An analysis of four different models for quantifying tomato transpiration in a soilless greenhouse

Abstract

Scientific literature points to several models used to calculate the transpiration of vegetable crops inside a greenhouse but the results are hard to compare since different experiment setups were used. This study aimed to analyse the accuracy of four different transpiration models, Penman-Monteith, Stanghellini, Priestley-Taylor, and Takakura for the calculation of tomato transpiration inside a soilless greenhouse. By addressing this knowledge gap the outcome of this study provides a step towards more autonomy in vegetable horticulture. During the summer a three month experiment was conducted in a Venlo-type soilless greenhouse in The Netherlands. The air temperature, relative humidity, wind speed and net radiation were measured inside a tomato greenhouse. Hourly transpiration data measured using a combination of sap flow sensors and lysimeters, was compared to the four different models. The Penman-Monteith and Priestley-Taylor models, intended for outdoor use, had a similar accuracy as the Stanghellini and Takakura models that are developed for Venlo-type soilless greenhouse. When comparing the models error range to the average hourly transpiration, it is found that the models are not accurate enough for irrigation scheduling. A model sensitivity analysis illustrates how changes in net radiation, temperature, humidity and wind speed affect modelled tomato transpiration. A recommendation on methods to improve model accuracy is made.