On Calibrating Curvature Data to Fracture Density: Causes

Volume curvature attributes have been used to infer fracture density in a variety of seismic data worldwide. Estimating accurate quantitative fracture density values from curvature data is of special importance because curvature belongs to a general class of attributes that infer fractures through causal relationships rather than through direct detection of them. This does not necessarily make curvature of greater or lesser use in our efforts to understand fractures, but rather adds a special perspective to strategies around its calibration. Beyond the obvious general utility of producing measureable numerical values, estimates of fracture density variations may be useful in assessing completion and production opportunities and risks. Calibrating curvature data extracted from 3D surface seismic data to actual fracture density values is unlikely to ever be a trivial task. This calibration will require both the acquisition of appropriate control data such as image log data, core data, and cross dipole log data. The sample size of this control data is unlikely to be generally sufficient to fully calibrate most of the curvature data by itself. We must bring forward geologic knowledge of the causes and behavior of fracture development to constrain our calibration and provide perspective to the sample data we do have. In particular, we must be aware of the effects of the fractured reservoir’s material properties, in-situ properties, and structural position. We will also have to consider other fracture inferring and detecting methods and find practical combinations with curvature that utilize the various attributes’ connection to the causes of fractures as well as their circumstances of validity. The most practical opportunities for meaningfulalthough probabilisticfracture estimation will likely come from the use of all available calibration data, a leveraged use of a variety of prediction attributes and a strong understanding of causal variables.