Vectors, Relative Motion & Wind Triangles · Guide · glidepathscience.org
GuideVectors, Relative Motion & Wind Triangles
Contents
This Guide explains the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round, in the order a textbook would take them. Part 1 prepares the Intro track and Part 2 the Advanced track; each takes about fifteen minutes to read. Each section ends with a worked example that is not one of the sim's own problems.
Part 1 · The essentials
Vectors add tip to tail
Some quantities are described completely by a single number and a unit: a mass of 2 kg, a time of 30 s, a temperature of 20 °C. These are scalars. Others also need a direction before they mean anything. "Walk 3 km" does not tell you where you end up; "walk 3 km north" does. A quantity with both a size and a direction is a vector. Displacement, velocity, acceleration and force are all vectors, and almost every problem in this sim comes down to combining them correctly.
We write a vector with an arrow, $\vec A$, and its size, its magnitude, as $|\vec A|$ or simply $A$. A magnitude is never negative. On paper a vector is an arrow whose length, to some scale, is its magnitude and which points in its direction. Sliding an arrow around without turning or stretching it does not change the vector it represents, and that freedom is what makes vectors easy to add.
To add $\vec B$ to $\vec A$, slide $\vec B$ until its tail sits on the tip of $\vec A$. The sum, called the resultant $\vec R$, is the arrow from the tail of the first to the tip of the last:
$\vec R = \vec A + \vec B = \vec B + \vec A$
The order does not matter: putting $\vec B$ first and $\vec A$ second walks around the other two sides of a parallelogram and reaches the same point, so the resultant is that parallelogram's diagonal. Several vectors add the same way, each tail on the previous tip. Subtraction is addition of the reverse: $-\vec B$ has the same length as $\vec B$ and points the opposite way, and $\vec A - \vec B = \vec A + (-\vec B)$.
The most common mistake is to add the magnitudes. Lengths add only when the vectors point the same way. In general $|\vec A + \vec B| \le |\vec A| + |\vec B|$, and two vectors that point in nearly opposite directions can add to something shorter than either one.
Drawing arrows tip to tail shows what a sum looks like, but it is only as precise as the drawing. For numbers, we break each vector into parts along fixed axes.
Components, and the quadrant trap
Choose an x axis pointing east and a y axis pointing north, and describe a direction by the angle θ measured counterclockwise from the +x axis. Any vector in the plane is then the sum of one vector along x and one along y. Their signed lengths are the vector's components, $A_x$ and $A_y$. Picture the shadows the arrow casts on the two axes. A component is positive when its shadow points along the axis and negative when it points against it, so unlike a magnitude it can have either sign. With the unit vectors $\hat{\imath}$ and $\hat{\jmath}$, arrows of length one along +x and +y, the vector is
$A_x = A\cos\theta$ $A_y = A\sin\theta$ $\vec A = A_x\hat{\imath} + A_y\hat{\jmath}$
Going the other way, the components give back the magnitude by Pythagoras and the direction by the tangent:
$|\vec A| = \sqrt{A_x^2 + A_y^2}$ $\tan\theta = A_y/A_x$
Here is the trap. A calculator's inverse tangent only returns angles between −90° and +90°, the first and fourth quadrants, because $A_y/A_x$ is the same for a vector and for its reverse. When $A_x$ is negative, the vector actually lies in the second or third quadrant, and you must add 180°. When only $A_y$ is negative, add 360° to land between 0° and 360°. Software usually offers $\mathrm{atan2}(A_y, A_x)$, which looks at both signs and does this check for you. It also handles the y axis itself, where $A_x$ = 0 and the tangent is undefined: the direction is 90° or 270°. A vector with both components zero has no direction at all.
Components turn addition into arithmetic. To add vectors, add their x components and their y components separately, $R_x = A_x + B_x$ and $R_y = A_y + B_y$, then rebuild the magnitude and the direction of $\vec R$ from them.
Mathematicians measure angles counterclockwise from east. People who navigate do not, and that is the next thing to sort out.
Bearings, and winds named for where they come from
Pilots, sailors and surveyors give a direction as a bearing β: the angle measured clockwise from north, written with three digits. North is 360° (or 000°), east 090°, south 180° and west 270°. A bearing and a math angle describe the same direction from different starting lines and in opposite senses, so converting between them takes a reflection as well as a shift:
$\theta = 90^\circ - \beta$ (add 360° if the result is negative)
Winds add a second convention. A wind is named for the direction it blows from: a north wind comes out of the north and blows toward the south. On a map, then, the arrow for a wind from WD points the opposite way, toward WD + 180°. With east and north as the x and y components, a wind from bearing β at speed W is
$\vec w = -W(\sin\beta,\ \cos\beta)$
The minus sign does the reversing. Reading a wind as the direction it blows toward is the classic slip here, and it puts every later correction on the wrong side.
With directions settled, we can ask the question at the heart of the sim: measured by whom?
Velocity belongs to a frame
A velocity always means velocity relative to something. A passenger who sits still on a moving ferry has zero velocity relative to the ferry and the ferry's velocity relative to the shore. Neither answer is wrong; they are measured in different frames of reference. To keep frames straight, give every velocity two subscripts: the first names what moves and the second what it is measured against. $\vec v_{PG}$ is the passenger relative to the ground, $\vec v_{PF}$ the passenger relative to the ferry, and $\vec v_{FG}$ the ferry relative to the ground. They are linked by the chain rule for relative velocity:
$\vec v_{PG} = \vec v_{PF} + \vec v_{FG}$
The inside subscripts match and cancel, leaving the outside pair, so the chain can be checked by reading its letters. The addition is a vector addition. Velocities in the same line add as signed numbers, but in two dimensions they add tip to tail, exactly as in the first section.
A river is a moving frame that you can see, which makes it the clearest place to use the chain.
Crossing a river
A boat moves through the water at speed u, the water flows at speed w, and the river is W wide. Over the ground, the boat's velocity is the chain $\vec v_{BG} = \vec v_{BW} + \vec v_{WG}$: its velocity through the water plus the current's. Aim the boat at an angle φ upstream of straight across. Only the part of its velocity pointing across the river, $u\cos\varphi$, carries it to the far bank, and the current minus the upstream part, $w - u\sin\varphi$, carries it a distance x downstream while it crosses:
$t = \dfrac{W}{u\cos\varphi}$ $x = (w - u\sin\varphi)\,t$
Two results follow, and each corrects a common belief. Aimed straight across (φ = 0) the boat crosses in the least possible time, $W/u$, whatever the current. The current changes where the boat lands, not when, because it adds nothing across the river. To land directly opposite the start, the upstream part of the boat's velocity must cancel the current, $u\sin\varphi = w$. That is possible only when $u > w$, a boat faster than the river. Here the boat's speed u is the hypotenuse of the velocity triangle, so the angle comes from the sine, not the tangent, and the crossing then takes longer than $W/u$.
Rain falling past a moving car is the same chain, turned on its side.
Rain seen from a moving car
Raindrops fall at a steady terminal speed $v_R$, about 4 m/s for a drop 1 mm across and faster for larger drops. Seen from a car moving at $v_C$, the rain's velocity is $\vec v_{RC} = \vec v_{RG} + \vec v_{GC}$. The ground's velocity relative to the car, $\vec v_{GC} = -\vec v_{CG}$, points backward at the car's speed, so the rain also moves backward past the car at $v_C$. Its streaks on a side window are tilted from vertical by an angle ϕ:
$\vec v_{RC} = \vec v_{RG} - \vec v_{CG}$ $\tan\phi = v_C/v_R$
The faster the car, the flatter the streaks, and they always run toward the back of the car, never the front. A walker meets the same slanted rain and tilts an umbrella forward by the same angle. The slant belongs to the frame, not to the rain: someone standing still beside the road sees the same drops fall straight down.
An airplane in a wind is a boat in a current with one change: it can be blown sideways as well as along, and the pilot has to plan for that before takeoff.
Heading, track and the wind triangle
An airplane moves through the air at its true airspeed TAS, in the direction its nose points, its heading. Call that velocity $\vec v_{PA}$. The air itself moves with the wind, $\vec v_{AG}$. Over the ground the airplane follows its track at its groundspeed GS, $\vec v_{PG} = \vec v_{PA} + \vec v_{AG}$. Fly the intended course as the heading and the wind carries the airplane downwind: a wind from the right drifts it to the left of its course.
To hold a true course TC, the pilot turns the nose into the wind by the wind correction angle WCA. Split the wind, from direction WD at speed W, into a crosswind XW across the course and a headwind HW along it. Positive XW comes from the right, and positive HW blows against the airplane:
$\mathrm{XW} = W\sin(\mathrm{WD} - \mathrm{TC})$ $\mathrm{HW} = W\cos(\mathrm{WD} - \mathrm{TC})$
Only the crosswind needs correcting. The nose turns until the part of the airspeed pointing across the course cancels it, and what is left of the airspeed along the course, less the headwind, is the groundspeed:
$\sin\mathrm{WCA} = \mathrm{XW}/\mathrm{TAS}$ TH = TC + WCA GS = TAS cos WCA − HW
A positive XW, a wind from the right, gives a positive WCA, a correction to the right. Turning away from the wind is the usual sign error, and drawing the triangle catches it. The triangle closes only when $|\mathrm{XW}|$ is less than TAS, and the airplane makes progress only when GS is positive. Even a pure crosswind lowers the groundspeed, because part of the airspeed is spent pointing into the wind. The FAA's handbook draws the triangle the way the sim does: the wind from the start point, then the airspeed swung from its tip to meet the course line.
Part 2 returns to these ideas with more tools: the dot product, rivers that are not uniform, and problems that run backward.
Part 2 · Going further
Components along any direction, and the chain both ways
Components along x and y are a special case. The component of $\vec A$ along any unit vector $\hat u$ is the dot product, the length of $\vec A$ times the cosine of the angle α between them:
$\vec A\cdot\hat u = |\vec A|\cos\alpha = A_xu_x + A_yu_y$
The second form, from components, is the one to compute with. Rotating the axes changes a vector's components but never its length, so $A_x^2 + A_y^2$ is the same in every orientation. For a wind and a course, take $\hat c$ along the course and $\hat n$ a quarter turn to its right. The headwind and the crosswind are then
$\mathrm{HW} = -\vec w\cdot\hat c$ $\mathrm{XW} = -\vec w\cdot\hat n$
Both minus signs keep Part 1's convention, a positive headwind blowing against you and a positive crosswind coming from the right. The raw component $\vec w\cdot\hat n$ is positive for a wind blowing toward the right, which comes from the left. Dropping that sign flips the correction to the wrong side.
The chain rule itself comes from positions. If $\vec r_{PA}$ locates P from A, then $\vec r_{PA} = \vec r_{PB} + \vec r_{BA}$, and differentiating gives $\vec v_{PA} = \vec v_{PB} + \vec v_{BA}$. Reversing the subscripts reverses the vector, $\vec v_{AB} = -\vec v_{BA}$. Differentiating once more shows that observers moving at a steady velocity relative to each other measure the same acceleration.
Back on the river, the question changes from "where will it land?" to "where is the best place it can land?"
Optimizing a crossing
When the current outruns the boat, $w > u$, no heading lands opposite: the upstream part of the boat's velocity is never enough. The useful question is which heading lands closest. The drift $x(\varphi) = (w - u\sin\varphi)W/(u\cos\varphi)$ has its minimum where its derivative is zero, which gives
$\sin\varphi = u/w$
The picture explains it. The boat's velocity through the water has a fixed length u but any direction, so its tip lies on a circle around the tip of the current vector. The ground velocity points closest to straight across when it just touches that circle, and then $\vec v_{BW}$ is perpendicular to $\vec v_{BG}$.
Real rivers flow fastest in the middle and slowest at the banks. For a profile such as $w(y) = 4w_{max}\,y(W - y)/W^2$, where y is the distance from the near bank, and a fixed heading, the drift adds up the current over the whole crossing. Only the mean current $\bar w$ sets where the boat lands, here $\bar w = \tfrac23 w_{max}$, and every formula above holds with w replaced by $\bar w$. The path is no longer straight but an S-curve, steepest at the banks and slanting most at mid-channel.
Currents and winds slow some trips and speed others. Over a round trip, which wins?
The out-and-back race
Fly a distance d out and back at airspeed v in a steady wind w, with $v > w$ so that both legs are possible. With the wind along the route, one leg goes at $v - w$ and the other at $v + w$. With the wind straight across, each leg is flown with a wind correction at the groundspeed $\sqrt{v^2 - w^2}$:
along: $t = \dfrac{d}{v - w} + \dfrac{d}{v + w} = \dfrac{2dv}{v^2 - w^2}$ across: $t = \dfrac{2d}{\sqrt{v^2 - w^2}}$
Both are longer than the still-air time $2d/v$. The tempting idea that the tailwind gives back what the headwind took is wrong: the slow leg lasts longer than the fast leg, so the wind spends more time hurting than helping. A steady wind lengthens a round trip from any direction. In 1887 Michelson and Morley ran this race with light as the airplane and the supposed "ether wind" as the wind. They found far less difference between the two directions than the ether predicted, a null result that helped lead to special relativity.
So far we have predicted what an observer sees from velocities we knew. Often the problem runs the other way.
Inverse problems
The chain can also be solved backward, to find a true velocity from relative ones. If you know your own velocity and see the whole relative velocity, its direction and its speed, the chain gives the true velocity in one step: $\vec v_{RG} = \vec v_{RC} + \vec v_{CG}$. Often you see less than that. A cyclist in the rain sees only the slant of the drops, a direction without a speed, while the rain's true velocity has two unknowns, how far it drifts sideways and how fast it falls. One observation cannot fix two unknowns; two can. Riding at $v_1$ and seeing the rain fall vertically says the rain drifts along the road at $v_1$. Riding faster, at $v_2$, the rain slants back at an angle α from vertical, and its fall speed follows:
$v_{fall} = (v_2 - v_1)/\tan\alpha$
A pilot finds the wind the same way. Holding a known heading and true airspeed gives $\vec v_{PA}$, the GPS gives the track and groundspeed $\vec v_{PG}$, and the wind is the difference, $\vec v_{AG} = \vec v_{PG} - \vec v_{PA}$. Reported as a wind, it is named for the direction it comes from, the reverse of where this vector points.
The same backward reasoning, applied to two aircraft, is how pilots judge whether traffic is a threat.
Closest approach and collision courses
Two aircraft on steady courses are easiest to judge from one of them. In your own frame, traffic now at position $\vec r_0$ from you moves in a straight line, $\vec r(t) = \vec r_0 + \vec v_{rel}t$, at its velocity relative to you, $\vec v_{rel} = \vec v_B - \vec v_A$ (written $\vec v_{BA}$ in the figure). The separation is smallest where that line passes closest to you, at the foot of the perpendicular. That moment may already be past, so the time of closest approach from now on is never negative:
$t^* = \max\!\left(0,\ -\dfrac{\vec r_0\cdot\vec v_{rel}}{v_{rel}^2}\right)$ $d_{min} = |\vec r_0 + \vec v_{rel}t^*|$
When $t^* > 0$ the miss distance is also $|\vec r_0\times\vec v_{rel}|/v_{rel}$. When $t^*$ = 0 the traffic is already opening and the closest approach is now. When $\vec v_{rel}$ = 0 the separation never changes. When the relative track points straight at you, $d_{min}$ = 0: a collision course. Then the bearing to the traffic never changes while the range closes, which is the handbooks' warning. Traffic that holds still on the windshield and only grows is on a collision course; traffic that drifts across it will miss.
How big it looks depends on the range r. An airplane of wingspan b subtends the angle $2\tan^{-1}(b/2r)$. A light trainer with an 11 m wingspan subtends only 0.32° at 2 km and 2.1° at 300 m, so it grows slowly at first and then very fast. The sim's view out the windows draws it at that angle, magnified, at its bearing.
The last section brings these tools to an actual flight plan, along with the shortcuts and corrections pilots use.
Real flight planning
Pilots often estimate before they calculate. The 1-in-60 rule, WCA ≈ 60 × XW/TAS degrees, comes from $\sin\theta \approx \theta$ in radians, since one radian is about 57.3°, near 60. It runs about 4.7 % high for small crosswinds, is exact near XW/TAS = 0.5, and falls low beyond that. The clock method estimates the crosswind itself, as a quarter, a half, three quarters or all of the wind when it is 15°, 30°, 45° or 60° and more off the course. It is a little low at 15°, exact at 30°, and high at 45° and 60°.
Directions add one more layer. Charts are drawn to true north, but a magnetic compass points to magnetic north, which differs by the variation printed on the chart's isogonic lines. The magnetic heading is MH = TH − east variation, or + west variation ("east is least, west is best"). The airplane's own deviation card then converts the magnetic heading to the compass heading. Winds aloft forecasts (FB), METARs and TAFs give winds in degrees true; a tower's winds and the ATIS give them magnetic, like the runway numbers. An FB group such as 3415 means from 340° at 15 kt, and 9900 means light and variable. From 100 to 199 kt, 50 is added to the direction digits, so 8622 is from 360° at 122 kt.
The model has edges, and a good navigator knows where they are. It uses one wind per leg, though the wind changes with height and time. It flies straight legs on a flat local chart, though over long distances the true course of a great circle changes along the way. And it adds velocities as vectors, which holds only at speeds far below the speed of light.
Reference
Common mistakes
- Adding magnitudes. Lengths add only for vectors in the same direction; otherwise add components.
- Trusting the calculator's arctangent. Check the quadrant: add 180° when $A_x$ is negative, or use atan2.
- Reading a wind as where it goes. Winds are named for where they come from.
- Letting the current slow the crossing. Aimed straight across, the time is $W/u$ whatever the current.
- Using the tangent to land opposite. The boat's speed is the hypotenuse: $\sin\varphi = w/u$.
- Turning away from the wind. Correct into it: a wind from the right needs a heading right of the course.
- Expecting a crosswind to leave the groundspeed alone. GS = TAS cos WCA − HW.
- Ignoring traffic that does not move. A steady bearing and a closing range is a collision course.
Equation sheet
| Part 1 | |
|---|---|
| components | $A_x = A\cos\theta$, $A_y = A\sin\theta$ |
| back again | $|\vec A| = \sqrt{A_x^2 + A_y^2}$, $\tan\theta = A_y/A_x$ (check the quadrant) |
| bearing | $\theta = 90^\circ - \beta$; wind $\vec w = -W(\sin\beta,\ \cos\beta)$ |
| frames | $\vec v_{PG} = \vec v_{PM} + \vec v_{MG}$ |
| river | $t = W/(u\cos\varphi)$, $x = (w - u\sin\varphi)t$ |
| opposite; rain | $\sin\varphi = w/u$ ($u > w$); $\tan\phi = v_C/v_R$ |
| wind parts | $\mathrm{XW} = W\sin(\mathrm{WD} - \mathrm{TC})$, $\mathrm{HW} = W\cos(\mathrm{WD} - \mathrm{TC})$ |
| triangle | $\sin\mathrm{WCA} = \mathrm{XW}/\mathrm{TAS}$ ($|\mathrm{XW}|$ below TAS); TH = TC + WCA; GS = TAS cos WCA − HW (> 0) |
| Part 2 | |
|---|---|
| component | $\vec A\cdot\hat u = |\vec A|\cos\alpha$; $\mathrm{HW} = -\vec w\cdot\hat c$, $\mathrm{XW} = -\vec w\cdot\hat n$ |
| chain | $\vec v_{PA} = \vec v_{PB} + \vec v_{BA}$, $\vec v_{AB} = -\vec v_{BA}$ |
| least drift | $\sin\varphi = u/w$ ($w > u$); any profile: the mean $\bar w$ |
| round trip | ($v > w$) along $2dv/(v^2 - w^2)$; across $2d/\sqrt{v^2 - w^2}$ |
| inverse | $v_{fall} = (v_2 - v_1)/\tan\alpha$; $\vec v_{AG} = \vec v_{PG} - \vec v_{PA}$ |
| traffic | $t^* = \max(0,\ -\vec r_0\cdot\vec v_{rel}/v_{rel}^2)$, $d = |\vec r_0 + \vec v_{rel}t^*|$ |
| headings | WCA ≈ 60 × XW/TAS; MH = TH − east variation |
Symbols: θ math angle (from +x, counterclockwise) · β bearing (from north, clockwise) · W width (m) or wind speed · u, w boat and current speeds (m/s) · φ heading upstream of straight across · ϕ the rain's angle from vertical · $v_R$, $v_C$ rain fall and car speeds · TC true course · TH true heading · MH magnetic heading · WD wind direction (from) · TAS true airspeed · GS groundspeed · XW, HW crosswind (from the right +) and headwind · WCA wind correction angle (right positive) · $\vec r_0$, $\vec v_{rel}$ traffic's position and velocity relative to you · b wingspan, r range. 1 kt = 1852/3600 m/s; 1 NM = 1852 m.
Further reading
OpenStax, University Physics Volume 1, chapter 2 (vectors) and §4.5 (relative motion), whose order and notation this Guide follows; the FAA’s Pilot’s Handbook of Aeronautical Knowledge (FAA-H-8083-25), chapter 16 on navigation. The Field Manual lists every source.
Vectors, Relative Motion & Wind Triangles v1.3 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org