In central Japan, a slope showed a much larger response at about 3.6 hertz than its peak shaking level suggested.
Researchers analyzed earthquakes recorded by a four-station strong-motion array on a ridge in central Japan. The stations covered a ridge crest, a valley bottom and two slopes with different orientations. They compared peak ground acceleration with measures that show how shaking varied by frequency.
One slope consistently showed a narrow frequency band of stronger motion at 3.5–4.0 hertz. Its five-percent-damped response-spectrum ratio reached 6.67 at 0.277 seconds, or about 3.61 hertz, compared with a full-record horizontal peak-acceleration ratio of 1.13.
A narrow band of stronger shaking
The north–south-facing slope showed a persistent peak in its shaking relative to the valley station at 3.5–4.0 hertz. This pattern remained across the 24 earthquakes and different data-processing choices. Measurements of the actual ground-motion spectra confirmed more energy at the slope in this band, although the large relative ratio was also partly due to weaker motion at the valley reference station.
The geometric-mean five-percent-damped response-spectral ratio reached 6.67 at 0.277 seconds. By comparison, the corresponding geometric-mean horizontal peak ground acceleration ratio for the full records was 1.13. Peak-acceleration ratios varied between earthquakes and were positively associated with epicentral distance, but the study could not separate distance effects from earthquake-size effects in this set of events.
Why peak shaking can miss it
The results show that peak ground acceleration alone can miss a strong, frequency-specific response on a mountain slope. Including frequency-dependent measures could therefore give a fuller picture of how earthquake shaking affects slopes and other structures in mountainous terrain.
The observations do not identify a single cause for the amplification. The pattern may reflect topography, shallow ground conditions or both, but the study says these explanations cannot be distinguished from the measurements alone.
Recordings and open questions
The study used direct field recordings from 24 regional earthquakes measured by four stations on one ridge in central Japan. The consistent narrow-band peak across events and processing choices supports a repeatable local response at that site.
The evidence comes from a single four-station array, so it does not establish that all mountain slopes behave the same way. The earthquake catalog also links larger magnitudes with greater epicentral distances, preventing the researchers from separating source-size effects from propagation-distance effects in the peak-acceleration results. In addition, the data cannot uniquely determine whether topography, near-surface structure or both caused the observed response.
