Slope as the fundamental measure of how things change through space and time
2026-02-26
You should know
How to plot points on axes, read a line on a graph, and solve a simple algebra equation.
You will learn
How slope measures change, how to read rate from a graph or equation, and why units like degrees per kilometre or dollars per hour matter.
Why this matters
Rate is the language of change. Once you understand slope, you are ready for almost every more advanced model in the book.
If this gets hard, focus on…
The meaning of “for every one unit of x, y changes by this much.” If that sentence makes sense, the rest of the chapter can grow from there.
The temperature at the surface of Calgary on a January morning might be −25°C. Drive an hour west into the Rockies, climb to 3,000 metres, and the temperature will be roughly −13°C — twelve degrees warmer. This apparent paradox — that the high mountain is milder than the city below — is an inversion, and it disappears under normal conditions: typically, air cools at about 6.5°C for every kilometre of altitude gained. That rate — 6.5°C/km — is constant enough that pilots, meteorologists, and mountain guides rely on it every day without thinking twice about the mathematics underneath it.
The environmental lapse rate is a linear relationship, and linear relationships are the simplest non-trivial tool in the geographer’s mathematical kit. They appear everywhere: river gradients, population growth, road grades, tax schedules. Each one is a straight line in disguise, governed by a single number — the slope. This model builds the language of slope from first principles, because slope is just rate of change, and rate of change is the foundation of every more complex model that follows.
How do we describe change that happens at a constant rate?
Temperature drops as you climb a mountain. Population grows year by year. A river’s gradient flattens as it approaches the sea. Some of these relationships are approximately linear over the range we care about. Others are exactly linear by design (a road built at 6% grade, a tax that increases by $50 per bracket).
The question is: How do we write down a linear relationship mathematically? What does the equation tell us? What does the slope mean, and why do the units matter?
A linear relationship is one where equal changes in input produce equal changes in output.
Walk up 100 meters in elevation → temperature drops 0.6°C
Walk up another 100 meters → temperature drops another 0.6°C
This is linearity: constant rate of change.
The general form is:
y = mx + b
Where: - y is the output (the thing that responds) - x is the input (the thing we control or observe changing) - m is the slope (the rate of change) - b is the intercept (the value of y when x = 0)
Slope is the engine of the model. It tells you how much y changes for each unit change in x.
Suppose we measure temperature at two elevations:
| Elevation (m) | Temperature (°C) |
|---|---|
| 1000 | 15.0 |
| 1500 | 12.0 |
How do we find the rate of change?
Change in temperature:
\Delta T = T_2 - T_1 = 12.0 - 15.0 = -3.0 \text{ °C}
Change in elevation:
\Delta z = z_2 - z_1 = 1500 - 1000 = 500 \text{ m}
Rate of change (slope):
m = \frac{\Delta T}{\Delta z} = \frac{-3.0 \text{ °C}}{500 \text{ m}} = -0.006 \text{ °C/m}
The negative sign tells us temperature decreases with elevation. The magnitude tells us by how much per meter.
We know m = -0.006 °C/m. We need the intercept b.
Using the point (z_1, T_1) = (1000, 15.0):
T = mz + b $15.0 = (-0.006)(1000) + b$ $15.0 = -6.0 + b$ b = 21.0 \text{ °C}
Full model:
T(z) = -0.006z + 21.0
Interpretation:
- At sea level (z = 0), temperature would be 21.0°C
- For every meter of elevation gained, temperature drops by 0.006°C
- At 2000 m: T = -0.006(2000) + 21.0 = 9.0 °C
Problem: A glacier is retreating. In 1990, its terminus was at 2400 m elevation. In 2020, it had retreated to 2700 m elevation. Assuming constant retreat rate:
(a) Rate of retreat
\Delta z = 2700 - 2400 = 300 \text{ m} \Delta t = 2020 - 1990 = 30 \text{ years} \text{rate} = \frac{300 \text{ m}}{30 \text{ yr}} = 10 \text{ m/yr}
The glacier terminus is moving upslope at 10 m/yr.
(b) Equation
Let z(t) be elevation, t be year.
z(t) = mt + b
We know m = 10 m/yr. Using the point (1990, 2400):
$2400 = 10(1990) + b$ b = 2400 - 19900 = -17500
Model:
z(t) = 10t - 17500
Check: z(2020) = 10(2020) - 17500 = 20200 - 17500 = 2700 ✓
(c) When will z = 3000?
$3000 = 10t - 17500$ $10t = 20500$ t = 2050
If the linear trend holds, the terminus will reach 3000 m in 2050.
Below is a simple interactive model. Adjust the parameters and watch the line respond.
<label>
Slope (m):
<input type="range" id="slope-slider" min="-2" max="2" step="0.1" value="0.5">
<span id="slope-value">0.5</span>
</label>
<label>
Intercept (b):
<input type="range" id="intercept-slider" min="-10" max="10" step="0.5" value="2">
<span id="intercept-value">2</span>
</label>Try this: - Set m = 0. What happens? (A horizontal line — no change with x) - Set m = 1 and b = 0. The line passes through the origin at 45°. - Set m = -1. The line tilts the other way. - Set m = 2 and b = -5. Where does the line cross the x-axis? (Solve $0 = 2x - 5 x = 2.5$)
The slope m is not just a number. It is a ratio of two physical quantities, and its units tell you what it measures.
Examples:
| Context | y | x | Slope m | Units | Meaning |
|---|---|---|---|---|---|
| Elevation profile | elevation | horizontal distance | \frac{\Delta z}{\Delta x} | m/m or % | gradient |
| Temperature lapse | temperature | elevation | \frac{\Delta T}{\Delta z} | °C/m | lapse rate |
| Population growth | population | time | \frac{\Delta P}{\Delta t} | people/year | growth rate |
| Cost function | total cost | quantity | \frac{\Delta C}{\Delta q} | $/unit | marginal cost |
Units are not decorative. They encode what the model means.
The absolute value |m| measures how fast the change happens: - |m| small: gradual change (prairie, slow warming, stable population) - |m| large: rapid change (cliff, thermal inversion, population boom)
Our glacier model predicted the terminus would reach 3000 m in 2050. But what if: - The glacier disappears entirely before then? - Climate policy reduces warming, slowing retreat? - A cold decade causes temporary advance?
Linear models are local approximations. They work well over the range where you measured them. Extrapolation is prediction, and prediction requires caution.
A common student error:
“The slope is 0.006.”
Wrong. The slope is -0.006 °C/m. Without units, you don’t know if you’re talking about temperature change, population change, or the cost of bananas.
The intercept b tells you where the line crosses the y-axis. It is often physically meaningful (sea-level temperature, initial population, fixed cost), but it is not the rate of change. That’s the slope’s job.
In T = -0.006z + 21, temperature depends on elevation. You don’t choose the temperature and look up the elevation. The convention is: - Independent variable (the one you control or measure): on the x-axis - Dependent variable (the one that responds): on the y-axis
Flipping these gives you a different slope. \frac{\Delta z}{\Delta T} is the inverse of \frac{\Delta T}{\Delta z}.
In the next model, we’ll ask: what happens when the rate of change itself changes? When growth accelerates or slows? When temperature doesn’t drop at a constant rate, but follows an exponential curve?
But before we get there, notice this: the slope of a line is the simplest possible derivative.
If y = mx + b, then:
\frac{dy}{dx} = m
The derivative is the rate of change. For a line, the rate is constant. For curves, it varies. That’s where calculus begins.
But you don’t need calculus to understand that slope is rate, and rate is how we measure change.
If you’re rusty on algebra, here’s the pattern:
Given: y = mx + b, and you know y, m, b — solve for x.
Steps: 1. Subtract b from both sides: y - b = mx 2. Divide both sides by m: x = \frac{y - b}{m}
Example: When does T = -0.006z + 21 reach T = 10°C?
$10 = -0.006z + 21$ -11 = -0.006z z = \frac{-11}{-0.006} = 1833.3 \text{ m}
At 1833 meters elevation, temperature would be 10°C.