpha\,dp = R_dT\,dp/p \), so \( d\ln T = \kappa\,d\ln p \). Integrating gives Poisson’s equation:

$$\boxed{\theta = T\left(\frac{p_0}{p}\right)^{\kappa}},\qquad p_0 = 1000\ \text{hPa}$$

The potential temperature \( \theta \) is the temperature the parcel would have if brought dry-adiabatically to 1000 hPa. It is conserved in dry adiabatic motion.

Dry adiabatic lapse rate. For a parcel in hydrostatic surroundings (\( dp = -\rho g\,dz \), parcel pressure equals environment pressure):

$$c_p\,dT = -g\,dz \quad\Rightarrow\quad \boxed{\Gamma_d = \frac{g}{c_p} = 9.8\ \text{K km}^{-1}}$$

Entropy and potential temperature

The specific entropy of dry air is

$$ds = c_p\,d\ln\theta$$

So isentropic surfaces are surfaces of constant \( \theta \), and \( \theta \) is a measure of entropy. Adiabatic, frictionless flow moves along \( \theta \) surfaces; diabatic heating moves air across them. “Isentropic analysis” follows air on \( \theta \) surfaces.

Water Vapour: The Humidity Variables

Water vapour behaves as an ideal gas: \( e = \rho_v R_v T \), with vapour pressure \( e \) and \( R_v = 461.5 \) J kg\(^{-1}\) K\(^{-1}\). The ratio \( \varepsilon = R_d/R_v = 0.622 \).

VariableDefinitionFormulaTypical surface values
Vapour pressure \( e \)partial pressure of vapour5–35 hPa
Mixing ratio \( w \)mass of vapour / mass of dry air\( w = \varepsilon e/(p - e) \approx \varepsilon e/p \)1–25 g kg\(^{-1}\)
Specific humidity \( q \)mass of vapour / mass of moist air\( q = \varepsilon e/[p - (1-\varepsilon)e] = w/(1+w) \)almost equal to \( w \)
Absolute humidity \( \rho_v \)vapour density\( e/(R_vT) \)up to about 25 g m\(^{-3}\)
Relative humidity\( e/e_s \) (nearly equal to \( w/w_s \))0–100%
Dew point \( T_d \)temperature at which air, cooled at constant \( p \) and \( w \), saturates (over water)
Frost pointas dew point, over icehigher than \( T_d \) below 0 °C
Wet-bulb \( T_w \)temperature reached by evaporating water into the air at constant \( p \) until saturation\( T_d \le T_w \le T \)

\( w \) and \( q \) are conserved in unsaturated motion (no condensation, no mixing), so they label air parcels just as \( \theta \) does. Relative humidity is not conserved: it rises as air cools.

Virtual temperature

Moist air is lighter than dry air at the same \( T \) and \( p \) (molar mass 18 against 29). Instead of a variable gas constant, define

$$\boxed{T_v = T\,\frac{1 + w/\varepsilon}{1 + w} \approx T(1 + 0.61\,w)}$$

so that \( p = \rho R_d T_v \). \( T_v \) is the temperature dry air would need to have the same density. With \( T = 303.15 \) K and \( w = 12.7 \) g kg\(^{-1}\), \( T_v \approx 305.5 \) K: 2.3 K warmer. Buoyancy and the hypsometric equation should use \( T_v \).

The Clausius–Clapeyron Equation

Derivation in brief

For two phases in equilibrium, the Gibbs functions per unit mass are equal along the coexistence curve. Moving along it, \( dg_1 = dg_2 \), so \( -s_1\,dT + \alpha_1\,dp = -s_2\,dT + \alpha_2\,dp \). With \( s_2 - s_1 = L/T \):

$$\frac{de_s}{dT} = \frac{L}{T(\alpha_v - \alpha_l)}$$

the Clapeyron equation. Now neglect the liquid volume (\( \alpha_l \ll \alpha_v \)) and treat the vapour as ideal (\( \alpha_v = R_vT/e_s \)):

$$\boxed{\frac{de_s}{dT} = \frac{L\,e_s}{R_v T^2}}\qquad\text{or}\qquad \frac{d\ln e_s}{dT} = \frac{L}{R_vT^2}$$

This is the Clausius–Clapeyron equation.

Fractional rate of increase

\( T \)\( L/(R_vT^2) \)
273 Kabout 7.3% per K
288 Kabout 6.5% per K
300 Kabout 6.0% per K

So saturation vapour pressure rises roughly 7% per kelvin, a little less in the warm tropics. This “Clausius–Clapeyron scaling” is why atmospheric water vapour, and to a first approximation extreme rainfall, increase with warming.

Integrated forms

Assuming constant \( L = 2.5\times 10^6 \) J kg\(^{-1}\):

$$e_s(T) = e_s(T_0)\exp\left[\frac{L}{R_v}\left(\frac{1}{T_0} - \frac{1}{T}\right)\right],\qquad e_s(273.15\ \text{K}) = 6.11\ \text{hPa}$$

\( L/R_v \approx 5{,}420 \) K. Because \( L \) actually decreases slightly with temperature, practical work uses empirical fits. A widely used one (Bolton’s form of the Magnus formula), with \( T \) in °C and \( e_s \) in hPa:

$$e_s = 6.112\exp\left(\frac{17.67\,T}{T + 243.5}\right)$$
\( T \) (°C)−20010203035
\( e_s \) over water (hPa)1.266.1112.323.442.556.3

Rule of thumb: \( e_s \) roughly doubles for every 10 K rise.

Over ice versus over water

For the ice–vapour equilibrium the latent heat of sublimation \( L_s \approx 2.83\times 10^6 \) J kg\(^{-1}\) (\( = L_v + L_f \), with \( L_f \approx 3.34\times 10^5 \)) replaces \( L_v \). Below 0 °C, \( e_{s,ice} < e_{s,water} \): at −10 °C, about 2.60 hPa against 2.87 hPa, roughly 10% lower. Air saturated with respect to supercooled water is supersaturated with respect to ice, so ice crystals grow at the expense of droplets: the physical basis of the Wegener–Bergeron–Findeisen process (see Clouds, Precipitation and General Circulation).

The triple point of water is 273.16 K and 611.657 Pa (about 6.11 hPa); at temperatures below it, liquid water is metastable (supercooled).

Moist Processes

Ways to saturate air

ProcessWhat changesExample
Isobaric cooling\( T \) falls to \( T_d \) at constant \( w \)dew, radiation fog, advection fog
Adiabatic ascent\( T \) falls at \( \Gamma_d \); \( T_d \) falls only about 1.8 K km\(^{-1}\)cumulus base at the LCL
Evaporation into the air\( w \) rises, \( T \) falls to \( T_w \)steam fog, rain-cooled outflow
Mixing of two near-saturated air massescurvature of \( e_s(T) \)breath condensation, contrails, mixing fog

Mixing fog: because \( e_s(T) \) is convex, mixing two saturated masses at different temperatures gives a mixture with more vapour than the mixture’s saturation value. Equal parts of saturated air at 0 °C (6.11 hPa) and 20 °C (23.4 hPa) mix to about 10 °C with \( e \approx 14.7 \) hPa, above \( e_s(10\ °C) = 12.3 \) hPa: about 120% relative humidity, so condensation follows.

Lifting condensation level

As a parcel rises dry-adiabatically its temperature falls about 9.8 K km\(^{-1}\) and its dew point about 1.8 K km\(^{-1}\) (because \( w \) is conserved but \( p \) falls). The spread closes at about 8 K km\(^{-1}\):

$$z_{LCL} \approx 125\,(T - T_d)\ \text{m}$$

Normand’s rule: on a thermodynamic diagram, the dry adiabat through \( T \), the saturation mixing-ratio line through \( T_d \) and the moist adiabat through \( T_w \) all meet at the LCL.

Psychrometry

A ventilated wet-bulb thermometer gives humidity through the psychrometric equation:

$$e = e_s(T_w) - \gamma\,p\,(T - T_w)$$

with \( \gamma \) the psychrometer constant (about \( 6.6\times 10^{-4} \) K\(^{-1}\) for a well-ventilated instrument). Example: \( T = 30 \) °C, \( T_w = 24 \) °C, \( p = 1000 \) hPa gives \( e \approx 29.8 - 0.66\times 6 \approx 25.9 \) hPa and relative humidity about 61%.

The saturated (moist) adiabatic lapse rate

Once saturated, a rising parcel condenses vapour and releases latent heat. Writing the first law with condensation of \( dw_s \):

$$c_p\,dT + g\,dz = -L\,dw_s$$

and expanding \( dw_s \) using Clausius–Clapeyron gives

$$\boxed{\Gamma_s = \Gamma_d\,\frac{1 + \dfrac{L w_s}{R_d T}}{1 + \dfrac{\varepsilon L^2 w_s}{c_p R_d T^2}}}$$
Level\( T \)\( w_s \) (g kg\(^{-1}\))\( \Gamma_s \) (K km\(^{-1}\))
1000 hPa30 °C27.6about 3.5
1000 hPa20 °C14.9about 4.2
850 hPa10 °C9.1about 4.9
500 hPa−20 °C1.6about 7.7

\( \Gamma_s \) is smallest in warm, moist air and approaches \( \Gamma_d \) in cold air with little vapour. Since \( \Gamma_s \) varies, moist adiabats on a diagram are curved, unlike dry adiabats.

Reversible versus pseudo-adiabatic. In a reversible saturated process all condensate is carried with the parcel (its heat capacity counts); in a pseudo-adiabatic process condensate falls out immediately. The difference is small, and diagrams usually draw pseudo-adiabats.

Equivalent potential temperature

The equivalent potential temperature \( \theta_e \) is the potential temperature a parcel would reach if all its vapour were condensed (releasing latent heat) and it were then brought dry-adiabatically to 1000 hPa. Approximately:

$$\boxed{\theta_e \approx \theta\exp\left(\frac{L\,w}{c_p\,T_{LCL}}\right)}$$
  • \( \theta_e \) is conserved in both dry and saturated (pseudo-)adiabatic processes, so it is the natural label for moist convection.
  • For \( \theta = 300 \) K, \( w = 15 \) g kg\(^{-1}\) and \( T_{LCL} \approx 300 \) K (illustrative), \( \theta_e \approx 300\exp(0.1245) \approx 340 \) K.
  • High \( \theta_e \) in the boundary layer signals warm, moist air that can drive deep convection; its vertical profile is the basis of convective instability (see Vertical Stability and Tropical Convection).
  • The wet-bulb potential temperature \( \theta_w \) is the temperature where the moist adiabat through the parcel crosses 1000 hPa; it is conserved in the same processes and labels the moist adiabats on a tephigram.
VariableConserved inNot conserved in
\( T \)nothing in vertical motion
\( \theta \)dry adiabaticcondensation, radiation, mixing
\( w, q \)unsaturated adiabaticcondensation, evaporation
\( \theta_e, \theta_w \)dry and moist adiabaticradiation, surface fluxes, mixing
Moist static energy \( h = c_pT + gz + Lq \)dry and moist adiabatic (approximately), hydrostaticradiation, surface fluxes

Thermodynamic Diagrams

A thermodynamic diagram plots a sounding with dry adiabats, moist adiabats, isotherms, isobars and saturation mixing-ratio lines. The best diagrams are area-preserving: the area enclosed by a cycle is proportional to the work (energy) involved, so CAPE and CIN can be read as areas.

DiagramAxesArea-preserving?Notes
Tephigram\( T \) against \( \ln\theta \) (entropy), usually rotated about 45° so isobars run nearly horizontalyestraditional in the UK and many Commonwealth weather services; name from “T–\( \phi \)”, \( \phi \) for entropy
Skew-T log-pisotherms skewed 45°, \( -\ln p \) verticalyesused widely in the USA
Emagram\( T \) against \( -\ln p \)yes“energy per unit mass diagram”
Stüve diagram\( T \) against \( p^{\kappa} \)nodry adiabats are straight lines, but areas are not energy

On any of them: the LCL follows from Normand’s rule; the LFC is where the parcel’s moist adiabat crosses the sounding to the warm side; the equilibrium level where it crosses back; the positive area is CAPE.

Memory Tricks

Don’t Confuse

PairDistinction
Mixing ratio vs specific humidityper kg of dry air vs per kg of moist air; \( q = w/(1+w) \)
Relative humidity vs mixing ratiodepends on \( T \) (not conserved) vs conserved in unsaturated motion
Dew point vs wet-bulb temperaturecooling at constant \( w \) vs cooling by evaporating water into the air (\( w \) rises)
Virtual vs potential temperaturedensity-equivalent dry-air temperature vs temperature after dry-adiabatic return to 1000 hPa
Clapeyron vs Clausius–Clapeyronexact, with both phase volumes vs approximation: vapour ideal, liquid volume neglected
\( e_s \) over water vs over iceice value lower below 0 °C; uses \( L_s \)
Reversible vs pseudo-adiabaticcondensate kept vs condensate removed
Tephigram vs Stüvearea-preserving vs not

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. Write the Clausius–Clapeyron equation for vapour over water.
  2. Which temperature is conserved in both dry and saturated adiabatic processes?
  3. Give the relation between specific humidity and mixing ratio.
Answers
  1. \( de_s/dT = L e_s/(R_vT^2) \).
  2. Equivalent potential temperature \( \theta_e \) (and the wet-bulb potential temperature \( \theta_w \)).
  3. \( q = w/(1+w) \).

Level 2: Part B Style

Practice Q1. At 300 K, using \( L = 2.5\times 10^6 \) J kg\(^{-1}\) and \( R_v = 461.5 \) J kg\(^{-1}\) K\(^{-1}\), the fractional increase of saturation vapour pressure per kelvin is closest to:

(a) 3% (b) 6% (c) 9% (d) 12%

Answer: (b) 6%

\( L/(R_vT^2) = 2.5\times 10^6/(461.5\times 90{,}000) \approx 0.060 \).

Practice Q2. At 1000 hPa the vapour pressure is 20 hPa. The mixing ratio is closest to:

(a) 8.2 g kg\(^{-1}\) (b) 12.4 g kg\(^{-1}\) (c) 12.7 g kg\(^{-1}\) (d) 20 g kg\(^{-1}\)

Answer: (c) 12.7 g/kg

\( w = 0.622\times 20/(1000 - 20) = 0.01269 \). (b) is close to the approximate \( 0.622e/p = 12.4 \) g kg\(^{-1}\); the specific humidity is 12.5 g kg\(^{-1}\).

Practice Q3. Air at 30 °C has relative humidity 50%. Using \( e_s = 6.112\exp[17.67T/(T + 243.5)] \), its dew point is closest to:

(a) 15.0 °C (b) 18.5 °C (c) 21.0 °C (d) 24.0 °C

Answer: (b) 18.5 °C

\( e_s(30) = 42.46 \) hPa, \( e = 21.23 \) hPa. Invert: \( a = \ln(21.23/6.112) = 1.245 \); \( T_d = 243.5a/(17.67 - a) \approx 18.5 \) °C.

Practice Q4. Which thermodynamic diagram is not area-preserving?

(a) Tephigram (b) Emagram (c) Skew-T log-p (d) Stüve diagram

Answer: (d) Stüve

The Stüve diagram uses \( p^\kappa \) as ordinate so dry adiabats are straight, but equal areas do not represent equal energy.

Level 3: Part C Style

Practice Q5. Two saturated air samples, one at 0 °C (\( e_s = 6.11 \) hPa) and one at 20 °C (\( e_s = 23.37 \) hPa), are mixed in equal masses at constant pressure. Treat the mixing as linear in \( T \) and \( e \). Which are correct?

  1. The mixture has \( T \approx 10 \) °C and \( e \approx 14.7 \) hPa before any condensation.
  2. The mixture is supersaturated, with relative humidity about 120%.
  3. Condensation will release latent heat, so the final temperature will be slightly above 10 °C.
  4. If the two samples were each at 50% relative humidity, the mixture would still be supersaturated.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 1, 2, 3 and 4 (d) 2 and 4 only

Answer: (b) 1, 2 and 3 only
  • 1 true: \( (0 + 20)/2 = 10 \) °C; \( (6.11 + 23.37)/2 = 14.74 \) hPa.
  • 2 true: \( e_s(10\ °C) = 12.27 \) hPa; \( 14.74/12.27 \approx 1.20 \).
  • 3 true: condensing the excess vapour warms the mixture a little, which raises \( e_s \) and reduces the excess until equilibrium.
  • 4 false: the mixture would have \( e \approx 7.4 \) hPa, about 60% relative humidity: unsaturated. Mixing fog needs both samples close to saturation.

Practice Q6. A ventilated psychrometer at 1000 hPa reads dry bulb 30 °C and wet bulb 24 °C. Take \( e_s(24\ °C) = 29.83 \) hPa, \( e_s(30\ °C) = 42.46 \) hPa and psychrometer constant \( 6.6\times 10^{-4} \) K\(^{-1}\). Which are correct?

  1. The vapour pressure is about 25.9 hPa.
  2. The relative humidity is about 61%.
  3. The mixing ratio is about 16.5 g kg\(^{-1}\).
  4. The dew point lies between 24 and 30 °C.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 1, 2, 3 and 4 (d) 2 and 4 only

Answer: (b) 1, 2 and 3 only
  • 1 true: \( 29.83 - 6.6\times 10^{-4}\times 1000\times 6 = 29.83 - 3.96 = 25.87 \) hPa.
  • 2 true: \( 25.87/42.46 \approx 0.61 \).
  • 3 true: \( 0.622\times 25.87/(1000 - 25.87) \approx 0.0165 \).
  • 4 false: the dew point is always at or below the wet-bulb temperature (here about 21.6 °C).

Practice Q7. Use \( \Gamma_s = \Gamma_d(1 + Lw_s/R_dT)/(1 + \varepsilon L^2w_s/c_pR_dT^2) \) with \( \Gamma_d = 9.77 \) K km\(^{-1}\). Which ranking of the saturated lapse rate is correct?

(a) 1000 hPa, 30 °C > 850 hPa, 10 °C > 500 hPa, −20 °C (b) 500 hPa, −20 °C > 850 hPa, 10 °C > 1000 hPa, 30 °C (c) 850 hPa, 10 °C > 1000 hPa, 30 °C > 500 hPa, −20 °C (d) all three are equal to within 0.5 K km\(^{-1}\)

Answer: (b)

Computed values: about 7.7 (500 hPa, −20 °C, \( w_s \approx 1.6 \) g kg\(^{-1}\)), about 4.9 (850 hPa, 10 °C, \( w_s \approx 9.1 \) g kg\(^{-1}\)) and about 3.5 K km\(^{-1}\) (1000 hPa, 30 °C, \( w_s \approx 27.6 \) g kg\(^{-1}\)). More vapour to condense means more latent heating and a smaller lapse rate.

Quick Revision Table

IdeaFact or formula
Constants\( R_d = 287 \), \( R_v = 461.5 \), \( c_p = 1004 \), \( c_v = 717 \) J kg\(^{-1}\) K\(^{-1}\); \( \kappa = 0.286 \); \( \varepsilon = 0.622 \)
First law\( \delta q = c_p\,dT - \alpha\,dp \)
Potential temperature\( \theta = T(1000/p)^{0.286} \); \( ds = c_p\,d\ln\theta \)
Dry adiabatic lapse rate\( g/c_p = 9.8 \) K km\(^{-1}\)
Humidity\( w = \varepsilon e/(p-e) \); \( q = w/(1+w) \); \( T_v \approx T(1 + 0.61w) \)
Clausius–Clapeyron\( d\ln e_s/dT = L/(R_vT^2) \): 7%/K at 273 K, 6%/K at 300 K
\( e_s \) values6.11 (0 °C), 12.3 (10), 23.4 (20), 42.5 hPa (30)
Latent heats\( L_v = 2.5\times 10^6 \), \( L_f = 3.34\times 10^5 \), \( L_s = 2.83\times 10^6 \) J kg\(^{-1}\)
Ice vs water\( e_{s,ice} < e_{s,water} \) below 0 °C (about 10% at −10 °C): Bergeron process
LCL\( 125(T - T_d) \) m; Normand’s rule
Psychrometer\( e = e_s(T_w) - \gamma p(T - T_w) \)
Moist lapse rateabout 3.5 (warm, humid) to 7–9 K km\(^{-1}\) (cold)
\( \theta_e \)\( \theta\exp(Lw/c_pT_{LCL}) \); conserved dry and moist
Diagramstephigram, emagram, skew-T are area-preserving; Stüve is not

Key Takeaways

What’s Next