4D seismic is a dynamic source of data that provides information about changes in reservoir-rock and -fluid properties over time. Seismic attributes are sensitive to variations in the fluid content, temperature, and pressure distribution; therefore, 4D-seismic images contain information on the nature of fluid flow within the reservoir. Perhaps the most-reliable and -important information that can be learned from 4D-seismic images is related to anomalies in fluid flow within the reservoir. During steam-assisted gravity drainage (SAGD), the steam-chamber propagation is fairly clear from 4D-seismic images, mainly because of higher gas saturation in the chamber. Therefore, anomalies are revealed by the absence or unexpected location of the steam chamber. In this paper, a practical methodology is proposed for consideration of anomalies identified from 4D-seismic images in geostatistical reservoir models. The geostatistical realizations are updated to enforce missing anomalies and improve reservoir characterization. The updated models are suitable for reservoir decision making and management.