How Accurate Does a Survey Need to Be? A Practical Tolerance Guide
Survey accuracy is not one number. A builder checking footing marks needs tighter data than a project manager checking bulk earthworks progress. An engineer certifying an as-built package needs evidence that matches the contract, authority or council requirement.
That is why the survey method changes from stage to stage. A feature and level survey gives the design team existing ground and site detail. A construction set-out survey marks the approved design on site. A volume pickup checks how much material moved. An as-built survey records what the contractor installed for handover.
Before you book the survey, confirm three things: what decision the survey will support, what tolerance applies, and what evidence the next person in the chain needs. Drone data can work well for progress records, broad surfaces and earthworks volumes. A total station or digital level is the better choice for column grids, drainage inverts, slab levels and other tight checks.
Accuracy, tolerance and uncertainty mean different things
On site, teams use “accuracy” to cover several different ideas:
| Term | Practical meaning on a project |
|---|---|
| Accuracy | How close a measured point, surface or model is to its true position. |
| Precision | How repeatable the measurement is under the same conditions. |
| Tolerance | The allowable difference between design and constructed position or level. |
| Uncertainty | A formal expression of doubt in a measurement or coordinate, used for survey control and reporting. |
| Evidence | The records, reports, check shots, point clouds, files or certificates that prove the work was measured to the required standard. |
The instrument specification gives you one part of the answer. Control quality, datum, observation method, site access, weather, operator process, processing workflow and check data all affect the final result. Manufacturer data sheets for instruments such as the Leica TS16 total station, Leica LS10 and LS15 digital levels and Leica RTC360 laser scanner help with method selection. They do not replace control, workflow or verification.
Australian survey control standards such as the ICSM Standard for the Australian Survey Control Network focus on survey quality and uncertainty. Construction and civil teams should use the same mindset: prove the result, not the brochure claim.
Accuracy by survey method
Use these ranges for planning. Contract documents, road authority requirements, council handover standards and engineering tolerances control the final survey brief.
| Survey method | Practical accuracy range to plan around | Best use | Watch-outs |
|---|---|---|---|
| Total station | Plan around 1-5 mm for controlled set-out and short-range verification; allow more where control or geometry is poor | Building set-out, structural checks, columns, walls, services, road furniture, detailed pickups | Needs reliable control and line of sight. Instrument accuracy does not remove setup or control error. |
| Digital level | Plan around 1-3 mm for site level transfer and checks with the right levelling procedure | Benchmarks, slab levels, drainage grades, structural level verification | Strong for height. It does not solve horizontal position. |
| RTK GNSS | Plan around 20-30 mm horizontal and 30-50 mm vertical in open sky for site work | Broad feature pickup, earthworks, control transfer, roads, corridors, large open sites | Degrades near trees, buildings, cranes, walls and reflective surfaces. Use another method for millimetre set-out. |
| Static GNSS / AUSPOS-style control | Centimetre-level coordinates with long static observations and good processing | Datum connection, project control, regional control, remote site control | Use it for control, then use another method for set-out. Observation time, antenna setup and processing quality matter. |
| Drone photogrammetry | Plan around 20-50 mm horizontal and 30-100 mm vertical with good GSD, control and checkpoints | Open sites, progress surveys, stockpiles, earthworks, orthomosaics, visual records | Accuracy varies across the model. Shadows, vegetation, shiny surfaces and weak control reduce reliability. |
| Drone LiDAR | Plan around 30-80 mm absolute accuracy for terrain models, depending on sensor, altitude, GNSS/IMU, control and classification | Vegetated corridors, mine sites, rough terrain, roads, large-area topography | Strong for surface capture. It still needs ground control and survey QA before engineers rely on it. |
| Terrestrial laser scanning | Scanner points can reach millimetres at short range; controlled registered deliverables may sit around 2-10 mm locally | Complex as-builts, plant rooms, refurbishments, heritage, facades, structural context | A dense point cloud does not certify set-out on its own. Registration and control carry the result. |
For GNSS work, Geoscience Australia’s GNSS guidance explains the satellite positioning context. AUSPOS supports static GPS processing for higher-confidence datum and control work, not quick construction set-out.
For drone work, the mapping workflow controls the result. Pix4D’s discussion of relative and absolute accuracy separates a model that holds together internally from one tied to project control. The ASPRS Positional Accuracy Standards give a formal framework for horizontal and vertical accuracy reporting. In Australia, commercial drone operations also need aviation compliance, including CASA requirements such as the Remotely Piloted Aircraft Operator’s Certificate.
For a deeper comparison of ground methods, see our Total Station vs GNSS guide. For aerial capture, see the Drone Surveying Australia guide.
Accuracy by project stage
Tolerance tightens as the project moves from investigation to construction. A broad model can work early, then cause trouble if someone reuses it for detailed set-out.
| Project stage | Main decision | Practical tolerance band | Common methods | Evidence to ask for |
|---|---|---|---|---|
| Feasibility and site due diligence | Is the site viable and what constraints exist? | 50-200 mm, unless a known risk needs tighter survey | Existing data review, GNSS, drone, broad feature pickup | Source notes, datum notes, limitations and recommended next survey |
| Feature and level survey | Can architects and engineers design from this existing-conditions base? | 10-50 mm for hard features; 50-100 mm for natural surfaces depending scope | Total station, GNSS, digital level, drone where suitable | CAD/PDF deliverables, datum, control notes, feature scope and contours |
| Civil design and road corridor survey | What levels, cross sections and tie-ins are needed for design? | 10-25 mm for pavement, structures and services; 50-100 mm for natural surfaces | GNSS, total station, digital level, drone photogrammetry or LiDAR | Control report, cross sections, surface model, check data and exclusions |
| Construction set-out | Where exactly does the design go on site? | 2-10 mm depending element and specification | Total station, digital level, sometimes GNSS for broader civil works | Set-out records, offsets, mark descriptions, control used and check observations |
| Structural verification | Has the work been built in the right position and level? | 1-5 mm for tight structural checks; 5-10 mm for many general checks | Total station, digital level, laser scanning for complex geometry | Pickup report, deviation table, marked-up plans, point cloud where useful |
| Earthworks and volumes | How much material moved, and is the surface close enough? | 20-100 mm depending material, surface and contract | GNSS, drone photogrammetry, drone LiDAR, machine control data, ground checks | Baseline surface, current surface, volume report, checkpoints and assumptions |
| As-built handover | What did the contractor build, and can the owner rely on the record? | Project-specific; tighter for hard assets than terrain | Total station, GNSS, digital level, laser scanning, drone where appropriate | As-built plans, asset files, point clouds, control metadata, QA records |
Queensland Transport and Main Roads’ public surveying standards assign different relative uncertainty expectations to different surfaces and features. Structures, survey marks and pavement get tighter values than broad natural surfaces. Project owners should expect the survey brief to name these differences.
The common mistake: using broad-area data for millimetre set-out
The expensive mistake is using broad-area data for a tight construction decision.
A drone model captured for progress reporting can show bulk earthworks movement and stockpile volumes. Do not use it to set out columns, hold-down bolts, kerbs, drainage pits or structural penetrations. A GNSS pickup may suit broad open-ground levels. A tight building grid beside scaffolding and steelwork needs a different workflow.
When the tolerance sits in millimetres, the survey workflow needs:
- Stable project control protected from site disturbance.
- A method suited to the tolerance, such as total station and digital level for set-out and verification.
- Check observations, not a single occupation.
- Clear records of what was set out, what was checked and which design revision was used.
- A handover format that the builder, engineer, certifier or asset owner can use.
If the survey cannot prove the result, the project team still carries the risk.
When to combine methods
Large construction, civil and development projects often need more than one method.
| Workflow | Why it works |
|---|---|
| GNSS for control, total station for set-out | GNSS connects the project to datum and broader site control; the total station handles millimetre-level marks and checks. |
| Feature survey plus later construction set-out | The first survey gives designers an existing-conditions base; later set-out translates the approved design into field marks from controlled coordinates. |
| Drone survey for progress, ground survey for compliance | Drone data gives fast site-wide coverage; total station, GNSS or digital level checks verify the critical positions and levels. |
| Drone LiDAR for terrain, total station for structures | LiDAR can capture difficult terrain and vegetation; total station work verifies built elements and hard tie-ins. |
| Laser scanning for as-built complexity, survey control for reliability | Scanning captures dense geometry; survey control makes the point cloud usable for design coordination and handover. |
If you need a design-ready existing-conditions survey, see our Feature and Level Survey Perth service. If work is moving into site marks and verification, see Construction Set-Out Survey Brisbane. For South Australian projects, the same staged logic applies through our construction set-out survey Adelaide and as-built survey Adelaide teams, where tolerance tightens from feature survey through set-out to council and asset-owner handover.
Tolerance guide by project role
Builders, designers, civil contractors and asset owners carry different risks, so they ask different questions of the survey.
| Project role | What they need | Tolerance mindset | Useful HR Surveyors service |
|---|---|---|---|
| Builder | Set-out, slab checks, footing marks, services, as-built confirmation | Millimetres where work is being built; proof of design revision and control used | Construction set-out surveys |
| Civil contractor | Road alignment, pavements, drainage, batters, volumes, cross sections | Tight for hard assets and tie-ins; broader for bulk surfaces and natural ground | Road alignment surveys WA |
| Architect or designer | Existing features, levels, constraints, datum and contours | Accurate enough to design without false assumptions; clear limitations on what was captured | Feature and level surveys |
| Developer or project manager | Scope, cost, risk and staged evidence | Match the survey to each decision instead of buying one survey and reusing it for everything | Mining construction surveys WA for large regional projects |
| Asset owner | As-built records, point clouds, asset data and handover packages | Evidence matters as much as measurement: datum, metadata, QA and usable files | 3D laser scanning surveys |
| Drone and progress reporting team | Large-area surfaces, imagery, volumes and change over time | Repeatability, checkpoints and surface assumptions matter more than a single headline accuracy figure | Drone surveying services |
What to ask before booking a surveyor
Before you commission a survey, ask:
- What decision will this survey support?
- What tolerance applies to that decision?
- Is the tolerance horizontal, vertical or both?
- What datum and coordinate system are required?
- What contract clause, council requirement, road authority standard or asset owner requirement applies?
- What features, surfaces or assets must be captured, and what can be excluded?
- Is the site open sky, obstructed, vegetated, indoors, hazardous or actively changing?
- Do we need independent check shots, checkpoints or a conformance report?
- Will this survey be reused later for set-out, volume reporting or handover?
- What files and evidence will be delivered: CAD, PDF, LandXML, point cloud, report, photos, control register or asset data?
- Which design revision or surface model will be used?
- Who signs off the deliverable, and what do they need to see?
A good survey brief is specific. “Survey the site” gives the surveyor too much room to guess. “Capture existing hard features and levels to support design, tied to MGA2020 and AHD, with control suitable for later set-out” gives the team a usable brief.
Practical decision rules
Use these rules when you decide what level of survey accuracy to request.
| If the project needs… | Plan for… |
|---|---|
| Millimetre set-out | Total station and digital level workflow, stable control and set-out records. |
| Broad open-ground pickup | GNSS, drone or hybrid survey, with clear expectations for surface accuracy. |
| Design base for a development site | Feature and level survey tied to the required datum, with enough detail for the design team. |
| Vegetated or difficult terrain | Drone LiDAR or ground survey, plus control and checkpoints. |
| Complex as-built geometry | Laser scanning controlled by survey marks, plus a deliverable format the design team can use. |
| Volumes over time | Repeatable method, comparable surfaces, consistent boundaries and documented assumptions. |
| Asset handover | Survey method selected around owner requirements, certification and handover files. |
Common questions
What is total station accuracy in mm?
Survey-grade total stations can measure distances at the millimetre level under controlled conditions. On a construction site, the set-out result depends on control marks, instrument setup, sight lines, prism choice, distance, weather and survey checks. Tight set-out needs the whole workflow, not the instrument brochure.
Is GNSS accurate enough for construction set-out?
GNSS can work for many civil and earthworks tasks in open sky, especially where the tolerance sits in centimetres. Use total station and digital level methods for millimetre building set-out, structural elements or obstructed sites. Many projects use GNSS for control and broad pickup, then total station for set-out.
How accurate is drone surveying?
Drone photogrammetry can produce centimetre-level mapping when the team handles flight planning, ground control, GSD and checkpoints well. Plan around 20-50 mm horizontally and 30-100 mm vertically for many controlled construction and earthworks applications. Site conditions and processing still decide the result. Drone LiDAR helps with vegetation and terrain, but it still needs survey control and QA.
Is laser scanning more accurate than a total station?
Use a total station for precise set-out of individual design points. Use laser scanning when the project needs dense measured context across complex geometry. The strongest laser scanning deliverables are tied to survey control and checked against project requirements.
What accuracy should an as-built survey have?
An as-built survey should match the handover requirement for the asset being recorded. A drainage pit, kerb line, bridge component, building column, natural surface and stockpile do not need the same tolerance. Before fieldwork starts, confirm the required format, datum, attributes, asset owner standard and evidence required for sign-off.
Summary
Survey accuracy starts with risk. The tolerance tells you which method to use, what to check and what evidence to keep.
Use broad-area methods for broad decisions. Use tighter ground methods for tight tolerances. Combine methods when the project needs coverage and precision. Define the handover evidence before the survey starts. For a broader look at how these decisions fit into a surveyor’s day-to-day role, read our guide to what engineering surveyors do.
Need help scoping the right tolerance for your project? Contact HR Surveyors and tell us what decision the survey needs to support.