Full-azimuth, high-density, 3D point-source/point-receiver seismic survey for shale gas exploration in a loess plateau: a case study from the Ordos basin, China

The Ordos Basin, located in the central part of China, contains abundant oil and gas shale resources. This study took place in an area southwest of Yan’an City, Shaanxi province, and its main exploration targets were the Mesozoic Yanchang Formation, with a maximum depth of around 1500 m, and the Upper Paleozoic Benxi Formation, with a maximum depth of around 3550 m. The tectonic structure at the target levels is almost flat, with dips of only around one degree. The objective of the survey was to identify ‘sweet spots’ in a heterogeneous gas shale reservoir beneath thick deposits of loess, which is sediment formed by the accumulation of wind-blown silt. Variations in the topography, thickness, and seismic velocity of the loess were identified as key factors that had led to inadequate imaging results from previous 2D exploration surveys in the area (Yao and Li, 2004). The new 3D survey, acquired by CNPC Sichuan Geophysical Company (SCGC) during 2013 using the WesternGeco UniQ integrated point-receiver land seismic system, delivered imaging results that improved the characterization of the gas shale. Several features of the project area presented challenges for seismic exploration, many of which are typical of a loess plateau. The area is covered by unconsolidated loess with severe variations in thickness and velocity that can result in serious statics problems. The loess also causes dramatic signal absorption, leading to poor signal-to-noise ratio and potentially contributing to low resolution in the seismic data. As shown in Figure 1, surface conditions include rapid variations in topography that further complicate near-surface statics correction and can be the cause of various types of interference and noise. Survey logistics needed to account for heavily forested areas, hills and gulches. Also, more than 500 oil wells were pumping and some rigs were drilling during survey acquisition, representing additional sources of noise (Figure 2).