The Lorenz system is a three-variable autonomous ODE originally derived from a severely truncated model of atmospheric convection: dx/dt = σ(y − x), dy/dt = x(ρ − z) − y, dz/dt = xy − βz with positive parameters σ, ρ, and β. For ρ > 1 the origin destabilizes; for sufficiently large ρ (with classic choices like σ = 10, β = 8/3, ρ = 28) trajectories settle onto a strange attractor—two lobes of fractal structure with sensitive dependence on initial conditions. The simulator integrates these equations to show the iconic butterfly-shaped set, illustrating stretching and folding in state space, bounded motion despite local instability (positive Lyapunov exponent), and the existence of a dissipative attractor volume. Idealizations include ignoring spatial structure of real fluids, stochastic forcing, and numerical truncation error, though the latter can be reduced with adaptive steps. Students connect the model to chaos, unpredictability in deterministic systems, and the limits of long-range weather forecasting.
Who it's for: Undergraduate dynamical systems, nonlinear dynamics, or computational science courses after linear systems and Runge–Kutta integration are familiar.
Key terms
Lorenz system
Strange attractor
Chaos
Sensitive dependence
Dissipative flow
Lyapunov exponent
Butterfly effect
Nonlinear ODE
How it works
The classic 1963 Lorenz system: three coupled nonlinear ODEs with sensitive dependence on initial conditions. Default parameters (σ,ρ,β) = (10,28,8/3) give the “butterfly” in the x–z projection.
Frequently asked questions
Is the Lorenz model realistic weather?
It captures qualitative features—nonlinearity, mixing, and sensitivity—but real forecasting models resolve millions of variables and physics absent here.
Why do two nearby trajectories diverge?
Local exponential instability along certain directions in state space amplifies tiny differences; nonlinear folding keeps trajectories bounded, producing the layered attractor geometry.
Does energy conserve like in Hamiltonian chaos?
The Lorenz system is dissipative (phase-space volumes contract). Hamiltonian chaos preserves volume; the mechanisms and tools differ.
Can I trust long runs with Euler's method?
Fixed-step explicit Euler quickly misrepresents trajectories; use higher-order integrators and awareness of timestep effects, as in the Euler-vs-RK4 pendulum demo.