Entries a–f: geocoding log

127 entries processed (every file in pinebarrens/entry/ starting a–f). Output: entries-a-f.json.

Counts by precision

  • site: 42
  • area: 37
  • locality: 32
  • none: 16

(Updated after a second Nominatim pass late in the session recovered several previously-failed lookups — see below.)

Sources used

  • USGS GNIS domestic-names extract for New Jersey (local file, 7,553 records) — primary source for populated places, civil units, streams and military installations. Cited in basis as “USGS GNIS domestic names gazetteer (NJ), local extract”.
  • NJ Pine Barrens community-map dataset (njpb-locations.json, 105 named Pines localities) and points-of-interest file (njpb-poi.json, 10 POIs, including the Carranza Memorial) — used for historic Pines place names not in GNIS (Atsion, Chairville, Elm, Flyat, Friendship, Hanover Furnace, Indian Mills, Pasadena, Tabernacle, Whitesbog, etc.).
  • OpenStreetMap Nominatim — attempted for ~20 specific addresses/landmarks. Nominatim was rate-limited/blocked (HTTP 429) for most of the session despite retries with backoff and a descriptive User-Agent. Two early queries succeeded (Cape May County Airport; Manasquan), then the block held through a large batch of retries, then a later batch (run in the background while the rest of this pass was completed) recovered and returned real results for: Coyle Field, Franklin Parker Preserve, Florida NY, Chatsworth Lake (cross-check only), Barnegat Lighthouse State Park, North Virginia Avenue (Atlantic City), Old Indian Mills Road (Tabernacle), Burns Street (Camden), Watchung Avenue (Plainfield, exact house number), Amatol Road, and Powell Place Road. Those entries were upgraded accordingly. Paulsdale, Highland Road (Cinnaminson), Center Street (Elizabeth), West Papoose Swamp, and Bodine Field still returned no results even in the recovered window and remain downgraded to the best GNIS/njpb-supported level, with a note explaining the gap. No coordinate was ever invented to stand in for a failed lookup.

Entries marked “none” (16) and why

  • Regional science/nature studies with no single site (9): atlantic-coast-leopard-frog-remapping, cedar-swamp-hardwood-succession-study, chaffseed-dormancy-long-census, corn-snake-headstarting-experiment, cranberry-scent-color-trap-experiment, cranberry-stream-invertebrate-recovery, cranberry-white-grub-nematode-trial, created-ponds-pathogen-study, fence-lizard-thermoregulation-comparison.
  • Folklore/travelogue spanning many places, no single defensible pin: captain-kidd-jersey-shore-legend (four separate legendary sites), friendship-etc and beck-tuckerton-road (Henry C. Beck route digests naming a dozen-plus stops each).
  • Unresolved/unlocatable specific place names: carolina-u-151-loss-1918 (“off the New Jersey coast,” no coordinates or distance given), della-penna-dolores-search-1972 (remains recovered somewhere in the “South Jersey Pine Barrens,” unnamed), dunham-jercy-1969-plane-crash (aircraft never found; only a general shore-area description).
  • No NJ Pine Barrens place at all: cool-ads (a Philadelphia-area car dealer’s advertisement).

black-saturday-forest-fires-1963 (Coyle Field) and florida-wildman-1881-monmouth-democrat (Florida, NY) were initially placed in this list but were later resolved once Nominatim recovered mid-session; both now carry real coordinates (see counts above).

Notable problems / likely errors found in the source entries (flagged for review)

  1. bear-newark-1906 — slug and page title say “Newark,” but the eyebrow and body both identify Kearny as the reported location. Pin uses Kearny; title/place mismatch should be fixed at the source.
  2. beast-1887-camden — the OCR’d place name reads “Owensburgh,” which does not match any place in the gazetteer; the eyebrow separately calls it “the Winslow-area predator.” These two labels don’t obviously reconcile. Pinned generally to Winslow Township with a wide radius; the specific place name is likely a garbled OCR transcription worth re-checking against the original clipping.
  3. brown-bradway-1937-murdered vs. brown-bradway-palmyra-shot-1933 — two separate entries appear to describe the same 1933 killing of Bradway Brown, but give conflicting place labels (Cinnaminson Township/Highland Road vs. Palmyra). The body text of the first reconciles this as “Highland Road in the Palmyra–Riverton area,” suggesting the two places are adjacent/overlapping, but the entries’ factbox “Place” fields disagree and should be reviewed together.
  4. elizabeth-kite-pineys-report-1913 — the factbox “Place” field names Vineland (the author’s institution) rather than the Pine Barrens communities the report actually describes; pinned to Vineland with a note clarifying this is the author’s base, not the report’s subject.
  5. A late-session Nominatim recovery resolved several street/landmark addresses that failed on the first pass (see Sources above): Virginia Avenue (Atlantic City), Old Indian Mills Road (Tabernacle), Burns Street (Camden), Watchung Avenue (Plainfield, exact house number), Amatol Road, Powell Place Road, Franklin Parker Preserve, and Coyle Field were all upgraded. A few addresses still returned nothing even after recovery and remain downgraded to the containing locality/area, flagged for a future pass: Paulsdale (Mount Laurel), the Red Lion Motel (Southampton), 232 Center Street (Elizabeth), and West Papoose Swamp.

Anchors worth double-checking later

  • bodines-tavern-tuckerton-stage-road uses the GNIS point where the “Oswego River” and “West Branch Wading River” stream records coincide (39.6581, -74.5277) as a stand-in for the historic river-forks tavern site — the two independent GNIS records land almost exactly on top of each other, which is a good sign, but this is an inferred confluence point, not a listed “Bodine(‘s) Tavern” or “Bodine Field” feature.
  • carslake-charles-l-1933 uses “Papoose Branch” (the nearest listed GNIS stream) as an anchor for “West Papoose Swamp,” which is not itself listed.

Validation (root fact-check)

Independent review run 2026-09-11 against the entry pages, the local USGS GNIS NJ extract, njpb-locations.json / njpb-poi.json, and OpenStreetMap Nominatim (which recovered part-way through and was then used at ~1 request / 2 s). Every one of the 42 site records was checked; all 111 records carrying a coordinate were additionally swept for wrong-county, wrong-same-named-place, in-water and transposed/sign-flipped coordinates.

Headline: of 42 site pins, 7 survived as sites. One pin was flat wrong by ~36 km. No transposed or sign-flipped coordinates were found, and every gazetteer-derived coordinate faithfully matched the feature its basis named – the systematic fault was not bad arithmetic but calling a settlement centroid a site: 34 records claimed site precision for a GNIS populated-place centroid or similar, when the thing the entry is actually about (a meetinghouse, a mill, a depot, a street address, a wildlife management area) was never located at all.

Worst error

cedar-bridge-tavern-battle-site – pinned in the wrong township, ~36 km off. The record used the GNIS ‘Cedar Bridge’ populated place at 40.0559496,-74.1281952, which reverse-geocodes to Brick Township. The Cedar Bridge Tavern of the 1782 skirmish is in Barnegat Township at 39.7832769,-74.3606379. This is precisely the trap the entry’s own source warns about – Ocean County’s history cautions that ‘numerous local places contain some combination of Cedar, Creek and Bridge’ – and the geocoder walked straight into it while marking the result site. Corrected to the tavern and set to area (r=2000), because the same source says the skirmish’s exact location is uncertain and the standing tavern post-dates the fighting.

Verified – kept at site (7)

  • atsion-mansion-fruitland – re-anchored from the njpb ‘Atsion’ locality centroid to the Nominatim ‘Atsion Mansion’ building (~40 m); stays ‘site’
  • atlantic-city-airlines-cape-may-1976 – Nominatim aerodrome match re-confirmed; note added that the crash was ~1.2 km short of the runway
  • carranza – njpb POI cross-checked against Nominatim, agreement ~60 m; genuine monument, stays ‘site’
  • barnegat-lighthouse-changing-coast – coordinate is the lighthouse, cross-checked to ~120 m; the leftover note still described a borough-centroid pin and was rewritten
  • atlantus-ferry-pier-wreck-1926 – GNIS beach feature reverse-geocodes correctly to Sunset Beach; note added that the wreck itself is offshore
  • escaped-monkey-plainfield-1948 – independently reverse-geocoded back to 63 Watchung Avenue, Plainfield; genuine street-address site pin
  • double-trouble-village – GNIS feature is the compact preserved village; stays ‘site’

Corrected (4)

  • cedar-bridge-tavern-battle-site – corrected: coordinate moved ~36 km from GNIS ‘Cedar Bridge’ in Brick Twp to the actual Cedar Bridge Tavern in Barnegat Twp; site -> area (r=2000) because the entry itself says the skirmish location is uncertain
  • friendship-bogs-school-preservation – label: record said ‘Tabernacle Township’ but the coordinate is in Washington Township; note added that the preserved building now stands at Tabernacle, away from the pin
  • burlington-bookmobile-rural-readers – basis: basis falsely cited the GNIS extract for a hand-rounded 39.85,-74.65 county-centre guess, via an incoherent ‘Camden County style civil centroid substitute’; rewritten to say plainly what it is
  • elizabeth-kite-pineys-report-1913 – note/name: pin marks the author’s institution, not the communities described; name and note now say so explicitly, per reviewer request

Downgraded (35)

site -> locality (22) – in every case the coordinate is a settlement/locality centroid and the building, address or feature the entry is about was never independently located:

  • alliance-colony-crystal-farm-diary – colony of dispersed farmsteads; GNIS point is the settlement centroid, not a located farm
  • arneys-mount-meeting-intermittent-use – meetinghouse is not separately listed; pin is the settlement centroid
  • barclay-farmstead-sale-subdivision – GNIS ‘Barclay Farm’ is the mid-20th-century residential development named after the farm, not the preserved Barclay Farmstead building the entry is about
  • batsto-iron-glass-community – pin is the village centroid (~230 m from Batsto Village Road per Nominatim), not a single building
  • beebe-david-1907-murdered – the Mathis house where the fight began is not located; pin is the Chairville settlement centroid
  • bender-craig-1940 – basis already conceded this is the borough centroid, not the Nautilus/Ocean avenues intersection; the remains were also found about two miles away
  • berlin-long-a-coming-depot-farmers – the Long-a-Coming depot building is not separately located (Nominatim returned no match); pin is the Berlin centroid
  • bivalve-oyster-shipping-sheds – the shipping sheds are not separately located; pin is the Bivalve settlement centroid
  • brotherton-land-statement-1780 – Brotherton/Edgepillock was a dispersed community; pin is the Indian Mills centroid
  • buddtown-schoolhouse-repair-demands – the schoolhouse is not separately located; pin is the Buddtown centroid
  • cedar-tree-rings-redate-granary – the granary building is not separately located, and the study also drew on Batsto and Port Norris timbers; pin is the Greenwich centroid
  • cranberry-false-blossom-research – the entry names three research stations (Whitesbog, Pemberton, Chatsworth); pin is the Whitesbog centroid only
  • crosswicks-meeting-school-reuse – the meetinghouse is not separately located (Nominatim returned no match); pin is the Crosswicks centroid
  • decatur-hanover-furnace-cannonball-legend – an undated legend at a vanished ironworks; pin is the njpb Hanover Furnace locality point (~305 m from the GNIS point of the same name)
  • eagle-hotel-barnhart-tract – the entry itself says operating dates are uncertain and follows a deed tract, not a located building; pin is the Eagle centroid
  • eayrestown-mill-rebuilding – the mill site at Landing and Eayrestown Roads is not separately located; pin is the Eayrestown centroid
  • elm-berry-pickers-wildman-1896 – a roving search across the turnpike, Murphy’s Crossing and Winslow Junction; pin is the Elm centroid. Note also that GNIS places Elm in Camden County while the record’s name says ‘near Hammonton’ (Atlantic County) – the two are adjacent
  • estellville-glassworks – the glassworks ruins are not separately located (Nominatim returned no match); pin is a settlement-level point
  • fenwick-manor-household-workers – Fenwick Manor is not separately located (Nominatim returned no match); pin is the New Lisbon centroid
  • franklinville-paddled-jersey-devil-1925 – the October 22, 1925 Monitor-Register names Morrell Langley and says he found the carcass in a small pine tree in woods a short distance from Harding Highway. His hotel residence was at the highway’s junction with the Franklinville–Iona Road; the article does not say the find was at that junction. The main map pin remains a broad Franklinville locality, not the tree or hotel.
  • domioni-giuseppe-1938 – basis already conceded this is the borough centroid, not 84 Pitman Street; the body was also recovered from the Delaware River, not at the address
  • friendship-bogs-school-preservation – pin is the Friendship settlement point, not the schoolhouse

site -> area (12) – the coordinate is defensible but the subject has real extent, so a point pin would mislead:

  • akron-airship-loss-1933 (r=3000 m) – NAS Lakehurst is a several-thousand-acre installation, and the loss itself was at sea; a point pin overstates it
  • belcoville-bethlehem-loading-1918 (r=3000 m) – the entry explicitly says the town, the plant and the later archaeological district had different footprints, the last about 12,947 acres
  • black-saturday-forest-fires-1963 (r=3000 m) – Coyle Field is correctly geocoded (reverse-geocode confirms Coyle Field, Woodland Township), but the entry describes a statewide outbreak of 127 fires over 183,000 acres; Coyle Field is the one specifically named place, not the extent of the event
  • bodines-tavern-tuckerton-stage-road (r=1500 m) – an inferred stream-confluence point standing in for a vanished tavern; no ‘Bodine’ feature was found in GNIS or Nominatim
  • body-found-in-stafford-forge-do-you-know-this-man-barnegat-nj-patch (r=3000 m) – the pin is the Stafford Forge settlement point used for a wildlife management area; the body was found in woods within its Eagleswood section (reverse-geocode confirms Eagleswood Township)
  • burr-haines-indigenous-household (r=3000 m) – the GNIS coordinate for a stream is its mouth, not the road-shoulder excavation somewhere along the run
  • chestnut-neck-privateering-1778 (r=2000 m) – a burned village plus wreck remains in the Mullica River; the commemorative Battle of Chestnut Neck Memorial is a further 1.7 km north-east at 39.5446004,-74.4614824
  • double-trouble-fire-weather-2002 (r=5000 m) – a roughly 1,300-acre wildfire across the state park and Ocean County was sharing a point coordinate with the compact Double Trouble Village entry
  • fort-dix-influenza-1976 (r=4000 m) – Fort Dix is a very large installation, and GNIS carries a second, conflicting ‘Fort Dix’ military point about 7 km away (40.0129485,-74.535118)
  • fort-dix-stockade-riot-1969 (r=4000 m) – same installation-scale problem; the stockade itself is not separately located
  • franklin-parker-restoration-soil-dna (r=4000 m) – the Nominatim coordinate is correct for the preserve, but the Franklin Parker Preserve is roughly 9,400 acres – not a point
  • fulton-minnie-1974 (r=2000 m) – the pin is the coordinate of a reservoir; the body was found in shallow water and woods near Chatsworth Lake, so a point pin risks placing a reader on open water

radius widened (1):

  • freyberger-howard-hermit-1931 – the record picked Gloucester City over Gloucester Township on a stated hunch; radius now spans both, since the source says only ‘Gloucester-area’

Totals

  before after
site 42 7
locality 32 54
area 37 50
none 16 16
total 127 127

No record was set to none in this pass and no coordinate was invented: the single coordinate change (Cedar Bridge Tavern) and the one refinement (Atsion Mansion) both come from named Nominatim features, cross-checked by reverse geocoding.

Sample check of locality / area records

Twenty-three non-site records were read against their entry pages, weighted towards thin basis text, round-looking radii and the log’s own unresolved list: alice-paul-paulsdale, amatol-shell-loading-town-1918, armstrong-lisa-2013, atlantic-city-drawbridge-wreck-1906, bass-river-fire-1936, beast-1887-camden, bear-newark-1906, bigfoot, black-saturday-forest-fires-1963, brown-bradway-1937-murdered, brown-bradway-palmyra-shot-1933, bryan-howard-g-1936-dix, burlington-bookmobile-rural-readers, butler-harriet-servant-woman-1909, calico-hamlet-oswego-river, carslake-charles-l-1933, chattin-murders-1917-brooksbrae, cohen, dorn-michael-1933-mob, elizabeth-burr-flyat-1895, elizabeth-kite-pineys-report-1913, franklin-parker-restoration-soil-dna, freyberger-howard-hermit-1931.

Most held up well, and several radii are genuinely well-reasoned rather than arbitrary – blue-comet-wreck-1939 (2,700 m, sized to the wreck’s documented 1.65-mile offset from Chatsworth station) and calico-hamlet-oswego-river (1,800 m, sized to the source’s ‘about a mile upstream’) are good practice. Three needed work and were fixed above (burlington-bookmobile-rural-readers, freyberger-howard-hermit-1931, elizabeth-kite-pineys-report-1913).

Entry-content problems noticed (flagged, NOT fixed)

  1. friendship-bogs-school-preservation – the record’s own label said Tabernacle Township but the coordinate is in Washington Township (label corrected here). Separately, the entry’s factbox reads ‘Place: Friendship Bogs; later Tabernacle’, and the preserved schoolhouse was physically moved to Tabernacle – so the building a reader would drive to is not at the pin and not in the place the pin sits. Worth resolving in the entry text.
  2. double-trouble-village vs double-trouble-fire-weather-2002 – a compact preserved village and a ~1,300-acre wildfire were sharing one identical coordinate at site precision. Fixed on the map side; the entries may want to distinguish the village from the park.
  3. barclay-farmstead-sale-subdivision – the GNIS feature ‘Barclay Farm’ is the mid-20th-century housing development named after the farm, not the preserved Barclay Farmstead the entry is about. The entry’s factbox ‘Place: Cherry Hill, Camden County’ does not flag the distinction.
  4. elizabeth-kite-pineys-report-1913 – factbox ‘Place’ names Vineland (the author’s institution) while the report’s subject is the Pine Barrens communities. Handled here by making the pin’s note say so outright, per reviewer instruction; the factbox itself still reads both ways.
  5. beast-1887-camden – the OCR’d place name ‘Owensburgh’ still matches nothing in the gazetteer, while the eyebrow calls it ‘the Winslow-area predator’. Unresolved; the 12 km area pin is honest about it but the underlying clipping needs re-reading.
  6. estellville-glassworks – the entry is titled Estellville but its factbox ‘Place’ says Estell Manor, and GNIS carries both names 1.8 km apart. Both points reverse-geocode to Estellville, so this is a labelling wobble rather than a wrong pin.
  7. Stale notes left by the second geocoding pass – barnegat-lighthouse-changing-coast and franklin-parker-restoration-soil-dna had their coordinates upgraded to real Nominatim features but kept the old notes describing the superseded centroid pins (‘Pin marks Barnegat Light borough…’, ‘Pin marks Chatsworth generally…’). Both notes rewritten. Worth checking the other eight records upgraded in that pass for the same pattern.

Two flags from the original log were checked and are non-issues: bear-newark-1906 is internally consistent (title, eyebrow and factbox all say Kearny; ‘Newark’ is only the wire dateline and the legacy slug), and the brown-bradway duplicate pair is being handled separately – both left alone.