A COMPARISON BETWEEN 2 MODELS FOR PREDICTING CORN YIELD IN SALINE STRESS CONDITIONS

In order to verify if salinity stress may be simulated as a water stre ss, two simulation models, widely applied in water-limited conditions, have been here described and used under saline conditions. The first one uses an empirical approach, the second model (CERES-Maize) is a fu nctional crop model, here conveniently modified. For both models input data set was collected from corn cropped in 30 drainage lysimeters, d ivided into two soil texture types (loam and clay) and three water qua lities (EC=0.9 dS m(-1) as a control. EC=2.3 and 3.6 dS m(-1)). Approp riate statistical analysis (linear regression between simulated and ob served data, ''lack of fit'' test) showed that both models provide a q uite good performance in estimating yield reduction in a saline enviro nment as a water stress. No significant, differences between the two t ypes of soil texture were noticed. A part from grain yield, the functi onal approach allowed for simulating evolution over time of LAI. bioma ss, and evapotranspiration on a daily scale. The agreement was very go od for biomass: as regards WI, simulated values didn't match maximum v alues: for saline treatments the fit in daily ET was acceptable. Direc tions for simulating maize growth under saline conditions are finally suggested.