From weather data to water fluxes simulation in Mediterranean greenhouses through a combined climate and hydrological modelling approach

In the Mediterranean basin, agricultural land covered by greenhouses has been surging in the recent decades. The main goal of this study is to provide estimates of water demand and fluxes in Mediterranean greenhouses starting from outdoor weather data. This is achieved by developing a novel agricultural water modelling framework that combines a greenhouse climate model with a Richards equation-based hydrological model. We improve and evaluate an existing greenhouse climate model with greenhouse data from an experiment using rocket ( Diplotaxis tenuifolia ) as the candidate crop in South Italy for its market importance. The first major improvement regards the iterative estimation of the potential crop evapotranspiration using the FAO56 Penman Monteith method, adapted for greenhouse conditions, at the hourly scale, rather than a locally calibrated formula. The second one concerns the full coupling between the heat balance equations of the air and the soil compartments. The greenhouse climate model was able to simulate with satisfying accuracy the measured indoor air temperature (r 2 =0.58 and KGE=0.76) and relative humidity (r 2 =0.47 and KGE=0.67). Importantly, the crop potential evapotranspiration estimated from climate data either measured indoor or simulated with the greenhouse model were identical. Next, the hydrological model CATchment HYdrology (CATHY) was evaluated in the same experimental setting but different period (rocket in autumn and spring growing conditions), under sprinkler and subsurface drip irrigation. The CATHY model, fed with irrigation data and crop potential evapotranspiration estimated from measured indoor climate, reproduced well the measured soil water content dynamics at five depths (10, 20, 30, 40, 50 cm), despite some bias due to the lack of soil-specific sensor calibration. While the proposed modelling framework is currently coupled in a one-way manner, it has the potential to unlock valuable knowledge for the enhancement of our understanding of greenhouse farming implications on water management at plot and larger scales. • A simple greenhouse climate model is combined with a process-based hydrological model. • The modeling framework simulates water fluxes in greenhouses given outdoor weather. • The modeling framework is tested in two greenhouse experiments. • Good agreement between simulated and measured soil water content at five depths.