Atmospheric Science

The Physics of Weather, Climate, and Planetary Atmospheres

Course Overview

Atmospheric science is the study of Earth's atmosphere and its processes, including weather, climate, and air quality. This interdisciplinary field combines physics, chemistry, and mathematics to understand atmospheric phenomena from local weather to global climate change.

100 km

Atmosphere Height

5.15×10¹⁸

kg Total Mass

78%

Nitrogen (N₂)

21%

Oxygen (O₂)

Atmospheric Layers

🌡️ Troposphere (0-12 km)

Contains 80% of atmospheric mass. All weather occurs here. Temperature decreases with altitude (~6.5°C/km). Ends at the tropopause.

🛡️ Stratosphere (12-50 km)

Contains the ozone layer. Temperature increases with altitude due to UV absorption. Very stable with little vertical mixing.

☄️ Mesosphere (50-85 km)

Temperature decreases with altitude. Meteors burn up here. Coldest temperatures in atmosphere (~-90°C at mesopause).

⚡ Thermosphere (85-600 km)

Temperature increases dramatically due to solar radiation absorption. Contains ionosphere. Aurora occur here.

Course Contents

🌍

Part 1: Atmosphere Fundamentals

Structure, composition, and solar radiation

🌡️

Part 2: Atmospheric Thermodynamics

Heat, moisture, stability, and adiabatic processes

💨

Part 3: Atmospheric Dynamics

Pressure, wind, geostrophic balance, boundary layer

🌀

Part 4: Synoptic Meteorology

Air masses, fronts, cyclones, and jet streams

☁️

Part 5: Clouds & Precipitation

Cloud physics, precipitation, severe weather

☀️

Part 6: Radiation & Climate

Energy budget, greenhouse effect, climate models

📡

Part 7: Observation & Prediction

Remote sensing, NWP, data assimilation

⚗️

Part 8: Atmospheric Chemistry

Air quality, aerosols, ozone, chemical reactions

🌡️

Part 9: Climate Science

Paleoclimate, climate change, extreme events

🪐

Part 10: Planetary Atmospheres

Mars, Venus, giant planets, exoplanets

Fundamental Equation Derivations

Complete step-by-step derivations of the core equations governing atmospheric physics, thermodynamics, dynamics, and radiative transfer.

1. Ideal Gas Law for Dry Air

1.1 Kinetic Theory Foundation

The ideal gas law is not merely an empirical rule; it emerges directly from the kinetic theory of gases. Consider a gas of $N$ identical molecules, each of mass $m$, contained in a cubic box of side $L$. A single molecule moving with $x$-velocity $v_x$ bounces off one wall, reversing its momentum. The impulse per collision is $\Delta p_{\text{mom}} = 2m v_x$. The molecule returns to the same wall after travelling a round trip of $2L$, so the collision rate is $v_x / (2L)$. The average force from this one molecule on the wall is therefore:

$$F_1 = \frac{2m v_x \cdot v_x}{2L} = \frac{m v_x^2}{L}$$

Summing over all $N$ molecules and noting that the wall area is $A = L^2$:

$$P = \frac{F}{A} = \frac{1}{L^3}\sum_{i=1}^{N} m\,v_{x,i}^2 = \frac{Nm\langle v_x^2\rangle}{V}$$

Since molecular motion is isotropic, $\langle v_x^2\rangle = \langle v_y^2\rangle = \langle v_z^2\rangle = \tfrac{1}{3}\langle v^2\rangle$, so the pressure becomes:

$$P = \frac{1}{3}\frac{Nm\langle v^2\rangle}{V} = \frac{1}{3}\frac{n_{\text{mol}}\,M\,\langle v^2\rangle}{V}$$

where $n_{\text{mol}}$ is the number of moles and $M$ is the molar mass. The equipartition theorem tells us that each translational degree of freedom carries $\tfrac{1}{2}k_B T$ of kinetic energy, so for three degrees of freedom: $\tfrac{1}{2}m\langle v^2\rangle = \tfrac{3}{2}k_B T$. Substituting $m\langle v^2\rangle = 3k_B T$:

$$P = \frac{N \cdot k_B T}{V} \quad\Longrightarrow\quad PV = Nk_BT$$

Since $N = n_{\text{mol}} N_A$ and $R^* = N_A k_B$, this is exactly the universal ideal gas law:

$$PV = n_{\text{mol}} R^* T$$

1.2 From Universal to Atmospheric Form

The universal ideal gas law for $n$ moles of gas in volume $V$ is:

$$pV = nR^*T$$

where $R^* = 8.314\;\text{J mol}^{-1}\text{K}^{-1}$ is the universal gas constant. Dividing both sides by volume and writing $n/V = \rho/M_d$ where$M_d = 28.97 \times 10^{-3}\;\text{kg mol}^{-1}$ is the mean molar mass of dry air:

$$p = \frac{\rho}{M_d}R^*T$$

Defining the specific gas constant for dry air:

$$R_d \equiv \frac{R^*}{M_d} = \frac{8.314}{0.02897} = 287.05\;\text{J kg}^{-1}\text{K}^{-1}$$

we arrive at the fundamental equation of state for dry air:

$$\boxed{p = \rho R_d T}$$

Dimensional check: $[\rho R_d T] = \text{kg m}^{-3} \times \text{J kg}^{-1}\text{K}^{-1} \times \text{K} = \text{J m}^{-3} = \text{Pa}$. Correct.

1.3 Extension to Moist Air: Virtual Temperature

The real atmosphere contains water vapour. By Dalton's law, the total pressure is the sum of the partial pressures of dry air and water vapour:

$$p = p_d + e$$

where $e$ is the vapour pressure. Each component satisfies its own equation of state:$p_d = \rho_d R_d T$ and $e = \rho_v R_v T$, where$R_v = R^*/M_v = 8.314/0.01802 = 461.5\;\text{J kg}^{-1}\text{K}^{-1}$ is the specific gas constant for water vapour. The total density is $\rho = \rho_d + \rho_v$. We want to write$p = \rho R_d T_v$ for some "virtual" temperature $T_v$. Starting from:

$$p = \rho_d R_d T + \rho_v R_v T = \left(\rho_d R_d + \rho_v R_v\right)T$$

Dividing by $\rho R_d$:

$$\frac{p}{\rho R_d} = T\left(\frac{\rho_d}{\rho} + \frac{\rho_v}{\rho}\frac{R_v}{R_d}\right) = T\left(1 - q + q\frac{R_v}{R_d}\right)$$

where $q = \rho_v/\rho$ is the specific humidity. Defining$\epsilon = M_v/M_d = R_d/R_v = 18.015/28.97 = 0.622$:

$$T_v = T\left(1 - q + \frac{q}{\epsilon}\right) = T\left(1 + \frac{1-\epsilon}{\epsilon}\,q\right) = T\left(1 + \frac{q}{0.622} - q\right)$$

Since $(1-\epsilon)/\epsilon = 0.378/0.622 \approx 0.608$:

$$\boxed{T_v = T(1 + 0.608\,q) \qquad\text{and}\qquad p = \rho R_d T_v}$$

Physically, $T_v > T$ because water vapour ($M_v = 18$) is lighter than dry air ($M_d = 29$). Moist air at temperature $T$ behaves like dry air at the higher temperature $T_v$. For typical tropospheric moisture ($q \approx 10\;\text{g/kg}$), $T_v - T \approx 1.8\;\text{K}$. In the tropical boundary layer with $q \approx 20\;\text{g/kg}$, this correction reaches $\sim 3.5\;\text{K}$ and cannot be neglected.

1.4 Numerical Example: Air Density at Sea Level

At sea level in the International Standard Atmosphere, $T = 288.15\;\text{K}$ and$p = 101\,325\;\text{Pa}$. Solving for density:

$$\rho = \frac{p}{R_d T} = \frac{101\,325}{287.05 \times 288.15} = \frac{101\,325}{82\,705} = 1.225\;\text{kg m}^{-3}$$

This is the standard value used widely in aerodynamics and meteorology. For moist air on a hot, humid day ($T = 305\;\text{K}$, $q = 0.020\;\text{kg/kg}$):

$$T_v = 305(1 + 0.608 \times 0.020) = 305 \times 1.0122 = 308.7\;\text{K}$$

$$\rho = \frac{101\,325}{287.05 \times 308.7} = \frac{101\,325}{88\,590} = 1.144\;\text{kg m}^{-3}$$

The hot, moist air is about 6.6% less dense than standard air — a significant buoyancy effect that drives tropical convection.

1.5 Limitations and Breakdown

The ideal gas law assumes: (1) molecules are point particles with no volume, and (2) there are no intermolecular forces except during elastic collisions. These assumptions break down when:

Very low temperatures: Near the condensation point of nitrogen ($77\;\text{K}$) or oxygen ($90\;\text{K}$), attractive van der Waals forces become significant. This is relevant for Mars ($T \sim 200\;\text{K}$ for CO$_2$) and Titan's atmosphere.
Very high pressures: At pressures above $\sim 100\;\text{atm}$, molecular volumes become non-negligible. The van der Waals equation $(p + a/\alpha^2)(\alpha - b) = RT$ provides a correction.
Near phase transitions: Where water vapour approaches saturation, the gas deviates from ideality; however, even saturated air at atmospheric pressures deviates by less than 0.2%.

For Earth's atmosphere, the ideal gas law is accurate to better than 0.1% under virtually all conditions. The compressibility factor $Z = pV/(nR^*T)$ remains within $1.000 \pm 0.001$ for$T > 200\;\text{K}$ and $p < 10\;\text{atm}$, covering the entire troposphere and stratosphere.

2. Hydrostatic Equation

2.1 Force Balance on an Air Slab

Consider a thin horizontal slab of air with cross-sectional area $A$, infinitesimal thickness $dz$, and density $\rho(z)$ at height $z$. We carefully enumerate every vertical force on this slab:

Upward pressure force on bottom face (at height $z$): $\;+p(z)\,A$
Downward pressure force on top face (at height $z + dz$): $\;-p(z+dz)\,A = -[p(z) + \frac{\partial p}{\partial z}dz]\,A$
Gravitational force (weight of slab): $\;-\rho(z)\,g\,A\,dz$

Note that the mass of the slab is $dm = \rho \cdot A \cdot dz$. For a stationary atmosphere in hydrostatic balance, Newton's second law requires the net force to vanish:

$$p(z)\,A - \left[p(z) + \frac{\partial p}{\partial z}\,dz\right]A - \rho g\,A\,dz = 0$$

The $p(z)\,A$ terms cancel. Dividing the remainder by $A\,dz$:

$$-\frac{\partial p}{\partial z} - \rho g = 0$$

$$\boxed{\frac{\partial p}{\partial z} = -\rho g}$$

Dimensional check: $[\rho g] = \text{kg m}^{-3} \times \text{m s}^{-2} = \text{kg m}^{-2}\text{s}^{-2} = \text{Pa m}^{-1}$. This matches $[\partial p/\partial z]$.

This states that pressure decreases with altitude at a rate proportional to the local air density and gravitational acceleration ($g \approx 9.81\;\text{m s}^{-2}$). Near sea level where $\rho \approx 1.225\;\text{kg m}^{-3}$:

$$\left|\frac{dp}{dz}\right| = 1.225 \times 9.81 = 12.02\;\text{Pa m}^{-1} \approx 1\;\text{hPa per 8.4 m}$$

2.2 Pressure as a Vertical Coordinate

Meteorologists frequently use pressure rather than geometric height as the vertical coordinate. This is natural because (1) pressure is what instruments directly measure, and (2) the equations of motion simplify dramatically. From the hydrostatic equation, rearranging:

$$dp = -\rho g\,dz \quad\Longrightarrow\quad dz = -\frac{dp}{\rho g}$$

Since $\rho > 0$ and $g > 0$, as $z$ increases, $p$ decreases, so pressure is a monotonically decreasing function of height. This one-to-one relationship means we can use $p$ to label heights. The key advantages in pressure coordinates are:

The continuity equation becomes exact and density-free: $\partial u/\partial x + \partial v/\partial y + \partial\omega/\partial p = 0$
The horizontal pressure gradient force involves geopotential gradients on constant-$p$ surfaces, avoiding density
Standard pressure levels (1000, 850, 700, 500, 300, 200 hPa) provide a universal height framework

2.3 Geopotential and Geopotential Height

The geopotential $\Phi$ is the work done per unit mass against gravity to raise an object from sea level to height $z$:

$$\Phi(z) = \int_0^z g\,dz' \approx gz$$

The geopotential height is $Z = \Phi/g_0$ where $g_0 = 9.80665\;\text{m s}^{-2}$is standard gravity. Using the hydrostatic equation in pressure coordinates with the ideal gas law ($\alpha = 1/\rho = R_d T_v/p$):

$$d\Phi = g\,dz = -\alpha\,dp = -\frac{R_d T_v}{p}\,dp$$

Therefore the hydrostatic equation in pressure coordinates takes the elegant form:

$$\boxed{\frac{\partial \Phi}{\partial p} = -\frac{R_d T_v}{p} = -\alpha}$$

2.4 Numerical Example: Pressure at 5 km

Method 1 — Isothermal approximation: Assuming $T = 255\;\text{K}$ throughout (scale height $H = R_d T/g = 287.05 \times 255/9.81 = 7460\;\text{m}$):

$$p(5000) = 101\,325 \times \exp\!\left(-\frac{5000}{7460}\right) = 101\,325 \times e^{-0.670} = 101\,325 \times 0.5117 = 51\,860\;\text{Pa} \approx 519\;\text{hPa}$$

Method 2 — Standard atmosphere with lapse rate: Using $\Gamma = 6.5\;\text{K/km}$,$T_0 = 288.15\;\text{K}$ (derived in Section 3):

$$p(5000) = 101\,325\left(1 - \frac{6.5 \times 5}{288.15}\right)^{g/(R_d\Gamma)} = 101\,325\left(1 - 0.1128\right)^{5.256} = 101\,325 \times (0.8872)^{5.256}$$

$$= 101\,325 \times 0.5334 = 54\,048\;\text{Pa} \approx 540\;\text{hPa}$$

The standard atmosphere value at 5 km is 540.48 hPa, so Method 2 is far more accurate. The isothermal method underestimates because it uses a temperature that is too cold for the lower troposphere.

2.5 When Hydrostatic Balance Breaks Down

Hydrostatic balance assumes that vertical accelerations are negligible compared to $g$. This fails when:

Intense thunderstorm updrafts: Peak vertical velocities of $w \sim 30\text{-}60\;\text{m s}^{-1}$ yield accelerations $Dw/Dt \sim 0.1g$, a 10% violation
Tornadoes: Extreme vertical motions with $Dw/Dt$ comparable to $g$
Small-scale turbulent eddies: In the convective boundary layer, vertical gusts of a few m/s produce transient non-hydrostatic pressure perturbations
Mountain waves: Vertically propagating gravity waves in flow over steep terrain can have significant non-hydrostatic effects

For synoptic-scale motions ($L > 100\;\text{km}$) the hydrostatic approximation is excellent, with errors less than 0.01%. Non-hydrostatic numerical weather prediction models are required to resolve convective-scale phenomena ($\Delta x < 4\;\text{km}$).

3. Scale Height & Barometric Formula

3.1 Combining Hydrostatics and the Ideal Gas Law

We substitute $\rho = p/(R_d T_v)$ from the ideal gas law into the hydrostatic equation:

$$\frac{dp}{dz} = -\frac{p\,g}{R_d T_v}$$

Separating variables:

$$\frac{dp}{p} = -\frac{g}{R_d T_v}\,dz$$

3.2 Isothermal Atmosphere and Scale Height

For the simplest case where $T_v = \text{const}$, integrating from the surface $z = 0$(pressure $p_0$) to height $z$:

$$\int_{p_0}^{p}\frac{dp'}{p'} = -\frac{g}{R_d T_v}\int_0^z dz'$$

$$\ln\frac{p}{p_0} = -\frac{gz}{R_d T_v} = -\frac{z}{H}$$

where we define the pressure scale height:

$$H \equiv \frac{R_d T_v}{g}$$

Exponentiating:

$$\boxed{p(z) = p_0\,\exp\!\left(-\frac{z}{H}\right)}$$

Physical meaning: $H$ is the height over which pressure drops by a factor of $e \approx 2.718$. It is also the height the atmosphere would have if it had uniform density equal to the surface value (since total mass per unit area $= \rho_0 H = p_0/g$).

3.3 Scale Height at Different Temperatures

The scale height varies with temperature:

$T = 220\;\text{K}$ (tropopause): $H = 287.05 \times 220 / 9.81 = 6435\;\text{m} \approx 6.4\;\text{km}$
$T = 255\;\text{K}$ (mean atmosphere): $H = 287.05 \times 255 / 9.81 = 7460\;\text{m} \approx 7.5\;\text{km}$
$T = 288\;\text{K}$ (sea level standard): $H = 287.05 \times 288 / 9.81 = 8426\;\text{m} \approx 8.4\;\text{km}$
$T = 300\;\text{K}$ (tropical boundary layer): $H = 287.05 \times 300 / 9.81 = 8777\;\text{m} \approx 8.8\;\text{km}$

3.4 The Hypsometric Equation

For the general (non-isothermal) case, integrate between two pressure levels $p_1$ (lower, higher pressure) and $p_2$ (upper, lower pressure):

$$\int_{z_1}^{z_2}dz = -\int_{p_1}^{p_2}\frac{R_d T_v}{gp}\,dp$$

Defining a layer-mean virtual temperature $\overline{T_v}$ such that$\int_{p_2}^{p_1}T_v\,d(\ln p) = \overline{T_v}\ln(p_1/p_2)$:

$$\boxed{\Delta z = z_2 - z_1 = \frac{R_d\,\overline{T_v}}{g}\,\ln\frac{p_1}{p_2}}$$

This is the hypsometric equation — one of the most important relationships in synoptic meteorology. It says the thickness of a pressure layer is proportional to the mean virtual temperature of that layer.

3.5 Application: 500-1000 hPa Thickness

The 1000-500 hPa thickness is a critical variable on weather maps. Using the hypsometric equation:

$$\Delta z_{1000-500} = \frac{R_d\,\overline{T_v}}{g}\ln\frac{1000}{500} = \frac{287.05 \times \overline{T_v}}{9.81} \times 0.6931 = 20.27\,\overline{T_v}\;\;\text{(in metres)}$$

For $\overline{T_v} = 256\;\text{K}$ (typical mid-latitudes):$\Delta z = 20.27 \times 256 = 5189\;\text{m} \approx 5190\;\text{m}$. The famous 540 dam (5400 m) thickness line roughly separates rain from snow regions, because it corresponds to $\overline{T_v} \approx 266\;\text{K}$ — indicating a column warm enough that precipitation falls as rain below the melting level.

3.6 Non-Isothermal Atmosphere: Constant Lapse Rate

A more realistic model uses a constant lapse rate $\Gamma$, so$T(z) = T_0 - \Gamma z$. Substituting into the separated equation:

$$\frac{dp}{p} = -\frac{g}{R_d(T_0 - \Gamma z)}\,dz$$

Integrating:

$$\ln\frac{p}{p_0} = -\frac{g}{R_d}\int_0^z\frac{dz'}{T_0 - \Gamma z'} = \frac{g}{R_d\Gamma}\ln\left(1 - \frac{\Gamma z}{T_0}\right)$$

Exponentiating:

$$\boxed{p(z) = p_0\left(1 - \frac{\Gamma z}{T_0}\right)^{g/(R_d\Gamma)}}$$

For the International Standard Atmosphere ($\Gamma = 6.5\;\text{K km}^{-1} = 0.0065\;\text{K m}^{-1}$,$T_0 = 288.15\;\text{K}$), the exponent is:

$$\frac{g}{R_d\Gamma} = \frac{9.81}{287.05 \times 0.0065} = \frac{9.81}{1.866} = 5.256$$

This formula reproduces the standard atmosphere pressure profile to within 0.1% up to the tropopause ($\sim 11\;\text{km}$). Above the tropopause, $\Gamma \approx 0$ and the isothermal formula applies. Setting $\Gamma \to 0$ in the constant-lapse-rate formula recovers the isothermal barometric formula (verify using $\lim_{\Gamma\to 0}(1 - \Gamma z/T_0)^{g/(R_d\Gamma)} = e^{-gz/(R_dT_0)}$).

4. First Law of Thermodynamics (Atmospheric Form)

4.1 Internal Energy Form

The first law of thermodynamics for a unit mass of gas (specific quantities) states that the heat added equals the increase in internal energy plus the work done by the gas in expanding:

$$dq = du + dw$$

For a simple compressible substance, the work of expansion per unit mass is $dw = p\,d\alpha$, where$\alpha = 1/\rho$ is the specific volume (m$^3$/kg). For an ideal gas, internal energy depends only on temperature: $du = c_v\,dT$, where $c_v$ is the specific heat at constant volume. Therefore the internal energy form is:

$$\boxed{dq = c_v\,dT + p\,d\alpha}$$

4.2 Deriving the Pressure Form

The internal energy form is inconvenient because atmospheric measurements are naturally in terms of$(T, p)$, not $(T, \alpha)$. We transform using the ideal gas law$p\alpha = R_d T$. Differentiating by the product rule:

$$p\,d\alpha + \alpha\,dp = R_d\,dT$$

Solving for the $p\,d\alpha$ term that appears in the first law:

$$p\,d\alpha = R_d\,dT - \alpha\,dp$$

Substituting into $dq = c_v\,dT + p\,d\alpha$:

$$dq = c_v\,dT + R_d\,dT - \alpha\,dp$$

$$dq = (c_v + R_d)\,dT - \alpha\,dp$$

Since $c_p = c_v + R_d$ (proven below), and $\alpha = 1/\rho$:

$$\boxed{dq = c_p\,dT - \alpha\,dp = c_p\,dT - \frac{dp}{\rho}}$$

4.3 Enthalpy Form

Define the specific enthalpy as $h = u + p\alpha$. Then:

$$dh = du + p\,d\alpha + \alpha\,dp = dq + \alpha\,dp$$

where we used $dq = du + p\,d\alpha$. Rearranging:

$$dq = dh - \alpha\,dp$$

For an ideal gas, $h = c_p T + \text{const}$, so $dh = c_p\,dT$, recovering the pressure form. The enthalpy form is especially useful because at constant pressure ($dp = 0$), the heat added equals the enthalpy change: $dq|_p = dh = c_p\,dT$.

4.4 Proof: $c_p - c_v = R_d$

From the two forms of the first law:

$$dq = c_v\,dT + p\,d\alpha \quad\text{and}\quad dq = c_p\,dT - \alpha\,dp$$

At constant pressure ($dp = 0$): $c_p\,dT = c_v\,dT + p\,d\alpha|_p$. From $p\alpha = R_dT$ at constant $p$: $p\,d\alpha = R_d\,dT$. Therefore:

$$c_p\,dT = c_v\,dT + R_d\,dT$$

$$\boxed{c_p - c_v = R_d}$$

Numerically: $c_v = c_p - R_d = 1004 - 287 = 717\;\text{J kg}^{-1}\text{K}^{-1}$. The ratio $\gamma = c_p/c_v = 1004/717 = 1.40$ for dry air (diatomic gas). The excess$c_p > c_v$ arises because heating at constant pressure requires extra energy to do expansion work against the surroundings.

4.5 Special Atmospheric Processes

Isothermal ($dT = 0$): $dq = -\alpha\,dp = -(R_dT/p)\,dp$. Heat must be added to keep temperature constant during ascent.
Isobaric ($dp = 0$): $dq = c_p\,dT$. Surface heating, radiative warming/cooling. This is the most common form for boundary-layer processes.
Isochoric ($d\alpha = 0$): $dq = c_v\,dT$. Rare in the free atmosphere; sometimes approximately realized in sealed containers.
Adiabatic ($dq = 0$): $c_p\,dT = \alpha\,dp$. The most important atmospheric process — governs rising/sinking air parcels. Leads to the dry adiabatic lapse rate (Section 5).

4.6 Poisson's Equation (Adiabatic Process)

For an adiabatic process ($dq = 0$), the first law gives:

$$c_p\,dT = \alpha\,dp = \frac{R_d T}{p}\,dp$$

$$\frac{dT}{T} = \frac{R_d}{c_p}\frac{dp}{p}$$

Integrating from state $(T_0, p_0)$ to $(T, p)$:

$$\ln\frac{T}{T_0} = \frac{R_d}{c_p}\ln\frac{p}{p_0}$$

$$\boxed{\frac{T}{T_0} = \left(\frac{p}{p_0}\right)^{R_d/c_p} = \left(\frac{p}{p_0}\right)^\kappa}$$

where $\kappa = R_d/c_p = 287.05/1004 = 0.2857$. This is Poisson's equation — the adiabat on a$(T, p)$ diagram. A parcel at 850 hPa with $T = 280\;\text{K}$ that descends adiabatically to 1000 hPa warms to $T = 280 \times (1000/850)^{0.286} = 280 \times 1.0487 = 293.6\;\text{K}$.

5. Dry Adiabatic Lapse Rate (DALR)

5.1 Derivation of the DALR

For an adiabatic process ($dq = 0$), the first law in pressure form becomes:

$$c_p\,dT = \frac{1}{\rho}\,dp$$

Substituting the hydrostatic relation $dp = -\rho g\,dz$:

$$c_p\,dT = \frac{1}{\rho}(-\rho g\,dz) = -g\,dz$$

Note that the density $\rho$ cancels — the dry adiabatic lapse rate is independent of the thermodynamic state of the parcel. Solving for the temperature change per unit height:

$$\frac{dT}{dz} = -\frac{g}{c_p}$$

Defining the lapse rate as the rate of temperature decrease with height ($\Gamma \equiv -dT/dz$):

$$\boxed{\Gamma_d = -\frac{dT}{dz} = \frac{g}{c_p} = \frac{9.81}{1004} = 9.77\;\text{K km}^{-1} \approx 9.8\;\text{K km}^{-1}}$$

Key insight: The DALR depends only on two fundamental constants — the gravitational acceleration and the specific heat capacity of air. It does not depend on pressure, temperature, or humidity (as long as the air is unsaturated). On Mars ($g = 3.72\;\text{m s}^{-2}$,$c_p \approx 735\;\text{J kg}^{-1}\text{K}^{-1}$ for CO$_2$),$\Gamma_d = 3.72/735 = 5.1\;\text{K km}^{-1}$.

5.2 Saturated (Moist) Adiabatic Lapse Rate

When a rising air parcel reaches saturation, condensation releases latent heat, which partially offsets the adiabatic cooling. The first law for a saturated parcel includes latent heating:

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

where $w_s$ is the saturation mixing ratio and $dw_s < 0$ during ascent (water vapour condenses out). Using the Clausius-Clapeyron equation and hydrostatic relation, after substantial algebra the saturated adiabatic lapse rate (SALR) is:

$$\boxed{\Gamma_s = \Gamma_d \,\frac{1 + \dfrac{L_v\,w_s}{R_d\,T}}{1 + \dfrac{L_v^2\,w_s}{c_p\,R_v\,T^2}}}$$

Since both numerator and denominator are greater than 1, and the denominator grows faster (because$L_v/(c_p T) \gg R_d/R_v$ at typical tropospheric temperatures), we always have$\Gamma_s < \Gamma_d$. Physically, condensational heating slows the cooling rate.

5.3 Numerical Comparison: DALR vs SALR

The SALR varies strongly with temperature because warmer air holds exponentially more moisture:

$T = -20°\text{C}$ (upper troposphere): $w_s \approx 0.8\;\text{g/kg}$, $\Gamma_s \approx 8.7\;\text{K/km}$ (nearly dry)
$T = 0°\text{C}$ (mid-troposphere): $w_s \approx 3.8\;\text{g/kg}$, $\Gamma_s \approx 6.5\;\text{K/km}$
$T = 15°\text{C}$ (lower troposphere): $w_s \approx 10.7\;\text{g/kg}$, $\Gamma_s \approx 5.0\;\text{K/km}$
$T = 30°\text{C}$ (tropical boundary layer): $w_s \approx 27\;\text{g/kg}$, $\Gamma_s \approx 3.6\;\text{K/km}$

In the tropics near the surface, the SALR can be less than half the DALR, making the tropical troposphere highly susceptible to conditional instability.

5.4 Static Stability and the Environmental Lapse Rate

The environmental lapse rate $\Gamma = -dT_{\text{env}}/dz$ is the observed temperature profile of the atmosphere (typically $\sim 6.5\;\text{K/km}$ in the troposphere). Comparing it with the parcel lapse rates determines stability:

Condition	Lapse Rate	Stability
$\Gamma < \Gamma_s$	$< 5\;\text{K/km}$ (typical)	Absolutely stable
$\Gamma = \Gamma_s$	Saturated neutral	Neutral (for saturated parcels)
$\Gamma_s < \Gamma < \Gamma_d$	$5\text{-}9.8\;\text{K/km}$	Conditionally unstable
$\Gamma = \Gamma_d$	$9.8\;\text{K/km}$	Neutral (for dry parcels)
$\Gamma > \Gamma_d$	$> 9.8\;\text{K/km}$ (superadiabatic)	Absolutely unstable

Conditional instability ($\Gamma_s < \Gamma < \Gamma_d$) is the most common situation in the lower troposphere. An unsaturated parcel is stable, but if lifted to its level of free convection (LFC), it becomes buoyant and accelerates upward — this is the mechanism triggering thunderstorms when a capping inversion is overcome. The standard atmosphere lapse rate of 6.5 K/km falls squarely in the conditionally unstable regime.

6. Potential Temperature

6.1 Derivation from the First Law

Starting from the adiabatic first law $dq = 0$: $c_p\,dT = \alpha\,dp$. Using the ideal gas law $\alpha = R_d T/p$:

$$c_p\,dT = \frac{R_d T}{p}\,dp$$

Dividing both sides by $T$:

$$c_p\,\frac{dT}{T} = R_d\,\frac{dp}{p}$$

Integrating from the current state $(T, p)$ to a reference state $(\theta, p_0)$where $p_0 = 1000\;\text{hPa}$ is the reference pressure:

$$c_p\int_T^{\theta}\frac{dT'}{T'} = R_d\int_p^{p_0}\frac{dp'}{p'}$$

$$c_p\,\ln\frac{\theta}{T} = R_d\,\ln\frac{p_0}{p}$$

Exponentiating both sides and defining $\kappa = R_d/c_p = 287.05/1004 = 0.2857$:

$$\boxed{\theta = T\left(\frac{p_0}{p}\right)^{\!\kappa}}$$

Physical meaning: Potential temperature is the temperature a parcel would have if brought adiabatically to the 1000 hPa reference level. It removes the effect of pressure variations, allowing direct comparison of air masses at different altitudes.

6.2 Conservation of Potential Temperature

We prove that $\theta$ is conserved in dry adiabatic processes ($d\theta/dt = 0$). Taking the logarithm of the definition:

$$\ln\theta = \ln T + \kappa\ln p_0 - \kappa\ln p$$

Differentiating with respect to time following a parcel (material derivative):

$$\frac{1}{\theta}\frac{D\theta}{Dt} = \frac{1}{T}\frac{DT}{Dt} - \frac{\kappa}{p}\frac{Dp}{Dt}$$

From the first law: $\dot{q} = c_p\frac{DT}{Dt} - \frac{1}{\rho}\frac{Dp}{Dt}$, so$\frac{1}{T}\frac{DT}{Dt} = \frac{\dot{q}}{c_pT} + \frac{1}{\rho T}\frac{Dp}{Dt}\cdot\frac{1}{c_p}$. Using $1/(\rho T) = R_d/p$:

$$\frac{1}{\theta}\frac{D\theta}{Dt} = \frac{\dot{q}}{c_pT} + \frac{R_d}{c_pp}\frac{Dp}{Dt} - \frac{\kappa}{p}\frac{Dp}{Dt} = \frac{\dot{q}}{c_pT}$$

since $R_d/c_p = \kappa$. Therefore:

$$\boxed{\frac{D\theta}{Dt} = \frac{\theta}{c_p T}\,\dot{q}}$$

For adiabatic processes ($\dot{q} = 0$), this confirms $D\theta/Dt = 0$ — potential temperature is materially conserved.

6.3 Connection to Entropy

From the second law of thermodynamics, the specific entropy satisfies $Tds = dq$. Using the first law in the form $dq = c_p\,dT - \alpha\,dp$:

$$ds = c_p\frac{dT}{T} - R_d\frac{dp}{p} = c_p\,d(\ln T) - R_d\,d(\ln p)$$

But from the definition of $\theta$: $\ln\theta = \ln T - \kappa\ln p + \kappa\ln p_0$, so$d(\ln\theta) = d(\ln T) - \kappa\,d(\ln p)$. Therefore:

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

Integrating:

$$\boxed{s = c_p\ln\theta + \text{const}}$$

This reveals that surfaces of constant $\theta$ are isentropic surfaces. In isentropic analysis, air parcels move along $\theta$-surfaces in the absence of diabatic heating, making $\theta$ the natural vertical coordinate for studying large-scale adiabatic transport. Cross-isentropic flow (parcels crossing $\theta$-surfaces) is driven exclusively by diabatic processes: radiation, latent heating, or turbulent mixing.

6.4 Equivalent Potential Temperature

For moist processes, $\theta$ is not conserved because latent heat release changes the entropy. The equivalent potential temperature $\theta_e$ accounts for this by including the latent heat of all water vapour. An approximate but widely used formula is:

$$\boxed{\theta_e \approx \theta\,\exp\!\left(\frac{L_v\,w_s}{c_p\,T}\right)}$$

where $w_s$ is the saturation mixing ratio. $\theta_e$ is conserved for both dry and saturated adiabatic processes (but not for precipitation removal, i.e., pseudoadiabatic ascent modifies it slightly). The wet-bulb potential temperature $\theta_w$ is the temperature reached when a parcel is lifted saturated-adiabatically to very low pressure (exhausting all moisture) and then brought dry-adiabatically to 1000 hPa. It is uniquely related to $\theta_e$ and is also conserved for moist adiabatic processes.

6.5 Numerical Example

Calculate $\theta$ for a parcel at $p = 850\;\text{hPa}$,$T = 280\;\text{K}$:

$$\theta = 280\left(\frac{1000}{850}\right)^{0.2857} = 280 \times (1.1765)^{0.2857}$$

Computing the exponent: $\ln(1.1765) = 0.1625$, so $0.2857 \times 0.1625 = 0.04644$, and $e^{0.04644} = 1.04754$:

$$\theta = 280 \times 1.0475 = 293.3\;\text{K}$$

This parcel, if brought dry-adiabatically to 1000 hPa, would have a temperature of 293.3 K (20.2 C). If this parcel also has $w_s = 7.6\;\text{g/kg}$ at saturation:

$$\theta_e \approx 293.3 \times \exp\!\left(\frac{2.5\times10^6 \times 0.0076}{1004 \times 280}\right) = 293.3 \times \exp(0.0676) = 293.3 \times 1.0699 = 313.8\;\text{K}$$

The 20.5 K excess of $\theta_e$ over $\theta$ represents the latent heat stored in the water vapour. In the tropics where moisture is abundant, $\theta_e - \theta$ can exceed 40 K, providing enormous energy for convective storms.

7. Clausius-Clapeyron Equation

7.1 Gibbs-Duhem Derivation

At phase equilibrium between liquid water and water vapour, the specific Gibbs free energies (chemical potentials) of the two phases must be equal: $g_l = g_v$. For an infinitesimal change along the coexistence curve (where both phases remain in equilibrium), we require$dg_l = dg_v$. The Gibbs-Duhem relation for a pure substance gives:

$$dg = -s\,dT + \alpha\,dp$$

where $s$ is specific entropy and $\alpha$ is specific volume. Applying this to both phases along the saturation curve (where $p = e_s(T)$):

$$-s_l\,dT + \alpha_l\,de_s = -s_v\,dT + \alpha_v\,de_s$$

Collecting terms:

$$(s_v - s_l)\,dT = (\alpha_v - \alpha_l)\,de_s$$

The entropy difference between the phases is related to the latent heat of vaporization:$s_v - s_l = L_v/T$ (since $L_v = T(s_v - s_l)$ is the heat absorbed during isothermal, isobaric vaporization). Therefore the exact Clausius-Clapeyron equation is:

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

7.2 Meteorological Approximation

Two key simplifications make this usable for atmospheric calculations:

Neglect liquid volume: At 20 C, $\alpha_v \approx 57.8\;\text{m}^3\text{/kg}$ while $\alpha_l \approx 0.001\;\text{m}^3\text{/kg}$, so $\alpha_v \gg \alpha_l$ and $\alpha_v - \alpha_l \approx \alpha_v$
Ideal gas for vapour: $\alpha_v = R_v T/e_s$ where $R_v = R^*/M_v = 461.5\;\text{J kg}^{-1}\text{K}^{-1}$

Substituting:

$$\frac{de_s}{dT} = \frac{L_v}{T \cdot R_v T/e_s} = \frac{L_v\,e_s}{R_v\,T^2}$$

$$\boxed{\frac{de_s}{dT} = \frac{L_v\,e_s}{R_v\,T^2}}$$

This is a separable ODE. Rearranging and integrating (treating $L_v$ as approximately constant):

$$\int_{e_{s0}}^{e_s}\frac{de_s'}{e_s'} = \frac{L_v}{R_v}\int_{T_0}^{T}\frac{dT'}{T'^2}$$

$$\ln\frac{e_s}{e_{s0}} = \frac{L_v}{R_v}\left(-\frac{1}{T} + \frac{1}{T_0}\right) = \frac{L_v}{R_v}\left(\frac{1}{T_0} - \frac{1}{T}\right)$$

$$\boxed{e_s(T) = e_{s0}\,\exp\!\left[\frac{L_v}{R_v}\left(\frac{1}{T_0} - \frac{1}{T}\right)\right]}$$

where $L_v \approx 2.501 \times 10^6\;\text{J kg}^{-1}$,$R_v = 461.5\;\text{J kg}^{-1}\text{K}^{-1}$,$e_{s0} = 6.112\;\text{hPa}$ at $T_0 = 273.15\;\text{K}$. The ratio $L_v/R_v = 5423\;\text{K}$, indicating the extreme sensitivity of saturation vapour pressure to temperature.

7.3 The ~7%/K Sensitivity

The fractional rate of change of $e_s$ is:

$$\frac{1}{e_s}\frac{de_s}{dT} = \frac{L_v}{R_v T^2}$$

At $T = 288\;\text{K}$ (15 C):

$$\frac{1}{e_s}\frac{de_s}{dT} = \frac{2.501 \times 10^6}{461.5 \times 288^2} = \frac{2.501 \times 10^6}{3.829 \times 10^7} = 0.0653\;\text{K}^{-1} \approx 6.5\%\;\text{per K}$$

At $T = 300\;\text{K}$ (27 C): $1/e_s \cdot de_s/dT = 6.0\%\;\text{per K}$. This near-constant ~6-7% per kelvin sensitivity is one of the most consequential relationships in climate science: for every 1 K of global warming, the atmosphere can hold roughly 6-7% more water vapour (assuming relative humidity stays approximately constant, which observations support).

7.4 Empirical Approximations

Because $L_v$ actually varies with temperature, several empirical formulas provide better accuracy than the constant-$L_v$ integrated form:

August-Roche-Magnus formula (accuracy $\sim 0.4\%$ for $-40°$C to $50°$C):

$$\boxed{e_s(T) \approx 6.1078\,\exp\!\left(\frac{17.27\,T_c}{T_c + 237.3}\right)\;\;\text{hPa}}$$

where $T_c$ is temperature in Celsius. Bolton's formula (1980), which is even more accurate ($\sim 0.1\%$ for $-35°$C to $35°$C):

$$e_s(T) \approx 6.112\,\exp\!\left(\frac{17.67\,T_c}{T_c + 243.5}\right)\;\;\text{hPa}$$

Quick numerical check at $T_c = 20°$C using August-Roche-Magnus:$e_s = 6.1078 \times \exp(17.27 \times 20/257.3) = 6.1078 \times \exp(1.343) = 6.1078 \times 3.831 = 23.39\;\text{hPa}$. The precise measured value is 23.37 hPa — excellent agreement.

7.5 Relative Humidity and Dewpoint

The relative humidity is defined as the ratio of the actual vapour pressure to the saturation vapour pressure at the air's temperature:

$$\text{RH} = \frac{e}{e_s(T)} \times 100\%$$

The dewpoint temperature $T_d$ is the temperature to which air must be cooled (at constant pressure and moisture) to reach saturation:

$$e_s(T_d) = e$$

Inverting the Magnus formula for the dewpoint:

$$T_d = \frac{237.3\,\ln(e/6.1078)}{17.27 - \ln(e/6.1078)}\;\;°\text{C}$$

The dewpoint depression $T - T_d$ measures how far the air is from saturation. A rule of thumb: $T - T_d \approx (100 - \text{RH})/5$ in Celsius for RH above 50%.

7.6 Lifting Condensation Level (LCL)

When an unsaturated parcel rises, it cools at the DALR ($\Gamma_d = 9.8\;\text{K/km}$) and its dewpoint decreases at a slower rate ($\sim 1.7\;\text{K/km}$, because the dewpoint depends on mixing ratio which is conserved during dry ascent). The height at which the parcel reaches saturation is the LCL:

$$z_{\text{LCL}} \approx \frac{T - T_d}{(\Gamma_d - \Gamma_{T_d})} \approx \frac{T - T_d}{8.0\;\text{K/km}} \approx 125\,(T - T_d)\;\text{m}$$

where the temperatures are surface values in Celsius (or Kelvin; the difference is the same). A more precise formula by Bolton (1980) using the Magnus equation gives:

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

For example, if $T = 25°\text{C}$ and $T_d = 15°\text{C}$:$z_{\text{LCL}} = 125 \times 10 = 1250\;\text{m}$. This is approximately the cloud base height for cumulus clouds formed by surface-driven convection.

The Clausius-Clapeyron equation is arguably the single most important equation in climate science: it governs the water vapour feedback (the strongest positive feedback in the climate system), the rate at which extreme precipitation intensifies with warming, the height of the tropopause, and the drying of the stratosphere. The roughly exponential dependence of $e_s$ on $T$ means that small temperature changes produce large moisture changes — amplifying both droughts (through increased evaporative demand) and floods (through increased atmospheric moisture capacity).

8. Brunt-Väisälä (Buoyancy) Frequency

The Brunt-Väisälä frequency $N$ is the fundamental measure of static stability in a stratified atmosphere. It describes the frequency at which a vertically displaced air parcel oscillates about its equilibrium level in a stably stratified environment. The derivation proceeds from parcel theory combined with Archimedes' principle.

Step 1: Parcel Displacement and Temperature Difference

Consider a parcel initially at height $z_0$ in equilibrium with its environment, both at temperature $T_0$ and pressure $p_0$. Displace the parcel vertically by a small distance $\delta z$. The parcel rises adiabatically (no heat exchange with surroundings), so its temperature changes at the dry adiabatic lapse rate $\Gamma_d = g/c_p$:

$$T_{\text{parcel}}(z_0 + \delta z) = T_0 - \Gamma_d\,\delta z$$

The environment at the new height has temperature:

$$T_{\text{env}}(z_0 + \delta z) = T_0 - \Gamma\,\delta z$$

where $\Gamma = -dT/dz$ is the environmental lapse rate (positive when temperature decreases with height). The temperature difference between parcel and environment is:

$$\Delta T = T_{\text{parcel}} - T_{\text{env}} = (\Gamma - \Gamma_d)\,\delta z$$

If $\Gamma < \Gamma_d$ (the typical case), then an upward displacement ($\delta z > 0$) gives $\Delta T < 0$: the parcel is cooler (denser) than its surroundings and experiences a downward restoring force. This is the stable case.

Step 2: Buoyancy Force from Archimedes' Principle

By Archimedes' principle, the net upward buoyancy force per unit mass on the displaced parcel is:

$$a = g\,\frac{\rho_{\text{env}} - \rho_{\text{parcel}}}{\rho_{\text{parcel}}}$$

Using the ideal gas law $\rho = p/(R_d T)$, and noting that the parcel and environment are at the same pressure (the parcel adjusts instantaneously to the ambient pressure), we get:

$$a = g\,\frac{T_{\text{parcel}} - T_{\text{env}}}{T_{\text{env}}} \approx g\,\frac{\Delta T}{T_0}$$

Substituting the temperature difference from Step 1:

$$a = \frac{d^2(\delta z)}{dt^2} = g\,\frac{(\Gamma - \Gamma_d)\,\delta z}{T_0} = -\frac{g(\Gamma_d - \Gamma)}{T_0}\,\delta z$$

Step 3: Simple Harmonic Oscillator Form

The equation of motion for the displaced parcel is:

$$\frac{d^2(\delta z)}{dt^2} = -N^2\,\delta z$$

where we identify the squared Brunt-Väisälä frequency:

$$N^2 = \frac{g}{T}(\Gamma_d - \Gamma)$$

Step 4: Potential Temperature Form

A more elegant derivation uses potential temperature. Recall $\theta = T(p_0/p)^\kappa$. Taking the logarithmic derivative:

$$\frac{1}{\theta}\frac{d\theta}{dz} = \frac{1}{T}\frac{dT}{dz} - \frac{\kappa}{p}\frac{dp}{dz}$$

Substituting the hydrostatic equation $dp/dz = -\rho g = -pg/(R_d T)$:

$$\frac{1}{\theta}\frac{d\theta}{dz} = \frac{1}{T}\frac{dT}{dz} + \frac{\kappa g}{R_d T} = \frac{1}{T}\!\left(\frac{dT}{dz} + \frac{g}{c_p}\right) = \frac{1}{T}(\Gamma_d - \Gamma)$$

where we used $\kappa/R_d = 1/c_p$ and $\Gamma_d = g/c_p$. Multiplying through by $g$ gives the fundamental result:

$$\boxed{N^2 = \frac{g}{\theta}\frac{d\theta}{dz} = \frac{g}{T}(\Gamma_d - \Gamma)}$$

Physical interpretation: $N^2$ is positive when $\theta$ increases with height (stable stratification). When $N^2 > 0$, the solution is oscillatory:$\delta z(t) = A\cos(Nt) + B\sin(Nt)$. When $N^2 < 0$, the solution is exponentially growing — convective instability.

Moist Brunt-Väisälä Frequency

For saturated air, latent heat release modifies the stability. The moist Brunt-Väisälä frequency uses the equivalent potential temperature $\theta_e$:

$$N_m^2 = \frac{g}{\theta_e}\frac{d\theta_e}{dz}$$

Since the moist adiabatic lapse rate $\Gamma_m < \Gamma_d$, saturated air has a lower effective stability. An atmosphere with $N^2 > 0$ but $N_m^2 < 0$ is said to be conditionally unstable — stable for unsaturated parcels but unstable for saturated ones. This is the most common stability regime in the tropical troposphere and is why deep convective storms develop when parcels reach their level of free convection (LFC).

Internal Gravity Waves

A stably stratified atmosphere supports internal gravity waves (buoyancy waves) with the dispersion relation:

$$\omega^2 = N^2\frac{k_h^2}{k_h^2 + m^2}$$

where $k_h$ is the horizontal wavenumber and $m$ is the vertical wavenumber. The buoyancy oscillation period is $\tau = 2\pi/N$, and the vertical wavelength for a wave with horizontal wavelength $\lambda_h$ propagating at frequency $\omega$ is:

$$\lambda_z = \frac{2\pi}{m} = \lambda_h\frac{\sqrt{N^2 - \omega^2}}{\omega}$$

A key property: the group velocity is perpendicular to the phase velocity, so energy propagates at right angles to the apparent wavefront motion.

Numerical Examples

Troposphere: With $\Gamma \approx 6.5\;\text{K km}^{-1}$,$\Gamma_d = 9.8\;\text{K km}^{-1}$, and $T \approx 250\;\text{K}$:

$$N^2 = \frac{9.81}{250}(9.8 - 6.5) \times 10^{-3} \approx 1.3 \times 10^{-4}\;\text{s}^{-2}$$

$$N \approx 0.011\;\text{s}^{-1}, \qquad \tau = \frac{2\pi}{N} \approx 9.2\;\text{min}$$

Stratosphere: With $d\theta/dz \approx 15\;\text{K km}^{-1}$ and$\theta \approx 400\;\text{K}$:

$$N^2 = \frac{9.81}{400} \times 15 \times 10^{-3} \approx 3.7 \times 10^{-4}\;\text{s}^{-2}$$

$$N \approx 0.019\;\text{s}^{-1}, \qquad \tau \approx 5.4\;\text{min}$$

The much higher $N$ in the stratosphere explains why this layer is so strongly stratified and resistant to vertical mixing — hence its name.

Mountain Waves

When stably stratified air ($N > 0$) with mean wind speed $U$ flows over topography, it generates mountain waves (lee waves). The vertical wavelength is:

$$\lambda_z = \frac{2\pi U}{N}$$

For $U = 20\;\text{m s}^{-1}$ and $N = 0.01\;\text{s}^{-1}$,$\lambda_z \approx 12.6\;\text{km}$. If the mountain height exceeds a critical value, wave breaking occurs, generating severe downslope windstorms and clear-air turbulence (CAT). Lenticular clouds mark the crests of these standing waves.

Connection to the Richardson Number

The gradient Richardson number connects buoyancy and wind shear:

$$Ri = \frac{N^2}{\left(\dfrac{dU}{dz}\right)^2}$$

When $Ri < 0.25$ (the Miles-Howard theorem), Kelvin-Helmholtz instability can occur and turbulence develops. When $Ri > 1$, shear-generated turbulence is suppressed by the stable stratification. The critical Richardson number $Ri_c = 0.25$ is one of the most important thresholds in atmospheric dynamics, governing the onset of clear-air turbulence (CAT) that affects aviation.

Dimensional Analysis Check

$[N^2] = \frac{[\text{m s}^{-2}]}{[\text{K}]} \cdot \frac{[\text{K}]}{[\text{m}]} = \text{s}^{-2}$ — confirming that $N$ has dimensions of inverse time (a frequency), as required. Typical buoyancy periods range from ~5 minutes in the stratosphere to ~10 minutes in the troposphere. These oscillations are the atmosphere's fundamental restoring mechanism against vertical displacements.

9. Equations of Motion in a Rotating Frame

The atmospheric equations of motion are derived from Newton's second law applied in a non-inertial reference frame rotating with the Earth. This transformation introduces two apparent forces — the Coriolis force and the centrifugal force — that profoundly shape large-scale atmospheric flow patterns.

Step 1: Newton's Second Law in the Inertial Frame

In an inertial (non-rotating) frame, the equation of motion for a fluid parcel of density $\rho$ is:

$$\frac{D\vec{v}}{Dt}\bigg|_{\text{inertial}} = -\frac{1}{\rho}\nabla p + \vec{g}^* + \vec{F}$$

where $\vec{g}^*$ is true gravitational acceleration (pointing toward Earth's center) and $\vec{F}$ represents friction. The material (Lagrangian) derivative is$D/Dt = \partial/\partial t + \vec{v}\cdot\nabla$.

Step 2: Transformation to the Rotating Frame

The relationship between time derivatives in inertial and rotating frames for any vector $\vec{A}$ is:

$$\left(\frac{d\vec{A}}{dt}\right)_{\text{inertial}} = \left(\frac{d\vec{A}}{dt}\right)_{\text{rot}} + \vec{\Omega}\times\vec{A}$$

Applying this to the position vector $\vec{r}$ gives $\vec{v}_{\text{inertial}} = \vec{v}_{\text{rot}} + \vec{\Omega}\times\vec{r}$. Applying it again to the velocity:

$$\frac{D\vec{v}}{Dt}\bigg|_{\text{inertial}} = \frac{D\vec{v}}{Dt}\bigg|_{\text{rot}} + 2\vec{\Omega}\times\vec{v}_{\text{rot}} + \vec{\Omega}\times(\vec{\Omega}\times\vec{r})$$

The two new terms are: (1) the Coriolis acceleration $2\vec{\Omega}\times\vec{v}$, which acts perpendicular to the velocity and deflects moving air (right in NH, left in SH); and (2) the centrifugal acceleration $\vec{\Omega}\times(\vec{\Omega}\times\vec{r})$, directed radially outward from the rotation axis.

Step 3: Absorbing Centrifugal Force into Effective Gravity

The centrifugal acceleration can be written as the gradient of a potential:

$$\vec{\Omega}\times(\vec{\Omega}\times\vec{r}) = -\nabla\!\left(\frac{1}{2}\Omega^2 R^2\right)$$

where $R$ is the distance from the rotation axis. We define effective gravityby combining true gravity and centrifugal acceleration:

$$\vec{g} \equiv \vec{g}^* - \nabla\!\left(\frac{1}{2}\Omega^2 R^2\right)$$

The magnitude of the centrifugal correction is $\Omega^2 R \sim 0.034\;\text{m s}^{-2}$ at the equator, about 0.35% of $g$. This is why Earth bulges at the equator (equatorial radius exceeds polar radius by ~21 km). The effective gravity $\vec{g}$ is perpendicular to geopotential surfaces (the "local vertical"), not toward Earth's center.

Step 4: The Coriolis Parameter and the f-plane

On a sphere, writing $\vec{\Omega} = \Omega\cos\phi\,\hat{j} + \Omega\sin\phi\,\hat{k}$(where $\hat{k}$ is the local vertical), the Coriolis acceleration for horizontal motion has components involving $f = 2\Omega\sin\phi$ (Coriolis parameter) and$\tilde{f} = 2\Omega\cos\phi$. For large-scale flow where the shallow atmosphere approximation holds, $\tilde{f}$ terms are negligible, and the horizontal equations become:

$$\boxed{\frac{Du}{Dt} - fv = -\frac{1}{\rho}\frac{\partial p}{\partial x} + F_x}$$

$$\boxed{\frac{Dv}{Dt} + fu = -\frac{1}{\rho}\frac{\partial p}{\partial y} + F_y}$$

Beta-Plane Approximation

The Coriolis parameter varies with latitude. Expanding about a reference latitude $\phi_0$:

$$f = 2\Omega\sin\phi \approx f_0 + \beta y$$

where $f_0 = 2\Omega\sin\phi_0$ and $\beta = df/dy = (2\Omega\cos\phi_0)/a$with Earth's radius $a \approx 6371\;\text{km}$. At $\phi_0 = 45°$,$\beta \approx 1.62 \times 10^{-11}\;\text{m}^{-1}\text{s}^{-1}$. The $\beta$-effect is responsible for the existence of Rossby waves and the westward propagation of planetary-scale disturbances.

Scale Analysis

The relative importance of terms is measured by dimensionless numbers. Let $U$ be typical wind speed, $L$ horizontal scale, and $T = L/U$ the advective time scale:

Number	Definition	Meaning	Typical Value
Rossby number	$Ro = U/(fL)$	Inertia vs. Coriolis	$\sim 0.1$ (synoptic)
Ekman number	$Ek = \nu/(fL^2)$	Viscosity vs. Coriolis	$\sim 10^{-8}$
Froude number	$Fr = U/(NH)$	Inertia vs. buoyancy	$\sim 0.01$

When $Ro \ll 1$, the flow is nearly geostrophic (Coriolis balances pressure gradient). When $Ro \sim 1$ (tornadoes, dust devils, tropical cyclone eyewalls), the full nonlinear equations are needed.

Pressure Gradient Force in Pressure Coordinates

On constant-pressure surfaces (isobaric coordinates), the pressure gradient force transforms to the gradient of geopotential $\Phi = gz$:

$$-\frac{1}{\rho}\nabla_z p \;\longrightarrow\; -\nabla_p \Phi$$

This eliminates density from the horizontal equations, which is why pressure coordinates are the standard framework in numerical weather prediction.

Friction Parameterization

The friction term $\vec{F}$ represents molecular and turbulent stresses. For molecular viscosity: $\vec{F} = \nu\nabla^2\vec{v}$ where $\nu \approx 1.5 \times 10^{-5}\;\text{m}^2\text{s}^{-1}$. For turbulent (eddy) viscosity in the boundary layer: $\vec{F} = K_m\nabla^2\vec{v}$ where$K_m \sim 1\text{-}100\;\text{m}^2\text{s}^{-1}$, many orders of magnitude larger than molecular viscosity.

Coriolis Parameter at Various Latitudes

Latitude	$f\;(\times 10^{-5}\;\text{s}^{-1})$	Inertial period $2\pi/f$
0° (Equator)	0	$\infty$
15°	3.77	46.3 hr
30°	7.29	23.9 hr
45°	10.31	16.9 hr
60°	12.63	13.8 hr
90° (Pole)	14.58	12.0 hr

The Primitive Equations (Full Set)

The complete set of primitive equations in pressure coordinates, used in all major weather and climate models, is:

$$\frac{Du}{Dt} - fv = -\frac{\partial\Phi}{\partial x} + F_x \qquad\text{(x-momentum)}$$

$$\frac{Dv}{Dt} + fu = -\frac{\partial\Phi}{\partial y} + F_y \qquad\text{(y-momentum)}$$

$$\frac{\partial\Phi}{\partial p} = -\frac{R_d T_v}{p} \qquad\text{(hydrostatic)}$$

$$\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y} + \frac{\partial\omega}{\partial p} = 0 \qquad\text{(continuity)}$$

$$\frac{DT}{Dt} - \frac{R_d T_v}{c_p p}\,\omega = \frac{Q}{c_p} \qquad\text{(thermodynamic)}$$

$$\frac{Dq}{Dt} = E - C \qquad\text{(moisture)}$$

Here $\omega = Dp/Dt$ is the pressure vertical velocity ($\omega < 0$ for rising motion),$Q$ is the diabatic heating rate, $q$ is specific humidity, and $E - C$ is evaporation minus condensation. These six equations plus an equation of state form a closed system for the seven unknowns $(u, v, \omega, \Phi, T, q, \rho)$.

10. Geostrophic Wind

The geostrophic wind is the theoretical wind that results from an exact balance between the Coriolis force and the pressure gradient force. It is the single most important diagnostic relationship in large-scale meteorology, providing a direct link between the pressure (or geopotential height) field and the wind field.

Step 1: Scale Analysis Leading to Geostrophic Balance

Start from the full horizontal momentum equations. For synoptic-scale flow, the typical scales are:$U \sim 10\;\text{m s}^{-1}$, $L \sim 10^6\;\text{m}$,$T \sim L/U \sim 10^5\;\text{s}$, $f \sim 10^{-4}\;\text{s}^{-1}$,$\delta p \sim 10^3\;\text{Pa}$. Scale each term:

Term	Scale	Magnitude
$Du/Dt$ (acceleration)	$U^2/L$	$\sim 10^{-4}\;\text{m s}^{-2}$
$fv$ (Coriolis)	$fU$	$\sim 10^{-3}\;\text{m s}^{-2}$
$(1/\rho)\partial p/\partial x$ (PGF)	$\delta p/(\rho L)$	$\sim 10^{-3}\;\text{m s}^{-2}$

The Rossby number $Ro = U/(fL) = 10/(10^{-4} \times 10^6) = 0.1 \ll 1$, so the acceleration term is an order of magnitude smaller than the Coriolis and pressure gradient terms. Dropping $Du/Dt$ and friction:

$$-fv_g = -\frac{1}{\rho}\frac{\partial p}{\partial x}, \qquad fu_g = -\frac{1}{\rho}\frac{\partial p}{\partial y}$$

Step 2: Solving for Geostrophic Wind Components

Solving for $u_g$ and $v_g$:

$$u_g = -\frac{1}{\rho f}\frac{\partial p}{\partial y}, \qquad v_g = \frac{1}{\rho f}\frac{\partial p}{\partial x}$$

In compact vector form:

$$\boxed{\vec{V}_g = \frac{1}{\rho f}\,\hat{k}\times\nabla_h p}$$

Step 3: Geostrophic Wind on Pressure Surfaces

On constant-pressure surfaces, the pressure gradient force becomes the gradient of geopotential $\Phi = gz$. Starting from the relationship between height and pressure gradients:

$$\left(\frac{\partial p}{\partial x}\right)_z = -\rho g\left(\frac{\partial z}{\partial x}\right)_p = -\frac{\rho}{1}\frac{\partial\Phi}{\partial x}\bigg|_p$$

Substituting into the geostrophic relation:

$$\boxed{\vec{V}_g = \frac{1}{f}\,\hat{k}\times\nabla_p\Phi}$$

In components: $u_g = -\frac{1}{f}\frac{\partial\Phi}{\partial y}\bigg|_p$ and$v_g = \frac{1}{f}\frac{\partial\Phi}{\partial x}\bigg|_p$. This form is especially useful because density does not appear — the geopotential height field on a pressure surface directly determines the geostrophic wind.

Natural Coordinate Form

In natural coordinates (along-flow $s$, cross-flow $n$ pointing left), the geostrophic wind speed is:

$$V_g = -\frac{1}{\rho f}\frac{\partial p}{\partial n}$$

Since $\partial p/\partial n > 0$ (pressure increases to the left of flow in NH) and$f > 0$, we get $V_g > 0$: the wind blows with low pressure to its left.

Gradient Wind: Extension to Curved Flow

For curved flow with radius of curvature $R$, centripetal acceleration adds a term. The gradient wind equation in natural coordinates is:

$$\frac{V^2}{R} + fV = -\frac{1}{\rho}\frac{\partial p}{\partial n}$$

Solving this quadratic for $V$:

$$V = -\frac{fR}{2} \pm \sqrt{\frac{f^2R^2}{4} - \frac{R}{\rho}\frac{\partial p}{\partial n}}$$

There are four solution types, classified by the sign of $R$ and $\partial p/\partial n$:

Regular low ($R > 0$, cyclonic): subgeostrophic flow ($V < V_g$)
Regular high ($R < 0$, anticyclonic): supergeostrophic flow ($V > V_g$)
Anomalous low ($R > 0$, anticyclonic PGF): rare, very fast
Anomalous high ($R < 0$, cyclonic PGF): physically unrealizable for large $R$

This explains why pressure gradients near low-pressure centers can be much tighter (stronger winds) than near high-pressure centers, which have an upper bound on wind speed.

Ageostrophic Wind and Vertical Motion

The actual wind deviates from geostrophic balance. The ageostrophic wind is:

$$\vec{V}_{ag} = \vec{V} - \vec{V}_g$$

It is the ageostrophic wind that drives convergence, divergence, and hence vertical motion. From the equations of motion: $\vec{V}_{ag} = \frac{1}{f}\hat{k}\times\frac{D\vec{V}_g}{Dt}$. The ageostrophic component is typically only ~10% of the total wind, but it is crucial for weather development because it controls ascent, clouds, and precipitation.

Numerical Example

Given: On a 500 hPa chart at $\phi = 45°\text{N}$, two height contours 5520 m and 5580 m are separated by $\Delta n = 300\;\text{km}$.

Find: Geostrophic wind speed.

$$V_g = \frac{1}{f}\frac{\Delta\Phi}{\Delta n} = \frac{g\,\Delta z}{f\,\Delta n} = \frac{9.81 \times 60}{1.03 \times 10^{-4} \times 3 \times 10^5}$$

$$V_g = \frac{588.6}{30.9} = 19.0\;\text{m s}^{-1} \approx 37\;\text{kt}$$

If the contours were only 150 km apart (double the gradient), the wind would double to ~38 m/s (74 kt).

Limitations: Geostrophic balance fails near the equator ($f \to 0$), in the boundary layer (friction is important), at small scales ($Ro \sim 1$), and in regions of strong curvature (use gradient wind instead). Despite these limitations, the geostrophic wind provides an excellent first approximation for synoptic-scale mid-latitude flow and is the starting point for more sophisticated analyses.

11. Thermal Wind Relation

The thermal wind relation connects horizontal temperature gradients to vertical wind shear. It is not an actual wind but rather the vector difference of geostrophic wind between two pressure levels. This relationship is one of the most powerful diagnostic tools in synoptic meteorology, linking the mass (temperature) field to the momentum (wind) field.

Step 1: Differentiate Geostrophic Wind with Respect to Pressure

Begin with the geostrophic wind on pressure surfaces:

$$\vec{V}_g = \frac{1}{f}\,\hat{k}\times\nabla_p\Phi$$

Take the partial derivative with respect to pressure:

$$\frac{\partial\vec{V}_g}{\partial p} = \frac{1}{f}\,\hat{k}\times\nabla_p\!\left(\frac{\partial\Phi}{\partial p}\right)$$

Step 2: Apply the Hydrostatic Relation

The hydrostatic equation in pressure coordinates gives:

$$\frac{\partial\Phi}{\partial p} = -\alpha = -\frac{R_d T}{p}$$

Substituting:

$$\frac{\partial\vec{V}_g}{\partial p} = \frac{1}{f}\,\hat{k}\times\nabla_p\!\left(-\frac{R_d T}{p}\right) = -\frac{R_d}{fp}\,\hat{k}\times\nabla_p T$$

Converting to $\ln p$ coordinates (since $p\,\partial/\partial p = \partial/\partial\ln p$):

$$\boxed{\frac{\partial\vec{V}_g}{\partial\ln p} = -\frac{R_d}{f}\,\hat{k}\times\nabla_p T}$$

Step 3: Finite-Layer (Integrated) Form

Integrating between pressure levels $p_1$ (lower, higher pressure) and $p_2$ (upper, lower pressure):

$$\vec{V}_T \equiv \vec{V}_g(p_2) - \vec{V}_g(p_1) = -\frac{R_d}{f}\,\hat{k}\times\nabla_p\overline{T}\;\ln\frac{p_1}{p_2}$$

where $\overline{T}$ is the mean temperature of the layer. In component form:

$$u_T = -\frac{R_d}{f}\frac{\partial\overline{T}}{\partial y}\ln\frac{p_1}{p_2}, \qquad v_T = \frac{R_d}{f}\frac{\partial\overline{T}}{\partial x}\ln\frac{p_1}{p_2}$$

The thermal wind vector is parallel to the mean isotherms with cold air to the left in the Northern Hemisphere — just like geostrophic wind has low pressure to the left.

Backing vs. Veering: Temperature Advection Diagnosis

The change in wind direction with height reveals the sign of temperature advection:

Veering (wind turns clockwise with height, e.g. S at surface to SW aloft): indicates warm air advection (WAA). The thermal wind has a component from the warm side, rotating the total wind clockwise with height.
Backing (wind turns counterclockwise with height, e.g. SW at surface to S aloft): indicates cold air advection (CAA). The thermal wind rotates the total wind counterclockwise with height.

This is an immensely practical tool: a single rawinsonde sounding reveals the horizontal temperature advection pattern through the depth of the troposphere.

Application: Jet Stream Structure

The subtropical jet stream exists at ~200 hPa (~12 km) above the strongest horizontal temperature gradients at the surface. The thermal wind relation demands that strong meridional temperature gradients (e.g., $\partial T/\partial y \sim -10\;\text{K per 1000 km}$ at the polar front) produce strong westerly shear that accumulates through the troposphere, concentrating into a jet core with speeds of 50-100 m/s. The jet exists directly above the frontal zone because that is where $|\nabla_p T|$ is greatest.

Thermal Wind and Frontal Zones

Fronts tilt with height toward the cold air side. The thermal wind relation explains this: the strong temperature gradient across a front produces strong vertical wind shear, and the geostrophic adjustment requires the frontal surface to slope at an angle given by the Margules formula:

$$\tan\alpha = \frac{f\,\Delta V_g}{g\,\Delta T/\overline{T}}$$

Typical frontal slopes are ~1:50 to ~1:300 (very shallow), which is why frontal zones extend hundreds of kilometers horizontally but only a few kilometers vertically.

Numerical Example

Given: At $\phi = 45°\text{N}$, the mean temperature of the 1000-500 hPa layer decreases northward at $\partial\overline{T}/\partial y = -5 \times 10^{-6}\;\text{K m}^{-1}$(5 K per 1000 km), with no east-west gradient.

Find: The thermal wind (vertical shear of geostrophic wind) in the layer.

$$u_T = -\frac{287}{1.03 \times 10^{-4}} \times (-5 \times 10^{-6}) \times \ln\frac{1000}{500}$$

$$u_T = -\frac{287}{1.03 \times 10^{-4}} \times (-5 \times 10^{-6}) \times 0.693 = +9.7\;\text{m s}^{-1}$$

The thermal wind is from the west at 9.7 m/s. If the 1000 hPa wind is westerly at 5 m/s, the 500 hPa geostrophic wind would be westerly at ~15 m/s — the wind strengthens with height above a region where temperature decreases poleward.

Barotropic vs. Baroclinic Atmosphere

In a barotropic atmosphere, density is a function of pressure alone, so constant-pressure and constant-density surfaces coincide, $\nabla_p T = 0$, and the thermal wind vanishes — the geostrophic wind is the same at all levels. In a baroclinicatmosphere, temperature varies on pressure surfaces, creating thermal wind shear, tilted fronts, and the potential energy that fuels extratropical cyclones through baroclinic instability. The real atmosphere is always baroclinic in mid-latitudes, which is why weather systems develop there.

12. Continuity Equation (Mass Conservation)

The continuity equation is the mathematical statement of mass conservation for a fluid. Mass can neither be created nor destroyed, so any change in density within a fixed volume must be balanced by a net mass flux through the volume's boundaries. This equation, combined with the momentum and thermodynamic equations, forms the backbone of all atmospheric models.

Step 1: Full Eulerian Derivation from a Control Volume

Consider a fixed (Eulerian) control volume $\delta V = \delta x\,\delta y\,\delta z$. The mass inside at time $t$ is $\rho\,\delta V$. The rate of change of mass is:

$$\frac{\partial}{\partial t}(\rho\,\delta V) = \frac{\partial\rho}{\partial t}\,\delta V$$

Now compute the net mass flux into the volume through its six faces. In the $x$-direction, the mass flux (mass per unit area per unit time) entering through the left face at $x$ is$(\rho u)|_x\,\delta y\,\delta z$, and leaving through the right face at $x + \delta x$ is$(\rho u)|_{x+\delta x}\,\delta y\,\delta z$. The net inflow in $x$ is:

$$\left[(\rho u)|_x - (\rho u)|_{x+\delta x}\right]\delta y\,\delta z = -\frac{\partial(\rho u)}{\partial x}\,\delta x\,\delta y\,\delta z$$

Adding contributions from all three directions and setting the total equal to the rate of mass change:

$$\frac{\partial\rho}{\partial t}\,\delta V = -\left[\frac{\partial(\rho u)}{\partial x} + \frac{\partial(\rho v)}{\partial y} + \frac{\partial(\rho w)}{\partial z}\right]\delta V$$

Dividing by $\delta V$:

$$\boxed{\frac{\partial\rho}{\partial t} + \nabla\cdot(\rho\vec{v}) = 0}$$

This is the flux form (conservative form) of the continuity equation.

Step 2: Lagrangian Form and Equivalence

Expanding the divergence of a product:$\nabla\cdot(\rho\vec{v}) = \vec{v}\cdot\nabla\rho + \rho\,\nabla\cdot\vec{v}$. Substituting:

$$\frac{\partial\rho}{\partial t} + \vec{v}\cdot\nabla\rho + \rho\,\nabla\cdot\vec{v} = 0$$

Recognizing the material derivative $D\rho/Dt = \partial\rho/\partial t + \vec{v}\cdot\nabla\rho$:

$$\frac{D\rho}{Dt} + \rho\,\nabla\cdot\vec{v} = 0 \qquad\Longleftrightarrow\qquad \frac{1}{\rho}\frac{D\rho}{Dt} = -\nabla\cdot\vec{v}$$

This Lagrangian form states: following a parcel, the fractional rate of density change equals minus the velocity divergence. If the flow is divergent ($\nabla\cdot\vec{v} > 0$), parcels expand and density decreases.

Step 3: Pressure Coordinate Form

In pressure coordinates $(x, y, p, t)$, the "vertical velocity" is $\omega = Dp/Dt$(units Pa/s). Because pressure surfaces are material in the hydrostatic limit, the continuity equation transforms to:

$$\boxed{\frac{\partial u}{\partial x}\bigg|_p + \frac{\partial v}{\partial y}\bigg|_p + \frac{\partial\omega}{\partial p} = 0}$$

This is exact under the hydrostatic approximation — there is no density or time derivative. This remarkable simplification is a major reason why pressure coordinates are preferred in numerical weather prediction. The atmosphere effectively behaves as an incompressible fluid in pressure coordinates.

Step 4: Divergence Implies Vertical Motion

Rearranging the pressure-coordinate form:

$$\frac{\partial\omega}{\partial p} = -\left(\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y}\right)_p = -D_h$$

where $D_h$ is the horizontal divergence on a pressure surface. Integrating from the surface ($p = p_s$) upward with boundary condition $\omega(p_s) \approx 0$:

$$\omega(p) = -\int_{p_s}^{p} D_h\,dp' = \int_{p}^{p_s} D_h\,dp'$$

This is the kinematic method for estimating vertical velocity. Low-level convergence ($D_h < 0$) produces upward motion ($\omega < 0$), while low-level divergence produces subsidence. Typical synoptic-scale values are$\omega \sim -0.1\;\text{Pa s}^{-1}$ (ascent ~1 cm/s), rising to$\omega \sim -10\;\text{Pa s}^{-1}$ (~1 m/s) in thunderstorm updrafts.

Mass-Weighted Circulation

The continuity equation in pressure coordinates is the foundation for understanding large-scale overturning circulations. The Hadley cell, for example, has low-level convergence ($D_h < 0$) in the ITCZ driving ascent, upper-level divergence ($D_h > 0$) driving poleward flow, and subsidence in the subtropics. The mass stream function $\Psi$ satisfying$\bar{v} = \frac{g}{2\pi a\cos\phi}\frac{\partial\Psi}{\partial p}$ quantifies the zonal-mean overturning, with $\Psi \sim 10^{10}\;\text{kg s}^{-1}$ for the Hadley cell.

Anelastic Approximation

For deep convection where density variations are important but sound waves are not, the anelastic approximation replaces the full continuity equation with:

$$\nabla\cdot(\rho_0\vec{v}) = 0$$

where $\rho_0(z)$ is a reference density profile. This filters out sound waves (which would require very small time steps in a numerical model) while retaining the effects of the decrease of density with height on vertical motions.

Boussinesq Approximation

For shallow convection (depth $\ll H$), the density variation in the background state can be neglected entirely, giving the Boussinesq form: $\nabla\cdot\vec{v} = 0$. Density variations are retained only in the buoyancy term of the vertical momentum equation. This is valid when the depth of the convection is small compared to the scale height ($H \approx 7.5\;\text{km}$). The Boussinesq approximation is widely used in ocean modeling (where compressibility effects are tiny), boundary-layer studies, and laboratory analogs of atmospheric flows.

13. Vorticity Equation

Vorticity — the curl of the velocity field — measures the local spin of fluid elements. The vorticity equation, obtained by taking the curl of the momentum equations, governs how this spin evolves. It is arguably the single most important equation in dynamic meteorology because it filters out gravity waves and focuses on the balanced, rotational component of atmospheric flow.

Step 1: Definition of Vertical Vorticity

The vertical component of relative vorticity is:

$$\zeta = \frac{\partial v}{\partial x} - \frac{\partial u}{\partial y} = (\nabla\times\vec{v})\cdot\hat{k}$$

Positive $\zeta$ corresponds to counterclockwise rotation (cyclonic in the NH). The absolute vorticity includes the planetary contribution:

$$\eta = \zeta + f$$

where $f = 2\Omega\sin\phi$ is the Coriolis parameter (the vertical component of Earth's rotation vorticity $2\vec{\Omega}$).

Step 2: Deriving the Vorticity Equation

Start with the horizontal momentum equations:

$$\frac{\partial u}{\partial t} + u\frac{\partial u}{\partial x} + v\frac{\partial u}{\partial y} + w\frac{\partial u}{\partial z} - fv = -\frac{1}{\rho}\frac{\partial p}{\partial x} + F_x$$

$$\frac{\partial v}{\partial t} + u\frac{\partial v}{\partial x} + v\frac{\partial v}{\partial y} + w\frac{\partial v}{\partial z} + fu = -\frac{1}{\rho}\frac{\partial p}{\partial y} + F_y$$

Take $\partial/\partial x$ of the $v$-equation minus $\partial/\partial y$ of the $u$-equation. After careful algebra (using the product rule repeatedly and the continuity equation to replace $\partial u/\partial x + \partial v/\partial y$), the result is:

$$\boxed{\frac{D}{Dt}(\zeta + f) = -(\zeta + f)\left(\frac{\partial u}{\partial x} + \frac{\partial v}{\partial y}\right) + \frac{1}{\rho^2}\left(\frac{\partial\rho}{\partial x}\frac{\partial p}{\partial y} - \frac{\partial\rho}{\partial y}\frac{\partial p}{\partial x}\right) + \text{friction}}$$

Physical Interpretation of Each Term

Divergence (stretching) term: $-(\zeta+f)(\partial u/\partial x + \partial v/\partial y)$. Horizontal convergence ($D_h < 0$) stretches vertical vortex tubes, increasing$|\zeta + f|$ — the ice-skater effect. This is how convergence in a developing cyclone spins up the low-level circulation. Quantitatively, if $f = 10^{-4}\;\text{s}^{-1}$ and$D_h = -10^{-5}\;\text{s}^{-1}$, then $D\eta/Dt \approx 10^{-9}\;\text{s}^{-2}$, doubling the vorticity in about one day.
Solenoidal (baroclinic) term: $\frac{1}{\rho^2}(\nabla\rho\times\nabla p)\cdot\hat{k}$. Generates vorticity when constant-density and constant-pressure surfaces are not parallel. This drives sea-breeze circulations (land-sea temperature contrast tilts density surfaces relative to pressure surfaces) and is the ultimate source of vorticity in baroclinic instability.
Tilting term: (often separated out) converts horizontal vorticity into vertical vorticity through differential vertical motion. Important near fronts and in supercell thunderstorms where tilting of horizontal wind shear generates mesocyclones.
Friction: Generally damps vorticity in the boundary layer through Ekman pumping. In a cyclone, frictional convergence in the boundary layer induces upward motion (Ekman pumping), which can paradoxically intensify the storm by driving ascent.

Ertel's Potential Vorticity

A far more powerful conserved quantity is the Ertel potential vorticity (PV). It combines the vorticity and thermodynamic equations into a single scalar:

$$\boxed{PV = \frac{1}{\rho}(\zeta_a\cdot\nabla\theta) = \frac{1}{\rho}(2\vec{\Omega} + \nabla\times\vec{v})\cdot\nabla\theta}$$

where $\zeta_a = 2\vec{\Omega} + \nabla\times\vec{v}$ is the absolute vorticity vector and $\theta$ is potential temperature. On isentropic surfaces (constant $\theta$), this simplifies to:

$$PV = -g\,(\zeta_\theta + f)\frac{\partial\theta}{\partial p}$$

where $\zeta_\theta$ is relative vorticity on an isentropic surface and$-\partial\theta/\partial p$ is the static stability.

PV Conservation

For adiabatic ($D\theta/Dt = 0$), frictionless flow:

$$\frac{D(PV)}{Dt} = 0$$

PV is conserved following parcels. This is one of the most profound results in geophysical fluid dynamics. It means the atmosphere "remembers" its PV even as it undergoes complex three-dimensional rearrangements. PV units are $1\;\text{PVU} = 10^{-6}\;\text{K m}^2\text{kg}^{-1}\text{s}^{-1}$. The dynamical tropopause is typically defined by the $PV = 2\;\text{PVU}$ surface.

Rossby Waves from Linearized Barotropic Vorticity Equation

For barotropic, non-divergent flow, the vorticity equation simplifies to conservation of absolute vorticity: $D(\zeta + f)/Dt = 0$. Linearizing about a mean zonal flow$\bar{U}$ with perturbation stream function $\psi'$:

$$\frac{\partial}{\partial t}\nabla^2\psi' + \bar{U}\frac{\partial}{\partial x}\nabla^2\psi' + \beta\frac{\partial\psi'}{\partial x} = 0$$

Assuming wave solutions $\psi' = A\,e^{i(kx + ly - \omega t)}$ with total wavenumber$K^2 = k^2 + l^2$, the dispersion relation is:

$$\boxed{c = \frac{\omega}{k} = \bar{U} - \frac{\beta}{k^2 + l^2}}$$

Rossby waves always propagate westward relative to the mean flow (since $\beta > 0$). Longer waves move faster westward. In the atmosphere, the mean westerly flow typically exceeds the westward phase speed, so Rossby waves appear to move eastward but more slowly than the jet.

PV Thinking: How PV Anomalies Induce Circulations

A positive PV anomaly (e.g., a stratospheric intrusion) at upper levels induces cyclonic circulation below it and lowers the tropopause locally. A negative PV anomaly induces anticyclonic circulation. When an upper-level PV anomaly moves over a surface temperature anomaly (which acts as a low-level PV anomaly), the two can phase-lock and mutually amplify, producing rapid cyclogenesis — the essence of the Hoskins-McIntyre-Robertson PV perspective on baroclinic instability.

Numerical Example: Rossby Wave

Given: A zonal wavenumber-5 Rossby wave at $45°\text{N}$ with$\bar{U} = 15\;\text{m s}^{-1}$, $l = 0$ (no meridional tilt).

The wavelength: $\lambda = 2\pi a\cos\phi/5 = 2\pi(6.371\times 10^6)(0.707)/5 = 5660\;\text{km}$

Wavenumber: $k = 2\pi/\lambda = 1.11 \times 10^{-6}\;\text{m}^{-1}$

$$c = 15 - \frac{1.62 \times 10^{-11}}{(1.11 \times 10^{-6})^2} = 15 - 13.1 = 1.9\;\text{m s}^{-1}$$

The wave moves eastward at only 1.9 m/s (nearly stationary). This explains why wavenumber-5 patterns in the mid-latitudes are quasi-stationary and can persist for days, producing persistent weather patterns (blocking).

The stationary wave condition $c = 0$ gives $k_s = \sqrt{\beta/\bar{U}}$. For$\bar{U} = 15\;\text{m s}^{-1}$, $k_s = 1.04 \times 10^{-6}\;\text{m}^{-1}$, corresponding to wavenumber ~5. This is why the large-scale stationary wave pattern in mid-latitudes typically has 4-6 troughs and ridges around a latitude circle.

14. Radiative Transfer Equation

Radiative transfer governs the propagation of electromagnetic radiation through the atmosphere. It is the physical basis for the greenhouse effect, remote sensing, and climate modeling. The key concepts involve Planck's radiation law, Beer-Lambert absorption, and the competition between absorption and emission that determines Earth's energy balance.

The Planck Function

A blackbody at temperature $T$ emits radiation with spectral radiance given by the Planck function:

$$\boxed{B_\lambda(T) = \frac{2hc^2}{\lambda^5}\frac{1}{e^{hc/(\lambda k_B T)} - 1}}$$

where $h = 6.626 \times 10^{-34}\;\text{J s}$ (Planck's constant),$c = 3 \times 10^8\;\text{m s}^{-1}$ (speed of light), and$k_B = 1.381 \times 10^{-23}\;\text{J K}^{-1}$ (Boltzmann's constant). The Sun ($T \approx 5778\;\text{K}$) peaks in the visible (~0.5 $\mu$m), while Earth ($T \approx 255\;\text{K}$) peaks in the infrared (~11 $\mu$m).

Wien's Displacement Law

The wavelength of peak emission is found by setting $dB_\lambda/d\lambda = 0$. Defining $x = hc/(\lambda k_B T)$, the maximum occurs when:

$$(5 - x)\,e^x = 5 \qquad\Longrightarrow\qquad x \approx 4.965$$

Giving Wien's law:

$$\lambda_{\max} = \frac{hc}{4.965\,k_B T} = \frac{2898\;\mu\text{m}\cdot\text{K}}{T}$$

For the Sun: $\lambda_{\max} = 2898/5778 = 0.50\;\mu\text{m}$ (green light). For Earth: $\lambda_{\max} = 2898/255 = 11.4\;\mu\text{m}$ (thermal infrared).

Stefan-Boltzmann Law

Integrating the Planck function over all wavelengths and all angles in a hemisphere gives the total emitted flux:

$$F = \int_0^\infty \pi B_\lambda(T)\,d\lambda = \sigma T^4$$

where $\sigma = 2\pi^5 k_B^4/(15 c^2 h^3) = 5.67 \times 10^{-8}\;\text{W m}^{-2}\text{K}^{-4}$is the Stefan-Boltzmann constant. The $T^4$ dependence means a 1% increase in temperature produces a 4% increase in emitted radiation — a strong negative feedback.

Kirchhoff's Law

At thermal equilibrium, a body's spectral absorptivity equals its spectral emissivity:

$$\alpha_\lambda = \varepsilon_\lambda$$

A good absorber at wavelength $\lambda$ is also a good emitter at that wavelength. This is why $\text{CO}_2$, which strongly absorbs 15 $\mu$m radiation, also strongly emits at 15 $\mu$m — and why the greenhouse effect simultaneously blocks outgoing radiation and adds downwelling radiation to the surface.

Radiative Transfer Equation

As a beam of monochromatic radiation of intensity $I_\lambda$ travels a distance $ds$ through an absorbing and emitting medium:

$$dI_\lambda = -k_\lambda\,\rho\,I_\lambda\,ds + j_\lambda\,\rho\,ds$$

where $k_\lambda$ is the mass absorption coefficient and $j_\lambda$is the mass emission coefficient. Defining optical depth $d\tau_\lambda = k_\lambda\,\rho\,ds$and source function $J_\lambda = j_\lambda/k_\lambda$:

$$\boxed{\frac{dI_\lambda}{d\tau_\lambda} = -I_\lambda + J_\lambda}$$

For pure absorption ($J_\lambda = 0$), this gives Beer-Lambert's law:$I_\lambda = I_{\lambda,0}\,e^{-\tau_\lambda}$. For LTE,$J_\lambda = B_\lambda(T)$ and the formal solution is:

$$I_\lambda(\tau_\lambda) = I_{\lambda,0}\,e^{-\tau_\lambda} + \int_0^{\tau_\lambda} B_\lambda(T)\,e^{-(\tau_\lambda - \tau')}\,d\tau'$$

Two-Stream Approximation

For atmospheric applications, the radiation field is split into upward ($F^\uparrow$) and downward ($F^\downarrow$) fluxes. In the infrared, assuming no scattering:

$$\frac{dF^\uparrow}{d\tau} = F^\uparrow - \pi B(T), \qquad \frac{dF^\downarrow}{d\tau} = -(F^\downarrow - \pi B(T))$$

The net flux is $F_{\text{net}} = F^\uparrow - F^\downarrow$, and the radiative heating rate is $Q_R = -(1/\rho c_p)\,dF_{\text{net}}/dz$.

Greenhouse Effect: Layer Models

Simple 1-layer model: The atmosphere is transparent to solar radiation but absorbs all infrared radiation (a single absorbing layer). Energy balance for the layer and the surface gives:

$$T_{\text{surface}} = 2^{1/4}\,T_{\text{eff}} \approx 1.19\,T_{\text{eff}}$$

N-layer model: With $N$ perfectly absorbing layers:

$$T_{\text{surface}} = (N+1)^{1/4}\,T_{\text{eff}}$$

For $T_{\text{eff}} = 255\;\text{K}$ and $N = 1$:$T_s = 2^{1/4} \times 255 = 303\;\text{K}$ (30°C). The actual Earth surface is ~288 K, between the zero-layer (255 K) and one-layer (303 K) cases, reflecting the partial transparency of the atmosphere in the infrared.

Radiative Equilibrium Temperature of Earth

Balance incoming solar radiation (absorbed) against outgoing longwave radiation:

$$\frac{S_0}{4}(1 - \alpha) = \sigma T_{\text{eff}}^4$$

where $S_0 = 1361\;\text{W m}^{-2}$ is the solar constant, the factor of 4 accounts for the ratio of Earth's cross-section to its surface area, and $\alpha \approx 0.30$is the planetary albedo. Solving:

$$\boxed{T_{\text{eff}} = \left[\frac{S_0(1 - \alpha)}{4\sigma}\right]^{1/4} \approx 255\;\text{K} \;(-18°\text{C})}$$

Numerical Example: Effective Temperature

$$T_{\text{eff}} = \left[\frac{1361 \times (1 - 0.30)}{4 \times 5.67 \times 10^{-8}}\right]^{1/4} = \left[\frac{952.7}{2.268 \times 10^{-7}}\right]^{1/4}$$

$$= \left[4.202 \times 10^{9}\right]^{1/4} = 254.6\;\text{K} \approx 255\;\text{K}$$

Earth's actual mean surface temperature is ~288 K, so the greenhouse effect provides ~33 K of warming. Water vapor contributes ~21 K, $\text{CO}_2$ ~7 K, and other gases (ozone, methane, $\text{N}_2\text{O}$) the remainder.

CO₂ Radiative Forcing

The radiative forcing from doubling $\text{CO}_2$ is approximately:

$$\Delta F_{2\times\text{CO}_2} = 5.35\,\ln(2) \approx 3.7\;\text{W m}^{-2}$$

The logarithmic dependence arises because the center of the 15 $\mu$m $\text{CO}_2$absorption band is already saturated (optically thick). Additional $\text{CO}_2$ broadens the wings of the band logarithmically. The resulting equilibrium warming (with feedbacks) is:

$$\Delta T = \lambda\,\Delta F$$

where the climate sensitivity parameter $\lambda \approx 0.8\;\text{K/(W m}^{-2}\text{)}$for the no-feedback case (Planck response). With positive feedbacks (water vapor, ice-albedo, lapse rate), the equilibrium climate sensitivity is estimated at 2.5-4.0°C per doubling of$\text{CO}_2$.

Radiative transfer is the foundation of climate science. The competition between solar shortwave absorption (mostly at the surface) and infrared longwave emission (from the atmosphere) determines Earth's temperature structure, drives the general circulation, and determines the sensitivity of climate to perturbations in greenhouse gas concentrations.

Quick Reference — All Equations

#	Equation	Key Result	Physical Significance
1	Ideal Gas Law	$p = \rho R_d T$	Relates pressure, density, and temperature for dry air
2	Hydrostatic Equation	$dp/dz = -\rho g$	Pressure decreases with height due to weight of air above
3	Barometric Formula	$p(z) = p_0\,e^{-z/H}$	Pressure drops exponentially; halves every ~5.5 km
4	First Law (atmos.)	$c_p\,dT - dp/\rho = dq$	Energy conservation: heating changes temperature or does expansion work
5	Dry Adiabatic Lapse Rate	$\Gamma_d = g/c_p \approx 9.8\;\text{K/km}$	Cooling rate of a rising unsaturated air parcel
6	Potential Temperature	$\theta = T(p_0/p)^\kappa$	Temperature a parcel would have if brought adiabatically to 1000 hPa; conserved tracer
7	Clausius-Clapeyron	$de_s/dT = L_v e_s/(R_v T^2)$	Saturation vapor pressure increases ~7%/K; controls moisture and precipitation
8	Brunt-Väisälä Frequency	$N^2 = (g/\theta)\,d\theta/dz$	Oscillation frequency of displaced parcels; measures static stability
9	Equations of Motion	$Du/Dt - fv = -(1/\rho)\,\partial p/\partial x$	Newton's 2nd law in rotating frame; Coriolis deflects moving air
10	Geostrophic Wind	$\vec{V}_g = (1/\rho f)\,\hat{k}\times\nabla p$	Wind from Coriolis-PGF balance; flows parallel to isobars
11	Thermal Wind	$\partial\vec{V}_g/\partial\ln p = -(R_d/f)\,\hat{k}\times\nabla T$	Links horizontal temperature gradients to vertical wind shear; explains jet streams
12	Continuity Equation	$\partial\rho/\partial t + \nabla\cdot(\rho\vec{v}) = 0$	Mass conservation; horizontal convergence implies vertical motion
13	Vorticity Equation	$D(\zeta+f)/Dt = -(f+\zeta)\,\nabla_h\cdot\vec{v} + \ldots$	Governs spin of air; convergence spins up cyclones; basis for Rossby waves
14	Radiative Transfer	$dI_\lambda/d\tau_\lambda = -I_\lambda + J_\lambda$	Absorption vs. emission of radiation; underpins greenhouse effect and remote sensing

Prerequisites

Physics

• Classical mechanics
• Thermodynamics
• Fluid dynamics basics

Mathematics

• Calculus (multivariable)
• Differential equations
• Vector calculus

Chemistry

• General chemistry
• Photochemistry basics
• Chemical kinetics