中文摘要：

低频成分缺失和地下速度强烈变化会导致严重的周期跳现象，是地震数据全波形反演的难题.通过对地震数据加时间阻尼和时间积分降主频处理，提出了一种可有效去除周期跳现象的多主频波场时间阻尼全波形反演方法.由浅到深的速度不准确会造成波形走时失配和走时失配的累积.浅部速度的准确反演可有效地减小深部波形走时失配与周期跳现象.对地震数据施加时间阻尼得到时间阻尼数据，利用不同阻尼值的时间阻尼地震数据实现由浅到深的全波形反演.低主频波场的周期跳现象相对高主频波场的要弱.对地震波场进行不同阶的时间积分以得到不同主频的波场，把低主频波场的全波形反演结果作为高主频波场全波形反演的初始模型.应用缺失4 Hz以下频谱成分的二维盐丘模型合成数据验证所提出的全波形反演方法的正确性和有效性，数值试验结果显示多主频波场的时间阻尼全波形反演方法对缺失低频成分地震数据和地下速度强烈变化具有很好的适应性.

英文摘要：

The lack of low-frequency components and the subsurface strong variations of velocity can lead to a severe cycle-skipping phenomenon, which is a big challenge to full waveform inversion of seismic data. Through the time-damping and the time-integral decreasing the dominant-frequency of the seismic wavefields, a time-damping full waveform inversion of multi-dominant-frequency wavefields is proposed. This method can efficiently eliminate the cycle-skipping phenomenon. Velocity errors from shallow to deep can lead to misfits of travel-time and their accumulations. The accurate inversion of shallow velocity can efficiently reduce the misfits of travel-time and the cycle-skipping phenomenon in the inversion of later waveforms. Applying the time-damping approach to seismic data can obtain time-damped data. The inversion of these time-damped data with different damping values can obtain the inversion results from shallow to deep. The cycle-skipping phenomenon is weak for the wavefields with lower dominant-frequency compared to higher dominant-frequency wavefields. The time integral with different orders to the seismic wavefields can obtain wavefields with different dominant frequencies. The inversion results of low dominant-frequency wavefilds are used as the starting models for the full waveform inversion of high dominant-frequency wavefileds. Numerical tests using synthetic data, the lack of low-frequency components below 4 Hz, of the 2D salt-dome model have demonstrated the validity and feasibility of the proposed method. The final results show that the time-damping full waveform inversion of multi-dominant-frequency wavefields has a proper flexibility to seismic data lacking low-frequency components and to the subsurface strong variations of velocity.