How to review Moenjodaro 3D
For archaeologists, heritage officers and anyone checking the work. About ten minutes. Last updated 13 October 2026.
What it is
A browser model of Mohenjo-daro built from published excavation plans (Marshall 1931, Mackay 1938, Wheeler 1953) and open terrain and satellite data. Seventeen pieces are traced, about 14 ha of plan-sheet extent, under 6 % of the 240 ha property. It has not been reviewed by an Indus archaeologist. That is what we are asking for.
Three things to open
- The model, in "Documented remains". Every wall is a rectangle traced from the black ink of a plate, drawn at one placeholder height of 1.2 m. Press Menu, then Evidence, then Evidence colours to see each piece coloured by how strong its evidence is.
- The evidence records. 125 records, each with a level and a source, in the Evidence panel (“All elements and their sources”) and as records.json.
- The method note. methods-note.pdf (four pages): sources, how the plates were traced, how pieces were placed and with what error, what the prediction view adds, and what changed after our own reviews.
What the levels mean
| Level | Meaning | Records |
|---|---|---|
| A | Surveyed or measured data (terrain and satellite imagery) | 2 |
| B | Excavation record: plan, section or photograph, with plate | 45 |
| C | Published description, a reading from a site sketch, a secondary summary, or a coarse volunteer-drawn outline | 26 |
| D | Analogy from other Indus buildings (none used yet) | 0 |
| E | Hypothesis or illustration with no direct evidence | 52 |
Where we most need an expert
- Levels B and C. Is each plate reading and each cited description fairly stated?
- Wall heights. None is measured; the plates are plans. The "How it may have looked" view uses 3.2 m and 4.2 m as hypotheses. Is there a published basis for better figures?
- Lanes and doorways. About 13 % of the gaps between traced walls are 0.5 to 1.0 m wide, and some of that is tracing error (±0.3 to 0.5 m per wall). Which are real?
- Positions. Pieces are placed from satellite imagery and Mackay's site plan: ±5 to 12 m for most, ±25 m for four DK sections. Do you hold better plans (ARPM, Aachen, Dales, Moneer, the 2025–26 mission)?
- Phases. The lower-town plans are Late-period plans, and three kilns belong to the last phase. The prediction view mixes phases. How should that be shown?
- The stupa. Marshall's Kushan-period reading and the Verardi and Barba Indus stepped-platform reading are both drawn; neither is chosen.
- Life in the past view. People, animals, boats, fields, trees, the river and sound are illustration (level E). Is anything misleading?
What not to take from it
- It is not an official record and not an exact replica.
- Only about 35 to 75 % of the black ink of each plate is traced; thin partitions, wells and street drains are missing.
- Walls of the Granary, the stupa mound and the two Fig. 6 outlines come from different kinds of plate, so their accuracy differs; the evidence panel says how.
How to send comments
Use the contact form at larkode.com ("Talk to us about heritage in 3D"), or name the record (for example hr.position) so we can find it. Every change we make is added to an open log, docs/paradata-log.md, with the date and the basis.
Developed by Larkode. Terrain: Copernicus DEM GLO-30; imagery: Sentinel-2; map data: OpenStreetMap through Overture; Esri World Imagery was viewed, not redistributed, to place pieces.