Case study 01 / QGIS + physical fabrication

North Carolina in 54 layers.

I used QGIS to turn statewide elevation data into nested contour sheets for a 50-inch laser-cut map. This page shows both sides of the project: building reliable geometry on the computer and working out how to cut, register, and stack it in the real world.

Back to survey projects
Horizontal scale
1:850,000
Physical layers
54 sheets
Finished footprint
50 × 15 in
Highest modeled layer
5,600 ft

Interactive model / The actual 54-sheet stack

See each contour sheet.

Move the slider to choose the highest visible sheet. The model keeps that contour and every sheet below it, just like the physical stack. Drag to rotate, scroll to zoom, or let it turn on its own.

QGIS view of the layered North Carolina topographic map.

Loading the 54 contour sheets…

54 of 54 · through 5,600 ft
Base

Showing the full stack.

5,600 ft
Coastal Plain / 25 ft Piedmont / 50 ft Mountains / 200 ft

The colors follow the project's approximate physiographic boundary polygons. A short gradient at each shared edge keeps the three regions distinct without drawing a hard stripe across the state.

Section view / W3 → W1 → W2 → E2 → E1 → E3

The registration-pin line, opened into a profile.

I cut the same six-point path through every contour sheet, then laid those cuts out in a straight line. The result works like a utility profile: each horizontal band is one physical sheet, and the gaps show where a registration hole still passes through that elevation.

Plan and profile diagram of the North Carolina contour stack along registration points W3, W1, W2, E2, E1, and E3.
The orange line in plan view is the section path. The highlighted profile band follows the sheet selected above.

Computer work / QGIS geometry

Building the contour stack.

01 / Source + projection

Set one statewide working grid

I paired a U.S. Census North Carolina boundary with USGS 3DEP one-arc-second elevation data, then worked in North Carolina State Plane coordinates (EPSG:32119). So the elevation surface was resampled to a consistent working grid before the actual contour extraction.

02 / Cumulative geometry

Keep every sheet inside the one below it

Each layer represents all ground at or above one elevation. I converted those thresholds into nested polygons, then checked validity, containment, slivers, and disconnected pieces. A small error low in the stack would carry through dozens of physical layers when represented in the real world.

03 / Regional interval logic

Use intervals that fit each region

The stack uses 25-foot intervals across the Coastal Plain, 50-foot intervals through the Piedmont, and 200-foot intervals in the mountains. That keeps the low relief on the coast visible while preventing the Appalachian Mountains from becoming a truly impractical number of sheets to cut.

04 / Two working files

Separate geography from fabrication changes

The geographic file preserves the mapped geometry. The fabrication file removes features that would not cut reliably, divides sheets to fit the machine, adds elevator-pin registration holes, and carries the physiographic boundaries for the Coastal Plain, Piedmont, and Mountains.

Manufacturing / cutting / From file to stacked board

Making 54 sheets line up.

Board size and split sheets

The finished map is 50 × 15 inches, so each statewide layer is divided into East and West cutting layouts sized for the available laser bed. Both halves keep the same coordinate origin and placement marks.

Elevator-pin heights

Each registration pin stops below the first contour that naturally closes over its location. The next intact sheet becomes its cover: from 150 feet at E3 to 4,200 feet at W3.

Cutting and stacking

The 1.297 mm white matboard sheets are cut as cumulative areas, not loose contour rings. Registration holes, seam logic, and consistent sheet order keep the stack from drifting as the layers build upward.

Checks before cutting

I kept source, geographic, fabrication, split-sheet, and layout outputs separate. That makes it possible to trace a cut change without quietly changing the mapped contour beneath it.

Data and references

The case study uses public elevation and boundary data. The web model is a simplified derivative of my QGIS fabrication layers.