Analysis of coda wave distribution and attenuation variation of the Binchuan Airgun experiment

Monitoring temporal variations of subsurface seismic velocity is one of the main tasks in modern seismology, which depends primarily on repeating sources. However, finding such sources in passive source studies is challenging. Active sources repeating in both space and magnitude have been proved to be capable of measuring the temporal variations accurately. Since 2012, a large volume airgun array has been deployed in Binchuan reservoir in Yunnan province of China. Several major faults are located in this region and the background seismicity is very high. Thus, the airgun provides ideal setup to study the temporal variation related to the earthquake. The energy excited by this airgun experiments were estimated to be equivalent to that of a magnitude 1 earthquake and receivers were densely distributed within the radius of 200km. This study focuses on (1) travel time analysis of different phase and (2) the variation of attenuation at different time window of the data. We first use the record of the reference station located on the shore of the reservoir as the source time function (STF). Then the STF are deconvolved from data to eliminate source effects such as water level in the reservoir. Secondly,we stacked daily records to enhance SNR due to the high similiarity among the individual records. We corss-correlated the stacked record with individual record through sliding thime window od 2s and pick high cross-correlation coefficients (CC's) across the whole records for each station. Following this approach we identified the main phases (Pg and Sg) and some high CC's stripes arrived much later than the main phase, which could be the coherent coda waves caused by scatterers. To further investigate these late arrival stripes, we applied a slowness to the stacked record of two stations correcting the travel time moveout and cross-correlated sliding time window between the two records. Then the slowness's with the highest CC's among different station couples are used to image the location of possible scatterers. The water distribution may strongly connect to the processes of the earthquakes. Because the attenuation is sensitive to water saturation, the attenuation will be a good indicator of the water distribution. We measured the dt* values between station couples, which display strong variations in both time and space.