宽角反射地震波走时模拟的双重网格法

在研究地壳结构的人工源宽角反射地震资料解释中，常规宽角反射波走时和射线路径计算大都假定地壳模型为层状块状均匀介质.为了逼近实际地壳结构模型，要求模型尺度较大，为了提高地震资料解释的可靠性，须减小模型离散单元的尺寸，但同时计算量大大增加，使资料解释的效率较低.为此，本文尝试同时提高宽角反射地震资料解释效率和可靠性的方法，即使用双重网格计算宽角反射地震波走时和射线路径的最小走时树方法.双重网格法在均匀介质内部仅计算大网格节点，在速度变化点、震源点和检波点区域，同时计算小网格节点；在界面边界点使用比介质内部节点更大的子波传播区域.模型计算结果表明，对于大尺度的层状块状均匀介质模型，在保证精度的条件下，本文所提出的双重网格射线追踪方法的计算效率比单网格方法显著提高.

A double-grid algrithm for calculating traveltimes of wide-angle reflection waves

Layered and blocked homogeneous models are usually used to calculated travel times and ray paths of wide-angle reflection waves in interpreting wide-angle reflection data of deep seismic sounding. In order to make the model approach the Earth, the size of the model has to be large. On the other hand, for obtaining accurate travel times and rays calculated with the minimum travel time tree algorithm, the model elements have to be small, and thus the efficiency of calculation is decreased. In this paper, a double-grid scheme is employed in calculation of travel times and ray paths in order to fast and reliably interpret deep sounding data. The double-grid algorithm covers a model with sparse and dense grids. Nodes in the sparse grid are calculated in the same way as in the single-grid algorithm. As for nodes in the dense grid, they are calculated only if they are in the neighborhoods of source and geophone nodes, and ones where velocity gradients are not zero, and the propagation areas of secondary waves are larger for interface boundary nodes than those within homogenous media. The numerical example shows that the double-grid algorithm has much higher calculation efficiency than the single-grid one under the same accuracy condition for large layered and blocked homogenous models.