The dissolved salt ions that are not absorbed during irrigation of greenhouse crops are gradually accumulated in the nutrient solution resulting in levels of salinity high enough to damage the crops. This water salinity presents operational and environmental challenges as the nutrient-rich greenhouse effluent should be discharged to the environment when deemed unsuited for irrigation. In this pilot-scale study, the potential of passive salt reduction (phytodesalination) in gravel and wood-chip flow-through reactors was evaluated using seven plant species including Schoenoplectus tabernaemontani, Andropogon gerardii, Typha angustifolia, Elymus canadensis, Panicum virgatum, Spartina pectinata and Distichlis spicata along with an unplanted control reactor. While the unplanted system outperformed the planted units with gravel media, the wood-chip bioreactors with S. tabernaemontani and S. pectinata improved the greenhouse effluent reducing the solution conductivity (EC) by a maximum of 15% (average = 7%). S. tabernaemontani and D. spicata showed higher accumulated contents of Na⁺ and Cl⁻ in comparison with T. angustifolia and S. pectinata. Overall, S. tabernaemontani was selected as the most capable species in the wood-chip bioreactors for its better salt management via EC reduction and salt accumulation. It was however concluded that further treatment would be required for the greenhouse effluent to meet the stringent irrigation water quality guidelines in order not to pose any adverse effects on sensitive crops. Finally, the present hydraulic residence time (HRT = 3.7 days) and the solution salinity concentration were identified as the potential factors that may be limiting the efficiency of plant salt uptake, emphasizing the need for conducting more research on the optimization and enhancement of passive desalination systems for the greenhouse effluent.

The microbial characteristics of four vegetated and one unplanted wood-chip bioreactors treating greenhouse effluent were investigated in a continuous experiment operated for over 2.5 years. The bioreactors were designed to reduce nitrate concentrations via naturally induced microbial denitrification. The vegetation type and reactor depth were both found to be significant factors in defining the mixed microbial activity. However, a consistent correlation between the abundance of the denitrifying communities and reactor depth could not be found across all reactors. The media samples from the unit planted with Typha angustifolia displayed higher microbial activities compared with the other reactors. This plant’s root-associated bacteria also demonstrated the greatest copies of the denitrifying genes nirK and nosZ. The most abundant denitrifier communities and those encoding the nosZ gene were found in the unplanted reactor, followed by the T. angustifolia unit. The T. angustifolia reactor demonstrated greater microbial activity and denitrification capacity at the depth of 20 cm, while the greatest denitrification capacity in the unplanted reactor was found at the depth of 60 cm. These findings indicated the importance of the T. angustifolia rhizosphere to support microbial community establishment and growth in the vicinity of the plant’s roots, although those populations may eventually develop in an unplanted environment.