Inter-crystalline pores, cavities, and fractures created from diagenetic shrinkage of dolomite are inter-connected each other, forming fine oil- and gas-bearing reservoirs. It is hard to predict these complex fracture-cavity reservoirs because of their random distribution, different growth timing, and so on. Taking the lacustrine dolomite fracture-pore reservoir in the Lower Cretaceous Xiagou Formation in the Qingxi oilfield within the Jiuquan basin as an example, we put forward a comprehensive geophysical method to predict carbonate fractures.
There are some limitations when we apply conventional methods to analyze the massive amounts of seismic data acquired with high-density spatial sampling since processors usually obtain the properties of raw data from common shot gathers or other datasets located at certain points or along lines. We propose a novel method in this paper to observe seismic data on time slices from spatial subsets. The composition of a spatial subset and the unique character of orthogonal or oblique subsets are described and pre-stack subsets are shown by 3D visualization. In seismic data processing, spatial subsets can be used for the following aspects： （1） to check the trace distribution uniformity and regularity; （2） to observe the main features of ground-roll and linear noise; （3） to find abnormal traces from slices of datasets; and （4） to QC the results of pre-stack noise attenuation. The field data application shows that seismic data analysis in spatial subsets is an effective method that may lead to a better discrimination among various wavefields and help us obtain more information.
