Seismic response analyses are performed using a minimum number of seismic records as input motions in order to achieve a statistically strong estimation. Unfortunately, the available information recorded from the current seismic networks is still scarce regarding events with considerable magnitude. In this context, the Random Vibration Theory (RVT) arises as an alternative tool for performing site response analyses without the need of seismic records, since it only requires adequate probabilistic seismic hazard assessment.
 In this study, RVT was applied to three shear-wave velocity profiles in Lima city with distinct geomorphological origin. These profiles are characteristic for gravelly, sandy and fine deposits so the influence of each soil type in their corresponding transfer function was taken into account. In that sense, the three RVT-based normalized response spectra show good agreement with the design spectra specified in the Peruvian code, despite some amplification in the short (below 0.10 s) and long (above 0.80 s) period ranges related to noise or far-field effects. Furthermore, RVT-based response spectra for La Punta and Villa el Salvador show good agreement with the time-series based analyses from a previous study.
 In addition, spectral acceleration values surpass those specified in the Peruvian code for a range beyond the corner period. This could suggest that the soil profile characterization based on the time-averaged shear wave velocity from the upper 30 m might be insufficient to evaluate the overall seismic behavior of a soil deposit. Therefore, additional parameters that account for the deeper soil substructure might be required.