A field experiment in the Western Canadian Sedimentary Basin, Alberta, Canada entails creation of a horizontal subsurface lens (fluid-driven fracture) through fluid injection from a vertical well. The initial lens creation is followed by cycling of smaller injection and flowback stages. After completion of these smaller injection/flowback cycles, the lens is (presumably) expanded through a larger scale injection followed by another series of smaller injection/flowback cycles. Tiltmeter data indicates subhorizontal lens orientation and nearly circular shape. Combining the lens radius inferred from tiltmeter inversion with measurements of downhole pressure and downhole caliper width (opening) measurements indicate the compliance is far below the linear elastic fracture prediction based on assumption that the parting between the fracture surfaces is complete and no intact rock bridges constrain opening of the lens. Additionally, the compliance is shown to increase with increasing maximum width attained by the lens. This history-dependence suggests intact rock bridges are progressively broken as larger injected volumes are used to attain larger widths. In this view, as the bridges break, the lens compliance increases. By considering a model wherein bridges are accounted for by springs with a prescribed ultimate breaking length, the observed crack compliance and dependence of the compliance on the width history are matched. Hence, the detailed field data combined with a suitable bridge constriction model comprises compelling evidence for rock bridges constricting fluid-driven crack apertures at field scale.

Acknowledgements. AB, AD and EP acknowledge support from the Australian Research Council through project DP190103260.