Universal neural networks for real-time earthquake early warning trained with generalized earthquakes

Deep learning enhances earthquake monitoring capabilities by mining seismic waveforms directly. However, current neural networks, trained within specific areas, face challenges in generalizing to diverse regions. Here, we employ a data recombination method to create generalized earthquakes occurring at any location with arbitrary station distributions for neural network training. The trained models can then be applied universally with different monitoring setups for earthquake detection and parameter evaluation from continuous seismic waveform streams. This allows real-time Earthquake Early Warning (EEW) to be initiated at the very early stages of an occurring earthquake. When applied to substantial earthquake sequences across Japan and California (US), our models reliably report most earthquake locations and magnitudes within 4 seconds of the initial P-wave arrival, with mean errors of 2.6–7.3 km and 0.05–0.32, respectively. The generalized neural networks facilitate global applications of real-time EEW, eliminating complex empirical configurations typically required by traditional methods.