GDTHF
ENRO

Sites & Maps

Sites & Maps — Archaeological Atlas

Starting the 3D model… your choices will be applied when it is ready.

GDTHF project model (revision 8). 14 m full rises are restored scenario estimates; central plan cores retained, other citadel cores schematic. Markers are reference positions on the model, not proof of the displayed hypothesis; RAN markers are generalized ~2 km cells, never exact site positions.

2D research map

Pețan 2026 overlays & RAN inventory

Loading map…

search area ±1–3 kmRAN ~2–4.5 km cellnot GPS · not terraces

Basemap caveat: satellite imagery has mixed capture dates and resolution and is not LiDAR; terrain relief is ~30 m SRTM. Credits for each basemap appear on the map. Imagery and map tiles are third-party services used under their providers' own terms (Esri, OpenTopoMap, OpenStreetMap); usage rights and required credits depend on those terms, and OpenStreetMap tiles are loaded only on demand in line with its tile usage policy (no bulk or offline downloading). Pre-launch note: the production-scale imagery provider and its terms must be verified before public launch.

Map evidence summary

Published terrace research and the RAN inventory

These numbers use different units and cannot be added, subtracted or turned into percentages.

  • >4,000

    LiDAR-identified terrace platforms (Fig. 10a, regional aggregate)

  • c.177

    excavated / trial-trenched points (Fig. 6, mixed units)

  • c.158

    survey-confirmed points (Fig. 6, mixed units)

  • >250

    architectural-platform anomalies (Fig. 12, not confirmed houses)

  • —

    official RAN records in the stored monthly snapshot (all periods; not terraces)

RAN-to-terrace exact intersection: unknown because terrace geometries are absent.

Sector comparison — published estimates

SectorCurrentEarlier
Grădiștii Hillc.373c.430 (earlier LiDAR count)
Fețele Albec.20020–30 (older estimate)

Published estimates (Pețan 2026), not GDTHF counts.

From the article

  • p. 18 — Paraphrase: the KDE at 300 m resolves local nuclei; the main hotspot spans Grădiștii Hill, Sarmizegetusa Regia, Fețele Albe and the central slopes of Valea Anineșului, with secondary cores at Costești and Piatra Roșie.
  • p. 19 — Paraphrase: at 800 m bandwidth the density surface shows regional continuity rather than isolated sites.
  • p. 25 — Paraphrase: platform anomalies concentrate near high-density cores; the social reading of this is offered as a working hypothesis.
  • p. 31 — Paraphrase (data availability): the full terrace vector GIS is not public; the 2011 high-resolution core and the Costești/Piatra Roșie LiDAR are not archived; the 1 m LAKI II DTM is available through ANCPI.

Original figures

Figure 10 — terrace distribution and density (a–c). © 2026 Aurora Pețan, Humans 6(2) 19. CC BY 4.0. Cropped for display; otherwise unmodified. Original ↗
Figure 13 — KDE 300 m with architectural-platform anomalies. © 2026 Aurora Pețan, Humans 6(2) 19. CC BY 4.0. Cropped for display; otherwise unmodified. Original ↗

Earlier maps & atlas

Earlier basemap view & research methodology

Open the earlier satellite atlas

Known discoveries vs the terraced landscape

What is already recorded — four separate datasets

These counts measure different things over different areas. They are shown side by side and are never added together or turned into an overlap percentage.

RAN snapshot · Șureanu study box

…

Official RAN site records, all periods, in the map's study box. Loading stored snapshot…

Source: CIMEC RAN GIS (layer 0), cached monthly.

Planning study · two communes

94 · 28

Orăștioara de Sus commune: 94; Boșorod commune: 28. All periods, whole commune territory — administrative totals, not overlap. County: 1,026.

Hunedoara PATJ heritage study, §2.4

Pețan 2026 · high-altitude core

≈335

177 excavated or trial-trenched + 158 survey-verified investigation points. Not 335 distinct terraces, not necessarily 335 RAN sites.

Humans 6(2) 19, Fig. 6

Pețan 2026 · LiDAR

more than 4,000

Anthropogenic terrace features mapped from LiDAR (regional total). Individual positions not published.

Humans 6(2) 19, Fig. 10

Exact RAN sites ∩ terrace feature geometry: not yet computable. The original GIS of the more than 4,000 individual features is not available, so the overlap is unknown — not zero.

RAN records by period (fallback)

  • Late Iron Age / La Tène (Dacian)21
  • Roman3
  • Prehistoric1
  • Medieval & later1
  • Undated / not stated1

0 of 21 dated from official RAN component fields; the rest inferred from the record title (caveat). A record with several phases counts in each.

By recorded component

  • Fortification0
  • Settlement0
  • Sanctuary0
  • Tower0
  • Workshop0
  • Other0

Registration precision: 11 located (COORD DA), 10 approximate locality (COORD NU). 0 records list research seasons on their RAN card.

Concentration near published terrace sectors

  • Grădiștii Hill / Sarmizegetusa Regia sector3
  • Fețele Albe sector2
  • Costești–Blidaru sector12
  • Piatra Roșie sector1
  • Costești–Cetățuie sector12
  • Grădiștii Hill western settlement slopes3

Generalized RAN records whose ~2 km cell centre lies within 3 km of each study-sector centre (cells may count for neighbouring sectors). Shows co-location with the terraced core, not intersection with terraces.

Geographic accuracy of RAN records (fallback)

  • Official mapped site boundary6
  • Official site point5
  • Locality-centre proxy10
  • Location unavailable0

RAN COORD: 11 DA, 10 NU. An official boundary (CIMEC RAN polygon layer) is matched by RAN code for 6 records; 0 of those touch a broad published terrace-study zone (1.5 km around the sector centre). That is an area comparison only — exact overlap with the more than 4,000 terraces remains unknown. Locality-centre records are not individual site positions. Only UNESCO-listed components are drawn with their boundary; all other boundaries are used on the server for counts and never published.

How we aggregate
  • One registered RAN site can contain many terraces; one survey point may cover several terraces, a terrace group or a sector.
  • A missing RAN record does not mean absence of ancient occupation — RAN is incomplete at altitude.
  • Location source: official mapped boundary (RAN polygon) where one exists, else the RAN site point (COORD DA), else the locality-centre proxy (COORD NU). COORD DA marks a located site centre; COORD NU only a locality centre. Both are further generalized here to ~2 km cells (single records ~4.5 km); no exact position is stored or published.
  • Commune totals are administrative and cover whole territories of all periods; they do not measure overlap.
  • Periods come from the dating of RAN components; where a card gives none, the period is inferred from the title and labelled as inferred.
  • Summaries are composed only from RAN card fields. Records are called excavated only when the card lists research. Hand-checked paraphrases were checked 2026-11-15.
  • Snapshot: refreshed monthly; a failed refresh keeps the last good copy.

Browse the RAN records

  • RAN 88500.02 · Late Iron Age / Dacian (RAN component dating (official card))

    Luncani – Piatra Roșie

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 88500.07 · Late Iron Age / Dacian (RAN component dating (official card))

    Luncani – Vârful Străin

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 88500.08 · Late Iron Age / Dacian (RAN component dating (official card))

    Luncani – Poiana Omului

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 88537.01 · Late Iron Age / Dacian (RAN component dating (official card))

    Târsa – Vârful Secului

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 88537.06 · Roman (RAN component dating (official card))

    Târsa – Platoul Târsa

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.01 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Cetățuie

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.02 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Cetățuia Înaltă

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.05 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Blidaru

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.07 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Dârlău

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.08 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Merișorul

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.12 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – Biserică / Gruiul Negru

    Record title only — RAN card details not yet retrieved; detailed finds not provided

  • RAN 90379.13 · Late Iron Age / Dacian (RAN component dating (official card))

    Costești – La Curți

    Record title only — RAN card details not yet retrieved; detailed finds not provided

Already investigated vs mapped by LiDAR

What has already been investigated?

In the high-altitude core, Pețan (2026) reports about 177 excavated or trial-trenched points and about 158 further points verified by surface survey — some 335 investigation points in all.

Pețan 2026 Figure 6: yellow points of excavated and survey-verified Dacian sites on a hillshade of the Șureanu high-altitude core
Figure 6 — investigated evidence. Pețan 2026, Figure 6: distribution of archaeologically investigated and survey-verified Dacian points in the high-altitude core area of the Șureanu Mountains, on a LiDAR-derived hillshade. Open original figure
Pețan 2026 Figure 10(a): distribution of over 4,000 LiDAR-mapped terrace platforms
Figure 10 — LiDAR terrace mapping. Pețan 2026, Figure 10 (a) distribution of identified terrace platforms (over 4,000 mapped terraces) and (b) density model, 300 m radius. Open original figure (incl. density maps b, c)

What the 335 are

Investigation points from more than a century of fieldwork. Mixed units: some points are individual terraces, others terrace groups, others broader investigated sectors. They are not 335 distinct RAN sites, and their individual coordinates are not published as data.

Dating

In this mountainous zone the archaeologically investigated record belongs exclusively to the Late Iron Age; earlier or later habitation is documented only in the lower valleys and adjacent low-altitude sectors. Most individual terraces have not been excavated and cannot be dated independently; their Dacian attribution is contextual, resting on association with directly dated sites.

How this map uses them

Not at all as dots. Figure 6 is shown as published and is not georeferenced or traced onto the map. The strongest high-altitude evidence is securely Late Iron Age, but not every one of the >4,000 terraces has been excavated.

Source: Aurora Pețan (2026), Humans 6(2), 19, https://doi.org/10.3390/humans6020019. © 2026 Aurora Pețan. Reproduced unmodified (reduced size) from Humans 6(2), 19 under CC BY 4.0. Licence.

Methodology

How to read this atlas

Official UNESCO references. The six components of World Heritage property 906, at the reference positions published by the UNESCO World Heritage Centre (DMS converted to WGS84). They mark each component as a whole, not a wall or gate. Listing is a heritage designation and implies no partnership. Sarmizegetusa Regia (Dacian) is not Ulpia Traiana Sarmizegetusa (Roman colony), which is outside this map's evidence layers.

Published study sectors. Approximate site-level centres of areas discussed in Pețan 2023, Pețan 2026 and Pețan & Hegyi 2023. The dashed circles are illustrative inspection zones, not excavated boundaries. No walls are drawn because none are georeferenced in the public record. The published aggregate of about 250 architectural anomalies is likewise not plotted.

RAN archaeological inventory. The National Archaeological Repertory (RAN, CIMEC) lists roughly 29,000 sites across Romania; it is a different dataset from the terraces. Twenty-one individual site records in Grădiștea de Munte, Costești, Luncani and Târsa were cross-checked code by code against the official entries and the official RAN map service; for heritage protection every RAN record, live or cross-checked, is generalized to a ~2 km grid cell and shown only as a numbered bubble listing codes and official links. The monthly sync generalizes positions before storing them, so precise RAN geometry is neither stored nor sent to the browser. The c. 250 architectural anomalies are not plotted, and the 76 visual samples are stored rounded to ~1 km. These are registry entries for sites, not excavation campaigns. For scale, the Hunedoara County planning study lists 94 RAN sites in Orăștioara de Sus and 28 in Boșorod (all periods, whole communes) and 1,026 in the county; these are shown as illustrative area circles, not boundaries. All records come from a stored snapshot of the official CIMEC map service for the Șureanu box, refreshed monthly (never fetched live while you move the map), with the RAN COORD field shown unchanged: per CIMEC, located points (DA) are site centres and approximate points (NU) are locality centres, not exact positions. The older RAN text for Fețele Albe (c. 20 terraces, approximately located) predates the LiDAR estimate of c. 200 terraces. RAN is incomplete in high-altitude terrain; absence of a record is not absence of remains.

Low-confidence visual samples. 76 points hand-estimated from Pețan 2026, Fig. 10(a), an image without registration. Their error is often 1–3 km or more. They do not reproduce or georeference that figure, and the more than 4,000 terraces it reports are a regional total, not 76 or 4,000 points here. No synthetic or simulated point set is used. Any future georeferencing of Fig. 10 would need control-point registration, a stated error estimate and the author's data permission.

Coordinate source / discrepancy

Older approximate positions used before this atlas have not been overwritten. Distances from those positions to the UNESCO references:

  • Sarmizegetusa Regia: 0.14 km
  • Costești–Cetățuie: 2.49 km
  • Costești–Blidaru: 3.68 km
  • Luncani–Piatra Roșie: 11.03 km
  • Bănița: 3.48 km
  • Căpâlna: 2.69 km

The Fețele Albe reference position comes from independent research and its provenance still needs verification.

Morphological signs vs confirmation. LiDAR shows shapes consistent with human terracing. Dating, function and phase need excavation, finds or field survey.

Privacy & anti-looting. We publish no precise positions for unknown or vulnerable archaeology, offer no directions, and accept no public submissions of find locations. Metal-detecting and unauthorised digging at these sites are illegal and destroy evidence; report damage to the site administration and local authorities.

Bibliography

Sources for this atlas

  1. Aurora Pețan (2 June 2026), “Shaping the Landscape in Late Iron Age Europe: The Terraced Mountains of the Dacians”, Humans 6(2), 19.

    Open access, CC BY 4.0. Reports about 177 excavated or trial-trenched points and about 158 surface-survey-verified points in the high-altitude core (Fig. 6; mixed units, all Late Iron Age), and separately more than 4,000 anthropogenic terraces as a regional aggregate. Fig. 10(a) shows their distribution; Fig. 10(b,c) show kernel-density maps at 300 m and 800 m. See the figures in the original article — they are not reproduced here.

  2. Aurora Pețan, “Settlements on terraces in the central area of the Dacian Kingdom. Some aspects revealed by airborne laser scanning”, Acta Musei Porolissensis 43 (2023), pp. 205–216.

    Public abstract describes raised foundations of about 7–10 m, most concentrated around Grădiștii, Fețele Albe and Anineș. It gives no coordinates of individual platforms.

  3. Aurora Pețan & Alexandru Hegyi (2023), “Freely available LiDAR-derived digital terrain model (DTM) uncovers the heartland of the Dacian Kingdom”, Digital Applications in Archaeology and Cultural Heritage 31, e00292.

    Method paper on the public LiDAR-derived terrain model used to identify terracing.

  4. UNESCO World Heritage Centre, Dacian Fortresses of the Orăștie Mountains (property 906), maps.

    Official reference positions of the six inscribed components. A heritage listing, not a partnership.

  5. Repertoriul Arheologic Național (RAN), National Archaeological Repertory of Romania (CIMEC).

    National inventory of roughly 29,000 archaeological sites. It is not the source of the >4,000 terraces or of the 76 exploratory points. The inventory is incomplete in high-altitude terrain: absence of a RAN record is not absence of ancient remains.

  6. Consiliul Județean Hunedoara, PATJ Hunedoara update — Studiu de fundamentare: patrimoniul natural și construit, peisaje (2024), section 2.4.

    Lists 94 RAN archaeological sites in Orăștioara de Sus commune and 28 in Boșorod (122 in two communes, all periods, whole commune territory) and 1,026 RAN sites in Hunedoara County. Not 122 Dacian sites and not 122 overlaps with the terrace distribution.

  7. CIMEC, Serverul cartografic pentru patrimoniul cultural național (map server and symbol key); official viewer map.cimec.ro/Mapserver.

    Map key: polygons are official boundaries; red points are site centres located geographically; green points are approximate locality-centre coordinates, not exact positions. This site stores a monthly snapshot of PatrimoniuWM layer 0 for the Șureanu box and shows only generalized ~2 km cells; it does not query the service while you browse.

  8. ANCPI — National Agency for Cadastre and Land Registration, terrain-data portal.

    Future source of high-resolution bare-earth LiDAR. Raster files have NOT been ingested; no LiDAR comparison is shown.

Beyond the core

Regional Dacian centres and Roman Dacia

Broad public locations only. We do not publish coordinates for vulnerable sites.

Thracian & Lower Danube

Thracian sites in Bulgaria

Broad public locations only. We do not publish coordinates for vulnerable sites.

Visiting responsibly

Leave nothing, take nothing

Do not use metal detectors, move stones or collect finds. Report damage or suspected looting to the site administration and local authorities.