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
On this page · 14 sections
Prerequisites
- First law, entropy, heat capacities and phase equilibrium, from Laws of Thermodynamics.
- Hydrostatic balance and the hypsometric equation, from Composition, Thermal Structure, Hydrostatics and the Standard Atmosphere.
- Descriptive humidity and precipitation, from Humidity, Precipitation and Water Balance.
The Hook: Why a Warmer World Is a Wetter World
Dry Air as an Ideal Gas
Equation of state and constants
$$p = \rho R_d T,\qquad R_d = 287\ \text{J kg}^{-1}\text{K}^{-1}$$| Constant | Value |
|---|---|
| \( c_p \) (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 \( \varepsilon = R_d/R_v = 0.622 \).
| Variable | Definition | Formula | Typical surface values |
|---|---|---|---|
| Vapour pressure \( e \) | partial pressure of vapour | 5–35 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 point | as dew point, over ice | higher 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 K | about 7.3% per K |
| 288 K | about 6.5% per K |
| 300 K | about 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) | −20 | 0 | 10 | 20 | 30 | 35 |
|---|---|---|---|---|---|---|
| \( e_s \) over water (hPa) | 1.26 | 6.11 | 12.3 | 23.4 | 42.5 | 56.3 |
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
| Process | What changes | Example |
|---|---|---|
| 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 masses | curvature 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 hPa | 30 °C | 27.6 | about 3.5 |
| 1000 hPa | 20 °C | 14.9 | about 4.2 |
| 850 hPa | 10 °C | 9.1 | about 4.9 |
| 500 hPa | −20 °C | 1.6 | about 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.
Related conserved variables
| Variable | Conserved in | Not conserved in |
|---|---|---|
| \( T \) | nothing in vertical motion | |
| \( \theta \) | dry adiabatic | condensation, radiation, mixing |
| \( w, q \) | unsaturated adiabatic | condensation, evaporation |
| \( \theta_e, \theta_w \) | dry and moist adiabatic | radiation, surface fluxes, mixing |
| Moist static energy \( h = c_pT + gz + Lq \) | dry and moist adiabatic (approximately), hydrostatic | radiation, 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.
| Diagram | Axes | Area-preserving? | Notes |
|---|---|---|---|
| Tephigram | \( T \) against \( \ln\theta \) (entropy), usually rotated about 45° so isobars run nearly horizontal | yes | traditional in the UK and many Commonwealth weather services; name from “T–\( \phi \)”, \( \phi \) for entropy |
| Skew-T log-p | isotherms skewed 45°, \( -\ln p \) vertical | yes | used widely in the USA |
| Emagram | \( T \) against \( -\ln p \) | yes | “energy per unit mass diagram” |
| Stüve diagram | \( T \) against \( p^{\kappa} \) | no | dry 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
| Pair | Distinction |
|---|---|
| Mixing ratio vs specific humidity | per kg of dry air vs per kg of moist air; \( q = w/(1+w) \) |
| Relative humidity vs mixing ratio | depends on \( T \) (not conserved) vs conserved in unsaturated motion |
| Dew point vs wet-bulb temperature | cooling at constant \( w \) vs cooling by evaporating water into the air (\( w \) rises) |
| Virtual vs potential temperature | density-equivalent dry-air temperature vs temperature after dry-adiabatic return to 1000 hPa |
| Clapeyron vs Clausius–Clapeyron | exact, with both phase volumes vs approximation: vapour ideal, liquid volume neglected |
| \( e_s \) over water vs over ice | ice value lower below 0 °C; uses \( L_s \) |
| Reversible vs pseudo-adiabatic | condensate kept vs condensate removed |
| Tephigram vs Stüve | area-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
- Write the Clausius–Clapeyron equation for vapour over water.
- Which temperature is conserved in both dry and saturated adiabatic processes?
- Give the relation between specific humidity and mixing ratio.
Answers
- \( de_s/dT = L e_s/(R_vT^2) \).
- Equivalent potential temperature \( \theta_e \) (and the wet-bulb potential temperature \( \theta_w \)).
- \( q = w/(1+w) \).
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?
- The mixture has \( T \approx 10 \) °C and \( e \approx 14.7 \) hPa before any condensation.
- The mixture is supersaturated, with relative humidity about 120%.
- Condensation will release latent heat, so the final temperature will be slightly above 10 °C.
- 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?
- The vapour pressure is about 25.9 hPa.
- The relative humidity is about 61%.
- The mixing ratio is about 16.5 g kg\(^{-1}\).
- 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
| Idea | Fact 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 |
| \( e_s \) values | 6.11 (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 rate | about 3.5 (warm, humid) to 7–9 K km\(^{-1}\) (cold) |
| \( \theta_e \) | \( \theta\exp(Lw/c_pT_{LCL}) \); conserved dry and moist |
| Diagrams | tephigram, emagram, skew-T are area-preserving; Stüve is not |
Key Takeaways
What’s Next
- Vertical Stability and Tropical Convection: use \( \theta \), \( \theta_e \) and the moist adiabat to decide whether air rises.
- Clouds, Precipitation and General Circulation: Köhler theory and precipitation growth.
- Radiation Laws, Scattering, Radiative Transfer and the Greenhouse Effect: water vapour as a feedback.
- Back to the chapter overview.