LiDAR and DEM: Terrain Data in Disaster Prevention

In short
- A DSM is terrain data that holds the heights of the tops of buildings and trees; a DEM holds the height of the bare ground.
- LiDAR lasers pass through gaps between leaves and capture reflections from the ground.
- If terrain accuracy is low, the predicted extent of floods and landslides can be badly wrong.
Imagine you are mapping a landslide risk area, but in the satellite image the forest is a dense green cover, and you cannot see at all how steep the slope below it is. Knowing only the height of the treetops, you cannot calculate where water will really flow or where soil will really fail. In disaster prevention, terrain data is not just a background map; it is the very input used to calculate flood extent and landslide paths. This post looks at the relationship between LiDAR (a sensor that measures distance with lasers) and the DEM (digital elevation model, data that holds the height of the ground as numbers), the key technologies that produce that input.

What is the difference between a DSM and a DEM?
A DSM (Digital Surface Model) records the height of the "topmost surface" as seen from a drone or aircraft. It includes everything: rooftops, treetops and even power lines. A DEM (Digital Elevation Model), on the other hand, removes such structures and vegetation and keeps only the height of the bare ground. As an analogy, a DSM is the outline of a person photographed under a blanket, and a DEM is the real outline of the body with the blanket removed. To calculate the path water takes or the direction a slope fails, you need information about the body (the terrain), not the blanket (buildings and trees). That is why flood simulation and landslide modeling almost always use a DEM, not a DSM, as the basic input.
How does LiDAR see the ground beneath trees?
LiDAR is equipment that fires laser pulses from a sensor several hundred thousand times per second and calculates distance by measuring the time the reflection takes to return. When it flies over a dense forest, some of the lasers reflect directly off the leaves, but lasers that pass through gaps between leaves reach the ground and reflect from there. When several reflections return from one laser pulse, these are called multiple returns. The first return usually comes from the treetop, and the last return usually corresponds to the ground. If you collect and interpolate only these last return points, you can make the ground height hidden by the forest, which satellite or drone photos cannot show, that is, a DEM. To calculate slope and slope direction accurately in landslide susceptibility analysis, terrain with the vegetation removed in this way is essential.
Why does a small terrain error change the flood extent?
Flood simulation rests on the simple principle that water flows from high places to low places, but the calculation is sensitive to small height differences in the terrain. For example, if an embankment or a small mound is shown just a few tens of centimeters too low in the DEM, the model can wrongly decide that water overflows at that point. Conversely, if a drainage ditch that is really low is smoothed out because the data resolution is coarse, the calculation may treat it as if the water has no way out. So flood analysis requires a DEM with a fine grid spacing (resolution) and a small vertical error, and this sensitivity grows in places with complex drainage structures, such as cities. In practice, keep in mind that even under the same rainfall conditions, the shape of the predicted flood map can change noticeably with the resolution and accuracy of the DEM used.
Why is terrain even more demanding in landslide simulation?
Landslide modeling combines terrain variables such as slope, slope direction and slope curvature (concave or convex) to calculate risk. These variables are derived by differentiating the DEM once or comparing it with neighboring cells, so if the original DEM has small noise or a stair-step effect, the slope calculation can come out more jagged than the real terrain. In heavily vegetated mountains in particular, the ground can hardly be seen without LiDAR, so in some cases a coarse DEM made by digitizing contour maps was used in the past, and risk zone boundaries were redrawn when it was replaced by a high-resolution LiDAR-based DEM. The landslide risk grade of the same village can change depending on which DEM was used, so when you receive a result map, make a habit of checking the source and production year of the terrain data it is based on.
How can you check whether the terrain data you use is trustworthy?
Before using terrain data as it is, there are several things you should check yourself. First, look at the metadata (descriptive information about the data) to see whether the DEM is based on LiDAR, on satellite stereo imagery, or converted from contour lines. If you make slope maps of the same area from two DEMs with different resolutions, you can see by eye how much the resolution difference affects the result. It is also a good check to compare against a satellite map or site photos to see whether drainage ditches or embankments that stand out at the site are properly represented in the DEM. If you use terrain data in a disaster response plan without such checks, you may miss real risk points, or conversely classify a safe place as a risk area.
Step by step on screen
The figure below is a diagram that summarizes the order of the steps. Menu positions are the same in the English interface.

- Check the terrain data
On the National Spatial Data Infrastructure Portal (Korea) or a public map service, search whether DEM or DSM data is available for your area - Check your QGIS installation
Start QGIS and check that the Raster entry appears in the top menu (its position may differ by version) - Load the DEM
Open the downloaded DEM file (.tif) with Layer > Add Layer > Add Raster Layer - Make a hillshade
Choose Raster > Analysis > Hillshade to see the relief of the terrain visually - Calculate slope
Choose Raster > Analysis > Slope to create a slope layer and check its colors - Compare resolutions
If you have DEMs of different resolutions for the same area, display them side by side and compare the slope differences by eye
Menu names may differ slightly between program versions.
Prompts you can copy
Draft an easy explanation of the difference between DEM and DSM for a report or teaching material
Your role is a writing assistant who prepares teaching material for public officials who are new to GIS data. Write an explanatory paragraph for a flood or landslide analysis report on [target area]. Explain the difference between DSM and DEM, focusing on the presence of buildings and vegetation, using an easy analogy. Summarize how LiDAR obtains terrain beneath vegetation in three sentences or fewer. Then summarize in one paragraph why a DEM was used for this analysis. When a technical term appears, add a short explanation in parentheses, and do not include speculative numbers.
Replace the bracketed parts for your situation. Check the facts in any answer before you use it.
Try it yourself
- In a public map service, switch back and forth between the satellite image and a terrain shaded relief of the same point, and check whether the buildings and trees disappear.
- In QGIS, load one DEM and compare the relief and slope by eye with the Hillshade and Slope tools.
- Find the production year and source (LiDAR or converted from contours) of the DEM for your home area or an area of interest in its metadata.
Common questions
Q. Which is more accurate data, a DSM or a DEM?
A. The idea of accuracy is different for each. For a DSM, what matters is how finely it captures the whole surface; for a DEM, how faithfully it extracts the ground height. The key is to choose data that fits the purpose.
Q. Can I make a landslide hazard map without LiDAR?
A. You can, but in heavily vegetated areas the terrain representation can become coarse. DEMs based on contour lines or on satellite stereo imagery are also used, but for seeing fine changes in slope in mountains, a LiDAR-based DEM is often better.
Q. Can ordinary people get LiDAR DEM data?
A. Whether it is public, and at what resolution, depends on the region and the agency. The most accurate way is to search yourself whether it is offered on a national spatial information portal or an open data platform run by a local government.
What comes next
In the end, both flood maps and landslide hazard maps rest on terrain data, and how truly that data captures the real shape of the ground determines the reliability of the result. A later post in this series looks at how terrain data made this way is actually fed into disaster simulation models and calculated.
This is an English adaptation of a post in the Korean GeoAI basics series. The original is here, and more learning materials are on GeoAI for Disaster Safety.
This post was drafted with the help of AI tools and reviewed by the author. Tool features and data portals change over time, so check the official documentation when you practice.
Comments
Post a Comment