Prosopis velutina (velvet mesquite) forests are one of many types of aridland riparian ecosystems that are threatened by groundwater pumping and other types of water development. Empirical models developed using both hydrological and vegetational data sets have potential uses in the management of these threatened ecosystems. To this end, we developed models for Prosopis velutina stands across a xeric-to-mesic moisture gradient. The models expressed canopy height, basal area, leaf area index, vegetation volume, and leaflet area as functions of plant water potential, and they expressed plant water potential and riparian stand structure as functions of water table depth. These data indicated that stand structure was strongly related to water availability. Management applications of the models include the ability (1) to identify minimum water-table depths for riparian stand maintenance and (2) to detect stressful hydrological conditions, via water potential measurements, before the onset of structural degradation.