Why sites die in a predictable order
Solar sites fail for a short list of reasons that is roughly the same in every US market. The nearest substation cannot take the power, the county will not permit the use, the land is wetter or steeper than the aerial suggested, or the owners will not lease. Solar resource is almost never the reason. Irradiance varies by a few percent between candidate parcels, while interconnection cost and permitting risk can vary by an order of magnitude.
That asymmetry should shape the screen. Criteria belong in the order of how often they end a project and how cheaply they can be checked, not in the order that is convenient to map. Fatal flaws come first because no amount of design work fixes them. Cost multipliers come second because they change the number, not the answer. Soft risks come last because they are real but negotiable.
Fatal flaw one: interconnection and substation headroom
Interconnection kills more utility-scale sites than any other single factor. The question is not whether a line runs past the parcel but whether a point of interconnection exists within an affordable distance that can absorb the project without upgrades it cannot carry. A site three miles from a constrained substation is usually worth less than a slightly worse site adjacent to one with headroom.
Screen for voltage level, substation ownership, existing queue volume in the zone, and any hosting capacity or headroom data the utility or ISO publishes. Distribution-voltage POIs cap project size quickly. Transmission taps and substation expansions add cost and time. For larger projects, an independent power flow screen separates POIs where headroom is plausible from those where a contingency overload is obvious in advance. If no plausible POI survives this pass, the parcel is not a solar site, regardless of how it scores on everything else.
Fatal flaw two: zoning, setbacks, and county moratoria
If the county will not permit the use, nothing else matters. Check the zoning district, whether utility-scale solar is permitted by right, by conditional or special use, or not at all, and whether the jurisdiction has an active or pending moratorium. Moratoria often appear with little notice after a contentious project and can stall a site for a year or more while a new ordinance is drafted.
Setbacks, height limits, fencing, screening, and glare rules do not usually kill a site by themselves, but they shrink it. A large setback from non-participating residences on an irregular parcel can remove a surprising share of the buildable area. Read decommissioning bond requirements and agricultural protection overlays closely, because they change the economics early. Ordinance language changes frequently, so every check must be dated.
Fatal flaw three: wetlands, floodplains, and protected habitat
Wetlands and floodplains subtract acreage from the center of a site, not the edges, which is what makes them fatal rather than expensive. Start with the National Wetlands Inventory and FEMA flood zones as a desktop pass, then treat the result as a hypothesis. NWI polygons miss wetlands and include upland, and hydric soil mapping often reveals a wetter site than the inventory shows. A delineation is the only reliable answer, and it is seasonal.
Federally protected species and critical habitat come next. Run the site through USFWS IPaC, note listed species and designated critical habitat, and check state natural heritage data. A listed species does not automatically end a project, but consultation adds time, survey windows, and design constraints, and some habitat designations are effectively no-build. Cultural resources, karst geology, and known contamination belong in the same pass because each can trigger a study that reshapes or removes the site.
Fatal flaw four: landowner control and parcel assembly
A site the developer cannot control is not a site. Ownership data, parcel boundaries, and outreach status belong in the screening phase, because the number of owners needed to assemble the buildable area predicts both transaction cost and failure probability. A 500-acre site with four owners is a different project from one with nineteen, and one holdout in the middle can sever the layout or the collector route.
Title and easements are the quieter version of the same risk. Existing pipeline, transmission, and access easements cross parcels in ways that limit racking and trenching, and mineral rights, life estates, and unresolved heirs can block a lease long after a handshake. Transmission line crossings in particular need early attention, since the utility will impose clearance and access conditions that the layout has to honor.
Cost multipliers: slope, soils, trees, distance, and access
Once the fatal flaws clear, the remaining physical criteria change cost rather than feasibility. Slope is the largest. Fixed-tilt and tracker systems each have practical grade limits, and terrain beyond them either needs grading, which triggers stormwater permitting, or gets excluded. Aspect matters less than slope in most of the country, but north-facing grades still reduce yield and complicate tracker layouts.
Soils drive foundation design. Shallow bedrock, karst, corrosive soils, and high water tables push projects from driven piles toward pre-drilling, ground screws, or ballast, each with its own cost curve. Prime farmland classifications carry both political and, in some states, regulatory weight. Tree clearing adds direct cost, stormwater complexity, and often permitting scrutiny, and it changes the visual profile that neighbors will react to.
Distance to the POI is a linear cost with a threshold. Gen-tie length adds conductor, structures, and right-of-way acquisition across intervening parcels, each of which needs its own easement. Access matters in the same way: a site without a road that can carry construction traffic and transformer deliveries needs one built, and the route may cross the same neighbors who oppose the project.
Soft risks: community opposition and permitting timelines
Community sentiment is a soft risk only because it is not printed on a map. Organized opposition delays hearings, invites moratoria, and hardens conditions of approval. Screen for it by reading recent public comment on nearby energy projects, checking whether the county recently revised its solar ordinance, and noting the density of non-participating residences within the viewshed. A technically clean site in a hostile county is not a clean site.
Permitting timelines are the other half. Some states run a siting board for large projects; others leave it to counties with varying capacity and appetite. Stormwater, wetlands, species, and cultural resource clearances each have their own clock and survey seasons. The screen should estimate a realistic critical path, because a site that permits in about a year and one that takes several years are not competing on the same economics.
Sequence the screen from cheap to expensive
The order of checks should follow cost. Desktop layers are nearly free: substation locations, zoning maps, NWI, FEMA flood zones, slope from elevation data, IPaC, parcel ownership, and ordinance text. Run every candidate through all of them before anyone visits a site. The output at this stage is not a score but a shortlist with a documented reason for each rejection.
The next tier costs money and time in modest amounts: a title search, an independent power flow screen, a conversation with the county planner, initial landowner outreach, and a site walk with someone who can spot wetland indicators and rock outcrops. Only the survivors of that tier justify the expensive work: wetland delineation, geotechnical borings, species surveys, cultural resource studies, and an interconnection application with its deposit.
Two habits make the sequence work. Record the date and source of every layer, because ordinances and queue positions change. And subtract constraints to a single buildable acreage number at each tier, since capacity, yield, and the financial model all flow from that figure rather than from gross parcel size.
When to walk away and when to keep working
Walk away when a fatal flaw is confirmed rather than suspected: no reachable POI with plausible headroom, a moratorium with no sunset, a delineated wetland through the buildable core, or an owner who will not lease and cannot be routed around. Time spent on a dead site is time not spent on a live one, and if the sequence above was followed the sunk cost is small by design.
Keep working when the problem is a cost multiplier with a known price or a soft risk with a known path. Slope can be excluded, trees can be cleared, a gen-tie can be lengthened, and a skeptical county can be engaged, as long as the financial model still closes with those costs included. Modern screening tools such as Basepoint put grid, ordinance, environmental, and ownership layers in one place so the fatal flaw check takes hours rather than weeks. The judgment about when a site stops being worth it still belongs to the developer.