Ten steps applied to development and evaluation of process-based biogeochemical models of estuaries
Environmental Modelling & Software
Environmental Modelling & Software
| Hi-index | 0.00 |
A composite model was developed for water and mineral relations of greenhouse tomato (Solanum lycopersicum L.) cultivated in semi-closed or open soilless (rockwool) culture. The model simulated on a daily basis: (i) the evolution of crop leaf area index and water uptake using empirical equations; and (ii) the variations of ion concentrations and electrical conductivity (EC) in the recirculating or drainage nutrient solution using a mass balance equation based on the concept of ion uptake concentration. The model was calibrated using measured data collected in previous works and validated in two independent experiments carried out in 2005 and 2007. In these experiments, different fertigation strategies were tested using nutrient solutions prepared with saline (9.5molm^-^3 NaCl) water. In semi-closed systems, the recirculating nutrient solution was discharged whenever EC exceeded a pre-defined threshold (4.5, 6.5 or 7.5dSm^-^1, depending on the experiment) and/or nitrate (NO"3^-) concentration was lower than 1.0molm^-^3. This value was selected because 20mgL^-^1 (1.43molm^-^3) is the limit imposed to NO"3^- concentration of wastewater discharged into surface water by the current Italian legislation. In open system, the crop was irrigated with full-strength nutrient solution without recirculation of drainage water. In both years, fertigation strategy did not affect significantly crop growth, fruit yield and water uptake. In most cases, simulations of seasonal crop water uptake were within the confidence interval of the measurements with a maximum deviation of -6%. The model predicted acceptably the time course of EC and ion concentration in recirculating (semi-closed systems) or drainage (open system) nutrient solution. A moderate discrepancy between observations and simulations was found for NO"3^- concentration, especially during the first weeks after planting. In general, there was a good agreement between simulated and measured values of total water and nitrogen (N) use. In 2005, simulated values of N uptake in semi-closed systems were -11% to +5% of measured values. Prediction of N uptake was less accurate in 2007, when simulated values were +17% of measured values. In open system, the model underestimated N uptake (-17%) mainly due to overestimation of N leaching (+6%). Applications of the composite model for operative management of soilless culture and for scenario analysis are discussed.