Project Kiva

Archaeological LiDAR Survey — American Southwest

---

A reproducible remote sensing pipeline for detecting and visualizing archaeological features across the ancient Southwest. Named for the kiva — the subterranean ceremonial chamber at the heart of Ancestral Puebloan architecture, and precisely the kind of subtle feature that LiDAR reveals after centuries of desert wind have erased it from the surface.

---

Study Areas

Site	State	Period	Key Features
Chaco Canyon	NM	850–1150 CE	Great houses, road network, kivas
Mesa Verde	CO	600–1300 CE	Cliff dwellings, mesa-top villages
Hohokam Phoenix	AZ	300–1450 CE	Canal network, platform mounds
Canyon de Chelly	AZ	2500 BCE–present	Cliff dwellings, petroglyphs

Switch between sites by changing active_site in config.yaml.

---

Quickstart

# Install environment (Windows)
pixi install

# Download DEMs for the active site
pixi run python scripts/download_dem.py

# Render 3D terrain
pixi run python scripts/render_site.py --site chaco

# Open notebooks
pixi run notebooks

---

Project Structure

project-kiva/
├── config.yaml                  ← Switch sites here
├── pixi.toml                    ← Python environment
│
├── notebooks/
│   ├── 01_data_acquisition      Download USGS 3DEP LiDAR tiles
│   ├── 02_render_forge3d        3D terrain renders via forge3d
│   ├── 03_lidar_processing      Point cloud → bare-earth DEM
│   ├── 04_archaeological_viz    SVF, LRM, hillshade products
│   └── 05_feature_detection     Automated mound/kiva detection
│
├── scripts/
│   ├── download_dem.py          Download GeoTIFF DEMs from USGS
│   └── render_site.py           forge3d 3D terrain renderer
│
├── data/
│   ├── raw/          LAZ point cloud tiles (gitignored)
│   ├── processed/    GeoTIFFs and raster products (gitignored)
│   ├── renders/      PNG render outputs (gitignored)
│   ├── vectors/      GeoJSON site features (committed)
│   └── assets/       Project imagery and branding
│
└── web/
    └── index.html    GitHub Pages interactive map

---

Pipeline

USGS 3DEP tiles (LAZ / GeoTIFF)
    ↓
Ground classification → bare-earth DEM
    ↓
┌─────────────────────────────────────────┐
│  Archaeological visualization suite     │
│  · Multi-azimuth hillshade              │
│  · Sky-View Factor (SVF)                │  ← best for kiva depressions, road berms
│  · Local Relief Model (LRM)             │  ← best for mounds, platform edges
└─────────────────────────────────────────┘
    ↓
Automated detection → candidate_features.geojson
    ↓
forge3d path_tracing → 3D terrain renders
    ↓
Leaflet web map → GitHub Pages

---

forge3d Rendering

Project Kiva uses forge3d for GPU-accelerated 3D terrain visualization via its headless path tracer — no Rust toolchain or viewer binary required.

from forge3d import path_tracing
from forge3d._png import save_png

tracer = path_tracing.create_path_tracer(1920, 1080, max_bounces=2)
camera = path_tracing.make_camera(
    origin=(ox, oy, oz), look_at=(cx, cy, cz),
    up=(0, 0, 1), fov_y=45.0, aspect=16/9, exposure=1.0
)
rgba = tracer.render_rgba(dem_array, camera)
save_png('render.png', rgba)

---

Web Map

Live at bdgroves.github.io/project-kiva

• Satellite, topo, and terrain basemap toggle
• Known great house and kiva locations with historical notes
• Automated candidate feature detections
• Schematic road and canal overlays

---

References

• Lekson, S.H. (1999). The Chaco Meridian. AltaMira Press.
• Chase, A. et al. (2011). Airborne LiDAR, archaeology, and the ancient Maya landscape. Journal of Archaeological Science, 38(2).
• Evans, D. et al. (2013). Uncovering archaeological landscapes at Angkor using LiDAR. PNAS, 110(31).
• Opitz, R. & Cowley, D. (Eds.) (2013). Interpreting Archaeological Topography. Oxbow Books.

---

Code: MIT  ·  Data: USGS 3DEP public domain  ·  Built in Tacoma, WA

Feature detections are unverified automated outputs and should not be cited as archaeological findings.