What georeferencing is and why it matters
Georeferencing aligns a scanned map or image to real-world coordinates so features on the sheet line up with modern basemaps and layers. For historical work, it lets a map’s sketches become spatial data that can be compared to today’s terrain, boundaries, and infrastructure.
What to look for on historic sheets
Methods I use: Control points and transformations Control points are pairs of “same place” locations—one on the historic map, one on a modern basemap. With enough well-spread points, a transformation warps the image into place.
Georeferencing aligns a scanned map or image to real-world coordinates so features on the sheet line up with modern basemaps and layers. For historical work, it lets a map’s sketches become spatial data that can be compared to today’s terrain, boundaries, and infrastructure.
What to look for on historic sheets
- The map’s datum and coordinate reference (e.g., Clarke 1866, a legacy latitude/longitude grid).
- Clear, stable control features that still exist (ridge lines, river bends, road intersections, building corners).
- Notes about scale or projection that affect distortion.
Methods I use: Control points and transformations Control points are pairs of “same place” locations—one on the historic map, one on a modern basemap. With enough well-spread points, a transformation warps the image into place.
- Spline (rubber-sheet): very flexible; ideal for large, uneven distortion across wide areas.
- Polynomial (1st–3rd order): good when distortion is moderate and a mathematically smooth fit is preferred.
- RMSE (root-mean-square error): summary of fit quality at the control points; lower is better, and “good enough” depends on purpose (regional context vs. site-scale analysis).
Spencer's Map, georeferenced.
Example: Regional campaign map (1880s). A large-area campaign sheet with uneven paper stretch was georeferenced with a spline using 200+ control points distributed across the full extent. The result yielded regionally reliable alignment for inter-post relationships and line-of-communication studies—appropriate for big-picture analysis, not parcel accuracy.
Example: Fort-level change detection. Three detailed site maps around a fort were aligned with a 3rd-order polynomial and 12+ well-chosen control points (building corners, road junctions, wall breaks). The RMSE ~6–7 meters supported precise interpretation of facility changes through time.
Example: Fort-level change detection. Three detailed site maps around a fort were aligned with a 3rd-order polynomial and 12+ well-chosen control points (building corners, road junctions, wall breaks). The RMSE ~6–7 meters supported precise interpretation of facility changes through time.
Source: Lieutenants and Light
Reading the results.
RMSE in context. An RMSE of a few meters can be excellent for site interpretation; several dozen meters may still be useful for regional questions (e.g., was a station on this ridge or the next?). Always judge error against the scale of the question.
Reality checks
“Accuracy is a question of purpose. Fit the method—and the error tolerance—to the decision you’re trying to make.”
Workflow:
Sources for Materials: National Archives, BLM’s GLO records, Library of Congress, and USGS holdings provide extensive historic map collections, plats, and topographic sheets suitable for georeferencing.
Glossary (short):
RMSE in context. An RMSE of a few meters can be excellent for site interpretation; several dozen meters may still be useful for regional questions (e.g., was a station on this ridge or the next?). Always judge error against the scale of the question.
Reality checks
- Toggle the georeferenced sheet on/off over a hillshade or modern orthoimage.
- Check a few independent features that were not used as control points.
- Look for local stretching near the edges or in sparsely controlled areas.
“Accuracy is a question of purpose. Fit the method—and the error tolerance—to the decision you’re trying to make.”
Workflow:
- Inspect the sheet (datum, projection, scale, legibility).
- Pick stable control features across the entire area.
- Start with a polynomial; switch to spline if distortion is uneven.
- Spread control points—avoid clusters.
- Review RMSE and, more importantly, visual alignment.
- Save the georeferenced image and document method, control count, and error.
Sources for Materials: National Archives, BLM’s GLO records, Library of Congress, and USGS holdings provide extensive historic map collections, plats, and topographic sheets suitable for georeferencing.
Glossary (short):
- Georeferencing — Aligning an image/map to real-world coordinates so it overlays modern spatial data.
- Control point — A location identifiable on both the historic sheet and a modern basemap.
- Spline transformation — A flexible “rubber-sheet” warp that handles uneven distortion.
- Polynomial transformation — A mathematical warp (1st–3rd order) for smoother, global fits.
- RMSE — Root-mean-square error; a single number summarizing control-point misfit.