Performance of Penman-Monteith Models in Predicting EvapoTranspiration in a Large Banana Screenhouse

Screenhouses are widely used in arid and semi-arid agriculture to protect crops from direct solar radiation and high wind speed and to increase water use efficiency. In this work, evapotranspiration (ET) of a banana plantation in a large shading screenhouse was measured using the Eddy Covariance (EC) technique, and measurements were compared with different variations of the Penman-Monteith model. The experiments were carried out in a flat-roof banana screenhouse located at the coastal plane of northern Israel. The woven screen was made of a transparent monofilament and had a nominal shading of 8% (manufacturer data). An eddy covariance system consisting of a 3-axis ultrasonic anemometer and an infra-red gas analyzer was deployed at the center of the screenhouse at a height of 5 m, and measured at 10 Hz. In addition, air temperature and humidity, net radiation, soil heat flux, global radiation and photosynthetic active radiation, were measured. The models examined were a classical Penman-Monteith (PM) model with an aerodynamic resistance based on a logarithmic wind profile, denoted as PMra; a similar model with the aerodynamic resistance replaced with a boundary layer resistance, PMrb; a PM model that was adapted in a previous study to screenhouse conditions using an additional boundary layer resistance, PMsc; two models where the boundary layer resistance was adjusted to the specific flow conditions, PMrbm and PMscrbm; and the PM model for reference evapotranspiration, ET0. Results show that in terms of the mean daily ratio, the ET0 model gave excellent prediction of the daily ET, then the PMrbm model whose prediction was 3% higher, the PMrb that was 4% higher, the PM(scrbm)that was 6% higher, the PMsc that was 7% higher and PMra that was 16% higher than the measurements.