The estimation of time-lapse time shifts between two, or several, repeated seismic surveys has become increasingly popular over the past eighteen years.These time shifts are a reliable and informative seismic attribute that can relate to reservoir production. Correction for these time shifts or the underlying velocity perturbations and/or subsurface displacement in an imaging sense also permits accurate evaluation of time-lapse amplitudes by attempting to decouple the kinematic component. To date, there are approximately thirty methods for time-shift estimation described in the literature.We can group these methods into three main families of mathematical development, together with several miscellaneous techniques. Here we detail the underlying bases for these methods, and the acknowledged benefits andweaknesses of each class of method highlighted. We illustrate this review with a number of time-lapse seismic examples from producing fields. No method is necessarily superior to the others, as its selection depends on ease of implementation, noise characteristics of the field data, and whether the inherent assumptions suit the case in question.However, cross-correlation stands out as the algorithm of choice based on the Pareto principle and waveform inversion the algorithm delivering best resolution. This is a companion study to the previous review of time-shift magnitudes and a discussion of their rock physics basis.

中文翻译：

---

4D 地震叠后时移：综述第 2 部分：测量方法

在过去的 18 年中，对两次或多次重复地震勘测之间的延时时间偏移的估计变得越来越流行。这些时间偏移是一种可靠且信息丰富的地震属性，可与储层生产相关。对这些时移或潜在速度扰动和/或成像意义上的地下位移的校正还允许通过尝试去耦合运动学分量来准确评估延时幅度。迄今为止，文献中描述了大约 30 种时移估计方法。我们可以将这些方法分为数学发展的三个主要系列，以及几种杂项技术。在这里，我们详细介绍了这些方法的基础，并强调了每一类方法的公认优点和缺点。我们用一些来自生产油田的延时地震实例来说明这一评论。没有任何方法一定优于其他方法，因为它的选择取决于实施的难易程度、现场数据的噪声特性以及固有假设是否适合所讨论的情况。然而，互相关作为基于帕累托原理和波形反演算法提供最佳分辨率。这是之前对时移震级的回顾及其岩石物理基础讨论的配套研究。以及固有假设是否适合所讨论的情况。然而，互相关作为基于帕累托原理的算法选择脱颖而出，而波形反演算法可提供最佳分辨率。这是之前对时移震级的回顾及其岩石物理基础讨论的配套研究。以及固有假设是否适合所讨论的情况。然而，互相关作为基于帕累托原理的算法选择脱颖而出，而波形反演算法可提供最佳分辨率。这是之前对时移震级的回顾及其岩石物理基础讨论的配套研究。