Tree Physiology 2, 205-214 (1986). 0 1986 Heron PublishinK-Vic,toria, Canada. Simulation model of a laboratory-grown pbreatophytic woody legume D. BACHELET’, W. M. JARRELL2 and R. A. VIRGINIA3 I Department of Biology, New Mexico State University, Las Cruces, NM 80523, USA ’ Dry Lands Research Institute, and Soil and Environmental Sciences, University of California, Riverside, CA 92521, USA ’ Biology Department and Systems Ecology Research Group, San Diego State University, San Diego, CA 92182, USA Summary A mechanistic mode1 was developed to simulate growth of mesquite Prosopis glandulosa Torr. trees under a phreatic (groundwater) moisture regime. Experimental data obtained in a greenhouse repro- ducing the phreatic environment (2 m soil columns with 10 cm of water-saturated soil at the bottom) were used to parameterize three submodels predicting carbon (C), nitrogen (N) and water dynamics in leaves, branches, roots and root nodules. In the column simulation model (COLSIM), photosynthesis was driven by air temperature and soil salinity. Water availability was nonlimiting. Nitrogen was absorbed by the roots from inorganic soil N and also fixed by root nodules. Comparison of the simulation with results from the greenhouse experiment showed that the model accurately reproduced shoot biomass and nitrogen content dynamics up to three years with or without a high soil salinity content. Root biomass was underestimated when soil salinity was high because the model did not account for the increased allocation of C to roots under conditions of high salinity. Observed annual cycles of water uptake during the three-year run were not reproduced because the mode1 did not include a phenological function which apparently drives these cycles. Introduction Few studies of desert plants have examined processes occurring at a soil depth of more than 1 m (Ehleringer and Mooney 1983). Phillips (1963) found roots of a woody legume, probably mesquite (Pros@ sp.), over 53 m below the soil surface in the Sonoran desert of Arizona. Because of the difficulty of making direct quantitative observations at depths greater than 2 m, our knowledge of belowground processes and their importance in the cycling of nutrients is limited. We have undertaken simulation modeling to provide insight into deep soil pro- cesses (Kowal 1971) of the mesquite system. Our objective in producing the model was to examine two hypotheses concerning the functioning of the deep root system: 1) mesquite productivity is determined by the availability of water and the ability of the soil to meet plant nutrient requirements; 2) phreatophytic (groundwa- ter-dependent) mesquite develops two functionally independent, lateral root sys- tems in the California Sonoran desert: a surface absorbing system in a soil zone where activity is limited by the environmental conditions at the ground surface, and a deep phreatic root system in a soil zone where conditions are optimal (constant temperature and moisture) for nodulation and NZ fixation. To test these hypotheses, outputs of a model (COLSIM) simulating the the C and N acquisition and water use by the woody legume, mesquite, under phreatophytic by guest on February 13, 2016 http://treephys.oxfordjournals.org/ Downloaded from