Choosing the right contour interval: turning QLD and NSW LiDAR into usable design contours
You've grabbed a DEM for a site and generated contours, and they look like a bowl of spaghetti. Or they're too sparse to tell you anything. Either way, they're not much use for the early site set-up you're actually trying to do. The raw elevation data is fine — the problem is usually the contour interval, the smoothing, and how the surface was built before you contoured it.
This is a small decision that quietly shapes the whole desktop review. Get it right and you can read fall, drainage lines and building platforms at a glance. Get it wrong and you spend an hour cleaning up before you can even start.
Where the elevation comes from in QLD and NSW
Most publicly available elevation for Queensland and New South Wales traces back to airborne LiDAR. In Queensland that's captured under state and council programs and served through Queensland Globe and the Spatial data catalogue; in NSW it's the Elevation Data (ELVIS) and Spatial Services holdings. Coverage and vintage vary block to block — inner-city and coastal corridors tend to be denser and newer than rural fringes.
What you download is usually a bare-earth Digital Elevation Model (DEM) — a raster grid where each cell holds a ground height, typically at 1 m or 5 m resolution. Bare-earth means buildings and vegetation have already been stripped out, which is what you want for design. A Digital Surface Model (DSM), by contrast, keeps the treetops and rooflines, so don't contour that one by accident.
Heights are almost always in AHD (Australian Height Datum), which is what your survey and design will reference too. Worth confirming before you build anything on top of it.
Matching the contour interval to the job
There's no single correct interval — it depends on the terrain and what you're using it for.
- Flat urban infill (say, under 3 m of fall across the lot): 0.25 m or 0.5 m contours. Anything coarser and the whole block reads as one flat line, hiding the subtle fall you need for drainage and slab set-out.
- Typical suburban and gently sloping sites: 0.5 m is the sweet spot. Readable, not cluttered.
- Steep or acreage sites, or a first regional look: 1 m or 2 m. On steep ground, tight intervals stack up and become unreadable.
A good rule of thumb: if the contours are so dense they merge into a solid band, widen the interval. If a whole area has no lines at all, tighten it.
Remember the source resolution caps what's meaningful. Contouring a 5 m DEM at 0.1 m intervals just draws the raster's stair-stepping as fake terrain. Keep the interval sensible relative to the grid — 0.25 m or coarser off a 1 m DEM is honest; finer starts inventing detail.
Cleaning the surface before you contour
Raw DEM contours look jagged because the grid is stepped. Two things fix most of it.
Smoothing the surface, not just the lines. A light low-pass or focal-mean smooth on the raster before contouring removes the worst of the pixel stepping without shifting heights meaningfully. Don't overdo it — heavy smoothing flattens real breaks like table drains and retaining walls.
Simplifying and smoothing the contour lines. Once generated, run a line smoothing (spline or Bezier) and a light vertex-thinning. In QGIS that's Smooth and Simplify under Vector geometry; in ArcGIS it's Smooth Line; in Civil 3D you can apply contour smoothing straight on the surface style.
The goal is contours that read like a survey drawing, not like a barcode.
Getting them into your software cleanly
What you export depends on where the data is going.
- Civil 3D: bring the DEM in as a surface (from raster or a points/TIN import) and let Civil 3D generate contours natively — you keep the surface for volumes and profiles later. If you only need linework, DXF with elevations on 3D polylines (so heights survive) is the safe bet.
- QGIS and ArcGIS: GeoPackage or Shapefile for the contour vectors, with an elevation attribute on each line. Keep the raster DEM itself as GeoTIFF if you want to re-contour or run analysis.
- Quick visual check or sharing: KMZ drops contours or the DEM straight into Google Earth for a fast sense-check against imagery.
Whatever the format, confirm the horizontal coordinate system (MGA2020 zones for current work) and that the vertical values are AHD before you trust a single number.
A sensible early-site-setup workflow
For a first desktop pass, you don't need much: pull the bare-earth DEM, generate contours at an interval that suits the fall, apply light smoothing, overlay the cadastre so you can see boundaries against the ground, and drop it all into your CAD or GIS. From there you can read drainage direction, spot low corners, and rough out a platform — all before anyone sets foot on site.
Common questions
What contour interval should I use for a residential lot? For most suburban lots, 0.5 m gives a readable picture. On very flat infill blocks, drop to 0.25 m so you can actually see the fall for drainage and slab levels. On steep sites, widen to 1 m so the lines don't merge.
Should I contour a DEM or a DSM? A DEM (bare earth) for design and drainage — it's the ground with buildings and trees removed. A DSM keeps vegetation and structures, so its contours will follow treetops and rooflines, which is rarely what you want for site set-up.
Why do my LiDAR contours look jagged? Because the raster grid is stepped, so contours trace those steps. Apply a light smooth to the raster before contouring, then smooth the contour lines themselves. Also check you're not contouring at an interval finer than the DEM resolution can support.
Want contours and cadastre ready to drop straight into your workflow? Export site data for any address or see a sample export.
LayeredGeo lets you download QLD and NSW site data — contours, cadastre and more — in DXF, Shapefile, GeoPackage, KMZ and raster formats, ready for Civil 3D, QGIS and ArcGIS.