Water-Smart Vertical Farming for Arid Regions – InCrop White Paper

Executive summary

Agriculture accounts for about 70% of global freshwater withdrawals, and FAO expects substantial water scarcity to persist in the Near East, North Africa and South Asia.²
A peer-reviewed comparison in the Arizona desert measured about 20 L of water per kg of hydroponic greenhouse lettuce versus about 250 L/kg in the field — roughly 92% less.¹
Fully sealed vertical farms can go further, because the water plants transpire stays inside the building and can be condensed and reused.
The same study found the hydroponic system used about 82 times more energy per kg.¹ In arid regions, saving water by spending energy is only sensible when that energy is affordable and, ideally, low-carbon.
InCrop combines a closed water loop with AI that models the farm’s water balance, optimises dosing and flags losses in minutes. We propose a pilot-first approach that measures litres and kWh per kg on site before any scale-up.

1. Water stress in arid regions

Fresh vegetables are among the most water-intensive foods to grow and the most perishable to import. Arid countries therefore face a double exposure: limited renewable fresh water to grow them locally, and long, fragile supply chains to buy them from elsewhere.

FAO’s 2017 report to the G20 identifies agriculture as the largest water user worldwide, at about 70% of total freshwater withdrawals on average. The same report expects substantial water scarcity to persist in the Near East, North Africa, South Asia and other areas, and notes that groundwater withdrawals already exceed natural recharge in parts of the Near East and North Africa.²

Established

Agriculture is the largest user of fresh water globally (~70% of withdrawals), and in parts of the Near East and North Africa groundwater is already being withdrawn faster than it is naturally recharged.²

Where fresh water comes from desalination, every litre also carries an energy cost. Producing more food per litre therefore matters twice: it conserves a scarce resource and the energy needed to make it.

2. Where field irrigation loses water

In open-field irrigation, only part of the water applied ends up in the harvested crop or is transpired by it. The rest is lost through mechanisms that intensify in hot, dry, windy conditions:

Evaporation from the soil surface and from sprinkler spray, highest under desert heat and wind.
Deep percolation below the root zone, especially on sandy soils common in arid regions.
Run-off from over-irrigation or poor scheduling, which also carries nutrients away.
Transpiration to open air: even the water the plant uses productively is released to the atmosphere and cannot be recovered.

Drip irrigation and good scheduling reduce the first three. Only an enclosed system can address the fourth.

3. Why closed systems use less water

Hydroponic systems deliver nutrient solution directly to roots and recirculate what is not taken up, removing most percolation and run-off. A sealed indoor farm adds a further step: the humidity created by plant transpiration must be removed by the air-conditioning system to keep the crop healthy, and that dehumidification produces liquid condensate — clean water that can be filtered and returned to the nutrient tank.

Source: Barbosa et al., 2015.¹ Hydroponic values are for a greenhouse, not a fully closed vertical farm.
Measure (lettuce, Yuma, Arizona)	Hydroponic greenhouse	Conventional field
Water use	20 ± 3.8 L/kg/yr	250 ± 25 L/kg/yr
Yield	41 ± 6.1 kg/m²/yr	3.9 ± 0.21 kg/m²/yr
Energy use	90,000 ± 11,000 kJ/kg/yr	1,100 ± 75 kJ/kg/yr

Industry claim

AeroFarms, a US vertical-farming company, has stated that its aeroponic farms use 95% less water than field farming and 40% less than hydroponics.⁴ We treat company figures like this as an upper bound to be verified, not a planning figure.

A comparative study of plant factories and greenhouses across the Netherlands, northern Sweden and the United Arab Emirates concluded that closed plant factories are most defensible where water or land is scarce, while greenhouses remain more energy-efficient with current technology.³ Arid regions sit squarely in that first category.

4. The InCrop water architecture

Figure 1. Water flow through an arid-region InCrop farm. Only water leaving in the harvested crop and in managed reject streams exits the loop.

The architecture has five functional blocks:

Source and polishing. Municipal, desalinated or treated grey water is polished by reverse osmosis (RO) to a consistent baseline so nutrient recipes are predictable. RO reject is a real waste stream and is designed for — blended, reused for non-crop purposes or discharged under permit — rather than ignored.
Nutrient reservoir and dosing. Automated dosing holds electrical conductivity (EC) and pH within crop-specific bands.
Recirculating grow system. NFT, deep-water culture or aeroponics, chosen per crop; solution not taken up by roots returns to the reservoir.
Condensate recovery. Water removed from the air by the dehumidifying cooling coils is collected rather than drained.
Treatment. Filtration and UV disinfection before recovered water rejoins the loop, limiting pathogen spread through shared solution.

5. What AI adds

A closed loop saves water by design; AI keeps it saving water in operation, when sensors drift, emitters clog, pipes leak and crops change.

AI function	Inputs	Output	Maturity
Water-balance anomaly detection	Intake meter, condensate meter, reservoir level, harvest weight	Alert when actual losses diverge from the model	Established technique
Dosing optimisation	EC, pH, dissolved oxygen, crop stage, growth images	Adjusted set-points and dosing schedule	Established / improving
RO membrane health prediction	Pressure, flow, permeate conductivity, temperature	Cleaning or replacement forecast	Established technique
Pathogen-risk early warning	Solution temperature, dissolved oxygen, root-zone images	Risk score and recommended action	Emerging

“Established technique” here means the underlying method (statistical process control, regression, anomaly detection) is well proven in industry; performance on a specific farm still has to be validated on that farm’s data.

6. Water balance and KPIs

Every InCrop farm is instrumented so that its water balance closes. Over a reporting period:

W_intake = W_crop + W_reject + W_unaccounted + ΔS_storage Water use (L/kg) = W_intake / saleable harvest (kg) Condensate recovery ratio = W_condensate_returned / W_transpired_estimate Unaccounted loss = W_unaccounted / W_intake (target: as close to 0 as measurement allows)

The AI water-balance model predicts expected transpiration from climate and crop data; a persistent gap between expected and recovered condensate, or a rise in unaccounted loss, triggers investigation.

Design target (to be validated)

For leafy greens in a sealed arid-region farm, we target water use well below the 20 L/kg measured for a hydroponic greenhouse in the Barbosa study, by recovering condensate. The actual figure will be set in feasibility and verified by metering during the pilot.

7. The energy trade-off

Water savings are not free. In the Barbosa comparison, the hydroponic greenhouse used about 82 times more energy per kg than field production, mainly for climate control and supplementary lighting.¹ Fully enclosed farms rely on artificial light for all photosynthesis and must reject that heat in a hot climate, so their energy demand is typically higher still.³

For arid regions this means three things:

The business case depends on electricity price and carbon intensity as much as on water price.
On-site or contracted solar generation is a natural fit, because peak solar output coincides with peak cooling demand.
Every design decision should be evaluated on both litres per kg and kWh per kg. White paper 02 covers the energy side in detail.

8. Implementation roadmap

Feasibility (desk + site). Water sources and quality, tariffs, grid and solar potential, crop and market fit, building options; a combined water–energy model.
Container or small-room pilot. Full metering of intake, condensate, reject and harvest; AI models trained and validated on pilot data; at least two full crop cycles.
Scale decision. Go/no-go based on measured L/kg, kWh/kg, yield and unit economics, not on projections.
Production farm. Build, commission and operate, with monthly KPI reporting and continuous model retraining.

9. Risks and limitations

Energy cost and carbon. The largest risk to viability; see section 7.
Pathogens in recirculating water. Shared solution can spread root diseases quickly; treatment, monitoring and hygiene protocols are essential.
Crop range. Economics favour leafy greens, herbs, microgreens and seedlings; staple crops are not viable indoors today.
Brine and reject management. Desalination and RO create concentrated waste streams that must be handled responsibly.
Skills. Operating a closed farm requires horticultural, mechanical and data skills; training is part of every engagement.
Model validity. AI models trained elsewhere may not transfer; all models are validated on site before they act.

10. Conclusion

For arid regions, closed-loop vertical farming offers a credible way to produce fresh greens with a small fraction of the water used in the field. The evidence for the water saving is strong; the evidence on energy is equally clear and must be planned for. InCrop’s approach is to make both visible from the first day of a pilot, use AI to keep the loop tight in operation, and scale only on measured results.

References

Barbosa G.L., Gadelha F.D.A., Kublik N., Proctor A., Reichelm L., Weissinger E., Wohlleb G.M., Halden R.U. (2015). Comparison of Land, Water, and Energy Requirements of Lettuce Grown Using Hydroponic vs. Conventional Agricultural Methods. Int. J. Environ. Res. Public Health 12(6), 6879–6891. https://www.mdpi.com/1660-4601/12/6/6879
FAO (2017). Water for Sustainable Food and Agriculture: A report produced for the G20 Presidency of Germany. Food and Agriculture Organization of the United Nations, Rome. https://www.fao.org/3/i7959e/i7959e.pdf
Graamans L., Baeza E., van den Dobbelsteen A., Tsafaras I., Stanghellini C. (2018). Plant factories versus greenhouses: Comparison of resource use efficiency. Agricultural Systems 160, 31–43. https://doi.org/10.1016/j.agsy.2017.11.003
WebUrbanist (June 2016). World’s Biggest Indoor Vertical Farm Near NYC to Use 95% Less Water (reporting AeroFarms’ own figures). https://weburbanist.com/2016/06/05/worlds-biggest-indoor-vertical-farm-near-nyc-to-use-95-less-water/

This white paper is provided for information. Figures labelled “Established” or “Peer-reviewed” are taken from the cited sources; figures labelled “Industry claim” have not been independently verified by LIDAVI; anything labelled “Design target” or “Hypothesis” is LIDAVI’s own and must be confirmed in a site-specific feasibility study. Nothing here is a performance guarantee.