The GEM model developed for soilless cultures consists of different submodels (A) for applications to crops grown on mats by drip irrigation, (B) for spray applications to crops grown on such mats, and (C) is for spray applications to crops grown in pots in an ebb/flood system (GEM-A, GEM-B, and GEM-C). The descriptions of the processes for pesticide behaviour in these submodels were reviewed, considering also their consistency with measurements available in the literature. For GEM-A it is recommended to include sorption to the mats, the foil surrounding the mats and the irrigation pipes and to include partitioning between the water in the mats and the plant roots. For GEM-B it is recommended to include direct contamination of the substrate mats and the troughs resulting from spray and Low Volume Mister (LVM) applications. For GEM-B and GEM-C it is recommended (i) to revise the procedures for calculating the initial concentrations in the air and the condensation water, (ii) to include deposition onto the roof by spray and LVM applications, (iii) to revise the procedure for calculating the volatilisation rates from the plant surfaces. For GEM-C it is recommended (i) to omit the sorption equilibration between the bottom 10 cm of the soil in the pots and the water on the ebb/flood tables, (ii) to revise the procedure for the flux in the gas phase between the greenhouse air and the top layer of the soil in the pots, and (iii) to use a crop-specific value for the fraction of the surface area covered by the pots.

An experiment was conducted to test the Greenhouse Emission Model (GEM) performance. The model simulates fate of plant protection products in soilless systems for a various combinations of application types, substrate and crop types in Dutch greenhouses. Pymetrozine and imidaclorid are applied withthe nutrient solution in sweet pepper growing on stone-wool. Measured and simulated concentrations were compared for (i) simulations with experimentally derived water flows, (ii) simulated water flows based on weather conditions and based on computer settings of the automatic control system and (iii) a predefined scenario in GEM3.3.2 for sweet pepper. GEM is able to simulate water flows well, when these flows are based on weather conditions and computer settings. Also, the concentrations in the mixing tank were simulated well. Simulated concentrations in the used water reservoir were higher than measured concentration. The model performance improves when the cultivation compartment is simulated with two reservoirs instead of one, with the rationale that no complete mixing occurs in the cultivation compartment. The effect of plant uptake and degradation could not be assessed. The concentration in the recirculation water in greenhouse is sensitive to the volumes of the various reservoirs in the greenhouse system. It is recommended to update the reservoir volumes according to the latest insight on commercial greenhouse systems.

Experiences with previous versions of the Greenhouse Emission Model (GEM) for soilless cultures prompted the need for developing a revised version. GEM consists of a submodel for the water flows in the greenhouse (the WaterStreams Model, WSM) and a submodel for the pesticide behaviour in the greenhouse (the Substance Emission Model, SEM). The resulting emission fluxes are used by the TOXSWA model to simulate concentrations in surface water. This report describes the changes in WSM and provides a full description of the concepts and equations in the revised SEM version The main changes in WSM are the extension of options: (i) emission norms for nitrogen up to 2027 can be used, (ii) sodium levels can be set by the user, (iii) options for managing the discharge to the surface water are extended, (iv) a waiting time can be prescribed between a pesticide application and the next discharge. Option (ii) and (iv) are only available is used as a standalone model, i.e. outside GEM. There are two types of SEM: SEM-S for crops grown on slabs and SEM-P for crops grown in pots on tables. Main changes in SEM-S include: (i) the water in the slabs is divided into two equal parts with root uptake restricted to the first part, (ii) sorption to the slab material and the irrigation pipes is included, (iii) the amounts present in the air and condensation water immediately after spray or Low Volume Mister (LVM) application are strongly reduced, (iv) for spray and LVM applications direct contamination of the slabs, the drainage-water troughs, and the roof is added. Main changes in SEM-P include: (i) the amounts present in the air and condensation water immediately after an application are strongly reduced, (ii) the sorption equilibrium between the bottom 10 cm of the pots and the water on the tables was removed, (iii) for spray of LVM applications, the deposition on the tables is increased from about 10 to about 40%.