Measuring and predicting evapotranspiration in an insect-proof screenhouse

This study addresses the current lack of accurate measurements and predictions of evapotranspiration below modern screening materials. Energy balance components, microclimatic and physiological parameters were measured in a 50-mesh insect-proof screenhouse cultivated with sweet pepper. A lysimeter-calibrated sap flow system and an eddy covariance system installed inside the screenhouse measured canopy transpiration and evapotranspiration, respectively. Sap flow (SF) and eddy covariance (EC) data showed good agreement (R-2 = 0.95) and although EC exceeded SF by 9%, this was less than the standard error of inter-specific sap flow, supporting previous findings that mid-day soil evaporation was negligible. Screenhouse bulk resistance was inversely correlated with external wind speed and simple mass balance estimations gave reasonable values of evapotranspiration on a daily basis. A one-dimensional screenhouse model was derived, based on a modified Penman-Monteith equation incorporating an additional boundary layer resistance. The model, along with a greenhouse model, was used to calculate evapotranspiration of the screenhouse crop. Half hourly predictions of both models yielded good agreement with measured SF (R-2 = 0.94-0.95). Daily crop water use, predicted with the greenhouse model was in better agreement with measurements than that predicted by the screenhouse model. Measured and predicted transpiration rates inside the screenhouse were approximately 1.8-2.1 mm day(-1) during the most active stage of growth, while those simulated for a similar crop grown outside would be 4.5-5.3 mm day(-1) on average. Model sensitivity analysis showed that reduced radiation and wind speed and modified vapour pressure deficit were the main factors influencing transpiration. Calculations of the decoupling factor Q suggest that the screenhouse evaporative climate is predominantly "decoupled". (C) 2004 Elsevier B.V. All rights reserved.