Following up on Nicolas’ update ( #67407) and the discussion during the last BCM, I report here the results of the glitch rate and population analysis over the past month, repeating the same procedure used in June.

Figures 1 and 2 show the glitchgrams of the strain channel for June and July, and for the past week. The lower plots display the rate split by SNR. These values are obtained by estimating the glitch rate every 30 minutes and applying a 6-hour moving average (so, maximum 12 points). Aside from an overall decreasing trend, there are notable drops in the rate observed between June 20-22 and on July 29. The latter may be correlated with the intervention on the Pstab performed on the same day  #67395 (?). If so, this could indicate that a glitch source, different from the (other) hypothesized one related to ear stabilizaition crackling, has been mitigated.

I then proceeded to identify glitch families with SNR > 50, clustering them based on the difference (1 – Pearson correlation) of their whitened time series, as described in #66972 ]. The resulting clusters are shown in Figure 3, and are available in the attached CSV file (renamedto  TXT format).

Clusters 1 and 2 correspond to step glitches, with some contamination in Cluster 1 from the 25-minute glitches.

Cluster 3 consists of the 25-minute glitches.

Cluster 4 includes a few other strong and similar glitches found in the data.

Cluster 5 mostly consists of unclassified glitches.

Figures 4 and 5 show the glitchgrams with the clusters labeled in different colors. As can be seen, step glitches are still present; see the example in Figure 6.

Lastly, I tried to verify whether the SNR of step glitches might also be decreasing. In Figure 7, I report the SNR of Clusters 1 and 2. The line of triggers with SNR between 400 and 550 corresponds to the 25-minute glitches. I applied a cut to exclude them. The orange line represents the daily average SNR. Also in this case, a weak decreasing trend is suggested.

In conclusion, over the coming days we should monitor whether the reduced glitch rate is maintained, and further understand whether the Pstab could have been responsible for part of the observed glitches (Piernicola suggested so in private conversations).