The median normalized VIM spectrogram of building microphones show a clear day/night/weekend modulation of the acoustic noise of experimental halls. The spectrograms are reported, for the week 21-28 December, in Figures 1 and 2 (NEB), Figures 3 and 4 (WEB), and Figures 5 and 6 (CEB). During the working hours the acoustic noise in enhanced by a factor up to ~2. This enhancement occurs at some frequencies between roughly 6 Hz and 60 Hz, which seem to correspond to acoustic modes of the halls (e.g. NEB modes).

Note that, as shown in Figure 7, during the week of 21-28 December the wind was not particularly intense and its contribution to acoustic noise in the halls was limited to below 5 Hz (see also: elog 57191).

The observed sound modulation is very similar to that of seismic noise measured by building seismometers in the 1-15 Hz band, which we know is associated to vehicle traffic in the nearby roads and FI-PI-LI. This correlation is shown in Figure 8, which compares the BRMS of the NEB seismometer in the 1-5 Hz band and the NEB microphone in the bands 0.1-4 Hz, 4-20 Hz, 20-40Hz, 40-100Hz.

Finally, Figures 9, 10 and 11 compare, respectively for NEB, WEB and CEB, the microphone average spectrum during the day-time (RED - December 21 6:30 UTC, 15 minutes) and during night-time (BLUE - December 21 00:00 UTC, 15 minutes).  The excess noise during day-time is evident at some frequencies. The bottom plot compares the microphone spectrum at night-time (where, we assume, the noise is dominated by the AHU) and the "anthropic noise" estimated as the quadrature difference of the microphone noise during day-time and the microphone noise during night-time (we assume AHU noise is more or less stationary). We see that, at some frequencies the "anthropic" noise is more or less as intense as the noise produced by the AHU at that frequency. Note that after 20 December the AHU of NEB and WEB were set at their final set-point for the holiday period.