Mathematics A-Level • IB • AP • University ●●● Advanced ★ Free

Lorenz Attractor

The butterfly effect visualized in 3D phase space

Simulation ready. Press Start to begin.

Edward Lorenz and Chaos Theory

In 1961, MIT meteorologist Edward Lorenz was running numerical weather simulations on an early computer. To save time, he restarted a simulation from the middle, entering data rounded to three decimal places instead of the full six. The result diverged completely from the original — weather forecasts became useless after just a few days.

This discovery led Lorenz to develop a simplified model of atmospheric convection in 1963, producing the now-famous Lorenz Attractor. His 1972 talk "Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?" popularized the term "butterfly effect" to describe sensitive dependence on initial conditions.

The Lorenz Equations

The Lorenz system consists of three coupled differential equations:

dx/dt = σ(y − x)
dy/dt = x(ρ − z) − y
dz/dt = xy − βz

Where:

  • σ (sigma) = Prandtl number (ratio of momentum to thermal diffusivity)
  • ρ (rho) = Rayleigh number (temperature difference)
  • β (beta) = geometric factor

Classic values: σ=10, ρ=28, β=8/3. These produce the iconic butterfly-shaped attractor.

Strange Attractors and Phase Space

The Lorenz Attractor is a strange attractor — a set in phase space with fractal structure. Unlike fixed points or limit cycles, strange attractors:

  • Never repeat — trajectories never pass through the same point twice
  • Remain bounded — all solutions stay within a finite region
  • Have fractal dimension — approximately 2.06 for the Lorenz attractor
  • Exhibit sensitive dependence — nearby trajectories diverge exponentially

The system alternates chaotically between circling the left and right lobes of the butterfly, with the choice unpredictable despite being deterministic.

Lyapunov Exponents and Predictability

The Lorenz system has a positive Lyapunov exponent, quantifying exponential divergence. If two trajectories start $\epsilon$ apart, after time $t$ they are approximately $\epsilon e^{\lambda t}$ apart, where $\lambda \approx 0.9$ for standard parameters.

This means:

  • Doubling time ≈ 0.77 time units
  • After 10 time units, trajectories 1000× farther apart
  • Finite precision (e.g., 10 decimal places) limits prediction to ~23 time units

This is why weather forecasts become unreliable beyond ~10 days, despite our equations being deterministic.

The Lorenz System: Where Chaos Theory Began

In 1963, meteorologist Edward Lorenz discovered that a simple 3-equation model of atmospheric convection produced unpredictable behaviour. Three coupled differential equations describe the system:

dx/dt = σ(y−x),   dy/dt = x(ρ−z)−y,   dz/dt = xy−βz

With classic parameters σ=10, ρ=28, β=8/3, the system traces the iconic butterfly-shaped strange attractor.

What is a Strange Attractor?

An attractor is a set of states towards which a system evolves. A strange attractor has fractal structure — it occupies a non-integer dimension (the Lorenz attractor has dimension ≈2.06). The trajectory never repeats exactly, though it stays confined to the butterfly shape forever. This is deterministic chaos: bounded, structured, yet never periodic.

Parameter Exploration

  • σ (Prandtl number): ratio of viscous to thermal diffusion
  • ρ (Rayleigh number): drives convection intensity — below ρ≈24.74, the system converges to fixed points; above, chaos emerges
  • β (geometric factor): relates to cell dimensions

Adjusting these parameters transitions the system between stable fixed points, periodic orbits, and chaotic behaviour.

Chaos Theory and Weather Prediction

Lorenz's discovery explains why weather cannot be predicted beyond approximately 2 weeks. Any measurement uncertainty — no matter how tiny — grows exponentially, making long-term prediction impossible. "Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas?" — Lorenz's 1972 lecture title that coined the term "butterfly effect."

Experiments to Try

  • Set ρ=28 for classic chaotic behaviour
  • Reduce ρ below 24 and watch the system settle into fixed points
  • Try ρ≈99.65 for a periodic orbit within chaos
  • Compare two trajectories with slightly different starting points to observe sensitive dependence

Key Equations

ConceptFormulaNotes
Lorenz system (x)dx/dt = σ(y − x)σ: Prandtl number (default 10)
Lorenz system (y)dy/dt = x(ρ − z) − yρ: Rayleigh number (default 28)
Lorenz system (z)dz/dt = xy − βzβ: geometric factor (default 8/3)
Chaos onsetρ > ρc ≈ 24.74Below: fixed points; above: chaos
Lyapunov exponentλ1 ≈ 0.906Positive ⇒ sensitive dependence on ICs
Attractor dimensiondH ≈ 2.06Fractal: non-integer Hausdorff dimension

Curriculum Relevance

LevelTopicRelevance
A-LevelDifferential equationsSystems of ODEs, numerical integration (Euler/RK4)
IB / APCalculus, mathematical modellingPhase portraits, parameter sensitivity
UndergraduateDynamical systems, fluid mechanicsNavier-Stokes approximation, chaos theory
PostgraduateNonlinear dynamics, ergodic theoryStrange attractors, Lyapunov spectra, fractal dimensions

Share this simulation

Send this page to students or colleagues.

Related simulations

Continue with similar experiments.

🔒 Unlock All 32 Simulations

Get unlimited access to all 32 simulations — including the Lorenz Attractor, Logistic Map, Mandelbrot Set, and more with MySimulator Premium.