We investigated the frictional properties of simulated fault gouges derived from the main lithologies present in the seismogenic Groningen gas field (NE Netherlands), employing in situ P–T conditions and varying pore fluid salinity. Direct shear experiments were performed on gouges prepared from the Carboniferous shale/siltstone substrate, the Upper Rotliegend Slochteren sandstone reservoir, the overlying Ten Boer claystone, and the Basal Zechstein anhydrite–carbonate caprock, at 100°C, 40 MPa effective normal stress, and sliding velocities of 0.1–10 μm/s. As pore fluids, we used pure water, 0.5–6.2 M NaCl solutions, and a 6.9 M mixed chloride brine mimicking the formation fluid. Our results show a marked mechanical stratigraphy, with a maximum friction coefficient (μ) of ~0.66 for the Basal Zechstein, a minimum of ~0.37 for the Ten Boer claystone, ~0.6 for the reservoir sandstone, and ~0.5 for the Carboniferous. Mixed gouges showed intermediate μ values. Pore fluid salinity had no effect on frictional strength. Most gouges showed velocity‐strengthening behavior, with little systematic effect of pore fluid salinity or sliding velocity on (a–b). However, Basal Zechstein gouge showed velocity weakening at low salinities and/or sliding velocities, as did 50:50 mixtures with sandstone gouge, tested with the 6.9 M reservoir brine. From a rate and state friction viewpoint, our results imply that faults incorporating Basal Zechstein anhydrite–carbonate material at the top of the reservoir are the most prone to accelerating slip, that is, have the highest seismogenic potential. The results are equally relevant to other Rotliegend fields in the Netherlands and N. Sea region and to similar sequences globally.

A thorough understanding of the mechanical properties of these faults is needed to assess the potential for induced fault reactivation and associated seismic hazard. Specifically, the data on the dynamic frictional behavior of the relevant fault rocks, obtained at realistic imposed displacement and velocity trajectories, are needed to constrain geomechanical modeling of fault rupture and associated seismicity (e.g.,

Fluid production from and injection into subsurface reservoirs can cause multiple (micro)seismic events near boreholes and in surrounding fracture networks, as well as causing creep or seismic slip events on pre-existing faults (Ellsworth, 2013). In the giant Groningen gas field (northeast Netherlands), induced seismic events have occurred since the 1990s, with the strongest occurring on 16 August 2012 with a magnitude of M W 3.6. Most events are interpreted to occur along the pre-existing, high-angle, normal faults that crosscut the Slochteren sandstone reservoir and associated claystones, the overlying basal Zechstein caprock and the underlying Carboniferous siltstone/shale source rocks (Van Eijs et al., 2006). Assuming a circular slip patch, and using the statistical relation between rupture area and earthquake magnitude given by Hanks and Bakun (2008), the larger M W 3-4 events correspond to rupture length-scales of 350-1,155 m, which do not necessarily rupture the entire fault length in the downdip or strike directions (Figure 1). From kinematic modeling of earthquake rupture processes in