Irrigation of grapevines with saline water - II. Mathematical simulation of vine growth and yield

Soil, water, atmosphere and plant (SWAP) model simulates deterministic transport of water and solutes, incorporating a semi-analytical sink function. It enables one to simulate detailed (SAWPd) or simple (SWAPs) crop growth patterns in response to flow patterns in the root zone. The objectives of this study were to evaluate the ability of SWAPd to account for various salinity effects in grapevines under and conditions, and to compare results from SWAPd and SWAPs growth models. A unique approach in our study was to use the same crop parameters for fresh and saline water while changing only the salinity of the irrigation water. We tested the effect of three salinity treatments (1.8, 3.3 and 4.8 dS m(-1)) on production parameters of grapevine. Vines in the fresh water treatment benefited from the better water quality in that they used water more efficiently than in the other treatments. Three objective criteria were used to test the validity of the two models. These were standard error of model estimation (SEE), root mean square error (RMSE) and the index of agreement (IoA). With respect to the measured variables SEE (expressed as percent of the maximal value) was 5% for 1.8 dS m(-1) treatment and 11% for 4.8 dS m(-1) treatment. RMSE was 7 for 1.8 dS m(-1) treatment and 15% for 4.8 dS m(-1) treatment. IoA was close to 1. It varied between a minimum of 0.8 for SWAPs to a maximum 0.99 for transpiration calculated by SWAPd for low salinity conditions, indicating acceptable agreements between the simulated and measured results. These simulated results were obtained when water quality was the only variable and hence they indicated the ability of SWAP's salinity models to generate realistic responses to salinity. In all calculations higher transpiration and LAI were simulated by SWAPd than by SWAPs. The deviations (SEE and RMSE) of SWAPs from the measured values were about 4% larger than the deviations of the values simulated by SWAPd. It probably resulted from the more realistic detailed model, which contains many more growth parameters than SWAPs that requires only LAI. The higher simulated transpiration by SWAPd model affected the differences in the entire water regime. For example, SWAPs simulated higher water content than SWAPd because less water was extracted from its soil profile by transpiration. It was concluded that for the first time SWAPd (and to a certain level also SWAPs) was validated for grapevine under saline conditions. (c) 2006 Published by Elsevier B.V.