用于火山岩预测的地震属性提取及有效性分析方法研究

The abstraction of seismic attributes and its effectiveness in volcanic prediction

（1.中国地质大学，北京100083；2.大庆物探公司，黑龙江大庆163357；3.大庆师范学院，黑龙江大庆163712）

China University of Geosciences，Beijing 10083，China

针对火山岩在地下分布成层性差的特点，采用固定时窗法对地震数据逐样点提取能量、频率及波形等类别的地震属性，通过对样点的逐一分析来提高火山岩描述的精度。在此基础上，采用井旁地震道逐样点地震属性对比法进行地震属性对比，通过比较井旁道火山岩与邻近围岩地震属性值的变化规律来分析此地震属性与火山岩关系的密切程度，从而判断它在火山岩预测中的有效性。经过多个地区的应用，发现自相关函数最大值、自相关函数脉冲度、曲线总长度、容量维及最大熵功率谱频率加权总能量等5种属性参数能较好地反映火山岩，它们在火山岩区一般呈现高值。实际应用中，首先形成这些属性参数的属性体，然后进行可视化或切片等处理，最后结合已知井情况来判断火山岩的分布情况。

According to the characteristic of volcanic distribution, the changeless time-window method was used to abstract seismic attributes sample point by sample point from seismic data so as to describe volcanic detailed. Based on the above, the method that contrasting one seismic attribute sample point by sample point in the seismic trace near a well was applied to search for the valid seismic attributes for volcanic prediction. It judges if the seismic attribute reflects volcanic by analyzing the changes of seismic attribute values of every sample point between the volcanic and nearby non-volcanic in the seismic trace near a well. Through the studying application in many work areas, the five seismic attributes were found out, which could reflect volcanic well and presented higher value generally in the position of volcanic. They are the maximum of autocorrelation function, the pulse degree of autocorrelation function, fractal dimension D0, curve length, and the total energy of maximum entropy power spectrum weighted by frequency. The good effects were achieved in real application. This is shown that the methods are feasible in volcanic prediction.