From October 2000 to April 2001 a local array of fourteen seismograph stations was deployed at the Trail Mountain Mine (TMM), an underground coal mine in Emery County, Utah, to locate mining-induced seismicity (MIS). Original analysis of the data provided epicentral locations that coincided closely with the active longwall panel. The original analysis yielded poor quality hypocenters, however, preventing a useful relation from being made between mining activity and the event depths. For the original analysis, only stations at elevations above mine level were used for locating MIS. As a result, P- and S-arrival times from four of the stations in the local array were not used: an in-mine station and three stations at elevations below mine level. This study compares mine workings and active mining to a revised TMM dataset that included P- and S-arrival times from the lower elevation stations that were not used in the previous study. In the revised locations, a modified master event methodology was used to calculate station delays. These station delays correct for errors in the shallow layers of the velocity model. The resulting hypocentral locations have improved depth resolution, determined by a lower root mean square (RMS) for the locations, compared to the original study. Most of the events occur in the roof of the mine and follow the longwall panel. When located using the velocity model determined in the original study, the elevations of the events increase as the longwall panel progresses. One possible explanation for the elevation increase is that it may be related to geology or the caving process. As presented here, a 10% decrease in the shallow velocity structure results in a more plausible location for the hypocenters. Consequently, it is hypothesized that progressive development of the TMM longwall panel(s) may have resulted in a progressive change in the velocity model.
展开▼