Many gas fields are currently being produced in the northern Netherlands. Induced seismicity related to gas production has become a growing problem in the Netherlands in the past two decades. To date, a few hundred induced seismic events occurred. Induced seismicity is generally assumed to be the result of induced reactivation of discontinuities in the subsurface. Field data of the Groningen and Annerveen gas fields as well as other Rotliegend gas fields in the Netherlands are analysed. A large amount of seismic cross sections through seismic events is studied. It is very likely that the seismic events are the result of reactivation of existing discontinuities (like faults) in or near the reservoirs. The objective of the research presented in this dissertation is to obtain a better understanding of the mechanisms of gas production induced reactivation of faults by means of 3D geomechanical modelling of gas reservoirs. It is a step towards future assessment of expected seismic energy release when (further) developing gas fields in the Netherlands. Furthermore, attention is given to the development of several quantification methods, used for the analysis of the calculation results. Quantification methods include relative shear displacements, seismic moment, stress paths, mobilised shear capacity and total stress changes per unit depletion. The geomechanical models represent the geometries found in the seismic cross sections. The models contain a disk-shaped gas reservoir in an extensional stress regime. A steeply dipping normal fault plane intersects the reservoir and divides it into two compartments: a footwall and a hanging wall reservoir compartment. Stress development and fault slip during gas depletion are analysed. Gas depletion can lead to both normal and reverse fault slip on the same fault plane. In the given setting of a steeply dipping normal fault in an extensional stress regime, normal fault slip due to differential reservoir compaction is the dominant mechanism, rather than reverse fault slip. The effect of differential reservoir compaction is most pronounced for a configuration, where the top of the hanging wall reservoir compartment is positioned exactly opposite to the bottom of the footwall reservoir compartment, resulting in a relatively large amount of fault slip over a narrow area. Normal fault slip is supported by equal depletion of both reservoir compartments. Reverse fault slip is supported by differential pore pressure development due to reservoir compartmentalisation. Especially the combination of a relatively stiff surrounding rock and differential pore pressure development due to reservoir compartmentalisation can result in relatively large amounts of reverse fault slip. Both normal and reverse fault slip are promoted by a Young's modulus or Poisson's ratio of the surrounding rock larger than those of the reservoir rock (Esur > Eres or νsur > νres). The initial state of stress is relatively closer to the failure line than in case of a smaller stiffness contrast. Esur < Eres and νsur < νres oppose the reactivation of the fault. Values of νsur lower than 0.2 seem to have no significant influence on the calculated maximum normal fault slip. Calculations indicated that the Young's modulus of the surrounding rock is a more important parameter influencing gas depletion induced fault slip than the Poisson's ratio of the surrounding rock. Calculations with a 3D anisotropic tectonic stress field show a strong dependency of the amount of calculated fault slip on the direction of the maximum horizontal stress with respect to the fault strike direction. Most normal fault slip occurs when the maximum horizontal stress is directed parallel to the fault strike. Minimum normal fault slip is calculated for a maximum horizontal stress direction perpendicular to the strike direction of the fault. A larger horizontal stress component parallel to the azimuth of the fault has a limiting effect on the fault slip. Total fault slip can consist of a dip slip and a strike slip component. In case of a horizontal reservoir, no significant strike slip is observed when the fault strike direction is a principal stress direction. A certain amount of strike slip is observed for calculations with an angle between the maximum horizontal stress direction and the fault strike direction of 31°, 45° and 59°. Strike slip contributes to both normal and reverse fault slip.