Abstract
The food–energy–water nexus is usually described as a set of interdependencies to be managed. This report takes a narrower and more operational view: disruptions propagate along specific couplings, in a direction that differs case by case, and the clean transition's real effect is to substitute one coupling for another rather than to remove coupling altogether. Getting that substitution right — choosing dependencies that are local, abundant and storable over ones that are imported, seasonal or priced in a foreign currency — is the imperative the title refers to.
Two recent disruptions show the mechanism working in opposite directions. In 2022 the chain ran from energy into food: natural gas reached as much as 90 per cent of the variable cost of European ammonia production, roughly 70 per cent of that capacity was shut or curtailed at the peak, and the FAO Food Price Index hit its highest level in the series' history in March 2022 at 160.2 points, with cereals at 170.1. In 2023 and 2024 the chain ran from water outward: drought cut Gatún Lake to its lowest level since at least 1965, the Panama Canal reduced daily transits from 36–38 in mid-2023 to about 18 by February 2024, and total FY2024 transits fell 29 per cent to 9,936 vessels. Water shortages also took power systems down directly — Ecuador, which draws around 77 per cent of its electricity from hydropower, imposed outages that grew from eight to fourteen hours a day during its worst drought in six decades.
Against that background the report examines four substitutions the transition is already making in the Global South. Diesel and grid-powered irrigation gives way to solar pumping, which removes the fuel-price and grid-reliability coupling and replaces it with a groundwater coupling made more dangerous by the disappearance of the marginal cost that used to ration extraction. Gas-based ammonia gives way to green ammonia, moving fertiliser from an imported-gas dependency to a domestic electricity one: India has allocated about 724,000 tonnes a year of green ammonia across 13 fertiliser units, at discovered prices well below the global benchmark. Rainfall-dependent hydropower is complemented by solar, wind and storage, trading drought exposure for land, minerals and capital exposure. And rainfall-dependent water supply gives way to desalination, which trades rainfall risk for an electricity dependency of roughly 2.5 to 4 kilowatt-hours per cubic metre — benign only where the power is clean and firm.
The report's practical conclusion is that each new coupling carries a governance requirement that the old one discharged automatically, usually through price. Free solar pumping needs metering and water allocation because the tariff no longer rations the aquifer; green ammonia needs firm clean power because an electrolyser idle half the year does not repay its capital; desalination needs the same. Substitutions that move a dependency onto a local, abundant resource are worth making even when they are not cheaper on day one, because they convert a foreign-exchange and geopolitical exposure into a domestic engineering problem. The report closes with recommendations for agriculture, energy and water ministries, for development finance, and for the utilities and regulators who will have to make the new couplings hold.
Executive Summary
Nexus disruptions do not strike three sectors at once. They enter through one and travel along whichever link is tightest — and the clean transition changes which links exist.
Shocks travel in both directions
In 2022 the chain ran from energy into food: gas prices closed European ammonia plants and fertiliser scarcity reached farmers everywhere. In 2023 and 2024 it ran outward from water: drought cut hydropower in Ecuador and shipping through Panama.
The transition substitutes couplings
Solar pumping, green ammonia, renewables with storage and desalination each break one dependency and create another. Nothing here is decoupled. The question is which dependency a country would rather hold.
Each new coupling needs new governance
The old couplings rationed by price. Free solar pumping does not ration an aquifer; an electrolyser idle half the year does not repay its capital. What price used to do, metering, allocation and firmness must now do deliberately.
The food–energy–water nexus is usually presented as a diagram of arrows: everything depends on everything. That is true and not very useful. What a planning ministry needs to know is narrower — which link transmitted the last shock, which link will transmit the next one, and what a given investment does to the set of links that exist at all.
This report answers those questions in three moves. It reconstructs two recent disruptions to show the mechanism running in opposite directions. It then argues that clean technologies are best understood as coupling substitutions: they exchange an imported, seasonal or foreign-currency dependency for a local, abundant one. And it sets out what each new coupling demands in return — because the substitution is only an improvement if the governance arrives with it.
Sections 1 to 3 establish how shocks propagate, with the 2022 energy-into-food chain and the 2023–24 water-outward chains. Section 4 states the substitution principle and Section 5 works through four instances of it. Section 6 sets out what each new coupling requires, and Sections 7 and 8 give recommendations and conclusions.
1. How Nexus Shocks Travel
A nexus disruption has an entry point, a transmission link and a terminus. Naming all three is what distinguishes an analysis from a diagram.
Food, energy and water are coupled through a small number of physical and commercial links, and only some of them carry a given shock. Nitrogen fertiliser couples food to natural gas, because ammonia synthesis uses gas as both feedstock and fuel. Irrigation couples food to electricity or diesel, because water at depth has to be lifted. Hydropower couples electricity to rainfall. Thermal generation couples electricity to cooling water. Inland waterways couple trade — including trade in food, fuel and fertiliser — to river and lake levels.
Each of those links has a characteristic direction and lag. A gas price shock reaches food prices in months, through fertiliser cost and planting decisions. A drought reaches electricity in weeks, through reservoir levels, and reaches food in a season. A canal closure reaches prices in days, through freight rates. The practical consequence is that a country's exposure is not "the nexus" in general but a specific list of links that are tight for it — and that list is what an investment programme can change.
Table 1 — Four recent disruptions, by link
| Disruption | Entered through | Travelled along | Arrived as |
|---|---|---|---|
| 2022 gas price shock | Energy | Ammonia synthesis — gas as feedstock and fuel | Fertiliser scarcity and record food prices |
| 2023–24 Panama drought | Water | Lock operation on Gatún Lake; freight capacity | Fewer transits, longer voyages, higher freight cost |
| 2024 Ecuador drought | Water | Hydropower reservoirs, 77 per cent of generation | Fourteen-hour daily blackouts and lost output |
| 2024 Zambezi drought | Water | Kariba's live storage; 86 per cent hydro capacity | First-ever halt at Kariba North Bank; 20-hour outages |
Compiled by the author from the sources cited in Sections 2 and 3. The Zambezi case is examined in a companion report and is included here for the pattern.
2. Energy into Food: the 2022 Chain
The clearest recent demonstration that food security is partly an energy-policy variable, and that the link is a chemical process rather than a metaphor.
Ammonia synthesis takes nitrogen from the air and hydrogen from natural gas. In Europe in the summer of 2022, gas reached as much as 90 per cent of the variable cost of ammonia production (European Commission). At those prices the plants were worth more shut than running: about 70 per cent of European ammonia capacity was closed or curtailed at the peak.
Fertiliser is traded globally, so the shortage did not stay in Europe. It arrived in importing countries as price, and in the poorest as quantity — less nitrogen applied, on the crop that followed. The FAO Food Price Index, which had been climbing since mid-2020 on pandemic-era supply disruption, reached 160.2 points in March 2022, the highest monthly value in a series that starts in 1990; the cereals sub-index hit 170.1 in the same month.
Figure 1 — The FAO Food Price Index, 2019–2026
Monthly index, 2014–2016 = 100, from FAO's published data file. The March 2022 peak of 160.2 is the current published value; FAO announced 159.3 at the time and has since revised the series. Prices fell back over 2023 and 2024 without returning to their pre-2021 level.
Figure 2 — Why the plants stopped
Per cent. Natural gas as a share of the variable cost of EU ammonia production in summer 2022, and the share of European ammonia capacity shut or curtailed at the peak of the price spike. Sources as cited in the text.
Two features of this episode matter for what follows. The first is that the transmission was industrial: no crop failed in Europe, and the mechanism was a chemical plant's marginal cost. The second is that the exposure was concentrated in a single input price denominated in a foreign currency — which is exactly the kind of dependency a substitution can address, and the subject of Section 5.2.
Where the burden landed
The State of Food Security and Nutrition in the World 2025 estimates that 673 million people faced hunger in 2024, 8.2 per cent of the world's population — down from 8.5 per cent in 2023 and 8.7 per cent in 2022, but still rising across most of Africa and western Asia (FAO). Aggregate improvement and regional deterioration are not in tension: the countries with the least fiscal room to absorb an input-price shock are the ones where it converts into hunger.
3. Water Outward: Power and Trade
Water shortage does not wait for agriculture to transmit it. It reaches electricity through reservoirs and global trade through lock chambers.
3.1 Drought into electricity
Ecuador draws roughly 77 per cent of its electricity from hydropower. In 2024, during what the government called the worst drought in 61 years, reservoir levels fell far enough to force rationing: scheduled outages that began at around eight hours a day were extended to fourteen hours a day from late October, across Quito, Guayaquil and Cuenca. Southern Africa's Zambezi drought produced the same failure in the same year, with Zambia — about 86 per cent hydro by installed capacity — halting generation at Kariba North Bank for the first time in its history.
Figure 3 — Ecuador's rationing escalates, 2024
Scheduled power cuts in hours per day at three points during the 2024 drought, from contemporary reporting. Ecuador draws about 77 per cent of its electricity from hydropower; Zambia, in the same drought year, is about 86 per cent hydro by installed capacity and reached outages of up to twenty hours a day.
This is the nexus at its most direct: no agricultural channel, no price mechanism, just less water behind a dam and therefore less electricity. And the loss is not confined to households. Ecuador's rationing was estimated to cost around US$12 million for every hour of outage — a figure that lands on exactly the industrial users discussed in an earlier report as reliability-bound.
3.2 Drought into trade
The Panama Canal is a freshwater machine: each transit consumes water from Gatún Lake, which is filled by rainfall. In 2023 and 2024 the lake fell to its lowest level since at least 1965, and the Canal Authority rationed the scarce input the only way it could — by rationing transits. Daily crossings fell from 36 to 38 vessels in mid-2023 to about 18 by February 2024, and total transits in FY2024 came to 9,936 vessels, 29 per cent below the previous year. Queues of waiting ships ran to about 135 vessels at the peak, and full capacity was not restored until August 2024.
Figure 4 — Panama Canal transits, rationed by rainfall
Vessels per day at three points in the drought, with the FY2024 annual total shown separately in the text. From Panama Canal Authority data as reported by the US Energy Information Administration and contemporary analysis.
What makes Panama a nexus case rather than a shipping story is the cargo. The vessels that could not transit were carrying food, fuel, fertiliser and equipment, including a substantial volume of liquefied petroleum gas and liquefied natural gas that rerouted or paid for slots. A rainfall deficit in one watershed became a freight-cost increase for importers on two oceans — the nexus operating at a distance of several thousand kilometres.
4. What the Clean Transition Actually Changes
Not the existence of coupling. Its address.
The rhetorical version of the transition promises independence: energy that is not imported, water that does not depend on the rains, food that does not depend on someone else's gas. The engineering version is more modest and more useful. Every technology in this report replaces one dependency with another:
- Solar irrigation pumping removes a diesel-price and grid-reliability dependency, and installs a dependency on the aquifer.
- Green ammonia removes an imported-gas dependency, and installs a dependency on domestic electricity and capital.
- Solar, wind and storage remove a rainfall dependency from the power system, and install dependencies on land, minerals and finance.
- Desalination removes a rainfall dependency from water supply, and installs an electricity dependency of a few kilowatt-hours per cubic metre.
Stated that way, the imperative is not "decouple" — which is not on offer — but choose the coupling deliberately. Three properties make one dependency preferable to another. Is the resource local, so that a shock in another jurisdiction does not transmit? Is it abundant, so that scarcity is an engineering constraint rather than an allocation fight? Is it storable or buffered, so that a bad week is not a crisis? Sunlight scores well on the first two and poorly on the third without storage; an aquifer scores well on the first and badly on the second once extraction exceeds recharge.
Figure 5 — The coupling substitution map
A conceptual schematic of the argument in Sections 4 and 5, not a quantitative model. The right-hand column is the claim that matters: none of these rows ends in independence.
5. Four Substitutions, Examined
What each one buys, what it costs, and what it newly requires.
5.1 Irrigation: diesel and grid to solar
Lifting water is the largest single energy demand in agriculture, and in many countries it is subsidised into invisibility. In India, groundwater irrigation on subsidised electricity accounts for a very large share of agricultural power use — in some states 40 to 60 per cent of total electricity consumption — and flat or zero tariffs have reduced the marginal cost of pumping close to nothing, with well-documented consequences for groundwater tables.
Solar pumping removes the fuel bill and the outage, which is why it spreads quickly wherever it is offered. But it should be understood precisely: it does not fix the groundwater problem, it removes the last remaining brake on it. A diesel pump has a marginal cost that rations extraction crudely but genuinely; a solar pump's marginal cost is zero, every day, forever. The substitution is worth making — and it is only safe alongside metering, allocation rules or buy-back arrangements that pay farmers for surplus generation instead of surplus water.
5.2 Fertiliser: imported gas to domestic electricity
Green ammonia is the direct answer to the 2022 chain. Electrolytic hydrogen made with domestic renewable electricity, combined with nitrogen from the air, produces the same ammonia without the gas import. The relevance to the Global South is not theoretical: under India's National Green Hydrogen Mission, about 724,000 tonnes a year of green ammonia has been allocated across 13 fertiliser units, with discovered prices in the range of roughly ₹50 to ₹65 per kilogram against a global benchmark around ₹110, and the government projecting foreign-exchange savings in the billions of dollars over the coming decade.
Figure 6 — India's green ammonia price discovery against the global benchmark
Indian rupees per kilogram. The discovered range is from the tender allocating supply across 13 fertiliser units; the benchmark is the approximate global reference cited alongside it. Prices are point-in-time tender outcomes rather than a market series.
The substitution here is unusually favourable, because the dependency being removed is an imported commodity priced in dollars and the one being installed is domestic electricity. But it is not free: electrolysers are capital-intensive, and capital-intensive plant needs high utilisation, which needs firm power. A green ammonia plant on an intermittent supply is an expensive way to make ammonia for part of the year.
5.3 Power: rainfall to solar, wind and storage
Sections 3.1 showed what a hydro-dominated system does in a drought year. The substitution is not to abandon hydropower — which remains valuable, and whose reservoirs are also water infrastructure — but to stop treating it as firm capacity when the climate that fills it is changing. Solar and wind diversify against rainfall because their resource is uncorrelated with it; storage converts them into something dispatchable; and hydropower, where it survives, becomes the balancing asset rather than the base.
What this installs in place of drought risk is land use, mineral supply chains and, above all, capital cost — a dependency examined in earlier reports in this series and not repeated here.
5.4 Water supply: rainfall to electricity
Desalination is the starkest substitution in the set, because it converts a water problem into an energy problem by design. Global installed capacity passed 100 million cubic metres a day in 2024, roughly fifteen times the year-2000 level. The energy intensity of seawater reverse osmosis has fallen by about an order of magnitude since 1970, from 20–30 kilowatt-hours per cubic metre to roughly 2.5 to 4 kilowatt-hours today.
Figure 7 — The energy intensity of seawater desalination
Kilowatt-hours per cubic metre of product water for seawater reverse osmosis, shown as published ranges for each period. The tenfold improvement is what makes desalination a plausible substitution at all; it remains an electricity dependency of a few kilowatt-hours per cubic metre.
That improvement is what makes the substitution defensible. It is still a trade: a city that desalinates has converted a rainfall risk into a power-system risk, and has done so for a service with no tolerance for interruption. On a clean, firm grid this is a good exchange. On a grid that is neither, it moves the failure from the reservoir to the substation.
Table 2 — The four substitutions
| Substitution | Dependency removed | Dependency installed | Is the new one local, abundant, storable? | Governance it requires |
|---|---|---|---|---|
| 01 Solar irrigation | Diesel price; grid reliability | Groundwater | Local yes; abundant no once extraction exceeds recharge; storable only as aquifer | Metering, allocation, surplus-power buy-back instead of surplus water |
| 02 Green ammonia | Imported gas, priced in dollars | Domestic electricity and capital | Local yes; abundant yes; storable yes, as ammonia itself | Firm clean supply to keep utilisation high; offtake contracts |
| 03 Solar, wind, storage | Rainfall, through reservoir levels | Land, minerals, capital | Local partly; abundant yes for the resource; storable with batteries | System planning that stops counting drought-exposed hydro as firm |
| 04 Desalination | Rainfall, through supply reservoirs | Electricity, 2.5–4 kWh per m³ | Local yes; abundant yes if clean generation is built; storable as treated water | Firmness standards; brine management; tariff design |
The framework and its assessments are the author's, applied to the evidence cited in this section. The middle columns are the analytical claim; the right-hand column is what Section 6 develops.
6. What the New Couplings Require
Each substitution transfers a job that price used to do onto an institution that has to do it deliberately. That transfer is where these programmes fail.
6.1 When the price signal disappears, something must replace it
This is the single most important line in the report. A diesel pump rations groundwater because diesel costs money. Remove the fuel cost and the rationing goes with it — not as a side-effect to be managed later, but immediately, on the day the panel is installed. The same logic applies wherever a clean technology has near-zero marginal cost: the resource it consumes is now free at the point of use, and whatever restraint the price provided has to be reconstructed as measurement, entitlement and enforcement.
The instruments exist and are well tested: volumetric metering, tradable or allocated abstraction rights, and — the most politically durable — buying farmers' surplus electricity so that the marginal hour of sunshine has a use other than pumping. Programmes that install pumps without one of these are converting a fuel subsidy into a water subsidy.
6.2 Capital-intensive substitutions need firmness, not just cleanliness
Electrolysers and desalination plants are mostly capital. Their unit cost is a function of utilisation, which makes firm supply an economic requirement rather than an engineering preference. A country that builds green ammonia capacity on an intermittent grid will discover that it has bought expensive ammonia; one that builds desalination on an unreliable grid has moved its water risk rather than reduced it.
6.3 Diversify against the hazard, not the fuel
A power system that is 77 or 86 per cent hydropower is low-carbon and singly exposed. The diversification that matters is against the hazard — in this case rainfall — which means the relevant question in system planning is not the fuel mix but the correlation of outages. Two clean technologies that fail in the same weather are one technology for planning purposes.
6.4 Buffers are cheaper than crises
Every disruption in Sections 2 and 3 was survivable with a buffer: fertiliser stocks, reservoir carryover rules, battery storage, alternative routing. Buffers lose in budget negotiations because they earn nothing in the years nothing happens — the asymmetry an earlier report examines in detail. In nexus systems the case for them is stronger than usual, because a buffer in one sector protects three.
7. Recommendations
By actor, in the order the decisions are actually taken.
7.1 Agriculture and water ministries
- Pair every solar pumping programme with volumetric metering and an allocation rule. The subsidy has moved from fuel to water; price it accordingly, or cap it.
- Buy the surplus electricity, not the surplus water. Feed-in or buy-back arrangements give the marginal sunny hour a use that does not deplete the aquifer, and give farmers an income that does not depend on pumping.
- Hold fertiliser and seed buffers sized against an input-price shock, not against a harvest failure; 2022 showed the former arriving first.
7.2 Energy ministries, utilities and regulators
- Stop counting drought-exposed hydropower as firm capacity. Adequacy assessments should use dry-year hydrology, and reservoir carryover rules should be set before the drought, not during it.
- Treat water-dependent generation and generation-dependent water as one planning problem, including desalination load, irrigation pumping and thermal cooling.
- Provide firmness to the capital-intensive new loads — electrolysers, desalination — through storage, contracts and reserve margin, because their economics depend on utilisation.
7.3 Development finance
- Finance the governance alongside the hardware. Metering, abstraction registries and hydrological monitoring are small line items that determine whether the hardware is a substitution or a depletion programme.
- Support domestic green ammonia where the power resource is strong, on the foreign-exchange and food-security case as much as the emissions one.
- Fund buffers explicitly, since no individual ministry's budget rewards them and every sector benefits from them.
7.4 Analysts and statistical agencies
- Publish the links, not just the levels. A country's nexus exposure is a short list of tight couplings — fertiliser import share, hydro share of generation, irrigation's share of electricity, food import dependence — and those indicators belong together in one place.
- Track the substitutions. Installed solar pumps against metered abstraction; green ammonia capacity against utilisation; desalination capacity against the firmness of the power that serves it.
8. Conclusions
The nexus is not a diagram of everything affecting everything. It is a short list of links, and the transition edits the list.
In 2022 a gas price reached the world's food supply through a chemical plant's marginal cost: gas at as much as 90 per cent of variable cost, roughly 70 per cent of European ammonia capacity shut, and the FAO Food Price Index at its highest ever monthly reading. In 2023 and 2024 rainfall reached global freight through a lake: Panama's transits cut from 38 a day to about 18, and 29 per cent fewer vessels across the fiscal year. In the same period, drought reached households through turbines: fourteen-hour blackouts in Ecuador, a first-ever shutdown at Kariba in Zambia. Different entry points, different links, the same structure.
What the clean transition does to that structure is more precise than the rhetoric around it suggests. It does not deliver independence. It substitutes couplings: gas for electricity in fertiliser, diesel and grid for aquifer in irrigation, rainfall for land and capital in power, rainfall for kilowatt-hours in water supply. Judged by the three tests that matter — local, abundant, storable — most of these exchanges are clearly worth making, and the fertiliser one is worth making on foreign-exchange grounds alone.
But each new coupling comes with a requirement that the old one discharged automatically. Price rationed the aquifer; metering and allocation now have to. Gas plants ran whenever they were needed; electrolysers need firm power to repay their capital. Reservoirs buffered a dry month; batteries and carryover rules now have to. The failures to expect in the next decade are not failures of the technology, which mostly works. They are failures to install the institution that the new dependency assumed was already there.
References
Every quantitative claim above is attributed inline. The principal sources are collected here.
- FAOFAO Food Price Index — the monthly series plotted in Figure 1, 2014–2016 = 100, including the March 2022 peak of 160.2 in the current published data. FAO's contemporaneous March 2022 release reported 159.3 before revision.
- European CommissionEnsuring availability and affordability of fertilisers (2022) — gas at up to 90 per cent of EU ammonia variable cost and the scale of curtailment. See also Fertilizers Europe and industry reporting on the 70 per cent figure.
- FAO, IFAD, UNICEF, WFP, WHOThe State of Food Security and Nutrition in the World 2025 — 673 million people, 8.2 per cent of the global population, facing hunger in 2024.
- US Energy Information AdministrationPanama Canal traffic to increase as drought conditions ease — transit restrictions and their effect on energy cargoes. Lake levels and transit counts also in PIIE and Woodwell Climate; attribution of the 2023 drought in World Weather Attribution.
- Contemporary reporting on EcuadorCBS News and The Watchers — hydropower share, the worst drought in 61 years, and outages extended from eight to fourteen hours a day.
- Research on India's groundwater–electricity linkWater Economics and Policy on subsidised electricity and groundwater depletion; see also Nature Communications on reconciling groundwater depletion with food security.
- Government of IndiaDecarbonising India's fertiliser sector and the National Green Hydrogen Mission — green ammonia allocation across 13 fertiliser units and the mission's 5 million tonne hydrogen target. Price discovery reported by IEEFA.
- Desalination dataGlobal energy, costs and emissions from reverse osmosis desalination under future water scarcity (Water Research) and industry summaries on installed capacity passing 100 million m³/day and modern SWRO energy intensity of 2.5–4 kWh/m³.
- H HeuristicsRelated reports in this series: integrating adaptation and mitigation, on protection as an electrical load and clean supply as climate-exposed; climate-resilient industrialization pathways, on reliability-bound industry; and the economics of prevention, on why buffers lose budget arguments.
Metadata
- Keywords
- food energy water nexusfertiliser crisisfood priceshydropower droughtPanama Canaldesalinationsolar irrigationgreen ammoniagroundwater governanceclean transitionGlobal Southsupply chain disruptionenergy securitywater security
- JEL classification
- Q25, Q18, Q42, Q54, O13 — water; agricultural policy and food; alternative energy sources; climate and natural disasters; agriculture and natural resources in development
- Data and method
- This report synthesises published institutional data and contemporary reporting rather than producing new modelling. The FAO Food Price Index series is taken directly from FAO's published monthly data file (2014–2016 = 100), which carries revisions relative to the figures announced at the time; the report uses the current published values and says so. European ammonia curtailment and the gas share of variable cost are from industry and European Commission sources for 2022; Panama Canal transit data from the Panama Canal Authority as reported by the US Energy Information Administration and analysts; Ecuador's blackout schedule and hydropower share from contemporary reporting; desalination capacity and energy intensity from recent industry and peer-reviewed summaries; India's green ammonia allocation and discovered prices from Indian government and trade sources; hunger figures from The State of Food Security and Nutrition in the World 2025. All figures were verified against their sources in September 2026. The four-substitution framework in Sections 4 and 5, Table 1 and the governance requirements in Section 6 are the author's analysis, assembled with the cited cases already known, and are offered as a planning checklist rather than a predictive model. Figure 5 is a conceptual schematic. Figures from different sources use different units and periods and are not additive. The report is analytical rather than predictive.
- Report
- H Heuristics Digital Report № 2026-14 · Published 17 September 2026
- Licence
- CC BY-NC-ND 4.0
- Cite as
- Hunter Hughes (2026). Food, Energy, Water: Disruption and Substitution: Nexus shocks in the Global South and what the clean transition actually changes. H Heuristics Digital Report 2026-14. https://digitalreports.hheuristics.com/reports/food-energy-water-nexus-clean-transition/