Hazard, Exposure and Vulnerability: The Risk Equation Behind Disaster Maps
A flood hazard map shows where water may reach. It does not show who lives there, or how badly they would be hurt. To answer those questions, you need a fuller idea of risk.
Disaster risk is usually described with three parts: hazard, exposure and vulnerability. This post explains each one, shows how each becomes a map layer, and walks through a simple overlay that combines them.
The risk equation
A common way to write the idea is:
Risk = Hazard × Exposure × Vulnerability
This is a concept, not a physical law. Different agencies and papers define the terms slightly differently, and some use addition or other combinations instead of multiplication. What matters is the logic: if any one part is missing, risk is low.
- A strong earthquake in an empty desert: high hazard, no exposure, so little risk.
- A town on a flood plain with a very high levee and a good warning system: high hazard, high exposure, but low vulnerability, so lower risk.
Keep this logic in mind. It explains why the same hazard can cause very different outcomes in different places.
Hazard: what can happen, and where
Hazard is the physical event or process that can cause harm. Examples are flooding, landslides, strong wind, heat or ground shaking. A hazard description usually includes where it can occur, how intense it can be, and how often it may happen.
On a map, hazard often becomes:
- A polygon layer of flood zones or landslide susceptibility classes.
- A raster layer (a grid of cells, each with a value) of flood depth, slope or wind speed.
Hazard layers are built from terrain, rainfall, soil, past events and models. If you want to see how these are made, see the earlier posts How Flood Hazard Maps Are Made and Where to Find Them and How Landslide Hazard Maps Are Made.
Exposure: what is in the way
Exposure means the people, buildings, roads, farmland and services located in a hazard area. It does not say how badly they would be damaged. It only says they are there.
Exposure layers are often:
- Points: hospitals, schools, shelters, power substations.
- Lines: roads, rail lines, pipelines.
- Polygons: building footprints, land-use areas, administrative districts.
- Grids: population per cell. The earlier post on grid population data covers day and night differences, which matter a lot here.
Good sources include national statistical offices, local government data and OpenStreetMap, which provides building, road and facility data in many regions. Check how complete the data is in your study area before you trust it.
Vulnerability: how easily harm happens
Vulnerability describes how likely people or assets are to be harmed, and how hard it is for them to cope and recover. It depends on the thing being exposed, not on the hazard itself.
Vulnerability can be physical or social:
- Physical: building age, structure type, number of floors, whether a basement or semi-basement dwelling is present.
- Social: share of elderly residents, people living alone, people with disabilities, low income, or limited access to transport and information.
Vulnerability is the hardest part to map. It is rarely a single clean dataset. You usually build it from several indicators and combine them into an index. The choice of indicators is a judgment, so write it down and be ready to explain it.
The earlier post on cold-wave vulnerable areas shows one example of this kind of indicator thinking.
Why a hazard map alone is not a risk map
Imagine two neighborhoods inside the same flood zone. The hazard layer gives them the same color.
- Neighborhood A is a park and a parking lot. Few people are there.
- Neighborhood B has dense housing, a care home for older adults, and many ground-floor apartments.
The hazard is the same. The risk is not. A hazard map answers "where can this happen?" A risk map answers "where would it matter most?"
This difference changes decisions. If you plan shelters, inspections or evacuation support using only a hazard map, you may spend effort on empty land and miss crowded or fragile places just outside the mapped zone.
The reverse also happens. A small hazard zone can still be a priority if a hospital sits inside it.
A simple overlay workflow
Overlay analysis means stacking layers and combining their values for each location. A simple, transparent version works well for learning and for first screening.
Step 1: Pick a common unit
Choose the unit where you will combine everything. Options are administrative districts, a regular grid (for example, 100 m cells), or individual buildings. Grids are easy to compare. Districts match how decisions are made.
Step 2: Align the layers
All layers must use the same coordinate system. If they do not, they will not line up. Use a projected system suitable for your area when measuring distance or area. See the earlier post on EPSG:4326 and EPSG:5186, and check codes on epsg.io.
Step 3: Score each component
Convert each layer to a simple scale, such as 1 (low) to 5 (high). For example:
- Hazard: flood depth classes become scores 1 to 5.
- Exposure: population per cell, split into five classes.
- Vulnerability: an index from elderly share and building age, split into five classes.
Be careful with the class breaks. Equal intervals, quantiles and expert thresholds give different maps. State which one you used.
Step 4: Combine
Multiply or add the scores. With multiplication, a zero in one part gives zero risk, which matches the equation. With a 1 to 5 scale, nothing is zero, so many people instead use a weighted sum. Both are valid if you explain them.
In QGIS, you can do this with the Raster Calculator for raster layers, or with a spatial join and the Field Calculator for vector layers. In Python, GeoPandas can do the same with a spatial join, as in this short sketch:
import geopandas as gpd
cells = gpd.read_file("grid.gpkg")
hazard = gpd.read_file("flood_zone.gpkg").to_crs(cells.crs)
joined = gpd.sjoin(cells, hazard, how="left", predicate="intersects")
# then score the columns and combine them
More on this tool is on the GeoPandas site, and QGIS is available at qgis.org.
Step 5: Check the result
Look at the map and ask if it makes sense. Compare it with local knowledge and past events if you have them. Change a weight and see what moves. If the top-ranked areas change a lot, your result is sensitive to your choices, and you should say so.
Common mistakes
- Mixing up exposure and vulnerability. Population count is exposure. Age or building quality is vulnerability.
- Using different dates. Old population data with a new hazard model can mislead.
- Ignoring scale. Combining a coarse district layer with a fine grid hides local detail.
- Showing a single score as certain. A risk index ranks places. It does not predict losses.
What to try next
Pick a small area you know, such as a district or campus. Download a flood or landslide hazard layer, and get building or population data for the same area. Build a 1 to 5 score for each of the three parts, combine them in QGIS, and compare the result with the hazard map alone.
Then change one weight and see which places move. Note what surprises you.
The next post, Satellite Imagery Basics: Resolution, Bands and Revisit Time, looks at the data that often feeds the hazard layer.
More materials (some in Korean) are on GeoAI for Disaster Safety.
This post was drafted with the help of AI tools and checked before publishing. Tool features and data portals change over time, so check the official documentation when you practice.
Comments
Post a Comment