DTM vs DSM in Viewshed Analysis | Certified Energy

Two viewshed analyses can use the same observation point and target height yet produce different visibility patterns. One reason is the type of elevation data used to represent the surrounding landscape.

Digital Terrain Models (DTMs) and Digital Surface Models (DSMs) describe different aspects of ground and surface elevation. The distinction can influence whether hills, buildings, trees and other elevated features are represented as potential obstructions in a visibility calculation.

For architects, planners, developers and infrastructure project teams, understanding the difference helps clarify what a viewshed model is testing and how confidently its results can be applied to a particular project.

A Digital Terrain Model represents the underlying ground surface, while a Digital Surface Model represents the elevation of the surfaces captured in the dataset. Neither automatically provides a complete representation of real-world visibility.

A Digital Terrain Model represents the ground elevation, generally excluding above-ground features such as buildings and vegetation.

A Digital Surface Model represents the elevations of surfaces captured by the source data. Depending on the dataset, these may include rooftops, tree canopies and other elevated features.

In viewshed modelling, a DTM is commonly used to examine terrain-based visibility. A suitable DSM may provide a more detailed representation of existing physical screening, although its usefulness depends on resolution, currency and how the data was processed.

The appropriate model depends on the project question. A more detailed surface model is not necessarily the better choice for every visibility assessment.

Modelling distinction: A terrain-based viewshed may identify locations with theoretical visibility in the absence of detailed above-ground screening. A surface-based viewshed may incorporate existing obstructions represented in its elevation data. Neither result should be interpreted as a complete visual impact assessment.

A Digital Terrain Model describes the elevation of the underlying ground surface. It represents landform features such as ridges, valleys, slopes and changes in terrain elevation.

In many GIS workflows, a DTM is produced from elevation measurements that have been classified or processed to represent bare-earth conditions. Above-ground features such as buildings and tree canopies are generally removed from the ground surface representation.

This makes a DTM useful when the assessment question concerns the influence of landform on visibility.

For example, a terrain-based viewshed can help identify whether a ridge is likely to obstruct a line of sight between a nominated observation point and a proposed structure.

However, a bare-earth model does not ordinarily represent screening from existing buildings or vegetation unless these features are introduced through additional modelling inputs.

A Digital Surface Model represents the elevations of surfaces captured by an elevation dataset. These can include the ground, rooftops, vegetation canopies and other above-ground features.

A DSM may be generated using LiDAR, photogrammetry or other remote-sensing methods. The features represented depend on the source data, acquisition conditions and processing approach.

For visibility modelling, a suitable DSM can introduce physical screening that would be absent from a bare-earth terrain model.

For example, an existing building may interrupt a modelled sightline across an urban site, while a mature tree canopy may influence visibility in a landscaped area.

A DSM does not necessarily represent every wall, branch, opening or other small-scale feature accurately. Its resolution and completeness must be considered before relying on it for detailed visibility conclusions.

Viewshed calculations evaluate the geometric relationship between observation locations, target locations and the elevations represented along a line of sight.

When a DTM is used, the calculation generally tests whether the landform itself obstructs visibility. Areas behind ridges or elevated terrain may be classified as non-visible, while areas without terrain obstruction may be classified as potentially visible.

When a DSM is used, represented buildings, vegetation and other elevated surfaces may create additional obstructions. This can change the extent and distribution of modelled visibility.

The difference does not mean that one model is correct and the other is incorrect. They may represent different assumptions or answer different questions.

The critical issue is whether the selected elevation surface corresponds with the purpose of the analysis and is clearly documented in the methodology.

Terrain-based viewshed modelling can be useful when the main objective is to understand the influence of topography on potential visibility.

This may be relevant to early-stage site selection, infrastructure corridors, renewable energy proposals or broader landscape visibility investigations.

By excluding detailed above-ground screening, a bare-earth model can help reveal the underlying geographic relationship between a proposal and surrounding observation areas.

However, a DTM-based result should not automatically be described as a guaranteed maximum visibility extent. The outcome also depends on target geometry, observer heights, model resolution, study boundaries and other calculation settings.

For project applications involving infrastructure and larger landscapes, see our article on viewshed analysis for renewable energy and infrastructure projects.

A DSM may be useful where existing buildings, vegetation or other surface features are relevant to the visibility question.

In developed urban environments, surrounding building forms can influence which locations have a potential line of sight towards a proposed development.

In vegetated landscapes, tree canopies may also influence modelled visibility, provided the dataset adequately represents the relevant vegetation.

However, the presence of a tree canopy in a surface model does not mean that the model accurately describes the transparency of foliage or seasonal changes in screening.

A surface-based result should therefore be interpreted in relation to the dataset's date, resolution, acquisition method and the permanence of the features represented.

Elevation datasets are commonly represented as grids or other digital surfaces. The spatial resolution determines the size of the cells or sampling intervals used to describe the landscape.

A coarser dataset may be sufficient for identifying broad terrain relationships across a regional study area, but it may not represent narrow ridgelines, local embankments or smaller obstructions adequately.

Higher-resolution data can represent more local detail, although finer cell size alone does not guarantee greater accuracy. Source measurement quality, processing, vertical accuracy and data currency remain important.

A detailed DSM may also contain irregularities or temporary features that require interpretation before being used in a visibility model.

The appropriate resolution should be selected in relation to the scale of the project, the assessment question and the reliability of the available elevation information.

The landscape represented in an elevation dataset may not match current site conditions.

Earthworks, new buildings, vegetation removal, tree growth and infrastructure construction can change the surfaces that influence visibility.

These changes can be particularly important for a DSM because the model may contain above-ground features that no longer exist or omit structures constructed after the data was acquired.

Terrain models can also become outdated where substantial grading, excavation or landform modification has occurred.

A viewshed methodology should identify the source and date of the elevation data, together with any material changes or limitations known at the time of assessment.

A viewshed workflow does not always need to rely on a single unmodified elevation dataset.

Depending on the assessment, a modeller may use terrain data together with separately represented buildings, proposed structures or other relevant obstructions.

This can help distinguish existing landform conditions from selected built features or proposed changes.

For example, a terrain model may establish the underlying topography while a proposed building is introduced at a specified location and height. Additional screening elements may be incorporated where suitable evidence is available.

The resulting model requires consistent elevation references, appropriate geometric inputs and clear documentation. Combining datasets without checking their alignment or accuracy can introduce new uncertainty.

Comparing terrain-based and surface-based viewsheds can help identify how represented above-ground features influence a modelled visibility pattern.

However, the comparison is most useful when observation locations, target heights, analysis extent and other relevant settings are held consistent.

If the datasets also differ in resolution, collection date or elevation accuracy, some differences may arise from data quality rather than physical screening alone.

A comparison should therefore explain the assumptions behind each model and avoid attributing every changed visibility cell to buildings or vegetation.

For guidance on reading the resulting map classes, see How to Read and Interpret a Viewshed Map.

The choice between a DTM, DSM or combined modelling approach should begin with the question the assessment needs to answer.

If the purpose is to investigate the influence of landform, a suitable terrain model may be appropriate. If existing buildings or vegetation are material to the question, a surface-based or supplemented model may be considered.

The project team should also consider the availability and quality of elevation data, the relevant observation and target locations, and the scale of the intended output.

For professional assessment work, the methodology should explain why the selected elevation representation is appropriate and identify limitations that affect interpretation.

Certified Energy's Viewshed Analysis Knowledge Pillar provides the wider context for visibility modelling and its applications in planning and design.

Not every project has access to recent, high-quality elevation data covering the full study area.

Where datasets contain gaps, inconsistent elevations or insufficient detail, the modeller may need to investigate alternative sources, adjust the assessment scope or document additional uncertainty.

A surface model collected before major construction activity may no longer represent existing buildings. A terrain model derived from coarse data may also omit smaller landform features relevant to local visibility.

Combining datasets from different sources requires attention to coordinate systems, vertical datums, resolution and alignment.

The appropriate response depends on the significance of the missing information. A model should not be presented as more precise or complete than its inputs support.

Consider a proposed development on elevated land overlooking a partly urbanised valley.

An initial viewshed calculation uses a Digital Terrain Model to identify locations with potential visibility based on the surrounding topography. The output shows several potentially visible areas along nearby roads and residential land.

A second model uses suitable surface elevation data representing existing buildings and tree canopies. With the same relevant observation and target settings, some locations are classified differently because additional obstructions are represented.

The project team reviews the acquisition date and resolution of the surface data before interpreting those differences. It also considers whether existing vegetation and buildings have changed since the dataset was collected.

The comparison helps explain the influence of modelled screening but does not establish the visual significance of the proposed development.

This is an illustrative example rather than a reported project outcome. It demonstrates how the choice of elevation representation can influence visibility modelling results.

What is the main difference between a DTM and a DSM?

A Digital Terrain Model represents the underlying ground elevation. A Digital Surface Model represents the elevations of captured surfaces, which may include buildings, tree canopies and other above-ground features.

Is a DSM more accurate than a DTM for viewshed analysis?

Not necessarily. A DSM may represent relevant above-ground obstructions, but its suitability depends on the assessment purpose, source accuracy, resolution and currency. A DTM may be more appropriate where the objective is to assess terrain-based visibility.

Does a DTM include trees and buildings?

A bare-earth DTM generally excludes trees and buildings. These features may be incorporated separately where the modelling approach and available information support their inclusion.

Does a DSM always include vegetation?

No. The vegetation represented depends on the data acquisition method, processing, resolution and conditions at the time of collection. A DSM should not be assumed to capture every tree or accurately represent foliage transparency.

Can DTM and DSM viewsheds produce different results?

Yes. A DSM may include surface obstructions that are absent from a bare-earth DTM. Differences may also arise from resolution, acquisition dates and other dataset characteristics, so the methodology should be reviewed before comparing results.

Which elevation model should be used for a planning project?

The choice depends on the visibility question, project setting, available data and assessment scope. A suitably qualified modeller should establish whether terrain-based, surface-based or combined modelling is appropriate and document the relevant limitations.