Researchers ran fully nonlinear three‑dimensional simulations of gravitational waves in a simple expanding universe and found clear signs that the waves transfer energy from one wavelength to others. In the simulations, random collections of waves started with most energy concentrated at a single scale. As the model universes expanded, that energy spread both to shorter wavelengths and to longer ones — a behavior called a cascade. The amount and speed of the cascade matched predictions from a four‑wave scattering model of wave turbulence.

To make the experiment concrete, the team built initial data on a compact spatial manifold shaped like a three‑dimensional torus (a cube with opposite faces identified). They represented the gravitational waves by the usual transverse‑traceless tensor harmonics and gave each mode a random phase. The chosen spectral shape f(κ)=16 κ^3/(64+κ^6) put the initial peak at a wavelength equal to half the size of the torus and cut off modes with κ>5. They tested three overall amplitude scales, As = 0.001, 0.01, and 0.1, solved Einstein’s constraint equations with a pseudo‑spectral code (SpEC) and then evolved the full Einstein vacuum equations with a covariant symmetric‑hyperbolic formulation until the spatial volume had grown by more than an order of magnitude.

What the authors measured was the evolving spectrum of the gravitational waves. Over time the peak in the spectrum lost energy that flowed both to higher wavenumbers (shorter wavelengths) and to lower wavenumbers (longer wavelengths). The flows depended on the initial amplitude in the way the four‑wave scattering theory predicts. In other words, the numerical results support the idea that nonlinear interactions among waves can produce a turbulent redistribution of energy, at least in the initial stages seen here.

This work is important because it tests, with full numerical relativity, analytic ideas about gravitational‑wave turbulence. Earlier studies were mostly analytical or limited to special spacetimes. By evolving the full nonlinear equations in three dimensions and observing amplitude scaling consistent with theory, the study gives direct evidence that wave interactions can drive spectral cascades in an expanding setting. That helps clarify when simplified wave‑turbulence models are a good guide to the complex behavior of Einstein’s equations.