Environmental Records Report - ASTM E1527-21 records-search appendix, one PDF per property

One address or coordinate in, one Environmental Records Report out - as HTML, as PDF and as a machine-readable JSON model - for $2.50 per report (list price) plus the child screeners' own per-row charges (measured $0.98-1.32 per property, never more than $2.40) - so one report fits inside the Apify free plan's $5 monthly credit. The report is the records layer a Phase I Environmental Site Assessment consultant attaches as an appendix or runs as a desktop pre-screen before ordering a full commercial database package: federal records at the ASTM E1527-21 Sec. 8.2.1 standard search distances, Sec. 8.3 prior industrial uses, Sec. 8.3.4 historical topographic map coverage, underground storage tank and spill registries, and a 20-layer site screen - with a radius map, per-layer data vintages, a live source-verification (drift gate) page and an explicit "what this report does NOT include" page.

It is composed by Actor-to-Actor calls to five published malonestar screeners on their latest builds, so every table in the report is exactly what those actors return today, and every one of them verifies its government upstream before it emits a row.

Who it is for

• Environmental consultants preparing Phase I ESAs (ASTM E1527-21 / 40 CFR Part 312) who want the federal and state records tables in a citable, dated, per-property document rather than as CSV rows.
• CRE lenders, brokers and acquirers doing a desktop pre-screen before paying for a full commercial environmental database report (published price lists put those at $250-415 - see the price section).
• Portfolio owners who need one honest one-page-per-layer document per site, including an explicit list of the record classes that were not searched.

What you get per property

Artefact	Where	Notes
report-NN-<label>.pdf	run key-value store, URL in report_pdf_url	Letter, ~15 pages, pure-JS render (no browser)
report-NN-<label>.html	run key-value store, URL in report_html_url	self-contained, inline SVG radius map, print-ready
report-NN-<label>.json	run key-value store, URL in report_json_url	the full report model every table is rendered from
one dataset row	default dataset	status, URLs, headline counts, per-layer status, child run ids, stacked cost

The run key-value store is private to your account: open the report_*_url links with your API token appended (?token=...), or click through report_console_url while signed in to the Console. A token-less request returns HTTP 403 - the report is never world-readable by URL.

Report pages

1. Cover - property, coordinates and how they were resolved, state and how it was resolved, radii, timestamp, run id, report status, headline findings, a layers-and-vintages table and the radius map (ASTM rings at 0.25 / 0.5 / 1 mi; every record the agency published a coordinate for is plotted; records without a coordinate are counted, not invented).
2. ASTM E1527-21 Sec. 8.2 federal records - the eight ASTM databases (NPL, delisted NPL / SEMS-CERCLIS, RCRA CORRACTS, RCRA TSD, RCRA generators, LUST, UST, Brownfields) each at its own standard search distance with count, nearest distance and result; whether the property sits inside a published Superfund site boundary or institutional-control boundary; the nearest-first records table (10 nearest per program, ECHO enforcement and penalty detail where EPA publishes it).
3. Sec. 8.3 prior uses - EPA FRS facilities within the adjoining-property radius carrying a contaminating SIC/NAICS classification (dry cleaners, wood preserving, plating, petroleum bulk storage, ...), the contaminants the cited authorities associate with each use, how many in-radius facilities carry no classification at all (never counted as clean), and the live inventory of state and tribal equivalent record systems EPA normalises for the property's state - including the twelve jurisdictions where that inventory is empty.
4. Sec. 8.3.4 historical topographic coverage - every USGS map year covering the property (1880s-present), the five-year interval strip for the review window, data gaps, longest uncovered span, and a chronology with a GeoPDF link per year.
5. Tank and spill registries - EPA UST Finder nationally plus NY DEC, CA GeoTracker, TX TCEQ and FL DEP, nearest first, open/closed with the basis for that status, and a plain statement of whether the property's state has a state-registry leg at all.
6. 20-layer site screen - the site-due-diligence-bundle verdict, score and flags, with each layer's status and data vintage.
7. Live source verification (drift gate) - the differentiator: per layer, the child's gate status, how many live checks ran and failed, which probes passed (positive canaries, negative controls, row-count floors, vocabulary checks), and what exactly was verified. A layer whose gate did not pass is shown as unavailable, never rendered as a clean result.
8. What this report does NOT include - historical aerials, Sanborn maps, city directories, state records beyond EPA FRS and the four state tank registries, coordinate-less records, site reconnaissance, a professional's interpretation, a signed E&O deliverable, AAI / 40 CFR Part 312 compliance.
9. Appendix - every child run id and build, rows returned per layer, and the estimated child charges for this property.

Honesty contract - three-valued everywhere

• Each layer is complete (the child answered and its own completeness flags say every source answered), partial (the child answered but disclosed a failed, truncated or deadline-cut source - its "none found" statements are unverified for those sources), unavailable (the child did not answer for this property - the layer was NOT screened and the report says so on a red page), or not_requested.
• A report is complete only when every requested layer is complete; partial when at least one layer answered; failed when no layer answered - then nothing is rendered and nothing is charged.
• Every headline count is null when its layer was not screened - never 0.
• If no report at all could be generated for any property, the run FAILS with a plain status message and no charge.
• The run checks its own charge limit before calling any child: the limit must cover the report fee plus the children's worst-case charges for each property ($2.50 + $2.40 = $4.90 for one property on the free plan). If it cannot, the run fails immediately and calls no child, instead of spending your budget on children for a report it cannot then bill. Each child call also carries its own charge cap, set just above the most that child can bill, so the stacked bill can never exceed what the pre-flight checked.

Pricing - read this first

Event	List price (FREE plan)	Bronze / Silver / Gold and above
report-generated - one per property with a complete or partial report whose HTML/PDF were stored	$2.50	$2.00 / $1.75 / $1.375 (Platinum and Diamond are billed at the Gold price)
dataset row	none - the row is the report's index, not a second charge	-

The bill is stacked, and this is stated plainly: the five child screeners run on your account and bill their own per-result prices in addition to the report fee. Measured on the prefill property (Denver, 1 mi, all five layers, 2026-09-26):

Child (latest build)	Rows	List price	Cost
malonestar/epa-contaminated-site-screener	68	$0.010/row	$0.68
malonestar/historic-land-use-sic-contaminant-screener	4	$0.012/row	$0.05
malonestar/usgs-historical-topo-records-review	1	$0.010/row	$0.01
malonestar/state-tank-spill-registry-screener	20	$0.010/row	$0.20
malonestar/site-due-diligence-bundle	1	$0.100/row	$0.10
Children total			~$1.04
Report fee			$2.50
All-in, one property, list prices			~$3.54

Measured again on 2026-10-06 across four dense industrial parcels (Denver, Newark NJ, Vernon CA, Houston ship channel): $0.98-1.32 in child rows per property. A rural parcel returns a handful of rows and costs about $0.30. The hard ceiling is $2.40 per property (every child at its own row cap, plus a 15% margin) - that is the figure the pre-flight budgets, so one property costs at most $4.90 all-in at list prices. Each dataset row carries child_rows_usd_estimate and total_usd_list_estimate for the property it describes. Paid Apify plans pay the tiered price on both the report fee and the children.

How this compares with a commercial environmental database report

A3 Environmental Consultants publishes a dated price list for the equivalent commercial packages (a3e.com, "What does an EDR report cost?", updated 2026-05-29): about $250 for a desktop environmental screen and $375-415 for a database report package, delivered in 48-72 hours. Those packages include things this report does not - historical aerial photographs, Sanborn fire-insurance maps, city directories, full state and tribal coverage and an analyst's review - and this report includes things they typically do not: the live source-verification page, per-layer data vintages measured at run time, the ASTM per-database table computed from EPA's own coordinate-bearing records, and a machine-readable model. It runs in about a minute. Read the not-included page before relying on it for anything the standard requires.

Input

{
  "properties": [
    { "lat": 39.8037, "lon": -104.9986, "state": "CO", "label": "Denver industrial parcel (prefill)" },
    { "address": "5980 Lipan St, Denver, CO 80221", "label": "Same parcel by address" }
  ],
  "layers": ["federal_records", "prior_uses", "historical_topo", "tank_spill", "site_screen"],
  "radiusMiles": 1,
  "priorUseRadiusMiles": 0.25,
  "topoReviewStartYear": 1940,
  "includePdf": true
}

• properties (required) - address (geocoded with the free US Census geocoder) or lat/lon, optional label and state hint. A property that cannot be resolved gets a not_screened row and is not charged.
• layers - default all five. Dropping a layer drops its child call and its page.
• radiusMiles - 0.25-5, default 1 (the widest ASTM Sec. 8.2.1 distance; the ASTM table always uses each database's own distance).
• priorUseRadiusMiles - 0.02-1, default 0.25 (subject and adjoining properties).
• topoReviewStartYear - 1880-2020, default 1940.
• includePdf - default true.
• maxProperties - optional cap; properties beyond it are not_screened and not charged.

The prefill runs one Denver property through all five layers in roughly 60-90 seconds and costs one report fee plus about $1 of child rows.

Output fields (one row per property)

Field	Type	Meaning
property_label	string	Your label for the property, echoed back (or the geocoded address when no label was supplied).
property_address	string	The address you supplied, or null when the property was given as lat/lon.
property_lat	number	Latitude actually screened (WGS84), after geocoding if an address was supplied. Null when the property could not be resolved.
property_lon	number	Longitude actually screened (WGS84). Null when the property could not be resolved.
property_state	string	Two-letter state the property resolved to (supplied hint, Census geocoder, or FCC block API). Null when unresolved - the tank/spill state-registry coverage is then unknown and says so.
property_resolution	string	"lat_lon_supplied" or "census_geocoded". Null when the property could not be resolved.
radius_miles	number	Radius used for the federal records, tank/spill and site-screen layers.
prior_use_radius_miles	number	Radius used for the ASTM Sec. 8.3 prior-uses layer.
generated_at	string	UTC timestamp (ISO 8601) the report was generated.
run_id	string	This Apify run id, printed on the report cover.
report_status	string	"complete" = every requested layer answered in full; "partial" = at least one layer answered but one or more layers are unavailable or partial (read layers_unavailable / layers_partial); "failed" = no requested layer answered, no report was rendered and NOTHING was charged; "not_screened" = the property could not be resolved or was beyond the cap, no report, no charge.
not_screened_reason	string	Why no report exists for this property (report_status failed or not_screened). Null when a report was generated.
report_html_url	string	API URL of the HTML report in this run's key-value store. The store is private to your account: open the URL with your API token appended as ?token=... or use report_console_url. Null when no report was generated.
report_pdf_url	string	API URL of the PDF report (same access rule as report_html_url). Null when includePdf was false or no report was generated.
report_json_url	string	API URL of the JSON report model the HTML and PDF were rendered from - every table in the report, machine-readable (same access rule).
report_console_url	string	Apify Console page of this run's key-value store, where the HTML, PDF and JSON records can be opened and downloaded while signed in.
report_urls_require_token	boolean	Always true: the run key-value store is private to the account that ran the report, so the report_*_url links need ?token=
report_html_key	string	Key-value store record key of the HTML report.
report_pdf_key	string	Key-value store record key of the PDF report.
report_json_key	string	Key-value store record key of the JSON model.
artefact_error	string	Set when the report could be composed but storing HTML/PDF failed. The report is then NOT charged.
layers_requested	array	Layer keys requested for this run.
layers_complete	array	Layers whose child answered in full for this property.
layers_partial	array	Layers whose child answered but disclosed a failed, truncated or deadline-cut source. Their "none found" statements are unverified for the sources named.
layers_unavailable	array	Layers whose child did NOT answer for this property. They were NOT screened; the report says so on a red page.
layer_status_federal_records	string	"complete" / "partial" / "unavailable" / "not_requested" for the ASTM Sec. 8.2 federal-records layer (malonestar/epa-contaminated-site-screener).
layer_status_prior_uses	string	Same vocabulary, for the ASTM Sec. 8.3 prior-uses layer (malonestar/historic-land-use-sic-contaminant-screener).
layer_status_historical_topo	string	Same vocabulary, for the historical topographic coverage layer (malonestar/usgs-historical-topo-records-review).
layer_status_tank_spill	string	Same vocabulary, for the tank/spill registry layer (malonestar/state-tank-spill-registry-screener).
layer_status_site_screen	string	Same vocabulary, for the 20-layer site screen (malonestar/site-due-diligence-bundle).
astm_databases_flagged	number	How many of the ASTM E1527-21 Sec. 8.2.1 databases have at least one record within their standard search distance. Null when the federal layer was not screened - never 0 in that case.
astm_databases_measured	number	How many ASTM databases the federal layer actually measured (8 when complete). Null when not screened.
federal_listings_within_radius	number	EPA program records within radius_miles, capped at 10 nearest per program. Null when not screened.
on_superfund_site	boolean	True when the property coordinate falls inside a published EPA NPL Superfund site boundary polygon; false when checked and outside; null when not checked.
superfund_site_name	string	Name of the containing Superfund site when on_superfund_site is true.
prior_use_verdict	string	The Sec. 8.3 verdict from the prior-uses child: "contaminating_use_identified" / "no_classified_contaminating_use_found" / "inconclusive" / "not_screened" ... Null when the layer was not screened.
prior_use_facilities_counted	number	Contaminating-use facilities with a site-specific coordinate that counted toward the Sec. 8.3 verdict. Null when not screened.
topo_coverage_pct	number	Share of the ASTM Sec. 8.3.4 review-window intervals covered by at least one USGS topographic map. Null when not screened or when a USGS collection did not answer.
topo_map_years	number	Distinct map years covering the property across the USGS collections. Null when not screened.
topo_data_gap_present	boolean	True when at least one review-window interval has no covering map; null when not determined.
tank_spill_records	number	Tank/spill records within radius (up to 20 nearest). Null when not screened.
tank_spill_open	number	Of those, how many the registry reports as open. Null when not screened.
tank_state_registry_covered	boolean	True when the property's state has a state-registry leg in the tank/spill child (NY, CA, TX, FL); false when only the national EPA UST Finder baseline was searched for it; null when the state is unknown.
site_verdict	string	The 20-layer bundle verdict: "proceed" / "caution" / "no-go". Null when not screened.
site_score	number	The bundle's composite 0-100 site score. Null when not screened.
child_run_ids	object	Apify run id of each child screener call, keyed by layer. Null for a layer that was not called.
child_builds	object	Build number each child ran on (always its latest build at run time), keyed by layer.
child_rows	object	How many child rows this property received from each layer - the basis of the child cost estimate.
drift_gate_summary	string	One line per layer summarising the live source verification each child reported (the drift-gate page of the report).
report_charged	boolean	True when one report-generated event was charged for this property. False when the report was not generated, artefacts could not be stored, or the charge did not go through (see report_charge_note).
report_charge_event	string	Always "report-generated" - the name of the pay-per-event event this actor charges.
report_charge_note	string	Why the report was not charged, when it was not. Null when charged.
report_fee_usd_list	number	The report fee at the FREE-tier list price when charged ($2.50), else 0. Paid Apify plans pay the tiered price.
child_rows_usd_estimate	number	Estimated charges the five child screeners billed to your account for this property: their rows x their FREE-tier list prices.
total_usd_list_estimate	number	report_fee_usd_list + child_rows_usd_estimate.
run_properties_requested	number	How many properties were supplied to this run.
run_reports_generated	number	How many reports (complete or partial) this run generated.
run_complete	boolean	True only when every supplied property produced a complete report.
disclaimer	string	Scope statement: records appendix / desktop pre-screen, not a Phase I ESA, not AAI.

How it is built

• geocode.js, children.js, compose.js, render.js, pdf.js, billing.js are pure modules; main.js is wiring only and is asserted as text.
• Fixtures are the five children's live datasets captured verbatim (run ids inside the fixture files); the offline suite renders the fixture report to HTML and PDF and exercises every degraded path (a child down, a child answering other properties but not this one, every child down, layers not requested).
• A mutation harness re-injects the defects this actor is built to prevent (charging a failed report, rendering an unavailable layer as clean, defaulting a null count to 0, skipping the budget pre-flight, pushing before the gate) and requires the suite to go red on each.

Use as an MCP tool

Published Apify Actors are callable by AI agents through https://mcp.apify.com (Claude Desktop, Claude Code, Cursor and any MCP client). An agent can pass a property and receive the report URLs plus the headline counts in one call. The report fee and the children's charges apply exactly as above, and a run that cannot answer fails without billing.

FAQ

Is this a Phase I ESA? No. It is a records-search appendix / desktop pre-screen. It does not perform AAI under 40 CFR Part 312, carries no professional interpretation and confers no CERCLA landowner liability protection. The not-included page lists what a full Phase I adds.

Why did my report come back partial? One of the five child screeners did not answer for that property - usually a government service outage. The report names the layer on a red page and in layers_unavailable, and the drift-gate page shows what the other children verified. You are charged the report fee for a partial report (the other layers are real work); you are never charged for a failed report.

Why is the state tank registry "NOT searched" for my property? The tank/spill child has state legs for New York, California, Texas and Florida only; elsewhere it searches the national EPA UST Finder baseline and the report says so rather than implying state coverage.

Can I get the tables as data? Yes - report_json_url is the full model every table is rendered from, and the five children's own datasets are linked by run id in the appendix.

Does the free plan cover a report? Yes - one property. The $2.50 report fee plus the children's $2.40 worst case is $4.90, inside the free plan's $5 monthly credit (typical all-in is about $3.50). A run whose remaining limit cannot cover that fails before calling any child and bills nothing.