10  Evidence Base and Research Needs

More grassland bird species use planted solar sites than the row-crop fields around them, and that is known because recorders listened for them through two summers (Walston et al. 2025). The claim that a seeded ground layer under an array keeps sediment and nitrogen out of the creek below is a different kind of claim. It is the prairie-strip result from corn and soybean fields (Schulte et al. 2017), carried over onto ground nobody has sampled. Both claims appear in this lexicon, and the sentences that make them look alike.

Every judgment this lexicon makes about how well something is known lives in this chapter. The other chapters describe what agrisolar design is and what it can do. This one says what is established, what is inferred, and what nobody has measured. Keeping it all in one place lets a reader weigh the whole lexicon at once, and keeps any chapter from quietly overstating its own confidence.

It is organized by what would fix each gap rather than by topic. Some gaps need a season of field instruments, some need somebody to go back to the original data, and some need a project that does not exist yet to be built and reported.

10.1 What the evidence supports

Read this as the confidence behind each chapter, stated once here rather than repeated as a caveat wherever a claim appears.

Water (§6.1). The link between shade and soil moisture in dry places is verified (Barron-Gafford et al. 2019; Pinos et al. 2026), though limited at some sites (Kannenberg et al. 2023), and grass grew about 20% more under a Colorado array in dry years (Sturchio and Knapp 2025). Field evidence on infiltration and runoff is still thin (Yavari et al. 2022), and recharge at depth has hardly been measured. Interception is documented for prairie strips in cropland (Schulte et al. 2017), but never yet from beneath an array. Avoided loading is arithmetic rather than measurement: it follows from fertilizer application rates on retired acreage, not from a monitored stream.

Ecology (§6.2). Verified for more grassland bird species (Walston et al. 2025) and the response of insect communities (Walston et al. 2024). That arrays planted for pollinators could support many bee species is an estimate from prairie data, not counts at arrays (Bruninga-Socolar et al. 2025). Scorecards write the choices down, though what each one gives credit for differs (Electric Power Research Institute 2021). The gain depends on context. It is largest where the surrounding landscape has little natural habitat left, so its size depends on what the site would otherwise have been, not on the planting alone.

Agriculture and soils (§6.3). On the farming side, more bees visiting the crop beside a planted array is verified (Walston et al. 2024), and the farmland within foraging distance has been mapped (Walston et al. 2018). Pest control is inferred, not measured at solar sites, and a change in yield or input costs that can be pinned on an array is not established anywhere. On the soil side, early gains are thin: at one ecovoltaic grassland restoration, soil carbon was slightly higher but not significantly so (Krasner et al. 2025), and long-run build-up is thinly measured (Carvalho et al. 2024). The size of the construction debt is the least measured part of the soil account, and among the most likely to decide whether the soil comes out ahead or behind, because the reported gains come from sites whose earthworks are not described.

Communities and economy (§6.4). The design principle is verified in the ecovoltaic frameworks themselves, where Tölgyesi et al. (2023) make stakeholder engagement within a viable business model one of five pillars. The costs to communities are better measured than the benefits (Bessette et al. 2024; Elmallah et al. 2023). Which arrangements deliver lasting, fairly shared benefit is still thinly evidenced.

The community scale, A3 (§5.3). Thin, and thinner than the confidence with which the middle sizes are usually discussed. The capacity arithmetic is solid and comes from federal sources. How pollination benefits travel is measured, at two sites, for one function. Everything else is inference. Nobody has compared several small arrays against one large one at the same capacity and management, and that comparison is the test of scale-led restorative work. Nobody has tested whether land quality varies with project size at a finer scale than the whole parcel. And nobody has evaluated whether the bonuses that state incentive programs pay for farming or pollinator habitat changed what actually got built. The outcomes footprint is under-evidenced at every zone on the transect, not only here. The A3 section argues about where the decision still lies. It does not report where restoration has already happened.

The regional calibration (§9.3). A set of hypotheses drawn from regional summaries, not from measured outcomes. The regional argument does not depend on any one row of it.

What the siting studies show. Wu et al. (2026) studies utility-scale solar projects of 1 MW and up, treating each project as a single data point. It does not test whether the factors driving site choice change with project size. So it shows what large projects are sited around, but not the contrast with smaller projects that the lexicon draws. That contrast still rests on Macknick et al. (2022) and the archetypes, which is weaker evidence than a study broken out by size would be. That is why where siting latitude peaks is still an open question. Owusu-Obeng et al. (2026) find prime farmland the cheapest ground for solar in the eastern United States, and contaminated land 14 to 33% more expensive.

The prime-farmland argument. Studies that score solar sites by soil quality find that distribution-scale projects sit disproportionately on good land, and it is tempting to read that as careless siting. Two things complicate that reading.

The first is a hidden third factor. Community-scale solar goes where the electricity is used, electricity is used where people live, and people live where the good soil was, because towns grew up on good farmland. That is why metropolitan areas hold an outsized share of the nation’s prime farmland, and why development there tends to land on the best soils (Dillman and Cousins 1982). A fleet built near where power is used will sit on prime soil whether or not anyone was careless.

The second is the level of detail. A parcel classed as prime farmland is not equally productive everywhere. Across 338 Midwest fields and 1,625 yield maps, about half of the field area yielded well year after year, 32% was persistently low-yielding, and 18% swung with the weather; in rainfed fields the unstable ground tended to sit in wetter, lower spots (Maestrini and Basso 2018). Roughly a third of a prime field never performs. That 32% is exactly what the freedom to choose a site at A3 can act on. A study that scores a site by the soil class of its whole parcel cannot see whether the array took the good part or the bad part. So the claim that mid-sized projects are sited carelessly and the claim that they are sited well are both, for now, untested at the level of detail that would settle them.

10.2 Measurements nobody has made yet

Table 10.1: The measurements that would settle claims this lexicon has to make on inference. Each is flagged where the chapters raise it, so a reader can check the claim rather than take the priority on faith.
Gap Flagged at What would close it
Nutrient and sediment interception under an array. Whether a seeded solar ground layer performs like a prairie strip. §6.1 Paired monitoring above and below an array sited on a flow path, against a strip or an unplanted control.
Recharge and the water balance at depth. Panel drip, altered cover, and what reaches the aquifer. §6.1 Measured drip and runoff, soil-moisture profiles, and coupled surface–groundwater modeling on instrumented sites.
Cumulative effect on one watershed. Whether arrays scattered through the same watershed add up, cancel out, or miss each other, and how much of the ground has to drain through them before a gauge can tell. §6.1 Gauged watersheds filling up with solar, against comparable ones without, and routing models built on those rather than one site at a time.
Soil carbon over decades. Early gains are barely measured, and the multi-decade curve not at all. §6.3 Re-sampling the early ecovoltaic sites on a decadal schedule, which needs a commitment longer than a grant cycle.
The construction debt, and how long it takes to repay. How much carbon mass grading costs a site, against how much a minimally graded build keeps. §6.3 Paired sampling before and after construction across projects with contrasting earthworks, which needs the earthworks recorded at the time.
Pest regulation from an array. Natural-enemy habitat is established for perennial refuge, never demonstrated from a solar site. §6.3 Natural-enemy and pest counts in fields at increasing distance from a planted array.
A yield or input-cost change attributable to the array. The step a farmer actually cares about. §6.3 Paired fields near and far from a planted array, with pollination and pest pressure measured alongside yield.
Whether a regional fleet of arrays functions as a connected network. Compatible seed mixes and calendars make one on paper; nothing establishes that dispersal actually happens between sites. §6.2 Mark–recapture or genetic work on a taxon whose dispersal distance is known, across arrays at a range of separations, with unbuilt patches as controls.
Whether co-located ground supports biodiversity the way unbuilt ground does. §6.2, and the stated gap in Brock et al. (2026) Comparative survey across co-located, conventional, and unbuilt sites in the same landscape.
Which benefit structures deliver durable, fairly shared benefit. The instruments are known; their outcomes are not. §6.4 Longitudinal work on host communities rather than one-time acceptance surveys, with landowner and tenant reported separately.

The first of these has the most riding on it. Interception is the water-quality approach most projects could actually use, since it asks only for a seed mix and a placement rather than for retiring a working field. Right now it is supported entirely by analogy.

10.3 Numbers that circulate without a traceable source

Some of this work is reading rather than measurement. The Sources chapter splits what has been checked from what has not, and the unverified block is not a list of doubts so much as a list of errands.

The grazing-cost figures are actively misquoted. Stewart et al. (2025) report McCall et al. (2023)’s $279/ha/yr as the median cost of sheep grazing. In McCall’s Table 4, that figure is actually the median mowing cost at sheep-grazed sites, and grazing itself ran about $50/acre/yr (§3.1). Both readings are now in circulation, and converting between acres and hectares makes the confusion worse. Anyone quoting a grazing cost should go straight to McCall’s table, and anyone who has already quoted the swapped figure should check which one they used.

The community-benefit and grazing-economics material sits in gray literature: it is real and widely practiced, but not yet pulled together anywhere citable.

10.4 Claims that a counterexample would test

The lexicon makes several claims that could be wrong, and it is more useful if it names what would prove them wrong.

  • Sparing does not assemble at scale. §4.2 holds that S4 thins toward A5 because idle remnants are scattered. A utility-scale project assembled from scattered marginal remnants on ordinary farmland, rather than from uniformly pre-disturbed ground, would reopen that corner. None is documented here, and one would be worth reporting.
  • Both hand-offs fall at A3. Figure 5.1 puts the crossing from structure to management and the crossing from ownership to contracts in the same zone, and that is the main support for treating A3 as the zone with the most design freedom. Whether the peak sits at A2 or A3 is argued from the archetypes rather than measured, and the lines are schematic, not data. Moving the peak one zone to the left would move two claims, not one, so real siting decisions that put either crossing somewhere else would be worth reporting.
  • Regions resolve the pressures differently. The calibration in §9.3 is a set of hypotheses. A region doing something its row says it should not is among the most useful contributions the lexicon can receive.
  • The six levers are complete and non-overlapping. Every decision a project makes should land in exactly one of scale, siting, structure, management, ownership, and contracts. A decision that lands in none of them would mean the list is short; one that lands in all six would mean the categories are not cutting anything. The edge is the known partial exception, split deliberately between structure and management because a berm and a hedgerow fail differently (§1.1).
  • The A3 ceiling rests on one untested number. The zone tops out near 5 MW because a restorative community-scale array is taken to occupy at most about a third of its parcel (§5.3). More rests on that fraction than on any other number in the scheme, and it has not been tested against built projects. At half a parcel, the same fields would hold 3 to 7 MW, and the case for a higher ceiling partly comes back. Measuring how much of its parcel each of a set of built restorative community-scale arrays actually occupies would settle it.
  • The vocabulary describes a site, and sparing is a claim about a landscape. Nothing here says where the array went instead. Statewide planning shows the gap matters. Siting New York’s 2050 solar build to protect farmland spares roughly 80% of the farmland a least-cost build would take, but opens up more than 41,000 hectares of forest in the process (Gallaher et al. 2026). Global models have the opposite problem. Brock et al. (2026) assign each map cell to a single land use and say plainly that they did not model co-location, so their land-demand figures are upper limits and the arrangements this lexicon describes fall outside their main results. Accounting at the landscape level that could handle both would test whether the axis needs a companion term.
  • Positions are exclusive. The two-question reading rule (§4.1) should place any real project in exactly one position. A project that genuinely will not fit is a flaw in the axis, not in the project.

10.5 What would make the evidence easier to build

These are recommendations about reporting, not about design. The lexicon prescribes an aim and leaves the dimensions to the site, and none of what follows is a dimension.

  1. Report acreage alongside capacity. Megawatts are how a project is filed; acres are how the ground experiences it, and the two come apart with tracking and clearance (§3.1).
  2. Name the benefit delivery when claiming a function. A hydrologic claim that does not say whether it means moisture under the panels or nitrogen kept out of a creek cannot be checked, and the two need different evidence (§6.5).
  3. Separate the landowner from the tenant. Nearly 40% of Central Valley agricultural land is leased, and a conversion that sustains an owner can end the operation of the person farming the ground. Aggregate farm-income figures hide exactly the party most at risk (§6.4).
  4. Publish the counterfactual behind a sparing claim. Sparing is a claim about land that was not taken, so it is only checkable against the local market and the sites genuinely in play (§4.1).
  5. State the prior land cover and the earthworks when reporting soil carbon. Putting perennial cover on ground that was tilled and converting existing grassland are opposite transactions, and whether soil carbon goes up or down depends on which one happened. How much the site was graded belongs in the report too, because a gain measured on ground that was never stripped and a gain measured on ground rebuilding from a stockpile are not the same number (§6.3).
  6. Give the region and the crop when reporting a shade-yield result. Whether shade helps or hurts flips with air temperature, so a yield figure with no location attached travels farther than the evidence behind it (§2.3).
  7. Report what did not work. Establishment failures, seed mixes that did not take, and grazing arrangements that fell apart are absent from the literature and present on the ground, and their absence is why cost expectations are unreliable (§3.1).