We used Alex Nitz's script from the LHO logbook to estimate the BNS range integrand and figure out where our Mpc are coming from.

Figure 1 is the cumulative sum of the BNS range for the 10Mpc lock from Monday night, up to 1kHz.  75% of the sensitivity is coming from below 200Hz.

Figure 2 is a zoom of the per-frequency-bin contribution up to 300Hz.  Most of the BNS sensitivity comes from a few narrowband spots in the 50-150Hz band.  Removing lines from this band might be a big deal.  WARNING: I am not sure about the normalization for this plot, the y-axis units could be totally off.  But the relative heights of different bands is the important part.

The script to generate these plots (and, generate the plot of AdV sensitivity vs aLIGO sensitivity) has been moved to /olusers/virgorun/hoak/sense_plot.  It's called plot_ASD_AdV.py.

It might be fun to estimate the sensitivity boost we would get from certain improvements, like removing the bump around 50Hz or removing a certain family of lines.  You'd have to fit these features in the spectrum (while assuming a certain noise floor), subtract the excess, and re-calculate the range integrand with the new PSD.  This could provide some guidance for costly commissioning activities, e.g. venting the PR, WI, or WE towers to mitigate lines.