This paper is included in the Special Publication entitled 'Faulting, fault sealing and fluid flow in hydrocarbon reservoirs', edited by G.Jones, Q.J. Fisher and R.J. Knipe. The primary focus of this paper is to emphasise the large volumes of information related to faults and fault systems that are present in, and extractable from, 3D seismic data. During most interpretations, this information is seldom included and transferred to the reservoir model so that their effects can be accounted for during reservoir simulation. Reasons why they are not included are generally related to the time constraints imposed on studies when commercial considerations are often given precedence above any scientific justification. The inadequacies of what are presently considered acceptable models are highlighted and methodologies that could lead to improved reservoir models are proposed. These methodologies are derived from an investigation of how fault systems will manifest themselves in seismic data, based on both theoretical concepts and the use of synthetic models. From these information sources, principles are derived for the identification of faults and fault systems in 3D seismic volumes. These principles are then tested in two case studies selected to emphasize the limitations imposed on seismic resolution by both target depth and seismic frequency content. After highlighting the diversity of fault related information, which is accessible and currently under-utilized by current reservoir modelling techniques, potential methods for automatic mapping and digital extraction of fault information are proposed. When these automated methods are implemented into seismic interpretation software, commercial reasons for ignoring small seismic-scale faults will be significantly reduced and as a result, reservoir performance predictions will become more realistic.