Visual walkthrough — Atmospheric drag — exponential atmosphere model, orbit decay

---

Step 0 — What an arrow (vector) means, and how to "dot" two of them

WHAT: A vector is a quantity with both a size and a direction — we draw it as an arrow. We put a little hat on the letter, like $\vec{v}$, to say "this is an arrow, not just a number". The plain letter $v$ (no arrow) means only its length — the speed, a single positive number.

WHY introduce this first? In Steps 3 and 5 we handle drag as an arrow that points backwards, and we ask "how much of one arrow lies along another". Those need two ideas: the arrow notation $\vec{v}$, and the dot product.

PICTURE: An arrow $\vec{v}$ with its length labelled $v$; the unit arrow $\hat{v}$ (length exactly 1) showing pure direction; and two arrows meeting at an angle, whose dot product is drawn as a shadow.

Two facts we will lean on: an arrow dotted with itself gives its length squared, $\vec{v}\cdot \vec{v}=v^2$ (angle $0$, $\cos 0=1$); and two opposite arrows give a negative dot product ($\cos 180^{\circ }=-1$). That negative sign is exactly why drag drains energy.

---

Step 1 — What is the satellite actually hitting?

WHAT: A satellite is a solid body of mass $m$ moving through very thin air of density $\rho$ (density = how many kilograms of gas sit in each cubic metre). It moves at speed $v$. Its front face has area $A$ — the shadow it would cast if you shone a light along its direction of travel.

WHY: Before any force, we must count how much gas the satellite meets per second. That mass of gas is what pushes back.

PICTURE: In a small time $dt$ (a tiny sliver of a second), the satellite moves forward a distance $vdt$. It sweeps out a tube: a cylinder of cross-section $A$ and length $vdt$. Look at the blue tube.

The gas mass caught inside that tube is density $\times$ volume:

$d{m}_{air}=\underset{{\text{kg per m}}^3}{\underbrace{\rho }}\times \underset{{\text{m}}^2}{\underbrace{A}}\times \underset{\text{m (tube length)}}{\underbrace{vdt}}$

Each piece earns its place: $\rho$ turns volume into mass; $A\cdot vdt$ is the tube's volume. Note the difference between $d{m}_{air}$ (a little parcel of gas) and $m$ (the fixed mass of the satellite) — we will need both.

---

Step 2 — Turning "gas hit" into a force

WHAT: Force is momentum given away per second. Momentum is mass $\times$ velocity. The satellite shoves that swept gas roughly up to its own speed $v$.

WHY force and not energy here? We want the push-back the satellite feels — that is a force. The cleanest route to a force from a moving mass is Newton's original form: $F=\frac{d(\text{momentum})}{dt}$. That is exactly the tool for "stuff arriving and getting accelerated".

PICTURE: Red arrow = the push-back on the satellite, pointing opposite to motion. Each second, a fresh slug of gas $d{m}_{air}/dt$ gets flung to speed $\sim v$.

$F\sim \frac{d{m}_{air}}{dt}\cdot v=\left(\rho Av\right)\cdot v=\rho A{v}^2$

Term by term: $\frac{d{m}_{air}}{dt}=\rho Av$ is gas-mass-per-second (from Step 1, divided by $dt$); multiplying by $v$ (the speed we give it) turns mass-rate into momentum-rate = force.

Real bodies don't hand all their speed to the gas — some slips around the sides. We fold that fudge into one dimensionless number $C_D$ (the drag coefficient) and the conventional factor $\frac{1}{2}$:

To get acceleration we divide the force by the satellite's own mass $m$ (Newton's $a=F/m$), and we group the fixed satellite properties into one symbol $B\equiv \frac{C_{D}A}{m}$:

$a_{drag}=\frac{1}{2}B\rho v^2,B\equiv \frac{C_{D}A}{m}\left(\text{units }\frac{{\text{m}}^2}{\text{kg}}\right).$

---

Step 3 — The direction: why $v\vec{v}$ and never ${\vec{v}}^2$

WHAT: Drag always points backwards along the path. Using the arrow notation from Step 0, we must write a vector whose length grows like $v^2$ but whose arrow points along $-\vec{v}$ (the minus flips the arrow to the opposite direction).

WHY: "${\vec{v}}^2$" is meaningless — you can't square an arrow and keep a direction. The trick: split $v^2$ into (a scalar length $v$) $\times$ (the arrow $\vec{v}$). Recall $\vec{v}=v\hat{v}$ from Step 0, so $v\vec{v}=v^2\hat{v}$ — size $v^2$, direction $\hat{v}$.

PICTURE: Velocity $\vec{v}$ in green; the drag acceleration in red, exactly anti-parallel, its length $\propto v^2$.

---

Step 4 — Two orbit facts we must import (built, not assumed)

WHAT: To talk about the orbit changing, we need two relations from the pure gravity problem (no drag yet). For a near-circular orbit of radius/semi-major-axis $a$ around a planet with $\mu =GM$:

$v=\sqrt{\frac{\mu }{a}}\text{and}\epsilon =-\frac{\mu }{2a}.$

WHY these two? Drag acts on speed $v$; but decay is measured as shrinking size $a$. We need a bridge between $v$, the energy $\epsilon$, and $a$. These come straight from the Two-Body Problem and are the heart of the Vis-Viva Equation.

PICTURE: A green circular orbit. As $a$ (radius) shrinks, the speed arrow (yellow) gets longer — a first glimpse of the paradox.

Read $\epsilon =-\mu /2a$: energy is negative (bound orbit) and gets more negative as $a$ shrinks — losing energy pulls the orbit inward. Read $v=\sqrt{\mu /a}$: smaller $a$ ⇒ bigger $v$.

---

Step 5 — How fast does drag drain energy? (Power)

WHAT: Power is energy changed per second. For a force, power $=$ force $\cdot$ velocity — and that "$\cdot$" is the dot product from Step 0. Since drag opposes $\vec{v}$, that dot product is negative — energy leaks out.

WHY the dot product? The dot product $\vec{a}\cdot \vec{v}$ picks out only the part of the force along the motion — the only part that does work. Drag is fully anti-parallel (angle $180^{\circ }$, $\cos 180^{\circ }=-1$), so it does maximum negative work.

PICTURE: Drag arrow (red) and velocity arrow (green) point opposite ways; the dot product of opposite arrows is negative — shown as a downhill energy slide.

$\frac{d\epsilon }{dt}={\vec{a}}_{drag}\cdot \vec{v}=(-\frac{1}{2}B\rho v\vec{v})\cdot \vec{v}=-\frac{1}{2}B\rho v(\vec{v}\cdot \vec{v})=-\frac{1}{2}B\rho v^3.$

Term by term: $\vec{v}\cdot \vec{v}=v^2$ (an arrow dotted with itself gives its length squared — Step 0); multiplied by the leftover $v$ gives $v^3$; the minus sign says energy only ever goes down.

---

Step 6 — Convert "energy lost" into "orbit shrinks"

WHAT: We have $d\epsilon /dt$. We want $da/dt$. Use $\epsilon =-\mu /2a$ and differentiate to link the two rates.

WHY differentiate? $\epsilon$ depends on $a$. The chain rule tells us: if $\epsilon$ drops by this much per second, how much does $a$ move? Differentiation is the exact tool that answers "rate of one thing in terms of the rate of another".

PICTURE: The $\epsilon$-vs-$a$ curve. A small slide down in $\epsilon$ (red) maps through the slope to a small slide inward in $a$ (blue).

Differentiate $\epsilon =-\frac{\mu }{2a}$ with respect to time:

$\frac{d\epsilon }{dt}=\frac{d}{dt}\left(-\frac{\mu }{2a}\right)=\frac{\mu }{2a^2}\frac{da}{dt}.$

Here $\frac{d}{da}\left(-\frac{\mu }{2a}\right)=+\frac{\mu }{2a^2}$ (the derivative of $1/a$ is $-1/a^2$; the two minus signs cancel), and the extra $\frac{da}{dt}$ is the chain rule.

Now set the two expressions for $d\epsilon /dt$ equal (Step 5 = Step 6):

$\frac{\mu }{2a^2}\frac{da}{dt}=-\frac{1}{2}B\rho v^3.$

---

Step 7 — Substitute the orbit facts and clean up

WHAT: Replace $v^3$ using $v^2=\mu /a$, so $v^3=(\mu /a{)}^{3/2}$, and solve for $da/dt$.

WHY: We want the answer purely in terms of $a$ (and constants), because $a$ is what we track as the orbit decays.

PICTURE: The algebra as a cancellation cascade — the $\mu$'s and $a$'s combine into a single clean $\sqrt{\mu a}$.

$\frac{\mu }{2a^2}\frac{da}{dt}=-\frac{1}{2}B\rho {\left(\frac{\mu }{a}\right)}^{3/2}.$

Multiply both sides by $\frac{2a^2}{\mu }$:

$\frac{da}{dt}=-B\rho \frac{a^2}{\mu }{\left(\frac{\mu }{a}\right)}^{3/2}=-B\rho \frac{a^2}{\mu }\cdot \frac{{\mu }^{3/2}}{a^{3/2}}.$

Collect powers: ${\mu }^{3/2}/\mu ={\mu }^{1/2}$ and $a^2/a^{3/2}=a^{1/2}$:

---

Step 8 — The runaway edge case (why the end is sudden)

WHAT: $\rho$ is not constant — from the exponential model, $\rho ={\rho }_0{e}^{-(h-h_0)/H}$. As $a$ shrinks, altitude $h$ drops, so $\rho$ climbs exponentially. Plug that back in: $|da/dt|$ contains $\rho$, so as the orbit falls, its own shrink-rate accelerates.

WHY show this case? A reader who assumes constant $\rho$ predicts a gentle linear glide. Reality is a nonlinear plunge — the model must not hide that.

PICTURE: Altitude vs time. Nearly flat for years, then a cliff at the end — the "catastrophic reentry".

---

The one-picture summary

WHAT this figure shows (read it top-left to bottom-left): each box is one step of the derivation, and each arrow is the logical move from one to the next. Box 1 — the satellite sweeps a tube of gas $dm=\rho Avdt$ (Step 1). Arrow to Box 2 — that swept mass, flung to speed $v$, becomes the drag force $\frac{1}{2}B\rho v^2$ (Step 2). Arrow to Box 3 — dotting force with velocity gives the power drain $-\frac{1}{2}B\rho v^3$ (Step 5). Arrow down to Box 4 — using $\epsilon =-\mu /2a$ and the chain rule links that energy loss to $a$ (Step 6). Arrow to Box 5 — substitute $v=\sqrt{\mu /a}$ (Step 7). Arrow to Box 6 — out drops the decay law $\frac{da}{dt}=-B\rho \sqrt{\mu a}$. Final arrow down — feeding an exponential $\rho$ makes it a runaway plunge (Step 8).

Recall Feynman retelling — the whole walk in plain words

A satellite scoops up a thin tube of air every second (Step 1). Shoving that air aside costs a backward push — a drag force that grows with the square of speed (Steps 2–3). That push always points against the motion, so it steadily drains the orbit's energy, second by second (Step 5). Now here's the twist: for a circle, less energy means a smaller circle, and a smaller circle means a faster satellite (Step 4). We translate "energy leaking out" into "circle shrinking" using a bit of calculus (Step 6), tidy the algebra (Step 7), and out pops one clean rule: the orbit shrinks at a rate $B\rho \sqrt{\mu a}$. Finally, since the lower it falls the thicker the air (altitude $h=a-R$ drops, density explodes), the whole thing snowballs — years of gentle drift, then a sudden fiery plunge (Step 8).

Recall One-line self-test

Why is the decay rate proportional to $\rho$, and why does that make reentry sudden? ::: $\frac{da}{dt}=-B\rho \sqrt{\mu a}$; $\rho$ grows exponentially as altitude falls, so the shrink-rate feeds itself into a runaway.

---

Connections

• Parent topic — the full theory and worked numbers.
• Two-Body Problem · Vis-Viva Equation — source of $v=\sqrt{\mu /a}$.
• Orbital Energy & Specific Mechanical Energy — the $\epsilon =-\mu /2a$ we differentiated.
• Hydrostatic Equilibrium (Atmospheres & Stars) — where $\rho (h)$'s exponential comes from.
• Perturbations in Orbital Mechanics · Reentry Aerodynamics & Ballistic Coefficient — the endgame.