LIST OF CHANGES TO NFM PAPER TITLE changed the title from"Environmentally-induced slow nonstationarity in LIGO science run data" to "Environmentally-induced nonstationarity in LIGO science run data" ABSTRACT changed "These analyses have identified time intervals in the gravitational wave channel that indicate non-stationarity due to seismic activity, and these intervals are being submitted as segments requiring data quality flags." to "These analyses have identified time intervals in the gravitational wave channel that indicate non-stationarity due to seismic activity, and these intervals are referred to as data quality flags." changed "The algorithm and its performance is presented here, and we discuss the potential for a fully automated, on-site pipeline that will generate daily data quality flags for future data runs." to "We present the algorithm and its performance, and discuss the potential for an on-site pipeline that automatically generates data quality flags for use in future data runs." INTRODUCTION changed "This science run encompassed 37,110.4 hours of observation among the three detectors." to "This science run encompassed approximately two years of observation among the two detectors of the LIGO Hanford Observatory (LHO) located in Hanford, Washington, and the single detector of the LIGO Livingston Observatory (LLO) located near Livingston, Louisiana." changed "Each of the different analyses will require knowledge of the state of the detector at any particular time." to "The search for gravitational waves in the LIGO data requires knowledge of the state of the detector at any particular time." changed "The search for gravitational waves divides into four general tracks according to standard models of the astrophysical sources of the waves: a track for compact binary coalescence, a track for unmodeled signals which last from a few milliseconds to hundreds of milliseconds (bursts), a track for continuous signals, and a track for the stochastic background of gravitational waves." to "The search for gravitational waves divides into four broad gravitational wave searches according to standard models of the astrophysical sources of the waves: a search for waves due to compact binary coalescence, a search for waves due to unmodeled signals which last from a few milliseconds to hundreds of milliseconds (bursts), a search for waves due to continuous signals, and a search for the stochastic background of gravitational waves." deleted the sentence "The effect of non-stationarity is more profound on the triggered searches since they are generally unmodeled, as opposed to the other gravitational wave searches which are based on astrophysical models." changed "Each of the different analyses will require knowledge of the state of the detector at any particular time. One aspect of detector characterization is the detector response to environmental conditions; another aspect is the detector response to its internal processes. This latter effect occurs when noise in a detector channel couples to the signals in other channels." to "One aspect of detector characterization is the detector response to environmental conditions; another aspect is the detector response to its internal processes, e.g., electronic signals, thermal vibrations of detector components, radiation pressure of the laser on the mirrors, etc. A residual effect of the internal processes is the production of noise in one or more detector channels that couples to the signals in the gravitational wave detection channel." changed "The combination of effects results in a time-dependent noise background that is classified as non-stationary." to "The combination of both environmental disturbances and the noise produced by the internal processes results in a time-dependent noise background that is classified as non-stationary." changed "One of the largest challenges in LIGO data analysis is to identify non-stationarities in the expansive repositories of detector data." to "One of the largest challenges in LIGO data analysis is to identify instances of non-stationarity present in the gravitational wave channel. Our work involves the identification of these instances in time as well as their correlation with channels designated as physical environmental monitors (PEM)." changed "Non-stationary noise can take the form of short-lived transients whose power exceeds the noise background,..." to "Non-stationary noise can take the form of short-lived transients whose power exceeds the noise background ("glitches"),..." changed "The identification of slow non-stationarity is an especially important consideration for certain triggered searches that involve the comparison of data around an event of interest with an electromagnetic signature (on-source data) with data from other time periods (off-source data)" to "The identification of slow non-stationarity is an especially important consideration for certain triggered searches that involve the comparison of data around times of external astrophysical triggers (on-source data) with data from other time periods (off-source data)." changed "The effect of non-stationarity is more profound on the triggered searches since they are generally unmodeled, as opposed to the other tracks which are based on astrophysical models." to "The effect of non-stationarity is more profound on the triggered searches since they are generally unmodeled, as opposed to the other gravitational wave searches which are based on astrophysical models." changed "The non-stationarity of the background noise in this scenario can produce the illusion of anomalous events in the data." to "The non-stationarity of the background noise in a triggered search for an unmodeled source can produce the illusion of anomalous events in the data." changed "The non-stationarity of the background noise in this scenario can produce the illusion of anomalous events in the data. The goals of the work presented here include the elimination of such aberrances due to non-stationary data as gravitational-wave candidates, and to characterize those noise floor segments used in off-source analyses." to "The non-stationarity of the background noise in a triggered search for an unmodeled source can produce the illusion of anomalous events in the data, necessitating the need to flag the non-stationary periods. The goal of the work presented here is to provide an efficient means of flagging non-stationary data in the gravitational-wave channel." made the original figure 2 the first figure Changed the caption on this figure to "This figure describes the pipeline developed to flag non-stationarity. The steps are detailed in the Overview section." OVERVIEW Completely rewrote the section entitled 'Overview' as per referee suggestion. OFFLINE ANALYSIS changed "The threshold is superimposed on these plots and points that cross the threshold are highlighted." to "The threshold is superimposed on these plots and points that cross the threshold are highlighted as can be seen in Figure 3." changed "The offline analysis was developed to highlight couplings between the gravitational wave and the PEM channels." to "The offline analysis was developed to highlight couplings between the gravitational wave and seismic activity measured by the PEM channels." changed "The most prominent events are considered those points at the greatest distances above the threshold." to "The most prominent events are considered those points at the greatest differences above the threshold." Split the paragraph beginning "The offline analysis was developed..." into two paragraphs Eliminated the following sentence from the section: "For the S5 analysis, a comparison of the findings was made with QScan, a prevalent utility in the LIGO data analysis community that generates time series and time-frequency plots for gravitational wave and auxiliary channels at specified times." RESULTS changed "The overarching goal of this work is to identify long-range correlations between the PEM channels and the gravitational wave channel." to "The overarching goal of this work is to identify correlations over years between the PEM channels and the gravitational wave channel." added the 2nd sentence "We anticipate that understanding the effect of seismic noise on the gravitational wave signal can lead to 1) possibly mitigating the effect through commissioning efforts, and 2) an explanation of the mechanism by which low-amplitude seismic disturbances are manifested as large glitches in the gravitational wave channel, a process known as up-conversion." changed "Lock loss or drops from science mode were also indicated by those crossings that showed the greatest deviation from the threshold." to "Lock loss or drops from science mode were also correlated to those crossings that showed the greatest deviation from the threshold." changed "Nearly all of the threshold crossings that occurred in the $32-64$ Hz band were not associated with noise transients but rather with oscillations." to "Nearly all of the threshold crossings that occurred in the $32-64$ Hz band were not associated with noise transients but rather with motion predominated by a particular frequency. It is unclear where the motion originated." changed "In the analyses conducted to date the greatest number of threshold crossings have occurred due to seismic activity in the y-direction in the PEM channel seismometers identified as the LIGO Vacuum Equipment Area (LVEA or corner station) and the x-arm end station (EX). As anticipated, a great deal of the crossings identified in the $0-16$ Hz band were associated with a drop from science mode (the optimal state of the interferometer for data collection) or lock loss (the state of the interferometer in which data acquisition is impossible due to excessive mirror displacements)~\cite{seismic}. Lock loss or drops from science mode were also indicated by those crossings that showed the greatest deviation from the threshold." to "In the analyses conducted to date the greatest number of threshold crossings have occurred due to seismic activity in the y-direction in the PEM channel seismometers identified as the LIGO Vacuum Equipment Area (LVEA or corner station) and the x-arm end station (EX) at both LHO and LLO. Lock loss (the state of the interferometer in which data acquisition is impossible due to excessive mirror displacements)or drops from science mode (the optimal state of the interferometer for data collection) were also correlated to those crossings that showed the greatest deviation from the threshold.~\cite{seismic}." changed "We hope that after the entire S5 analysis is complete that we will be able to contribute to a model of seismic up-conversion in the gravitational-wave channel." to "After the entire S5 analysis is complete our goal is to contribute to a model of seismic up-conversion in the gravitational-wave channel." and moved the sentence to the CONCLUSION section. changed the last paragraph of results from "Lock loss due to seismic activity is to some degree inevitable and there is an expected correlation between the seismic PEM channels and the gravitational wave channel for energetic events. We are especially interested in non-linear responses in the gravitational-wave channel to seismic activity. Nearly all of the threshold crossings that occurred in the $32-64$ Hz band were not associated with noise transients but rather with oscillations. Threshold crossings due to noise transients that did not result in lock loss occurred primarily in the $0-16$ Hz band. We hope that after the entire S5 analysis is complete that we will be able to contribute to a model of seismic up-conversion in the gravitational-wave channel." to "Lock loss due to seismic activity is to some degree inevitable and there is an expected correlation between the seismic PEM channels and the gravitational wave channel for energetic events. We are especially interested in non-linear responses in the gravitational-wave channel to seismic activity. Nearly all of the threshold crossings that occurred in the $32-64$ Hz band were not associated with noise transients but rather with motion predominated by a particular frequency as seen in Figure 4. It is unclear where the motion originated. Threshold crossings due to noise transients that did not result in lock loss occurred primarily in the $0-16$ Hz band. Figure 5 shows an example of such a transient." FIGURES divided the original two figures of Figure 3 to Figure 3 and Figure 4. Changed the captions to "Sample output of the offline analysis indicating lock loss." and "This plot shows the same lock loss as it appears in the gravitational-wave channel time series in a two-minute segment centered around the threshold crossing time. Subsequent investigation indicated that this lock loss event was caused by seismic activity." respectively. changed original Figure 4 to Figure 5 with the caption "The bottom plots shows signals with strong components around 32 Hz centered around the threshold crossing. The top plot shows the crossing." changed the original Figure 5 of Figure 6 with the caption "The bottom plot shows a glitch centered around the threshold crossing time. There was no lock loss in this case. The top plot shows the crossing."