Application and evaluation of Stanghellini model in the determination of crop evapotranspiration in a naturally ventilated greenhouse

Stanghellini model is one of the few models primarily developed to predict the evapotranspiration of crops (ETc) in naturally ventilated greenhouses. However, there are insufficient data on the model regarding its use, particularly in China where solar greenhouses without heating systems are fast spreading for vegetable growth and production. The application of Stanghellini model and the evaluation of its performance using meteorological and tomato plant data generated inside an unheated and naturally ventilated multi-span Venlo-type greenhouse is exploited in this study. Model capability was evaluated by utilizing data from sap flow measurements, meteorological and crop data. Measured meteorological data included solar radiation (R-s), air temperature (T-a), relative humidity (RH) and net radiation (R-n). Average leaf area index (LAI) values measured during the experimental period were 1.00, 3.30, 4.05 and 2.93; while determined crop coefficients (K-c) changed from 0.40, 0.62, 1.12 to 0.83 for the initial stage, development stage, mid-season stage and late-season stage, respectively. Results from the study indicated that the average hourly ETc values of tomato plants using sap flow measurements were 0.165 mm/h, 0.148 mm/h, 0.192 mm/h and 0.154 mm/h for the initial stage, development stage, mid-season stage and late-season stage, respectively. Meanwhile, the ETc values obtained from calculation using Stanghellini model were 0.158 mm/h, 0.152 mm/h, 0.202 mm/h and 0.162 mm/h for the initial stage, development stage, mid-season stage and late-season stage, respectively. These ETc values calculated by the Stanghellini model were close to the measured values within the same period. The coefficients of correlation (R-2) based on hourly ETc for the calibration data was 0.94 and that of the validation dataset was 0.90. Scatter plots of the estimated and measured hourly ETc revealed that the R-2 and the slope of the regression line for May, June and July were 0.94, 0.90, 0.96 and 1.15, 0.97, 1.10 respectively. These data were well represented around the 1:1 regression line. A model sensitivity analysis carried out illustrates how the changes in R-s and T-a affect greenhouse ETc. Stanghellini model was therefore proven to be suitable for ETc estimation with acceptable accuracy in unheated and naturally ventilated greenhouses in the Northeast region of China.