A topographic survey records the three-dimensional shape of a site: the elevation of the ground, the features on it, and the visible infrastructure serving it. Where a boundary survey answers a legal question, a topographic survey answers an engineering one. It becomes the base drawing on which a grading plan, a drainage design, a road profile or a building pad is developed.
The single most useful thing a client can do is specify the deliverable properly. Two surveys of the same field can differ by a factor of five in cost because one was collected for a conceptual site plan and the other for a construction-grade grading design, and nothing on the face of the drawing announces which one you received.
What gets collected
Ground shots define the surface. The crew or the sensor collects points dense enough to describe every meaningful change in slope: tops and toes of banks, ridges, swales, the crown and edge of pavement, the lip of a curb, the invert of a ditch. Density is not uniform. A flat parking lot needs few points; a steep eroded bank needs many. A survey judged only by point count is being judged by the wrong measure.
Breaklines matter more than point density. A breakline is a line along which the surface changes slope abruptly, and it forces the triangulated surface to bend where the ground actually bends. Without breaklines along a curb face or a ditch centerline, the interpolated surface will cut corners and the computed volumes will be wrong even though every individual point is correct.
Planimetric features are collected alongside the elevations: buildings, pavement edges, walks, walls, fences, trees above a specified size, signs, poles, and the visible evidence of underground utilities such as manholes, valve boxes, inlets and meters. Where rim and invert elevations are needed, the crew opens structures and measures inside them, which is a separate and slower operation than shooting a rim from the surface.
Vertical control is the part people skip
Every elevation on the drawing is relative to a datum, and a topographic survey on an assumed datum is nearly useless for anything that must connect to the world: a sewer that ties into an existing main, a floor elevation referenced to a flood study, or a site that has to match an adjoining phase surveyed last year.
State the vertical datum in the request and require it on the drawing, along with the benchmarks used and their published values. Require that the crew set at least two permanent benchmarks on site that will survive construction, so the contractor and the as-built surveyor can work from the same reference. Two benchmarks rather than one, because a single benchmark that gets destroyed or disturbed leaves nothing to check it against.
Specifying the deliverable
Write these six items into the request and most disputes never arise.
- Purpose and the design that will consume the data, so the surveyor can judge what density and accuracy the design actually requires.
- Limits of survey, as a mapped boundary rather than a description, including how far beyond the property the mapping must extend to pick up connecting roads and outfalls.
- Horizontal and vertical datum, including the coordinate system, the geoid model, and whether coordinates are grid or ground.
- Contour interval and the accuracy standard the surface must meet, stated as a tolerance rather than as a single word like accurate.
- The feature list: exactly which features are to be located, the minimum tree size, and whether structure inverts are required.
- File formats and drafting conventions: the CAD or BIM version, layer or level naming, whether a surface object and breaklines are delivered rather than only points and lines, and whether a point file with codes is included.
Choosing a contour interval
Contour interval should follow the terrain and the design tolerance, not habit. Too large an interval hides the features the designer needs; too small an interval on rough ground produces a drawing so dense it cannot be read and implies a precision the collection method does not support.
| Situation | Common interval | Reasoning |
|---|---|---|
| Flat urban site, drainage design | 0.5 ft or 0.1 m | Slopes are shallow and small elevation differences decide where water goes. |
| Rolling suburban site, grading plan | 1 ft or 0.25 m | Enough resolution for pads and driveways without unreadable density. |
| Rural corridor, preliminary alignment | 2 ft or 0.5 m | Adequate for route selection; detailed design areas get resurveyed later. |
| Steep or mountainous terrain, planning | 5 ft or 1 m | Closer contours would overlap on the sheet and add no usable information. |
A common rule of thumb is that a surface should be reliable to roughly half the contour interval, so a one-foot interval implies elevations good to about half a foot. If the design needs better than that, say so directly and specify the tolerance rather than shrinking the interval and hoping.
Collection methods and where each fits
| Method | Best suited to | Main limitation |
|---|---|---|
| Total station | Small sites, dense urban detail, anywhere under canopy or between buildings | Slow over large areas; needs line of sight |
| Network or base-rover satellite positioning | Open sites, corridors, large fields | Degraded or unusable under trees and beside tall buildings |
| Terrestrial laser scanning | Complex structures, facades, congested industrial sites | Large data volumes; occlusion behind objects; registration effort |
| Aerial imagery from a small uncrewed aircraft | Open sites needing surface and imagery together | Cannot see ground through vegetation; needs ground control |
| Airborne lidar | Large areas, forested terrain where bare earth is required | Cost is only justified at scale; ground point density varies with canopy |
Most real projects mix methods. A site might be flown for the open areas, walked with a total station under the tree line, and levelled conventionally along the sewer to get invert elevations that no remote method can produce.
How the survey is checked
Ask what independent check was performed. A surface can be internally consistent and still be wrong, typically because of an error in the control or the datum rather than in the collection. Reasonable checks include occupying a published benchmark not used in the adjustment and comparing the result, collecting a set of independent check shots on hard surfaces and reporting the differences, and closing a level circuit rather than running it one way.
allowable misclosure = C × √KThe deliverable should include a short statement of what was checked and what the residuals were. That paragraph is worth more to a designer than another thousand points.