Transpirational supply and demand: Plant, soil, and atmospheric effects evaluated by simulation

The assumption that transpiration is the lesser of an atmospheric demand function and a water supply function was tested by simulation with Federer's (1979) soil‐plant‐atmosphere model. The best estimate of atmospheric demand is called unstressed transpiration, defined as the transpiration that would occur in ambient conditions if stomata were unaffected by plant‐water potential. For practical purposes the Penman equation provides a good estimate of unstressed transpiration for short vegetation but not for forests. Even when atmospheric variables and the Penman estimate are held constant among forest canopies, unstressed transpiration can vary by a factor of two because of variation both in the maximum value of leaf conductance and in the ratio of canopy conductance to leaf conductance. The best water supply function incorporates depth variation of soil water potential and of root and soil properties. A more practical supply function uses the ratio of available water in the root zone, W , to maximum available water W M . The maximum available water is soil water held at potentials less than that at which the hydraulic conductivity is 2 mm/d and greater than the critical leaf water potential at which stomata are completely closed. Using a mature hardwood forest as a standard, various parameters were varied to examine their effects on a water supply function defined as a supply constant times W / W M . The supply constant was found to be independent of soil texture and physical properties. Root density and the internal resistance of the plant to water flow were the most important determinants of the supply constant. Reasonable variation of root density and internal resistance produced variation in the constant from 1.9 mm/h, which implies that supply is less than demand only when soil is very dry, to 0.5 mm/hr, which implies that supply cannot meet the demand even when the soil is wet.