EPA Brownfield-to-Solar Site Finder — RE-Powering Mapper

Find contaminated land, closed landfills and abandoned mine sites that EPA has already pre-screened as renewable-energy redevelopment candidates. Screen a state, county or city against all 190,976 sites in EPA's RE-Powering America's Land Mapper, filter by technology, acreage, solar and wind resource and grid context, and get back a ranked list of development candidates with every EPA caveat stated on the row.

This is a finder, not a screener. You give it a region and criteria; it tells you which sites qualify. That is a different question from "what is at my coordinates", which is what the site-screening actors in this portfolio answer.

---

Who it is for

• Brownfield-to-solar and landfill-solar developers doing origination: which closed landfills in Pennsylvania are over 20 acres, near transmission, with decent irradiance?
• Renewable site-selection and land-acquisition teams building a pipeline of pre-vetted parcels in a target market.
• Abandoned mine land redevelopment specialists working AML and coal-community reuse programmes.
• Project-finance and diligence analysts who need to know what a site's contamination programme actually is before underwriting it.
• AI agents doing grounded siting research — this returns a record class no language model holds.

What makes it different

EPA published this screening layer and nobody has transformed it. The moat is not the fetch; it is the honest labelling. EPA's RE-Powering data contains at least six traps that turn into confident wrong answers if you read the columns literally, and every one of them is measured, flagged and disclosed on the row. See "What EPA's data actually says" below.

---

Example input

{
  "states": ["PA"],
  "landfillOnly": true,
  "minAcreage": 20,
  "maxTransmissionDistanceMiles": 2,
  "sortBy": "score_desc",
  "maxResults": 200
}

Or, for a broad regional sweep:

{
  "states": ["CO"],
  "technologies": ["utility_solar"],
  "minAcreage": 20,
  "sortBy": "score_desc",
  "maxResults": 300
}

---

⚠ What EPA's RE-Powering data actually says — and the six traps

EPA RE-Powering is a DESKTOP PRE-SCREEN. It is not an entitlement, not a feasibility study, not a site-control position, and not a determination that a project can be built or interconnected. Every row carries epa_screening_only: true, is_entitlement: false and a full-text screening_disclaimer. Cleanup status, ownership, zoning, offtake and interconnection all have to be confirmed independently.

Everything below was measured live against EPA's service on 2026-08-18.

1. Acreage = 0 is a null sentinel, not zero acres. 108,888 of 190,976 sites (57.0%) read 0, and 108,886 of them also carry EPA's own "No Acreage PV" flag — EPA could not size an array because acreage was unknown. Acreage IS NULL returns zero rows, so 0 is the only way this layer expresses "unknown". Reading it as zero acres would rank a viable mine site as unbuildable. This actor publishes acreage_acres: null plus acreage_raw_value and acreage_is_null_sentinel so nothing is hidden.

2. PV capacity sits on a 600 MW PLATEAU — but 600 MW is not a ceiling, and 16 sites exceed it. EstPVCap reads exactly 600 MW on 362 sites whose acreage spans 4,144 to 2,669,225 acres — a 644x range of site sizes mapped to one identical number. Fort Carson at 137,400 acres and a 4,144-acre site both read "600 MW". But 16 sites publish MORE than 600 MW, up to 11,788.08 MW (Hudson River PCB Sediments, NY, at 81,337.78 acres), so never filter on est_pv_capacity_mw <= 600 — that would silently drop the largest solar-redevelopment candidates in the layer, which is exactly what this actor exists to find. Of the 375 sites over 4,200 acres with a published capacity, 359 read exactly 600, 16 read above it, and none read below.

EstWindCap is a genuine ceiling by contrast: 1,600 MW is never exceeded anywhere in the layer, and 49 sites sit exactly on it.

Underneath the plateau, capacity is simply acreage times a single national constant — measured at ~0.1449 MW/acre (1 MW per 6.9 acres) in every state checked, including CO, TX, CA, AZ, NY, VA and MA. (The 0.10–0.25 MW/acre spread you see in pv_density_mw_per_acre_derived on small sites is EPA publication rounding, not regional variation: restrict to sites of 100 acres or more and it collapses to 0.14489–0.14496.) So EPA's capacity figures carry no site-specific engineering information. This actor flags them with est_pv_capacity_is_capped (true at or above the 600 MW plateau) / est_wind_capacity_is_capped, publishes the derived pv_density_mw_per_acre_derived so you can see the constant for yourself, and excludes EPA capacity from the ranking score entirely. All three figures — the at-plateau count, the above-plateau count and the observed maximum — are re-asserted against the live layer by the pv_plateau_integrity drift probe on every run, so this paragraph cannot go stale without the run failing.

3. EPA's summary technology columns are constants. Solar = 'Y' on all 190,976 sites and Solar IS NULL matches nothing; GeoHeatPmp = 'Y' on 190,956 of 190,976 (99.99%). An actor that offered "filter for solar-viable sites" and filtered on EPA's Solar column would return the entire national inventory while appearing to have filtered. The real discriminators are the scale-specific columns — utility-scale PV 17,649 sites, distributed 64,268, off-grid 82,090, utility-scale wind just 601 — and that is what this actor filters on.

4. The single-space " " sentinel is the null on every flag column. Landfill IS NULL returns 0 rows while Landfill = ' ' returns 172,678. EPA's own RE-Powering Site Profile page renders that sentinel as -. Site names, addresses, cities and ZIPs use it too: 1,675 sites publish no name at all, 22,261 no address, 19,675 no city, 61,533 no ZIP. And EPA publishes no explicit negative value anywhere in the layer — so false on any flag here means "EPA did not flag this", which is weaker than "EPA determined this is not the case". Every flag ships as both a boolean and a verbatim _raw column, with flag_semantics_note on the row.

5. The grid fields are a FROZEN 2022 PRECOMPUTATION and are internally inconsistent. Substation distance, substation voltage, transmission distance, transmission kV and transmission status are baked into the layer, not measured live. 16,803 sites report the nearest line status as NOT AVAILABLE while also publishing a voltage — a contradiction on the face of the record. Every one of those fields is named *_2022_screening_estimate, interconnection_data_is_live is always false, and the contradiction is surfaced on transmission_status_contradicts_voltage. For a live measurement, run hifld-grid-proximity-screener on the coordinates in the row. That is the intended workflow: this actor finds the candidate, that one verifies the interconnection.

6. Some acreage values are not the site footprint. The largest is 2,669,225 acres — 4,170 square miles for a single "site" — and an urban Denver brownfield is published at 43,000 acres, 67 square miles inside the city. The value is published as EPA wrote it, with acreage_implausible: true attached, because which figure EPA intended is unknowable from the data.

---

How this actor protects you

• A live drift gate runs before a single billable row. Six probes: a national row-count floor, a positive canary (a known site must still carry its recorded values), a sentinel-integrity check (EPA's null convention must not have changed), a negative control (a bogus state must return 0), a strictly-less-than check (a filtered count must be genuinely smaller than the unfiltered total, so an ignored filter cannot return a national answer under a regional label), and a documentation check — the EstPVCap distribution this actor describes in prose on every row (362 at the 600 MW plateau, 16 above it, maximum 11,788.08 MW) is re-measured against the layer, so a stale factual claim fails the run rather than being quietly published.
• Each probe has its own error boundary and a three-valued outcome — pass, fail, or unavailable. A probe that could not complete is not a probe that failed. If a load-bearing probe is unavailable the run fails and bills nothing; if only a corroborating probe is unavailable the run proceeds and says so on every row (drift_gate_status: verified_degraded).
• No silent zero. states is required, so a bare {} call gets an HTTP 400 before the run starts and can never bill for a 190,976-site national sweep. An upstream outage fails loudly rather than returning an empty dataset. Zero rows only ever means EPA genuinely holds no matching site.
• A capped answer can never look complete. Every row carries matched_sites_total (read from EPA's free exact count endpoint before any row is fetched) and results_truncated.
• null means not checked; false means checked and negative. Nothing is coerced. There is no Number(x) || 0 anywhere in this actor.

---

Related actors — the brownfield-to-solar workflow

Step	Actor	What it adds
1. Find candidates	this actor	EPA's 190,976 pre-screened contaminated / landfill / mine sites
2. Verify the grid	hifld-grid-proximity-screener	Live transmission line and substation distance, voltage, utility, ISO/RTO
3. Check the IRA bonus	ira-energy-community-bonus-screener	10% ITC/PTC energy-community adder eligibility for the coordinates
4. Model the yield	nrel-pvwatts-solar-resource-scorer	NREL PVWatts production estimate at the site
5. Read the contamination	epa-contaminated-site-screener	Phase I ESA record search — SEMS, RCRA, ECHO, brownfields findings
6. Check the queue	interconnection-queue-tracker	Live ISO/RTO interconnection queue positions nearby
Public land alternative	blm-solar-siting-area-screener	BLM Western Solar Plan designated leasing areas
Soil and terrain	ssurgo-soil-suitability-screener	NRCS ground-mounted solar array soil interpretations

---

Use as an MCP tool

This actor is callable by AI agents through Apify's hosted MCP server. Billing is unchanged when it is called as a tool, and a run that cannot answer fails without billing.

Endpoint: https://mcp.apify.com/?actors=malonestar/epa-repowering-brownfield-to-solar-site-finder

Claude Desktop / Claude Code / Cursor config:

{
  "mcpServers": {
    "epa-repowering": {
      "url": "https://mcp.apify.com/?actors=malonestar/epa-repowering-brownfield-to-solar-site-finder",
      "headers": { "Authorization": "Bearer YOUR_APIFY_TOKEN" }
    }
  }
}

Worked agent prompt: "Find closed landfills in Pennsylvania over 20 acres that EPA has screened for utility-scale solar and that sit within 2 miles of transmission. Rank them, and tell me which capacity estimates sit on EPA 600 MW plateau rather than being real site-specific estimates."

Chain several tools in one connection:

https://mcp.apify.com/?actors=malonestar/epa-repowering-brownfield-to-solar-site-finder,malonestar/hifld-grid-proximity-screener,malonestar/ira-energy-community-bonus-screener

---

Pricing

$10 per 1,000 results, billed as the standard Apify Result event — one row, one result. Graduated discounts apply on paid plans. A run that fails emits nothing and bills nothing.

states is required and capped at 6 per run, and maxResults defaults to 500 with a hard ceiling of 5,000, so there is no way to accidentally trigger a large national bill.

---

Output fields

Every field this actor can emit is listed below. null always means "not checked or not published by EPA", never "zero" and never "no".

Site identity

Field	Type	What it means
site_ref	number	EPA's RE-Powering cross-reference number (the Ref column). Stable identifier for the site inside the RE-Powering Mapper and the key used in the site profile URL.
site_id	string	The identifier assigned by the originating cleanup program, for example an EPA RCRA handler ID or a state program number. Not unique across programs.
site_name	string	Site name as published by EPA. Null when EPA published the single-space sentinel instead of a name, which happens on 1,675 sites (mostly state abandoned coal mine areas).
site_name_withheld_by_epa	boolean	True when EPA published no site name for this record. Distinguishes a genuinely unnamed site from a fetch problem.
program	string	The originating cleanup program exactly as EPA publishes it. 49 distinct values, from short federal codes such as SUPERFUND and RCRA to full state program names.
program_category	string	Coarse category derived from the verbatim program name: abandoned_mine_land, brownfield, landfill, superfund, rcra_corrective_action, underground_storage_tank, institutional_control or state_cleanup_program.
program_is_federal	boolean	True when the program is one of EPA's six federal programs, false when it is a state program. Null when no program was published.
epa_region	number	EPA regional office number (1-10) covering the site.
objectid	number	The ArcGIS feature OBJECTID for this row. Useful for reproducing a query; not stable across EPA republications.

Location

Field	Type	What it means
state	string	Two-letter US state or territory postal code.
county	string	County name as published by EPA.
city	string	City name as published by EPA. Null on the 19,675 sites carrying the single-space sentinel.
address	string	Street address as published by EPA. Null on the 22,261 sites carrying the single-space sentinel.
zip_code	string	ZIP code as published by EPA. Null on the 61,533 sites carrying the single-space sentinel. EPA does not zero-pad, so New England ZIPs may be four digits.
latitude	number	Site latitude in decimal degrees, read from EPA's Latitude column. Null when missing or when the value failed a coordinate-range check.
longitude	number	Site longitude in decimal degrees, read from EPA's Longitude column. Verified to agree with an outSR=4326 geometry to nine decimal places.
coordinate_status	string	ok, missing, null_island, or out_of_range_possible_projection_drift. A coordinate outside plus or minus 90/180 is rejected rather than published, because Web Mercator metres are in the millions.

Site character

Field	Type	What it means
acreage_acres	number	Site acreage. NULL when EPA published 0, because 0 is a null sentinel on this layer meaning acreage is unknown (108,888 sites), not zero acres.
acreage_raw_value	number	EPA's raw Acreage value before the null-sentinel rule is applied, so the 0 sentinel remains visible on the row.
acreage_is_null_sentinel	boolean	True when EPA published acreage 0, meaning acreage is unknown. On 108,886 of those 108,888 sites EPA also sets No Acreage PV, confirming it could not size an array because acreage was missing.
acreage_implausible	boolean	True when acreage exceeds the plausibility threshold. Some EPA acreage values are clearly not the site footprint, for example an urban Denver brownfield published at 43,000 acres. Flagged rather than dropped, because the intended figure is unknowable. Null when acreage is unknown.
acreage_implausible_threshold_acres	number	The acreage above which this actor flags acreage as implausible. Published so the flag can be re-derived.
is_known_landfill	boolean	True when EPA flags the site as a known landfill (18,298 sites). False means EPA did not flag it, which is weaker than EPA determining it is not a landfill.
known_landfill_raw	string	EPA's Landfill column verbatim: Y, or null where EPA published the single-space sentinel.
is_abandoned_mine_land	boolean	True when EPA flags the site as known Abandoned Mine Land (17,756 sites).
abandoned_mine_land_raw	string	EPA's AML column verbatim: Y, or null where EPA published the single-space sentinel.

Solar screening

Field	Type	What it means
utility_scale_pv	boolean	True when EPA's screen flagged the site for utility-scale photovoltaics (17,649 sites). This is the meaningful solar discriminator; EPA's summary Solar column is Y on all 190,976 sites and carries no information.
utility_scale_pv_raw	string	EPA's UtilPV column verbatim: Y, or null for the single-space sentinel.
distributed_scale_pv	boolean	True when EPA flagged the site for distributed-scale photovoltaics (64,268 sites).
distributed_scale_pv_raw	null	EPA's DistribPV column verbatim: Y, or null for the single-space sentinel.
offgrid_pv	boolean	True when EPA flagged the site for off-grid photovoltaics (82,090 sites).
offgrid_pv_raw	string	EPA's OffgridPV column verbatim: Y, or null for the single-space sentinel.
no_acreage_pv	boolean	True when EPA recorded that it could not size a PV array because site acreage was unknown. Co-occurs with acreage_is_null_sentinel on 108,886 sites.
est_pv_capacity_mw	number	EPA's estimated PV capacity. Read the capped and usable flags before using it: it sits on a 600 MW plateau (362 sites read exactly 600 MW) but 600 is NOT a ceiling — 16 sites exceed it, up to 11,788.08 MW, so never filter on est_pv_capacity_mw <= 600. Underneath the plateau the figure is a flat national multiple of acreage at roughly 0.1449 MW/acre, so it carries no site-specific engineering information.
est_pv_capacity_is_capped	boolean	True when EPA's estimate is at or above 600 MW. Measured live 2026-08-19: 362 sites read exactly 600 across an acreage range of 4,144 to 2,669,225 acres, and 16 more exceed 600 (maximum 11,788.08 MW), so 600 is a plateau rather than a hard ceiling. On all of those sites the number is formulaic, not a site-specific estimate. Null when EPA published no capacity.
est_pv_capacity_usable_for_ranking	boolean	False when the capacity is missing or capped, and therefore must not be used to rank or compare sites. This actor's score never uses EPA capacity for that reason.
est_pv_capacity_note	string	Why the capacity is or is not usable: published, at_or_above_600mw_plateau, or not_published.
pv_density_mw_per_acre_derived	number	EPA's estimated PV capacity divided by acreage, computed by this actor. Published so a buyer can see for themselves that EPA's capacity figure is a constant-density desktop calculation. Null when either input is unknown.
ghi_kwh_m2_day	number	Maximum annual Global Horizontal Irradiance modelled at the site. Measured range across the layer is 2.0 to 6.9.
solar_resource_class	string	Band derived from GHI: excellent (5.75+), strong (5.25+), moderate (4.75+), marginal (4.25+) or weak. Null when EPA published no GHI.

Wind screening

Field	Type	What it means
utility_scale_wind	boolean	True when EPA flagged the site for utility-scale wind (601 sites nationally).
utility_scale_wind_raw	string	EPA's UtilWnd column verbatim: Y, or null for the single-space sentinel.
community_scale_wind	boolean	True when EPA flagged the site for community-scale wind (9,770 sites).
community_scale_wind_raw	string	EPA's CommunWnd column verbatim: Y, or null for the single-space sentinel.
facility_scale_wind	boolean	True when EPA flagged the site for facility-scale wind (18,851 sites).
facility_scale_wind_raw	null	EPA's FacilWnd column verbatim: Y, or null for the single-space sentinel.
no_acreage_wind	boolean	True when EPA recorded that it could not size a wind project because site acreage was unknown.
est_wind_capacity_mw	number	EPA's estimated wind capacity. Capped at 1600 MW, which is never exceeded anywhere in the layer. Read the capped flag before using it.
est_wind_capacity_is_capped	boolean	True when EPA's estimate is at or above the 1600 MW ceiling (49 sites). Null when EPA published no wind capacity.
est_wind_capacity_usable_for_ranking	boolean	False when the wind capacity is missing or capped and therefore must not be used to rank sites.
est_wind_capacity_note	string	Why the wind capacity is or is not usable: published, at_or_above_epa_ceiling, or not_published.
wind_speed_40m_ms	number	Modelled annual average wind speed at 40 m.
wind_speed_80m_ms	number	Modelled annual average wind speed at 80 m.
wind_speed_100m_ms	number	Modelled annual average wind speed at 100 m, the hub height modern turbines actually use and the basis for wind_resource_class.
wind_resource_class	string	Band derived from the 100 m wind speed: excellent (8.5+), strong (7.5+), moderate (6.5+), marginal (5.5+) or weak. Null when no wind speed was published.

Biomass screening

Field	Type	What it means
biopower_potential	boolean	True when EPA flagged the site as a candidate for a biopower facility (30,909 sites).
biopower_potential_raw	string	EPA's BioPwrPot column verbatim: Y, or null for the single-space sentinel.
biorefinery_potential	boolean	True when EPA flagged the site as a candidate for a biorefinery (3,667 sites).
biorefinery_potential_raw	string	EPA's BioRefPot column verbatim: Y, or null for the single-space sentinel.
cumulative_biopower_residues_tpy	number	Cumulative biopower residues available within 50 miles of the site, in metric tons per year.
cumulative_biorefinery_residues_tpy	number	Cumulative biorefinery residues available within 50 miles, in metric tons per year.
crop_residues_tpy	number	Crop residues available within 50 miles, in metric tons per year.
forest_residues_tpy	number	Forest residues available within 50 miles, in metric tons per year.
primary_mill_residues_tpy	number	Primary mill residues available within 50 miles, in metric tons per year.
secondary_mill_residues_tpy	number	Secondary mill residues available within 50 miles, in metric tons per year.
urban_waste_wood_tpy	number	Urban waste wood residues available within 50 miles, in metric tons per year.

Geothermal screening

Field	Type	What it means
geothermal_heat_pump_potential	boolean	EPA's geothermal heat pump flag. Note that this is Y on 190,956 of 190,976 sites (99.99%), so it is effectively a constant and is not offered as a filter.
geothermal_heat_pump_raw	string	EPA's GeoHeatPmp column verbatim: Y, or null for the single-space sentinel.
surface_temperature_c	number	Modelled annual average surface temperature at the site, used by EPA in its geothermal heat pump screen.
permafrost_present	boolean	True when EPA flagged permafrost at the site. Only 20 sites nationally, all in Alaska.
permafrost_raw	null	EPA's PermFrost column verbatim: Y, or null for the single-space sentinel.

Infrastructure (EPA 2022 screening estimates, NOT live)

Field	Type	What it means
substation_distance_miles_2022_screening_estimate	number	EPA's 2022 precomputed distance to the nearest substation. A FROZEN SCREENING ESTIMATE, not a live measurement. Run hifld-grid-proximity-screener on this row's coordinates for the current figure.
substation_voltage_volts_2022_screening_estimate	number	EPA's 2022 precomputed nearest substation voltage, published by EPA in the Volts column though the values are in kV magnitude. Null on sites where EPA published none.
substation_voltage_kv_derived	number	The substation voltage restated for convenience. EPA's column is labelled Volts but publishes values such as 115 and 230, which are kV, so this actor does not rescale it and says so.
transmission_distance_miles_2022_screening_estimate	number	EPA's 2022 precomputed distance to the nearest transmission line. A frozen screening estimate, not a live measurement.
transmission_line_kv_2022_screening_estimate	number	EPA's 2022 precomputed nearest transmission line voltage in kilovolts.
transmission_line_status_2022_screening_estimate	string	EPA's 2022 precomputed status of the nearest transmission line. Measured vocabulary: IN SERVICE, INACTIVE, NOT AVAILABLE, UNDER CONST.
transmission_status_contradicts_voltage	boolean	True when EPA reports the nearest line status as NOT AVAILABLE while also publishing a voltage above zero. This internal contradiction affects 16,803 sites and is surfaced rather than silently resolved. Null when either input is missing.
nearest_interconnection_distance_miles_2022	number	The smaller of EPA's 2022 substation and transmission distances. A frozen screening estimate used for ranking only.
road_distance_miles	number	EPA's precomputed distance to the nearest graded road.
rail_distance_miles	number	EPA's precomputed distance to the nearest rail line.
interconnection_data_is_live	boolean	Always false. Every interconnection figure on this row comes from EPA's frozen 2022 precomputation, never from a live grid query.
interconnection_note	string	Full-text warning that EPA's grid fields are a 2022 precomputation, are internally inconsistent on 16,803 sites, and must not be used to size an interconnection.
verify_interconnection_with	string	The sibling actor to run on this row's coordinates for a live line and substation measurement.

Ranking

Field	Type	What it means
technologies_flagged	array	Every renewable technology EPA's screen flagged for this site, from the scale-specific columns: utility_solar, distributed_solar, offgrid_solar, utility_wind, community_wind, facility_wind, biopower, biorefinery.
technologies_flagged_count	number	Count of technologies EPA flagged for this site.
best_flagged_technology	string	The flagged technology with the greatest typical project value, ordered utility_solar, utility_wind, distributed_solar, community_wind, facility_wind, biopower, biorefinery, offgrid_solar. Null when EPA flagged none.
development_candidate_score	number	An EDITORIAL 0-100 ranking score computed by this actor from EPA's screening inputs: resource, site size, 2022 interconnection proximity, road access and reuse archetype. Rescaled to the components that were actually available. NULL, never 0, when no component could be computed. This is a ranking aid, not a feasibility determination.
score_components_available	array	Which of the five scoring components had inputs on this site.
score_components_missing	array	Which scoring components could not be computed because EPA published no input.
score_weight_available	number	How much of the 100-point weight was actually computable on this site. A score built on 45 of 100 available weight is far weaker evidence than one built on 100.
score_is_incomplete	boolean	True when at least one scoring component could not be computed, so the score rests on partial inputs.
score_basis	string	Plain-language statement of exactly how this row's score was computed and what was excluded from it.

Honesty and provenance

Field	Type	What it means
epa_screening_only	boolean	Always true. Everything on this row is EPA's desktop pre-screen output.
is_entitlement	boolean	Always false. An EPA RE-Powering screening result confers no permit, no land right, no interconnection position and no approval of any kind.
screening_disclaimer	string	Full-text statement that EPA RE-Powering is a desktop pre-screen and not an entitlement, feasibility study or site-control position.
flag_semantics_note	string	Full-text explanation that EPA publishes no explicit negative value, so a false on any flag means EPA did not flag it rather than EPA determining the negative.
capacity_ceiling_note	string	Full-text explanation of EPA's 600 MW PV plateau (which 16 sites exceed, so it is NOT a ceiling), the genuine 1600 MW wind ceiling, and the constant-density national basis of both estimates. Every figure in this sentence is asserted live against the layer by the pv_plateau_integrity drift probe before any row is billed.
epa_site_info_url	string	EPA or state program page for the underlying cleanup site.
epa_repowering_profile_url	string	EPA's own RE-Powering Site Profile page for this site, where every value on this row can be independently checked against EPA.
data_vintage	string	The vintage of EPA's RE-Powering Mapper layer. EPA states it in the service name itself.
layer_title	string	The EPA ArcGIS layer this row came from.
source_url	string	The EPA ArcGIS FeatureServer layer endpoint queried for this row.
checked_at	string	UTC timestamp when this run queried EPA.

Run context (repeated on every row)

Field	Type	What it means
matched_sites_total	number	The exact number of sites matching the query, read from EPA's free returnCountOnly endpoint before any row was fetched. Compare with the row count to see whether the answer was capped.
results_truncated	boolean	True when more sites matched than were returned, because maxResults capped the run. A capped answer can therefore never be mistaken for a complete one.
max_results_requested	number	The maxResults cap in force for this run.
filters_applied	array	Every filter this run actually applied, so a buyer can confirm the query they thought they asked for is the query that ran.
where_clause	string	The exact SQL WHERE clause sent to EPA's FeatureServer. Published so the result is reproducible.
drift_gate_status	string	verified when all six live drift probes passed, or verified_degraded when every load-bearing probe passed but a corroborating probe could not complete. A run never emits rows in any other state.
drift_probes_verified	array	Which drift probes completed and measured what was expected on this run.
drift_probes_unavailable	array	Which drift probes could not complete (an outage, not a wrong measurement). A probe that could not complete is not a probe that failed, and only corroborating probes may be unavailable on a run that emits rows.
drift_gate_note	string	Plain-language summary of what the live drift gate verified before any billable row was produced.
layer_status	string	ok when every drift probe passed, ok_degraded when a corroborating probe was unavailable. A failed layer never reaches this field because the run fails first.

---

FAQ

Where does this data come from? EPA's RE-Powering America's Land initiative, published as the ArcGIS layer RE-Powering Mapper Sites 2022 (services.arcgis.com/cJ9YHowT8TU7DUyn/.../RE_Powering_Mapper_Sites_2022/FeatureServer/0). 190,976 sites, 55 published columns, verified live.

How current is it? The layer is EPA's 2022 vintage and EPA states that in the service name. Every row carries data_vintage: "2022". There is no newer EPA-published copy at the time of writing. The grid fields in particular are a 2022 precomputation — see trap 5 above.

Does a site appearing here mean I can build a solar farm on it? No. It means EPA's desktop screen did not rule the site out. Nothing here confers a permit, a land right, an interconnection position or a cleanup determination. That is why every row carries is_entitlement: false.

Why is est_pv_capacity_mw sometimes exactly 600? Because 600 MW is a plateau EPA applies to large sites, not a site-specific estimate. 362 sites read exactly 600 across a 644x range of acreage. Those rows carry est_pv_capacity_is_capped: true and est_pv_capacity_usable_for_ranking: false, and the ranking score ignores EPA capacity entirely.

Is 600 MW a hard ceiling on est_pv_capacity_mw? No — and this is the one number to be careful with. 16 sites publish more than 600 MW, up to 11,788.08 MW. If you filter with est_pv_capacity_mw <= 600 you will silently discard the 16 largest solar-redevelopment candidates in the layer. Filter on est_pv_capacity_is_capped instead, which is true at or above the plateau and covers both groups. (EstWindCap's 1,600 MW ceiling is genuine — zero sites exceed it. That asymmetry is why the PV figure is checked live on every run.)

Why is acreage_acres null on a site that clearly has land? Because EPA published acreage 0, which on this layer means "unknown", not "zero". Check acreage_raw_value and acreage_is_null_sentinel.

Why is the site name blank? 1,675 sites publish no name — mostly state abandoned coal mine areas identified only by a program number. Those rows carry site_name_withheld_by_epa: true and still have a county, coordinates and an EPA profile URL.

Can I filter on EPA's Solar column? No, and you should not want to: it is Y on every one of the 190,976 sites. Use the technologies input, which filters on EPA's scale-specific columns.

How do I get a real interconnection distance? Run hifld-grid-proximity-screener on the latitude/longitude in the row. Every row names it in verify_interconnection_with.

What happens if EPA changes or retires the layer? The run fails loudly with a message naming the probe that detected it, and bills nothing. The gate catches a collapsed row count, a moved canary site, a change to EPA's null convention, a filter that stopped being applied, a retired service, and a change to the EstPVCap distribution this actor describes in prose on every row.

Can I verify a row against EPA directly? Yes — every row carries epa_repowering_profile_url, EPA's own RE-Powering Site Profile page for that exact site.

Is a zero-row result a bug? Only if you did not expect it. This actor never returns an empty dataset because something broke; it fails instead. Zero rows means EPA holds no site matching your filters, and matched_sites_total will read 0 to confirm it.

---

Data source and licence

US EPA RE-Powering America's Land, Cleanups in My Community (CIMC). US federal government data, public domain. This actor adds normalisation, sentinel handling, ranking and disclosure; it does not alter EPA's underlying values, all of which are re-published verbatim in the *_raw columns and are checkable against epa_repowering_profile_url.