Could olfactory displays improve data visualization?

Smell (olfaction) can be critical to our daily living - for example, smelling smoke from a fire in time to leave a burning building - but it is seldom used in data visualizations or virtual reality (VR) systems. Could olfaction displays (devices that output scented air) augment data visualization, that is, communicate information relevant to many fields? Perhaps incorporating underutilized modalities such as haptics (touch), olfaction, and gustation (taste) as data visualization aids is the next logical step to optimizing human information processing. It seems reasonable that adding the sense of smell to a virtual environment (VE) would enhance the environment's presence or "realness". Attempts have been made to include olfactory displays in VEs (for example, John Cater's Deep Immersion Virtual Environment Laboratory at the Southwest Research Institute), but most have been unsuccessful. One reason is the lack of a standard to represent and playback smells. Olfactory effects could play a crucial role in certain training environments, such as those for fire fighters and medical personnel. Current VEs include advanced visual and audio outputs, but smell is either very limited or absent. Very few studies focus on this subject, with most discussing ambient (whole-room and long-duration) rather than specific (localized and short-duration) odors, which are relevant to data visualization and VR.