An aggregated model for water requirements of greenhouse tomato grown in closed rockwool culture with saline water

The paper reports an aggregated model for the water requirements (W) of greenhouse tomato grown in spring or summer�autumn season in closed (recycling nutrient solution) rockwool culture. A series of experiments were conducted between 2000 and 2003. In each experiment, two different nutrient solution treatments were established in a randomized block design, with three replicates. The nutrient solutions differed only for the concentration of NaCl and the electrical conductivity (EC), which were, respectively, 10 mol m-3 and 3.0 dS m-1 in T10 treatment, and 20 mol m-3 and 3.9 dS m-1 in T20 treatment. Fruit yield was 10.1 ± 0.3 and 9.6 ± 0.2 kg m-2 in T10 and T20 treatments, respectively, and the difference was not significant. Two diverse sub-models were developed, for the daily crop water uptake (WU) and for the leaching requirement (LR) in dependence of a maximum tolerable sodium (Na+) concentration in the recycling nutrient solution. WU was simulated using a simple regression model, which considered the radiation actually intercepted by the crop. The determination coefficient for the linear regression between predictions (P) and measurements (M) of daily WU was 0.853 (n = 581), with the slope (0.963) and the intercept (-0.053) of the linear regression not significantly different, respectively, from 1 and 0. There was a good agreement between P and M values of WU, with a mean deviation between P and M of 2 ± 10%. The modelling of LR was less satisfactory, especially in spring crops; the average value for P and M values of LR was 0.30 and 0.28, respectively. The P�M residual averaged 11 ± 25% for LR and 13 ± 26% for the volume of drainage water or runoff (WR). Nevertheless, the aggregated model accounted well for W with an average deviation of 4 ± 11% between P and M. The sensitivity analysis showed the great influence of radiation on W and suggested that the quality of irrigation water (namely, NaCl concentration) is more relevant for crop water use efficiency than any other factor considered by the LR model, including the affinity of the plant for the salts contained in the nutrient solution.