Water use of two desert landscape tree species in Tucson, Arizona.

Abstract

Although water conservation programs in the arid Southwest have prompted prudent landscaping practices such as planting low water use trees, there is little data on the actual water use of most species. Few methods or models have been developed for measuring tree water use. The stem heat balance method is one such method. Predictive models of tree water use have been limited to applications of the Penman-Monteith (PM) equation with varying degrees of success. The purpose of this study was: to validate stem flow gauge accuracy in a greenhouse and a desert environment; to determine the actual water use of two landscape tree species in Tucson, Arizona; to determine water use coefficients for two tree species based on the crop coefficient concept; and to test and develop a predictive model of tree water use based on the Penman-Monteith equation. Water use of oak (Quercus virginiana 'Heritage') and mesquite (Prosopis alba 'Colorado') trees in containers was measured using a precision balance and stem flow gauges. Water use coefficients for each tree species were calculated as the ratio of water use per total leaf area and per projected canopy area to reference evapotranspiration (ETo) using the Penman combination equation. After accounting for tree growth, water use coefficients on a total leaf area basis were calculated to be 0.48 and 0.97 for the oaks and mesquites, respectively, and 1.36 and 1.56 for the oaks and mesquites, respectively, on a projected canopy area basis. These coefficients indicate that mesquites (so called xeric trees) use more water than oaks (so called mesic trees) under non-limiting conditions. Stomatal resistances (r(s)) were calculated using the PM equation, and ranged from 20 to 200 s cm⁻¹. Calibrations were developed between r(s) and net radiation for both species. Results of the PM model to predict daily tree water use ranged from -15 to +150 percent error, depending on the tree, indicating the need for accurate measurements of stomatal resistance in order to use the PM model. Results indicate that a shortened form of the PM equation requiring only vapor pressure deficit and r(s) would be sufficient to predict tree water use.