Login
Donate
Join us
Donate
Community

Under the Panels at Spadra Farm 

June 7, 2026

What We Are Learning About Agrivoltaics in Southern California

By Rose Cesena Olivas, M.S., Robert Redford Conservancy at Pitzer College, Co-Principal Investigator and Dr. Eshwar Ravishankar, Cal Poly Pomona, Assistant Professor of Plant Science

Cal Poly Pomona’s Spadra Farm Agrivoltaics Site At-A-Glance

This agrivoltaic research site on Cal Poly Pomona’s Spadra Farm is testing how PV panels reshape crop performance, soil, and microclimate under hot, semi-arid Inland Empire conditions. Here is what the site looks like, what we are measuring, and how it lines up with what other groups have reported. 

Figure 1. The Spadra Agrivoltaic Research Area on Spadra Farm atCal Poly Pomona. Six fixed-tilt arrays sit at the western edge of California’s semi-arid Inland Empire, with the San Gabriel Mountains behind.

A working research farm in the Inland Empire 

The Spadra Agrivoltaic Research Area occupies a corner of the historic Spadra Farm at Cal Poly Pomona, about 35 miles east of downtown Los Angeles. Six fixed-tilt photovoltaic structures sit over active row-crop beds. The site is instrumented with weather stations, a soil-sensor network, and a drone fleet that flies multispectral imagery on a regular cadence through each cropping cycle. 

The array geometry mirrors what a California grower could realistically deploy on working farmland: south-facing modules at a 22-degree tilt, 2 meters of ground clearance, and a 3 meter pitch. Each pair of panels frames a strip of cultivated bed. Through our first three cycles the crop has been romaine lettuce, cv. ‘Salvius,’ transplanted at roughly nine plants per square meter, which matches commercial coastal practice. 

Every seasonal cycle runs as a 2 × 2 factorial: solar versus open field, crossed with conventional versus regenerative management, with multiple replicates inside each panel bay. That layout separates the effect of the panels from the effect of soil management and lets us test whether the two interact. 

Figure 2. Strip-cropping pattern under a panel bay at the start of a cycle. The flagged center rows mark replicate positions; each bay contains paired conventional and regenerative beds.

The questions we set out to answer 

Six questions shape the work, and they are the ones growers and developers in California tend to ask first. 

  1. How does production under the panels compare with the open-field control, and does that comparison change between cooler and warmer cycles?
  2. What is actually driving any difference? Shading on its own, or also temperature, humidity, ventilation, and soil moisture?
  3. Which design and management choices give a grower the most leverage: panel height, ventilation, irrigation, cultivar?
  4. Does the soil behave differently under the panels in ways that change how nutrients should be managed?
  5. Does the amendment (compost/biochar or synthetic fertilizer) react and supply nutrients differently?
  6. Do the soil microbial levels differ under panels?

The reason these questions are not independent is that the answer to the second one constrains the third and fourth. If a yield difference turns out to be driven by humidity and ventilation rather than by photosynthetically active radiation, the design lever is mounting height and not cultivar shade tolerance.

How the site answers them 

Spadra was built as a measurement-rich site rather than a demonstration plot. Replicated harvests give the agronomy data. The soil sensor network gives continuous microclimate data. The drones give spatial canopy data through the season. Soil samples taken at transplant and harvest at two depths are characterized by bacterial and fungal targeted DNA sequencing, so we can track how the soil microbial community responds to the array. 

Behind the field work, a physics-based light-and-energy-balance crop model is calibrated against the open-field controls and then run under the panel geometry. The model exists to separate the shade contribution from everything else. If the model predicts a yield penalty of X% from light alone and the field shows a larger or smaller penalty, that gap is the part of the answer that microclimate and soil moisture have to account for. 

 Figure 3. Soil-sensor nodes and the irrigation backbone going in alongside a fresh transplant. Each replicate carries co-located moisture, temperature, and EC probes at two depths.

What the first cycles are showing 

The qualitative picture in our first cycles is consistent with the published lettuce-under-PV record, which generally indicates compatibility. Lettuce yields differ between under-panel and open-field plots, and the difference is more pronounced in the warmer cycle than the cooler one. The magnitudes fall within the range reported for romaine, butterhead, and salad mixes in comparable climates by Marrou et al. (2013) in southern France, Kavga et al. (2018) in Greece, Carreño-Ortega et al. (2021) in southern Spain, and most recently Dal Prà et al. (2024) in northern Italy. 

What is more interesting than the average penalty is the seasonal pattern. If light were the only mediator, the under-panel yield gap would track seasonal solar availability. Across our cycles it varies more strongly than that, which is the signature that other groups have used to argue the response is not a pure shading response. We are holding specific yield numbers and the formal mechanism attribution until our peer-reviewed analyses are out; a follow-up post will return with the figures. 

Figure 4. Romaine ‘Salvius’ approaching harvest under one of the panel bays. Pairing the modelled physics decomposition with replicated field harvests is what lets us identify the mechanism behind yield differences, beyond simply measuring their magnitude.

Why “look beyond shade” is the working hypothesis 

Two field studies anchor the working hypothesis that PV arrays act as microclimate modifiers, not pure shaders. 

Barron-Gafford et al. (2019) reported that peppers and tomatoes in arid Arizona produced more fruit under panels than in the open, and attributed the response to reduced heat stress and lower evaporative demand rather than to light. Hassanpour Adeh et al. (2018), working on semi-arid Oregon pasture, reported substantially higher soil moisture and biomass under panels. Weselek et al. (2021) extended this to celeriac, potato, clover-grass, and winter wheat under a tracker array in southwestern Germany, and reported crop-specific microclimate signatures in canopy temperature and humidity. Jung et al. (2024), in semi-arid Chile, quantified the linkage: global horizontal irradiance fell by roughly 42% under the array, potential evapotranspiration by roughly 31%, and soil moisture rose by roughly 29% relative to open ground. 

The common thread is that panels cut wind, retain humid air near the canopy, redirect longwave radiation, and slow soil drying. For some crops in some climates that bundle is a net positive. For others, in particular leafy greens in mild coastal climates, it can swing the other way. The research question Spadra is structured to answer is which microclimate effects matter most, and which design and management knobs let a grower bias the bundle in a favorable direction. 

Why the soil matters as much as the canopy 

Soil moisture, soil temperature, and the root-zone microbial community are coupled. If panels modify the first two, the third should follow, and the microbial groups most likely to shift are the ones driving nitrogen cycling and nutrient availability. 

Direct evidence is starting to appear. Luo et al. (2024) reported that an agrivoltaic peanut / ryegrass system in southwestern China significantly raised soil organic carbon, total nitrogen, microbial biomass, and urease activity, with inter-row gap zones showing stronger gains than under-panel zones. Choi et al. (2020), on a Colorado solar site seven years post-construction, documented soil-moisture heterogeneity driven by panel-edge runoff, with associated nutrient patterning. Moretta et al. (2026), on Mediterranean pasture, reported that inter-row zones supported legume-rich communities while under-panel zones supported stress-tolerant forbs, implying differential biological nitrogen inputs across short spatial scales. 

There is also a crop-quality dimension that does not show up in yield alone. Hsiao (2023) found that pak choi and rape under a 38% PV-shaded greenhouse roof in Taiwan accumulated more leaf nitrate than under control plastic, consistent with reduced light-driven nitrate reductase activity. Khudhair (2025) reported a similar shade-by-fertilization interaction for lettuce in Iraq. The implication for a California grower is that fertilizer rates set for full sun may produce a different product, not just a smaller one, under panels. 

The Alomari et al. (2026) review of agrivoltaic nutrient dynamics states the broader gap explicitly: comprehensive studies of plant mineral nutrition under PV remain scarce. The only published California agri-photovoltaics crop trial we are aware of, Hudelson & Lieth (2021) at UC Davis, reported yield responses across shade levels for kale, chard, broccoli, peppers, tomatoes, and spinach, but did not measure soil nitrogen, nitrogen leaching, or nitrogen budgets. That is the gap our pending CDFA Fertilizer Research and Education Program (FREP) project is built to address. 

What this already implies for growers 

A few practical orientations follow from the published record and from what the first Spadra cycles are showing.

  1. Shade tolerance is the wrong sole rubric for crop selection. A crop’s sensitivity to canopy heat, humidity, and ventilation matters about as much as its photosynthetic light response.
  2. Panel clearance height and cross-ventilation are real design levers. Raising clearance and improving air movement under the array reduces thermal and humidity buildup without sacrificing energy yield.
  3. Nutrient management probably needs to be position-aware. The Luo (2024), Choi (2020), and Moretta (2026) evidence on within-bay heterogeneity, combined with the Hsiao (2023) nitrate-accumulation result for leafy greens, makes a uniform field-wide fertilizer rate hard to defend on the under-panel strip.
  4. Land-use efficiency stays favourable. Even where the crop yields modestly less under panels, the combined energy plus crop output per hectare exceeds what either single use delivers alone, as documented in the Renno (2026) and Mehta (2025) systematic reviews.

What’s next at Spadra 

Cycle 4 is in the ground. The new layer is direct microclimate measurement under the canopy: temperature, humidity, wind, soil moisture at two depths, co-located with each replicate. With those data we can close the loop on the mechanism question that the model sets up. 

With anticipated support from CDFA FREP we plan to extend the work to position-specific fertigation, deeper soil-nitrogen and microbial sampling, drone-based nitrogen-status mapping, and a grower-facing decision-support tool. Our existing collaboration with the U.S.–Israel Binational Agricultural Research and Development (BARD) program supports the modelling thread; peer-reviewed publications from that work are in progress, and a follow-up post will return with full results once those are out.

Figure 5. Spadra Agrivoltaic Research Area site fact sheet (Redford Conservancy at Pitzer College, in collaboration with Cal Poly Pomona)

The team at SPADRA 

Dr. Rose Olivas (Pitzer College, Spadra Director) and Dr. Eshwar Ravishankar (Cal Poly Pomona), with co-investigators Dr. Sai Chandra Kosaraju, Dr. Aaron Fox, and Dr. Brianna Posedas, and student researchers from CPP and Pitzer. 

Get involved 

Spadra hosts grower visits, field days, and student tours. To learn more, contact Eshwar Ravishankar (eravishankar@cpp.edu) or Rose Olivas (rose_olivas@pitzer.edu).

Selected references 

Tagged

Leave a Reply

Your email address will not be published. Required fields are marked *