Differences between DEM, DSM and DTM?
I am listing some of the acronyms used in our work and had to stop to think about whether to list these together or separately:
- DEM: Digital Elevation Model
- DSM: Digital Surface Model
- DTM: Digital Terrain Model
In some circles they seem synonymous, in others they seem to have distinct meanings.
Is there an agreed-upon definition for each that enables them to be distinguished?
In my experience, DEM is most of the time used as a generic term for DSMs and DTMs. I think this image on Wikipedia depicts the differences between DSMs and DTMs well:
- DSM = (earth) surface including objects on it
- DTM = (earth) surface without any objects
A different definition is found in [Li et al., DIGITAL TERRAIN MODELING - Principles and Methodology]:
DEM is a subset of DTM and the most fundamental component of DTM.
In practice, these terms (DTM, DEM, DHM, and DTEM) are often assumed to be synonymous and indeed this is often the case. But sometimes they actually refer to different products. That is, there may be slight differences between these terms. Li (1990) has made a comparative analysis of these differences as follows:
- Ground: “the solid surface of the earth”; “a solid base or foundation”; “a surface of the earth”; “bottom of the sea”; etc.
- Height: “measurement from base to top”; “elevation above the ground or recognized level, especially that of the sea”; “distance upwards”; etc.
- Elevation: “height above a given level, especially that of sea”; “height above the horizon”; etc.
- Terrain: “tract of country considered with regarded to its natural features, etc.”; “an extent of ground, region, territory”; etc.
Yes, there are numerous different definitions. I added a different one plus it's source.
I also thought that DEM was used as a generic name for a DSM/DTM. Whats a DEM then?
@Underdark I find this quotation deeply confusing, because it does not reveal what distinctions are intended between "height," "elevation," and "terrain." (Indeed, the description of terrain seems to imply they are talking about LULC data, not DEMs or DTMs!) Do you have a link that would provide fuller context to understand this?
@Underdark Thank you. The correct link appears to be http://www.torabzadeh.bloghaa.com/files/2010/03/1.pdf, which exists but seems to run at a glacial pace... I'll keep trying.
@Underdark I found it at http://tehran.academia.edu/aliasgharheidari/Books/194823/DIGITAL_TERRAIN_MODELING_Principles_and_Methodology. The fundamental distinction they make between DEM and DTM is that the DTM "will attempt to incorporate specific terrain features such as rivers, ridge lines, ..." Of all the answers so far, this distinction is clearly brought out only by @WolfODrade. In particular, the Wikipedia article on DEMs is useless in this respect.
@underdark I do not think this is quite right. DTM has been misidentified by Li 1990 or it changed from 1990 to 2000 when I entered grad school to do GIS. AS 1990 is pre say lidar this may be the reason.
Digital elevation models (DEM) are a superset of both digital terrain models (DTM) and digital surface models (DSM). Remote sensing generally captures the surface height, so the top of the tree canopy or buildings is returned, not the bare ground elevation. If this data is corrected to remove elements which extrude above the terrain height, you're left with a DTM.
In general, most people use DEM interchangeably with the other two terms, but it can matter: I once built a hydrology model using SRTM data in South America in very flat arid terrain, but because of the canopy height along the river itself, the true river location became the highest point on the terrain, causing a ruckus.
The Wikipedia article on digital terrain models also includes some useful background and examples you may find helpful.
That article says TINs are 3 dimensional. I'd always thought they were 2.5 dimensional, meaning they cannot model solids, only surfaces. Would you say DEM's in general are 2.5 dimensional?
In my experience, a more complete answer to this question lies in defining the difference between a DEM, DTM and a DSM. A DTM is NOT a generic name covering both DEMs and DSMs. So...
A DEM is a 'bare earth' elevation model, unmodified from its original data source (such as lidar, ifsar, or an autocorrelated photogrammetric surface) which is supposedly free of vegetation, buildings, and other 'non ground' objects.
A DSM is an elevation model that includes the tops of buildings, trees, powerlines, and any other objects. Commonly this is seen as a canopy model and only 'sees' ground where there is nothing else overtop of it.
A DTM is effectively a DEM that has been augmented by elements such as breaklines and observations other than the original data to correct for artifacts produced by using only the original data. This is often done by using photogrammetrically derived linework introduced into a DEM surface. An example is hydro-flattening commonly seen in elevation models done to FEMA specifications
Incidentally, a DEM is far cheaper to produce an a DTM.
Just picky semantics, but when you say "unmodified from its original data source", does that not include the filtering of the non-ground returns?
Only in very recent years have dem been related to lidar. The usgs has been creating dem for many years.
Experts in the field (http://bloglidar.wordpress.com/2010/12/06/dem-dsm-dtm/) make kind of is same distinction that a 'DTM' is more than a generic term for 'some sort of surface in digital form'
@WolfOdrade this is the best answer due to the DTM component.
@WolfOdrade this answer seems to provide the best distinction between the three terms, however it would be best if backed up with a peer-reviewed source (the blog post above is probably not sufficient). Are you aware of any?
@WolfOdrade this whitepaper seems to back up your definitions, if you agree, I suggest adding it to your answer as a reference: http://www.nrcs.usda.gov/internet/fse_documents/stelprdb1047930.pdf
It is not uncommon to find in the LiDAR/forestry literature the acronyms DTM and DEM being used as synonyms.
However, according to Behrendt (2012; paragraph 6 on page 15), there is a difference between them:
DTM: bare-earth representation with irregular spaces between points (non-raster).
DEM: gridded raster representation of the DTM.
Still according to Beherendt (2012; paragraph 4, on page 15):
DSM as a raster. This represents the first echo the laser received for each laser pulse sent out, and represents the tops of buildings, trees, and other objects, or the ground, if unobstructed.
And to complete, a CHM (Canopy Height Model) or nDSM (normalized Digital Surface Model) is the DEM subtracted from the DSM, i.e., CHM = DSM - DEM.
Behrendt, R. Introduction to LiDAR and forestry, part 1: a powerful new 3D tool for resource managers. The Forestry Source, p. 14-15, set. 2012. link.
digital surface model (dsm) – a first-reflective-surface model that contains elevations of natural terrain features in addition to vegetation and cultural features such as buildings and roads.
digital terrain model (dtm) – a bare-earth model that contains elevations of natural terrain features such as barren ridge tops and river valleys. Elevations of vegetation and cultural features, such as buildings and roads, are digitally removed.
In my experience a DEM refers to a raster based elevation model used in programs like ESRI's ArcGIS. A DTM is a vector based elevation model used by programs like Autodesk CivilCad.
In this context raster and/or vector are merely the storage format or representation method, not a defining characteristic. A DEM can be represented as points or a triangular irregular network (TIN) and a DTM as seamless raster surface. Downvoters: please take the time to state why you think an answer is inappropriate so posters have an opportunity to learn -- especially when they're new here.
@Matt I'm confused. If the DTM/DEM distinction is not based on format, then it must depend on a difference in meaning between the words "terrain" and "elevation." Exactly what difference is being emphasized, given that both (apparently) apply to "bare earth" models?
@whuber We built drainage enforced elevation models for our region from contours, watercourses, and other ancillary data. Following WolfOracle's definitions, this product would properly be called a DTM. However in the 12 years since generated we've referred to it in publications etc as a Digital Elevation Model, perhaps in error. It's always been distributed in a raster format but some have converted it to TIN etc. to work better in their native environment. In general use I don't know that the elevation/terrain distinction is truly useful without prefacing the discussion with definitions.
@Matt Thank you. It appears @WolfOdrade has captured the distinction best (and better than the Wikipedia article). This also helps me appreciate the reply here by @Chris, because typically, when you are incorporating terrain features explicitly, you can best do that with a vector format. If that DTM later gets converted to a grid we no longer have an *explicit* representation of rivers, ridges, and so on. So there seems indeed to be an intimate, if somewhat nebulous, relationship between the underlying format and being a DEM or a DTM.