LAUSR

S U M M A R Y Secondary microseisms are ubiquitous ambient noise vibrations due to ocean activity, dominating worldwide seismographic records at seismic periods between 3 and 10 s. Their origin is a heterogeneous distribution of pressure fluctuations along the ocean surface. In spherically symmetric earth models, no Love surface waves are generated by such a distributed surface source. We present global-scale modelling of three-component secondary microseisms using a spectral-element method, which naturally accounts for a realistic distribution of surface sources, topography and bathymetry, and 3-D heterogeneity in Earth’s crust and mantle. Seismic Love waves emerge naturally once the system reaches steady state. The ergodic origin of Love waves allows us to model the horizontal components of secondary microseisms for the first time. Love waves mostly originate from the interaction of the seismic wavefield with heterogeneous Earth structure in which the mantle plays an important role despite the short periods involved. Bathymetry beneath the source region produces weak horizontal forces that are responsible for a weak and diffuse Love wavefield. The effect of bathymetric force splitting into radial and horizontal components is overall negligible when compared to the effect of 3-D heterogeneity. However, we observe small and well-focused Love-wave arrivals at seismographic stations in Europe due to force splitting at the steepest portion of the North Atlantic Ridge and the ocean–continent boundary. The location of the sources of Love waves is seasonal at periods shorter than about 7 s, while seasonality is lost at the longer periods. Sources of Rayleigh and Love waves from the same storm may be located very far away, indicating that energy equipartitioning might not hold in the secondary microseism period band.