A canopy of plants may become a vital component of advanced controlled ecological life support systems (CELSS). The interactions of the canopy with its environment need to be modeled so that designers can properly assess alternate configurations and operating strategies. Collective behavior of an entire canopy can sometimes be more expeditiously modeled than microscopic processes while preserving the robustness of the model for analysis of a CELSS.Water transpiration is a particularly important canopy process for which it is possible to link underlying microscopic processes and arrive at a description of canopy-level aggregated behavior. The underlying fundamental processes driving transpiration are relatively well understood. Unfortunately, the usual characterization of transpiration relies on parameters such as stomatal and boundary layer conductivities that are not directly measurable in typical CELSS designs.We have developed a model of water transpiration in a plant canopy that combines two approaches. The first approach is to account for underlying physical processes, while the second is to empirically incorporate transpiration data now being generated at the Johnson Space Center Variable Pressure Growth Chamber, a bioregenerative CELSS test-bed. The two approaches, physical modeling and data analysis, allow us to produce amodel that is more robust than either the standard first-principles model or a straightforward empirical model. The former requires knowledge of difficult to measure parameters, while the latter ignores known plant behavior.We show that our transpiration model is able to efficiently capture the dynamic behavior of the plant canopy over the entire range of environmental parameters now envisioned to be important in an operating CELSS. We also present examples of the use of this model in assessing plant canopy dynamics and CELSS design options.