Integrated use of surface and subsurface geologic data, structural deformation analysis from second-derivative maps, and dimensionless type curves in identifying potential horizontal well locations in the fractured Niobrara Shale of northwest Colorado is described. The analysis included construction of detailed structure maps from surface geology with tie-ins to seismic surveys and well control and of second-derivative maps that were overlaid on production data to determine the minimum radius of curvature required for commercial production. Reservoir properties were calculated for known fields from type-curve matching with type curves for dual-porosity, anisotropic reservoirs with stress-sensitive permeability. The curves were constructed for vertical and horizontal wells by use of reservoir-simulation models. Spacing and economics of drilling horizontal Niobrara wells in northwest Colorado were evaluated. The analysis suggests that recoverable oil from the fractured Niobrara in northwest Colorado averages 1,392 bbl oil/acre on the basis of a productive fracture area identified by second derivatives of the structural surface with values > 1 x 10 - 4.5 ft - 1. It is also indicated that horizontal well spacing in the fractured Niobrara should be > 640 acres. In some instances, it may be more economical to develop the fractured Niobrara with vertical rather than horizontal wells.
Oil and gas production from Upper Cretaceous calcareous shales is widespread in northwest Colorado. Cumulative production from 34 separate pools in Moffat, Routt, Rio Blanco, and Jackson counties exceeds 26 million bbl oil. Reservoir-quality natural fractures have been generated in these fields by a combination of folding and faulting. Historically, most exploration has been directed toward fractures produced by faulting because faults provide specific, more easily identified drilling targets. To account for the large oil recoveries from certain fractured-shale wells, it is our thesis that a much more extensive, fold-generated, 3D network of fractures must be present. Furthermore, we believe that fold- or curvature-generated fracture systems are both predictable and mappable. The magnitude and extent of the fracture network can be quantified with maps of second derivatives or rate of change of dip over the structural surface.