We report here of the test performed on March 23 (see Operator report 45360). On that day we hired one tractor (John Deer 5600 140CV, 2ton) with grass cutting tool to simulate grass cutting operation and evaluate the impact on the interferometer.

While ITF was locked (sensitivity around 45Mpc, SQZ on) the tractor went along the whole West arm cutting grass close to the tunnel, then some test grass cutting was done next to WEB, CEB and NEB, as detailed below. We monitored the noise looking at dark fringe, suspensions and seismometers.

Figure 1 and Figure 2 show trend data over the test period: BNS range, rms of SUSP F0 LVDT and 10-40Hz rms of seismometers (we notice that the peak seismic noise is produced in the 10-40Hz band). Figures 3, 4 and 5 are spectrograms, respectively of WEB, CEB, and NEB, with annotations of the tractor location when noise is recorded. Figures 6,7,8 are maps, respectively WEB, CEB, NEB with indicated the approximate tractor path and a drawn circle of about 150m radius within which we recorded some excess seismic activity. Figures 9,10 and 11 shows spectral excess noise in F0 lvdts and seismometers, resp. WEB, CEB and NEB.

Here some observations:

1. We saw no effect on DARM,  WE SA suspensions nor seismometers cause by the tractor while it was grass-cutting along WE arm, utill  it was 150m or so away from the terminal building.
2. The ITF unlocked once (11:39 UTC)  because of the tractor, while working very close to NEB (20m approx. from building).
3. When the tractor was working within 100m or so from the virgo buildings we saw and increase of 1-40Hz seismicity, excess noise in the suspension F0 lvdts (see comment n.4 below). The impact on sensitivity is not so evident, except for CEB (see n.5 below) and NEB (see n.6 below) I could not spot clear evidence of noise in the DARM spectrogram, and Range fluctuations were within the typical non stationarity.
4. The WE and NE suspensions responded a bit diffentry to the tractor: while F0 of WE (Figures 3 and 9) was excited down to 1Hz, the F0 of NE (Figures 5 and 11) looked less disturbed overall, just excitation of narrow modes at 6, 10 and 15Hz.
5. There was definitely an impact at CEB, limited to when the tractor was stepping on the hill side (see Figures 4 and 7).
6. There was definitely an impact at NEB (that concluded with unlock) when the tractor was grass-cutting at the location closest to the building, about 20m away on West side (see Figures 5 and 8)
7. When approaching WEB (still 350m away) we spotted an excess noise on seimometers and on WE F0 signals. We realized it was not us (indeed we went back 100m or so and approached again without effect) but if was (very likely) the farmer's tractor that was working outside EGO site property but close to the fence at about a 150m distance from WEB (!!). So, this is a potential offender, we can do nothing to prevent.

Some conclusions:

• Grass cutting looks safe along the arms from about 150m away of Virgo buildings.
• Safely, we set a limit at 200m from end buildings not accessible to the tractor as well as other vehicles when in science mode.
• The grass cutting of all other areas (terminal areas and all central area) must be done during maintenance. For the central hill area the small grass-cutter vehicle is going to be used.