宾川气枪地震信号发射台以水库大容量非调制气枪阵列为震源，可以连续不断地激发重复性极高的地震信号。本文利用其2013年的数据，对地下介质进行了观测。由于水库水位的变化对震源波形及远处台站接收到的地震波形会产生影响，因此，为了区分介质变化与震源变化，研究了陆上水库环境下激发气枪震源所产生的地震波形的特征和合适的处理流程。结果表明：①水库水位变化较大时，同一台站记录的地震波形之间的互相关系数过小，不能直接进行叠加处理，需进行震源聚类分析，而震源聚类受水位变化量的控制；②为消除水位的影响，进行了反褶积处理，当震中距小于10km时，对不同水位激发的信号进行叠加再进行反褶积的效果优于直接进行叠加的效果；③通过互相关计算和信噪比分析发现，4个台站的波形变化可能是由地下介质变化引起的。

Binchuan Transmitting Seismic Station usesa large-volume air gun array as seismic source，which can continuously activate highly repeatable signal. We analyzed the seismic data collected in 2013 and did observations on the subsurface structure. Due to the influence on the waveform of the source and far-field air-gun signals by the water level change of the reservoir，the characteristics of source signals from an air-gun array in a reservoir and proper data processing flow were studied for the need of differentiating the structure variation and source variation. The results show that: (1)When the reservoir water level variation is large，the cross-correlation coefficient between the seismic waveforms recorded in the same station is too low. Under this condition，we can't do direct stacking to the data，but should do source cluster analysis first. The result of the analysis is governed by the amount of water level change. (2)In order to eliminate the influence of water level change，we conducted a deconvolution process. When the epicentral distance is less than 10km，stack method for multi-shot air-gun data is used to activate in different water levels. Deconvolutionhas a better result than direct stacking. (3)By cross-correlation and SNR analysis，we found that 4 stations have waveform variations which may be caused by subsurface structure variations.

Research Grant Council (grant 2191093)、国家自然科学基金(41561164018)、云南省陈颙院士工作站专项经费(2014IC007)联合资助