Where the power line ends: can a field keep a village working?
Journey to Resilience, Part 2: Agrivoltaics
Solstrides Projects · Büşra Yılmaz-Depondt
In Morogoro, in eastern Tanzania, the Farmer Training Centre run by Sustainable Agriculture Tanzania used to get its electricity from diesel generators. Today it runs on 106 solar modules mounted on a steel frame three metres above the vegetable beds, with gutters along their lower edges that carry rainwater into a 10,000-litre tank. Underneath, the farm's technicians grow onions, sweet peppers, Swiss chard, beans, maize, eggplant and kale, and next to the array is an open field of the same size where the same crops grow in full sun, so that the two can be compared.
The system is not connected to any grid. It has its own bank of lead-acid batteries, and between June 2022 and May 2023 it produced 12.55 MWh, measured by researchers from the University of Sheffield, World Agroforestry and the host farms alongside every harvest. The centre uses that power for charging points, lights for evening classes and fridges. Over the three growing seasons of the trial, the beds under the panels needed 12.6% less irrigation water than the open field, and the Swiss chard and beans grown there gave significantly bigger harvests.
I chose Morogoro to open this article because of its location. Rural sites like this one sit at the end of the power line, or beyond it, and the land that could carry solar panels is usually the land people grow their food on. Agrivoltaics tries to let the same land do both jobs.
Last in line
Part 1 ended with a question for building owners: which loads must keep running when the grid goes down? In much of rural Africa and South Asia that question comes earlier, because the grid was never reliable in the first place. A 2023 report by the WHO, the World Bank, IRENA and Sustainable Energy for All estimated that close to a billion people in low- and lower-middle-income countries are served by health facilities with unreliable electricity or none at all. Based on facility surveys, more than one in ten health facilities in South Asia and sub-Saharan Africa had no electricity whatsoever, power was unreliable for half of those in sub-Saharan Africa, and rural clinics and primary health centres were much less likely to have power than urban hospitals. The report concluded that decentralised solar is often the most practical way to close that gap, because it does not have to wait for the grid to arrive.
For resilience, I find this more important than the land-use argument usually made for agrivoltaics. In a village the loads that must keep running are few and well known, such as a vaccine fridge, a light for a birth at night, a water pump and a cold room that keeps the harvest from spoiling before it reaches the market. They sit in farming communities, so the obvious place for the panels that would run them is farmland, the land people are least willing to give up.
Two harvests from one field
Agrivoltaics means growing crops, or grazing animals, on the same land that produces solar electricity, with the system designed so that farming can carry on underneath. In practice that usually means raising the modules high enough for people and machines to work below them, and spacing them so that enough light reaches the plants. The best-known example in Europe is at Heggelbach, a Demeter farm community near Lake Constance, where Fraunhofer ISE and its partners mounted 194 kW of modules on a structure five metres high over a third of a hectare of arable land.
There is no single agrivoltaic design. Modules can sit high above the crop on a fixed frame or on rows that turn with the sun, stand upright in rows facing east and west, run in low rows with crops growing between them, shelter grazing animals, or form part of a greenhouse roof. Germany's pre-standard DIN SPEC 91434 sorts the open systems into two groups. Category I leaves more than 2.1 m of clearance, so the farming happens underneath; Category II is lower, and the crops grow between the rows. Both must keep at least 66% of the usual harvest, and greenhouses fall outside the standard.
A system like this has to be judged on the crop and the electricity together. Researchers do this with the land equivalent ratio, which compares what the shared field produces with what the same land would give if it were split between a solar park and a farm. At 1 nothing is gained, and above 1 the shared field produces more in total than the two uses side by side.
What happens under the panels
Heggelbach's second season fell in the very hot, dry summer of 2018, which turned it into an unplanned test. Three of the four crops under the panels beat the reference field, with celeriac 12% higher, winter wheat 3% higher and potatoes also ahead, while clover grass yielded 8% less. With the electricity included, Fraunhofer ISE reported a land equivalent ratio of 1.86 for the potato field, up from 1.6 the year before. The researchers measured higher soil moisture under the panels in the wheat that summer and concluded that the shade had helped the crops get through the heat and drought.
In the Arizona desert, a team led by Greg Barron-Gafford at the University of Arizona grew chiltepin peppers, jalapeños and cherry tomatoes under a solar array and in an open plot next to it. Under the panels, chiltepin fruit production was three times higher and tomato production twice as high. The jalapeños gave a similar harvest with 65% less water lost through their leaves, and soil moisture stayed about 15% higher than in the open plot. The plants also kept the panels slightly cooler, which the researchers calculated would raise electricity output by about 3% over the summer.
The East African results agree with these, although the conditions were very different. The arrays at Morogoro and at Latia Agribusiness Solutions near Isinya in southern Kenya both cover half the ground with modules, and at midday the light directly under a panel in Morogoro was 77.5% lower than in the open field, far more shade than most European trials would use. Even so, Swiss chard and beans did significantly better under the panels, and maize kept its yield. Peppers, eggplants and the 2022 onions did worse, yet they were still the crops that earned the farm the most money. Combining harvests with electricity, the researchers calculated land equivalent ratios averaging 1.88 in Morogoro and 1.77 in Kenya, and above 1 for every crop in every season.
Where it doesn't fit
Not every crop copes with shade, of course. The most useful study I have found on this is a 2022 meta-analysis by Moritz Laub and colleagues at the University of Hohenheim, which combined 58 field and pot experiments from temperate and subtropical regions into yield curves for nine crop types. Every data point behind it is a measured harvest, while the curves themselves are a statistical estimate. The figure below shows what they predict at 40% less sunlight, a shade level within the range the authors give for agrivoltaic systems with fixed panels.
Berries, fruit and fruiting vegetables such as peppers came out ahead at moderate shade, forage grasses and leafy vegetables lost a little, and maize and grain legumes lost about half their harvest. A large share of farmers grow maize and beans, so that last result is hard to set aside. Berries do not all behave alike either. A Wageningen team found that blueberries in sunny conditions can benefit from up to 50% shade, while other berries hold their yield only up to about 35%, with larger losses where sunlight is weaker to begin with.
The maize result from Tanzania does not fit the temperate curve, and I think heat explains the difference. Where the sun is strong and water is short, shade relieves a stress that a German or Dutch field does not have, which is good news for hot, dry regions. It also means that a design proven in one climate cannot simply be copied to another.
How the system is built can change the result as much as the choice of crop. In a model of maize in Italy, covering 13% of the ground with panels cut the light reaching the crop by 12–15%, while covering 36% cut it by 28–32%. Panels that track the sun take less light than fixed ones, about 15–36% against 28–50% according to Laub and colleagues, and a 2022 modelling study found that tracking also spreads the light more evenly. The Tanzanian arrays run north to south so that the shadow moves across the beds during the day, whereas east–west rows can leave the same strip in shade all day. Upright two-sided panels facing east and west are another option, and at Foulum in Denmark, Aarhus University measured almost the same harvest of wheat and grass-clover between them as in the open field.
The largest review so far, by Wei Zhang and colleagues in 2025, gathered 264 yield results from 20 countries. Of these, 158 were lower under the panels, 72 unchanged and 34 higher, with no significant loss below about 20% shade and clear losses above 50%. The map below compares the meta-analyses with individual field trials, crop by crop. Taken together, I read the evidence as saying that agrivoltaics seldom increases a harvest, though a well-designed system keeps most of it while adding electricity, and in hot, dry places it can protect the crop.
Cost is the other limit. Raising modules several metres in the air takes more steel and stronger foundations, and the special structures are not mass-produced. A 2020 Fraunhofer ISE study modelled the electricity cost of agrivoltaics at €0.0828 per kWh, 38% more than a conventional ground-mounted plant, and Fraunhofer's agrivoltaics guideline gave a range of €0.07 to €0.12 per kWh for German projects, depending on the type of farming. A cost comparison by Germany's Technology and Support Centre (TFZ) estimated about €1,234 per kW for an elevated system against €572 per kW for a ground-mounted one. These are German figures from 2020 and 2021, so I use them only as an order of magnitude. Even so, for cheap electricity on land nobody farms, an ordinary solar park is the better choice. The extra cost of agrivoltaics is worth paying where the farming, the water or the shade has real value, or where the land is too valuable to give up.
What policy is doing
Europe is starting to write rules for agrivoltaics, and so far they put the farm first. France requires an agrivoltaic project to keep at least 90% of the agricultural yield, and Germany at least 66%. When SolarPower Europe published its agrisolar policy map in October 2025, it found that only five of the 18 member states it assessed had a legal definition at all, and called for one at EU level.
Italy has spent the most. Its recovery plan funded a dedicated tender, in which the energy agency GSE awarded about 1.5 GW of agrivoltaic capacity, with most of the bids coming from agricultural companies. Construction has been slow. The deadline of June 2025 was pushed back to 30 June 2026, with up to 18 months more to connect to the grid, and a draft decree law in early 2026 proposed relaxing the timeline again, so I would not trust any figure for how much has actually been built until the final numbers are out. For this series, the more telling detail comes from PV Tech, which reported that only a few dozen of the awarded projects plan to use any of their electricity on site. That suggests to me that most European agrivoltaics is being built to feed the grid. It will help to decarbonise the grid while doing little to keep a farm or a village running when that grid fails.
East Africa has a different problem. There is plenty of interest, but agrivoltaics sits between the ministries of agriculture and energy, and participants at a learning event in Kenya warned that this can mean nobody takes ownership of it. In West Africa, the APV-MaGa project, run by a consortium including UNU-EHS and Fraunhofer ISE with partners in Mali and The Gambia, planned five demonstration systems for uses such as an irrigated rice farm, a processing platform and a cold store, but I have not yet found published results from them.
From the field to the clinic
I have not found a published case of an agrivoltaic system powering a rural clinic, so I cannot answer the title question with an example. The parts needed to build one already exist, though.
Morogoro is already an off-grid system with batteries, running a training centre that used to depend on diesel, and the researchers point out that a system of that size is in the range used for rural mini-grids in sub-Saharan Africa: enough, by their estimate, for a small agribusiness or a village of around 500 people. The grid-connected array at Latia in Kenya covered 56% of the farm's electricity over a year, but it has no batteries, so on the day the grid fails it will go quiet for the same reason the rooftops in Spain did. The Sheffield team also suggests that farming communities without a grid connection could build their mini-grids as agrivoltaic systems instead of clearing farmland for a conventional array.
What follows is my own inference. A clinic or school next to a farmers' cooperative's array could share its power, if the system has batteries, an inverter that can run as an island, and a clear agreement on whose loads come first when the batteries run low. In that agreement the vaccine fridge has to come before the irrigation pump and the cold room, and it should be settled in the contract and the wiring before the first outage. Ownership matters as well. A clinic can only rely on an array on a farmer's land if the farmer is a real partner. Rules like those in France and Italy, which formally link the farmer to the project, matter more here than the size of any tender. I will come back to batteries, controls and microgrids in Part 4.
One question worth asking
If you work with a rural clinic, a school or a farmers' cooperative, ask one question before anyone talks about modules or tenders: if there were a solar array over the fields next door, which single load would you put on it first, and who would decide when there is not enough power to go round? Let that answer drive the design, and choose the panels afterwards.
Next time I will leave both the roof and the field. Part 3 looks at solar on water, on vehicles and on the infrastructure we already have.
The views expressed are those of Solstrides Projects and do not represent the positions of ETIP PV, IEA PVPS or any of its Tasks.
Sources
Randle-Boggis R.J. et al., Harvesting the sun twice: Energy, food and water benefits from agrivoltaics in East Africa, Renewable and Sustainable Energy Reviews 208, 2025 (open access). Morogoro and Latia systems, crop yields, light, irrigation, electricity, land equivalent ratios, the mini-grid estimate, panel orientation, French maize trials. Measured, except the land equivalent ratios and village estimate, which are calculated.
WHO, Close to one billion people globally are served by health-care facilities with no electricity access or with unreliable electricity, 14 January 2023, on the report Energizing Health by WHO, World Bank, IRENA and SEforALL. Global figure, rural and primary-care facilities.
IRENA, Decentralised renewables solutions can improve healthcare services for almost 1 billion people, January 2023. Facilities without electricity in South Asia and sub-Saharan Africa, reliability, decentralised solar.
Fraunhofer ISE, Agrophotovoltaics: High harvesting yield in hot summer of 2018, press release, 12 April 2019. Heggelbach system, 2017 and 2018 yields, soil moisture, land use efficiency.
Barron-Gafford G.A. et al., Agrivoltaics provide mutual benefits across the food–energy–water nexus in drylands, Nature Sustainability 2, 2019, with the University of Arizona's summary. Arizona yields, water loss and soil moisture.
Laub M. et al., Contrasting yield responses at varying levels of shade suggest different suitability of crops for dual land-use systems: a meta-analysis, Agronomy for Sustainable Development 42:51, 2022 (open access). Crop types, yields at 40% less sunlight (used in the figures), shade under fixed and tracking panels.
Hermelink M.I., Maestrini B. and de Ruijter F.J., Berry shade tolerance for agrivoltaics systems: A meta-analysis, Scientia Horticulturae, 2024. Blueberries and other berries.
Schindele S. et al., Implementation of agrophotovoltaics: Techno-economic analysis of the price-performance ratio and its policy implications, Applied Energy 265, 2020 (open access). Modelled cost, 38% above ground-mounted PV.
pv magazine, Agrivoltaics have an average LCOE of €0.093 per kWh in Germany, 24 February 2021, on Fraunhofer ISE's guideline Dual Yield on Arable Land: Guideline for Agrivoltaics. Modelled cost range.
pv magazine, Cost comparison between agrivoltaics and ground-mounted PV, 26 March 2021, on the TFZ study. Estimated cost per kW, mounting structures.
pv magazine, SolarPower Europe unveils agrisolar policy map, 28 October 2025. Five of 18 member states with a legal definition.
RE Global, Agri-PV regulatory developments across Europe, on the SolarPower Europe policy map. Yield thresholds in France and Germany, farmer links in France and Italy.
PV Tech, Italy awards 1.5GW in oversubscribed agriPV tender. Awarded capacity, projects with self-consumption.
pv magazine, Italy extends construction deadline for agrivoltaic tender, 26 June 2025. Deadlines, applicants.
ADVANT Nctm, The race against time to safeguard PNRR incentives, 2026. Draft decree law on PNRR deadlines.
CIFOR-ICRAF Forests News, Working on sunshine: Agrivoltaics in East Africa, 9 February 2024. Learning event, policy ownership, Latia without batteries.
United Nations University, Agrophotovoltaics for Mali and The Gambia (APV-MaGa). Project partners and demonstration sites.
Zhang W. et al., Climatic and design tipping points in agrivoltaic crop production systems. A meta-analysis, Agronomy for Sustainable Development 45:69, 2025. 264 yield results, shade thresholds, maize, beans and lettuce.
Amaducci S., Yin X. and Colauzzi M., Agrivoltaic systems to optimise land use for electric energy production, Applied Energy 220, 2018. Modelled maize yields, ground coverage and light, tracking and fixed panels.
Sekiyama T., Performance of agrivoltaic systems for shade-intolerant crops, master's thesis, Harvard Extension School, 2019. Corn under sparse and dense panels in Japan.
Weselek A., The impact of agrivoltaics on crop production, doctoral thesis, University of Hohenheim, 2022. Heggelbach yields by crop and year, light under the system, Montpellier lettuce.
Zainali S. et al., Direct and diffuse shading factors modelling for the most representative agrivoltaic system layouts, preprint, 2022. Modelled light under tracking and vertical systems.
Aarhus University, Double harvest: vertical solar panels and crops thrive side by side, 12 September 2025, on Victoria M. et al., Energy Nexus, 2025. Foulum yields and output timing.
Gajewski M., Fraunhofer ISE, Agrivoltaics – Introduction, ISES webinar, 10 February 2022. DIN SPEC 91434 categories and requirements.
PV Tech, Agrivoltaics across the globe – combining solar power and food production, 30 May 2023. System types, solar grazing.
Willockx B. et al., KU Leuven, A standardized classification and performance indicators of agrivoltaic systems. Open and closed systems, greenhouses.
Comments