Lorenz Attractor
Three coupled ODEs, three parameters, and a strange attractor that never repeats. An interactive simulation of deterministic chaos — the butterfly effect made visible.
In 1963, MIT meteorologist Edward Lorenz simplified atmospheric convection to three coupled differential equations. He discovered that their solutions — for all their apparent simplicity — never settled into periodic orbits. Instead they traced a strange attractor: a fractal structure of infinite topological complexity, bounded in space but aperiodic in time.
This article examines the Lorenz system as a prototype of deterministic chaos: a system where the future is fully determined by the present, yet long-term prediction is impossible in practice. The interactive simulation above allows exploration of the parameter space — σ (Prandtl number), ρ (Rayleigh number), β (aspect ratio) — and direct observation of sensitive dependence via the Ghost trajectory.
Historical Context — Convection and Unpredictability
Lorenz was studying a simplified model of Rayleigh-Bénard convection: a fluid layer heated from below. The variable x is proportional to the rate of convective turnover; y is the temperature difference between ascending and descending currents; z is the deviation of the vertical temperature profile from linearity. The equations:
dx/dt = σ(y − x)
dy/dt = x(ρ − z) − y
dz/dt = xy − βz
were obtained by truncating the Navier-Stokes equations to the first Fourier modes. Lorenz discovered their chaotic behaviour by accident: restarting a computation from a printout (rounded to three decimal places instead of six) produced a completely different trajectory. A difference of 0.000127 had grown to span the entire attractor in less than two simulated days.
Fixed Points and Bifurcation
The Lorenz system has three fixed points: the origin (0, 0, 0), and the two symmetric points C± = (±√(β(ρ−1)), ±√(β(ρ−1)), ρ−1) — the two „lobes“ of the butterfly. The origin is always unstable for ρ > 1. The fixed points C± are stable for ρ < ρ_H (the Hopf bifurcation value, ≈ 24.74 for σ = 10, β = 8/3) and unstable for ρ > ρ_H. At the classic value ρ = 28, all fixed points are unstable — trajectories wander chaotically between the two lobes without converging to any equilibrium.
Strange Attractor Geometry
The Lorenz attractor is a strange attractor: a subset of phase space with fractal dimension d_KY ≈ 2.062, attracting nearby trajectories at an exponential rate but allowing no convergence to a fixed cycle. It is strange because it is neither a point, nor a curve, nor a surface — it is a geometric object of fractional dimension.
The largest Lyapunov exponent λ₁ ≈ 0.906 s⁻¹ quantifies the average rate of exponential divergence. The Lyapunov time τ_L = 1/λ₁ ≈ 1.1 s is the characteristic time after which a perturbation of size ε has grown to order 1. Errors in knowledge of initial conditions double approximately every τ_L · ln 2 ≈ 0.76 s — making any prediction beyond about 30–40 Lyapunov times operationally meaningless.
∞ Lorenz Attractor — Interactive Simulation
The system evolves under dx/dt = σ(y−x), dy/dt = x(ρ−z)−y, dz/dt = xy−βz. Classic parameters σ = 10, ρ = 28, β = 8/3 produce the strange attractor. The Ghost (red) starts with Δx = 0.01 — initially indistinguishable, diverging exponentially.
Drag on the canvas to orbit. The colour encodes velocity magnitude (blue = slow, red = fast). Ghost trajectory (red line) starts Δx = 0.01 from the primary — watch exponential divergence.
λ Strange Attractor — Topology and Chaos
The attractor is a fractal surface of dimension ≈ 2.06 — not a 2D surface and not a 3D volume, but something between. Nearby trajectories separate at rate e^{λ₁t} while the attractor as a whole is dissipative (sum of Lyapunov exponents < 0): volume in phase space contracts at rate σ + 1 + β ≈ 13.7.