Production data from many North Sea chalk fields have indicated moderate to considerable contribution to production from natural fractures. This paper illustrates a detailed study of natural fracturing in the Albuskjell field, where gas/condensate hydrocarbons are contained in Upper Cretaceous (Maastrichtian) and lower Tertiary (Danian) chalks. The field is a large halokinetically induced dome located on the northern limits of the productive chalk region known as the Greater Ekofisk area.

Examination of core material from wells 1/6-3, 2/4-F10, and 1/6A-10 (Albuskjell), has revealed the presence of two types of fractures. The first appear to be early, are commonly compacted, predominantly healed, and resemble conjugate shear fractures. The second type is related mainly to the tips of stylolites; these are vertical, preferentially open, and are interpreted as tension fractures.

Tension fractures form, when the minimum effective stress is reduced to the tensile strength of the chalk, as a result of increased pore-fluid pressure and/or decreased total confining stress due to relative extension. Construction of effective stress versus progressive burial-depth profiles, using three models of overpressure generation, suggests that high pore-fluid pressures alone could not have formed the tension fractures in Albuskjell. Measurements and calculations from depth-converted seismic sections have shown that halokinesis occurred throughout chalk deposition and continued, in a series of pulses, until early Miocene time. Incremental stress values, associated with halokinesis, predict optimum conditions for tension fracture formation at 4,500-ft (1,370-m) burial, owing to a combination of halokinetic doming and overpressuring, with an important contribution from source rock maturation. These fractures possibly formed, therefore, post-hydrocarbon emplacement (~3,000-ft [1,000-m]) burial ^identity mid-Oligocene), and have been preserved by hydrocarbon invasion.

The healed, shear fractures formed prior to significant pressure solution and hydrocarbon emplacement and are thought to be associated with extra-dome processes, possibly involving graben tectonics or halokinetic reactivation of earlier northwest-southeast-trending major faults. Their distribution is expected to be fairly uniform over the field, with broadly northwest-southeast orientations.

Shear-fracture density is moderately constant between the Danian and Maastrichtian chalk sequences, but is significantly reduced, to absent, in the argillaceous base-Danian.

Model studies predict that tension fractures should be concentrated in crestal areas of the field where near-random orientations may occur. Preferentially radial, with some concentric, orientations relative to the structural dome, may occur off the crest, with radial fractures persisting farther downflank. Predicted tension fracture orientations agree well with those measured from the oriented cores of well 1/6-A10. Open tension fractures are preferentially absent from tight (early-cemented or argillaceous) chalks. They are also poorly developed in "soft" lithologies and predominantly occur in chalks with moderate to high porosities with stylolites.

The spacing of open tension fractures is assumed to be constant over the field, whereas fracture width decreases from crest to flank; this is in line with the concept of lower horizontal stresses in the structural crest. A plot of optimum fracture spacing versus fracture width has been constructed using strain values calculated from the halokinetic models. Analysis of production data supports the proposed model and agrees well with a strong halokinetic influence on fracture development. Although significant halokinesis ceased during the early Miocene, residual strain may still exist in Albuskjell, with larger horizontal effective stresses occurring in the flanks of the structure. Calculations suggest that tension fractures are probably open under initial field conditions to a potentia depth of 11,500 ft (3,500 m).