Kirsty Holstead, Alessio Belmondo Bianchi, Thomas Wagner, Huub Rijnaarts, Arjen van Nieuwenhuijzen and Tiza Spit highlight a Dutch research programme to improve water management in the horticultural sector.
Greenhouse horticulture is a cornerstone of Dutch agriculture. In South Holland thousands of hectares of greenhouses produce vegetables, flowers, and ornamentals for global markets, requiring large volumes of high-quality irrigation water.
Traditionally growers have relied on rainwater collected from greenhouse roofs and stored in open basins, supplemented by groundwater abstraction during dry periods. But this system is increasingly under pressure. Limited basin capacity, combined with rising groundwater abstraction, contributes to land subsidence, salinisation and growing dependence on external freshwater sources. At the same time, urban areas, industry and nature conservation compete for water from the same regional system. Climate change further exacerbates these pressures. Under these conditions, the horticultural sector’s traditional water management approach cannot guarantee reliable freshwater availability, making a transition unavoidable.
AquaConnect is a transdisciplinary Dutch scientific and applied research programme that aims to contribute to a paradigm shift in freshwater provision in the Netherlands from ‘discharging precipitation surplus through the rivers to the sea as quickly as possible’ to ‘storing and reusing water for regional self-sufficient water provision’ (www.aquaconnect.nu). As part of the AquaConnect project, researchers, growers, water companies and public authorities explored alternatives to conventional freshwater supply for Dutch greenhouse horticulture. These discussions highlighted the potential of water storage, sharing and subsurface buffering to build a more resilient circular water system.
The dynamic water sharing model emerged from these explorations as a vision for a circular and resilient water system (see Figure 1), which integrates innovative physical, digital and institutional components. This concept brings together multiple alternative water sources and storage options into one coordinated system: rainwater harvesting, brackish and fresh groundwater extraction and treatment, wastewater reclamation, and both surface and subsurface storage. Four interconnected pillars are central to making this model function effectively: 1. Managing groundwater and storage systems; 2. Ensuring water quality and availability; 3. Optimised storage system design and monitoring; and 4. Enabling governance and legal frameworks.
Figure 1. Conceptual illustration of the dynamic water-sharing model
Managing groundwater and storage systems
Groundwater extraction can trigger saltwater upconing, drawing saline water upward into shallow freshwater layers and wells, risking irrigation water quality and crop damage. Managing this risk is essential for subsurface storage and recovery systems.
Thijs Hendrikx, Utrecht University, Netherlands, has developed a three-dimensional, variable-density groundwater flow and salt-transport model (3D-VD-FT) to simulate these processes. The model was calibrated using field data from the Freshman-project, an initiative led by the drinking water company Dunea. Hendrikx’s simulations demonstrated how extraction influences the movement of the fresh-saltwater interface in the aquifer and revealed distinct patterns of downconing of freshwater and upconing of saline water, depending on pumping depth and rate. These findings allowed researchers to identify maximum safe pumping thresholds, reducing the risk of salinisation of the shallow freshwater aquifer (see Figure 2).
Figure 2. Normalised salt loads to the surface and area contribution percentage per spatial resolution in the Delfland region (Adapted from Farias Gutierrez et al. (2024).
Complementary research by Ignacio Farias, Utrecht University, examined how grid resolution affects predictions of saltwater upconing. By testing spatial resolutions from 10 to 250 metres in a regional groundwater model of the Delfland area, he found that fine resolution models better captured freshwater lens formation and localised salt intrusions, features that coarser grids missed (see Figure 2). This work highlights the importance of model precision in designing safe extraction strategies for shared subsurface water systems.
Ensuring water quality
Water quality is a central concern in circular horticulture. Greenhouses reuse irrigation water multiple times to conserve resources, but over time sodium ions (Na+) accumulate, hindering plant growth. To extend water reuse, harmful ions must be selectively removed while preserving essential nutrients.
Alaaeldin Elozeiri, Wageningen University, Netherlands, studied electrodialysis membranes coated with polyelectrolyte multilayers (PEMs) that regulate ion transport. The coatings selectively block divalent ions (Ca2+, Mg2+) while allowing monovalent ions (Na+, K+) to pass, enabling better control of toxic Na+ versus nutrient ions. Laboratory tests showed that transport is direction dependent. When ions move from the coated side (see Figure 3), divalent ions are hindered and accumulate near the interface, improving selective removal. However, this effect is weaker in the reverse direction. A 1D model explained this by slow ion movement and charge build-up within the PEM, which drives the observed selectivity. Overall, the results inform membrane designs that remove growth inhibiting ions (especially Na+) from irrigation water while retaining beneficial nutrients, supporting more efficient greenhouse water recycling.
Figure 3. Schematic of ion transport through PEM-coated electrodialysis membranes, illustrating direction dependent ion selectivity (Adapted from Elozeiri et al. (2025)
Optimising network efficiency and water storage
Implementing a water sharing system at a regional scale is inherently complex. It involves multiple actors, uncertain water availability, fluctuating demands and strong interactions with energy systems. As a result, decision-making must be supported by mathematical optimisation. Such decision support can be structured across three interconnected levels (see Figure 4): 1. Design or long-term planning level, where infrastructure such as storage, pipeline and treatment systems is sized and located; 2. Scheduling or short-term planning level, where daily or day-ahead decisions determine how infrastructure is operated; and 3. Real-time control level, where operations are continuously adjusted based on sensor data and forecasts. Researchers have developed optimisation algorithms tailored to each of these levels to support robust, efficient, and water-sharing decisions.
Figure 4. Multi-level optimisation framework for a regional water-sharing system.
At the design (long-term planning) level, Alireza Shefaei, Delft University of Technology, Netherlands, developed a stochastic multi-objective optimisation framework for hybrid water storage networks that combine centralised reservoirs, decentralised tanks and aquifer storage. The model accounts for uncertainty in rainfall, irrigation demand and energy costs, enabling planners to test storage configurations across different climate scenarios. Applied to a botanical garden in Amsterdam, Netherlands, it increased water system efficiency by 16% (risk-neutral) and 29% (risk-averse) during dry years. The framework can be extended to greenhouse regions such as Westland, in the Netherlands, to assess how shared storage and aquifers can buffer water availability and enhance regional climate resilience.
At the short-term scheduling level, Alessio Belmondo Bianchi, Wageningen University, developed day-ahead optimisation algorithms that model water pumps as flexible energy assets. These methods combine convex optimisation, learning-based approximations and electricity price forecasts to shift pumping to periods with low electricity prices or high renewable energy availability. Within a water-sharing system, this approach can be adapted to coordinate pumping across multiple users and shared storage facilities, enabling fair and cost-efficient water allocation while minimising collective energy costs and emissions.
At the real-time control level, Peter Verheijen, Eindhoven University of Technology, Netherlands, designed model predictive control (MPC) strategies for regional water networks. Using sensor data and short-term weather forecasts, the controller updates pump operations every 10-15 minutes, anticipating demand up to 24 hours ahead. This approach reduces overflows, limits the need for large storage buffers and lowers electricity consumption while ensuring a reliable water supply for greenhouses.
Water governance and legal frameworks
AquaConnect research shows that without appropriate institutional arrangements, even technically viable circular water solutions struggle to scale beyond pilots.
International comparative research on wastewater reuse governance demonstrates that successful reuse implementation does not depend on a single enabling factor, nor is there a universal governance model. Contrary to common assumptions, neither acute water scarcity nor strong financial incentives alone are necessary or sufficient conditions for success. Instead, implementation is shaped by governance pathways, specific combinations of regulatory clarity, social acceptability, perceived water stress and innovation capacity that work together in a given context (see Figure 5).
Figure 5. Summarises three governance pathways identified in successful reuse cases (Holstead et al., [7]. In the first, clear and stable regulation provides certainty for investment and operation. In the second, strong social acceptability compensates for limited financial incentives. In the third, water stress combines with social acceptance and innovation capacity, including technical expertise and local experimentation.
These findings are directly relevant to the Dutch greenhouse sector. The Netherlands has advanced technical knowledge and pilot experience in reuse and subsurface storage, but governance conditions for regional-scale water sharing remain fragmented. Research by Melchers (2024) shows that European wastewater directives provide only a harmonised baseline, leaving significant discretion to national and regional authorities. Compared with Spain and Malta, where standards, permitting procedures and allocation rules for treated effluent are clearly defined, the Dutch framework remains cautious and less explicit, particularly on ownership, liability and long-term rights to reused water.
Together, these findings resonate with AquaConnect’s findings that while technical solutions and pilots exist, scaling them requires building a general, hospitable institutional environment, including trust and legitimacy among society, growers, utilities and regulators. To create favourable conditions for dynamic water sharing, examples of small-scale practical circular water systems need to be demonstrated, supported by intensive monitoring of water quality parameters set by regulators and water specialists, and the direct involvement and supervision of permitting authorities, with a clear institutional model for scaling up.
The AquaConnect programme demonstrates how interdisciplinary research can transform water management in the Dutch horticultural sector. Through dynamic water sharing, greenhouses might collectively access alternative sources, such as rainwater, treated effluent and brackish groundwater, while reducing reliance on freshwater extraction.
AquaConnect research indicates that: 1. 3D groundwater modelling can safeguard against salinisation and define safe extraction limits; 2. Selective ion-removal membranes can maintain irrigation quality with lower energy use; 3. Stochastic optimisation can design efficient hybrid storage systems under uncertainty; 4. Predictive and flexible pump control can align water transport with renewable energy availability and; 5. Scaling circular water solutions depends as much on governance pathways, clear rules, social acceptance, and innovation capacity, as on technology, with monitored pilot-to-scale models helping to build trust and develop momentum towards institutional clarity.
Together, these advances provide the basis for a digital water grid – a connected network that stores, treats and distributes water across the horticultural landscape. However, technical feasibility will not translate into regional implementation unless institutional structures evolve in parallel. A coordinating institution to manage access, responsibilities and safeguards will need to be designed and tested. Dutch water governance was built for abundant rainfall and flood control, not for drought and competition over freshwater. Progress, therefore, requires not only new rules but also new relationships and coordination among growers, utilities, regulators and wider society. Building on AquaConnect, follow-on initiatives such as STURDI-Water (NWO) and Circular Water Governance (I4CS-EWUU) provide opportunities to develop and trial such arrangements at scale. l
Further reading:
Hendrikx, T.L., et al.; Optimising scavenger well strategies under parameter uncertainty to maximise allowable freshwater pumping rates in a coastal aquifer, The Netherlands. Journal of Hydrology (2025) https://doi.org/10.1016/j.jhydrol.2025.134554.
Farías, I. et al.; Effects of grid resolution on regional modelled groundwater salinity and salt fluxes to surface water. Journal of Hydrology (2024), 643, 131915.
Elozeiri, A. A., et al.; Current direction regulates ion transport across layer-by-layer one-side-coated ion-exchange membranes in electrodialysis. ACS Applied Materials & Interfaces (2025) 17(14), 22004-22013.
Shefaei, A., et al.; Optimising rainwater harvesting systems under uncertainty: A multi-objective stochastic approach with risk considerations. Resources, Conservation & Recycling Advances (2025) 26, 200254.
Bianchi, A. B., et al.; Neural network-informed optimal water flow problem: Modeling, algorithm, and benchmarking. Water Research X (2026), 100479.
Verheijen, P. C. N., et al.; Multi-resolution model predictive control with real-time demand forecasting for water distribution networks. Urban Water Journal (2025), 1-16.
Holstead, K., et al.; (submitted). Environmental Policy and Governance.
Melchers, S.; (2024). European approaches to wastewater reuse regulation: A comparison between Spain, Malta and the Netherlands. Review of European, Comparative and International Environmental Law (2024) 33(3), 350-366.
The authors:
Kirsty Holstead, Alessio Belmondo Bianchi, Thomas Wagner, and Professor Huub Rijnaarts undertake research at Wageningen University and Research, Arjen van Nieuwenhuijzen is Associate Professor at Wageningen University and Research, R&D and innovation project director circular and climate positive solutions at Witteveen+Bos and industry principal investigator at AMS Institute, and Tiza Spit is team lead drinking water technology, Witteveen+Bos.
