Accuracy of locally constant field approximation in radiation-reaction simulations
In plain words
Computer codes for intense laser plasmas assume that light emission at each instant happens as if the field were constant. The question is how large the resulting errors are in the regime current experiments reach.
Precise statement
Particle-in-cell QED codes compute emission and pair creation with the locally constant field approximation (LCFA: constant-crossed-field rates evaluated at the local $\chi$), which fails for low-energy photons and for $a0 \sim< 10$. For head-on electron-laser collisions with $a0 = 1-20$ and $\chi = 0.1-1$, quantify the error of LCFA-based Monte Carlo relative to full strong-field QED (including interference within a formation length and multi-photon emission) in mean electron energy loss, energy spread and photon spectrum, and identify the corrections needed for $1\%$ accuracy in these observables.
What would settle it
Benchmarks of LCFA and corrected emission models against exact plane-wave QED calculations for the stated parameter ranges, published as an error table.
Status in the literature
Unverified note
Beyond-LCFA and locally monochromatic emission models have been developed in recent years; I do not have a verified source for a percent-level error budget as of 2026.