Mapping Population Exposure To Flood Hazard
Mapping Population Exposure To Flood Hazard
Mapping Population Exposure To Flood Hazard
This guide was prepared by Aahlaad Musunuru and Ayodele Marshall, under the supervision
of Rikke Munk Hansen.
We thank Daniel Clarke, Linda Li and Soheil Rastan for inspiring us to develop this guide
and the authors of the operational manual1 “Exposure to Hazards Assessment based on
‘POP-to-GUF’ Methodology”,2 upon which the methodology of the present guide builds. We
also thank Amin Shamseddini, Anthony Dvarskas and Borra Himasri for testing the steps in
the guide and providing feedback. Further, the development of the guide benefitted from
insights provided in the 2019 Asia-Pacific Disaster Report3 and the 2019 UN Global
Assessment Report on Disaster Risk Reduction.4
1 Developed by Jean-Louis Weber (Consultant) with Daniel Clarke (UNESCAP), Trevor Clifford (Consultant), and
Gao Xian Peh (Consultant), with feedback from the Disaster Management Agency of Indonesia (BNPB), the Dept.
of Disaster Prevention and Mitigation (DDPM) of Thailand and the National Statistical Office (NSO) of Thailand.
https://www.unescap.org/kp/2021/exposure-hazards-assessment-based-pop-guf-methodology
https://stat-confluence.escap.un.org/download/attachments/16155653/poptoguf_manual_final1.pdf?version=1&
modificationDate=1629883284705&api=v2
2 POP-to-GUF Methodology assigns an estimated average population density to a selected geographical area by
integrating population census data with Global Urban Footprint gridded data.
3 The Asia-Pacific Disaster Report is a biennial flagship publication of the United Nations Economic and Social
on worldwide efforts to reduce disaster risk. The GAR is published biennially by the UN Office for Disaster Risk
Reduction (UNDRR). https://gar.undrr.org/report-2019
For illustrative purposes we have focused on one type of flood hazard, riverine floods 5 and
one geographical area, Lao People’s Democratic Republic. The map generated is shown in
Figure 1 below together with explanation of the applied methodology. The guide explains how
to generate similar maps for hazard types and geographical areas of your own choosing.
This exercise uses a very specific version of QGIS (3.10.2). We advise that you download
and use the same version of QGIS when going through the steps of this guide. Appendix
A. QGIS gives instructions on how to download and set up this version of QGIS.
Appendix B. Available datasets provides a list of open-source datasets that may be useful
for your own purposes.
The guide uses Copernicus Global Land Cover data, Census data and Administrative
Boundaries data and Flood Hazard data to generate the map in Figure 1 below. The guide
explains the use of open-source software, namely Quantum Geographic Information System
(QGIS) to construct the map. The guide takes you through 3 steps:
Step 1: Download the data you want to work with from available open sources – Data
Downloading.
Step 2: Strip and clean the data downloaded to filter the layers relevant to your chosen
geographical area – Data pre-processing
Step 3: Produce the Percentage of Population exposed to flood hazard map – Data
processing
In total, these steps should take approximately 3 hours to complete from start to finish.
Before starting step 1, make sure you have QGIS installed (3.10.2) on your computer.
5 Various climatic and non-climatic processes can result in different types of floods: riverine floods, flash floods,
urban floods, glacial-like outburst floods and coastal floods. Riverine flooding occurs when excessive rainfall over
and extended period of time causes a river to exceed its capacity. It can also be cause by heavy snow melt and
ice jams.
The land cover data contains land cover pixels, with each pixel representing an area of land
of 100 square metres. We first classify the Land Cover data for Lao People’s Democratic
Republic into 2 categories – built-up areas and non-built-up areas.
We then assign a population density pixel level to the administrative boundaries data and
census data, based on a relative proximity to accumulated built-up areas. Population density
is higher in built-up areas where land is scarce, than in non-built-up areas areas where land
is more abundant. The accumulated built-up areas with the maximum population density are
placed in the centre of a grid and the further away the administrative boundary is from the
centre, the lower its assigned value.
Population density is estimated by applying Gaussian Filtering to the datasets using QGIS.
Gaussian Filtering is an image smoothing technique which helps to transform pixel population
density into normalized values based on the described grid pattern, shown in the picture
below.
The map below illustrates the percentage of population exposed to riverine flood hazard.
Different levels of exposure are indicated by differences in colour. The darker the colour, the
greater the exposure to riverine flood hazard. The parameters for threshold exposure can be
determined by the user.
We start with downloading the Administrative Boundaries data for Lao People’s Democratic
Republic from HDX Data (Humanitarian Data Exchange) which provides shapefile6 boundaries
for Lao People’s Democratic Republic. Please note that the Appendix outlines examples of the
most frequently used administrative data and we encourage you to investigate these examples.
6The shapefile format is a geospatial vector data format for geographic information system (GIS) software. It is
developed and regulated by Esri as a mostly open specification for data interoperability among Esri and other GIS
software products. The shapefile format can spatially describe vector features: points, lines, and polygons,
representing, for example, water wells, rivers, and lakes. Each item usually has attributes that describe it, such
as name or temperature.
Now we are going to Download sub-national level population data for Lao People’s Democratic
Republic from Humanitarian Data Exchange (HDX)
5. Copy the columns “ADM1_EN, ADM1_PCODE, T” into a new excel sheet and save as
“lao_admpop_adm1_2020_data.csv”
This information is used to calculate the total population in the admininstrative boundaries and
used to show how exposed the population is to flood hazard.
If the link was updated and 2020 data sets are not available, users can use the file produced in
the step above (“lao_admpop_adm1_2020_data.csv”), and continue the exercise using this file
Users can also use population data sets from other sources, such as City Population
(https://www.citypopulation.de/en/laos/admin/)
Now we are going to download Built-up Area datasets for the year 2019 from Copernicus Land
Monitoring Service website. The service is designed to monitor the variance of Built-up area,
Forest and Cropland Land cover categories. The data on the Copernicus Land Monitoring Service
To download the Built-up Area land cover data used in this exercise:
3. In the new window that appears, under “Cover Fractions,” select the “Built-up area” and
under “Click on the year to start the download,” select “2019” as shown in the picture
below:
Now we are going to download the Global Flood Hazard dataset from United Nations Environment
Programme Global Resource Information Database (UNEP GRID). The flood hazard assessment
is done using a probabilitistic approach for modelling riverine floods from the major river basins
around the globe. This guide uses the flood hazard data for 200 years as suggested by the GAR
Global Risk Assessment Report 8.
8
https://www.preventionweb.net/english/hyogo/gar/2015/en/gar-pdf/Annex1-GAR_Global_Risk_Assessment_Data_
methodology_and_usage.pdf
5. Click “Download” and save the downloaded data in a folder named “Downloaded Data”
or the original folder created.
More information on how to download and install and use QGIS is provided in the Appendix to
this guide.
The steps in this section prepare the downloaded data from Step 1 for the population exposure
analysis to be done in Step 3.
For this exercise, we created a new folder and named it “Data Pre-Processing”. We encourage
the user to do the same or choose a folder name and location that is convenient.
1. Click on “Layer”
2. Go to “Add Layer”
3. Select “Add Vector Layer 9” as shown in the picture below:
9 Vector layers are, along with raster layers, one of the two basic types of data structures that store data. Vector
layers use the three basic GIS features – lines, points, and polygons – to represent real-world features in digital
format.
The next steps add the Lao People’s Democratic Republic Admin Boundaries to QGIS:
10A raster consists of a matrix of cells (or pixels) organized into rows and columns (or a grid) where each cell contains
a value representing information. Rasters are digital aerial photographs, imagery from satellites, digital pictures, or
even scanned maps.
In this step, we clip the Lao People’s Democratic Republic built-up area from the raster files
downloaded.
1. For first-time QGIS downloads, click the button highlighted in the picture below, as
the “Processing Toolbox” may not appear automatically
11
A single-precision floating-point format is a computer number format, usually occupying 32 bits in computer
memory
1 thru 100 = 1
12 Make sure to open QGIS with the GRASS plug-in. More information is given in the Appendix to this guide
These steps show how to generate a Gaussian filter to the Lao People’s Democratic Republic
Built up area Layer in QGIS. The Gaussian filter is used to estimate the population density of each
administrative area by considering the characteristics of each area.
Gaussian Filtering is an image smoothing technique which helps to transform pixel population
density based on the described grid pattern, shown in the picture below.
As seen above, pixels in the centre of the grid are considered more densely populated, and these
areas are assigned a higher weight than those pixels in the corners of the grid, which are
considered less densely populated. The image shows how the population density is estimated
based on the weights assigned by the Gaussian filter.
These next steps show how to further pre-process the Built-up area layer in QGIS
13 https://communities.unescap.org/system/files/poptoguf_manual_final.pdf
These next steps show how to preprocess the flood inundation area data using QGIS
The next steps show how to clip the Lao People’s Democratic Republic Flood Inundation data
from the Global Flood Inundation data
8. Click on the “Processing Toolbox” and search for “Clip raster by mask layer”
9. In the window that appears, complete the Parameters as follows:
a. Input layer = “flood_hazard_200_yrp”
b. Mask layer = “lao_admbnda_adm1_ngd_20191112 [EPSG:4326]”
c. Source CRS [optional] = “Project CRS: EPSG:4326 – WGS84”
d. Assign a specified nodata value to output bands [optional] = “-99999.000000”
e. Select “Match the extent of the clipped raster to the extent of the mask layer”
These next steps show how to resample the Flood_Hazard_Laos raster layer. Resampling refers
to the process of preparing the raster files for display as maps, based on desired map scale and
screen capacity. QGIS determines how to render the raster in map form and still maintain the
character of the full-resolution raster. For this exercise, we resample the pixels in the
Flood_Hazard_Laos raster layer to the dimensions of the pixels in the Built up
Area_LAOS_Reclassifiey raster layer.
These next steps show how to layer the Lao People’s Democratic Republic Census Data into the
administrative boundaries in QGIS
15. On the Attribute Table Taskbar, unselect “Edit” by clicking the “Edit” button
16. In the Layer Panel, right-click on “lao_admbnda_adm1_ngd_20191112”
17. Click “Properties”
18. Click “Joins”
In this section, we will show how to generate the population raster by distributing the census data
to the estimated pixels:
The steps in this section generate the flood hazard calculation layer using the raster calculator
in QGIS
The steps in this sub-section show how to calculate the percentage of population that is
exposed to the flood hazard area:
10. On the “Layers Panel”, right-click and remove all other layers except
“lao_admbnda_adm1_ngd_20191112”
11. Click “Remove Layer”
These steps produce the visualization of the percentage of population exposed to hazard using
maps:
b. Value = “Affected%”
c. Legend format = “%1 - %2%”
d. Mode = “Natural Breaks (Jenks)”
e. All other parameters are kept as default
f. Classes = 5
g. Click “OK”
7. Define “Population Exposed to Flood Hazard” as the print layout title in the window that
appears
8. Click “OK”
10. For the map to appear on the white screen, click on the white screen and draw a rectangle
with the cursor. This image below reflects the selection as was made in this exercise, and
yours may appear differently:
14. On the toolbar at the left of the interface, click “Adds a new Scale Bar to the layout”
15. For the scale to appear on the map, click on the map and draw a rectangle with the
cursor, and the following image will appear:
Under “Appearance”, edit “Font” and other characteristics as outlined in the picture below:
We see the “Population exposed to Flood Hazard” as shown in the picture below:
22. On the taskbar at the top of the interface, click “Layout” button and Select on the “Export
as Image” as shown in the picture below:
You have completed Step 3 of 3! Well done on producing your visualizations! We wish
you good luck as you continue the journey ahead.
A. QGIS
QGIS is a cross-platform desktop geographic information system that supports viewing, editing
and analysis of geospatial data.
We used the most recent version of QGIS available, and we advise that you do the same.
However, while most versions of QGIS will work, we recommend that you use QGIS version 3.10
and above, with the following plugins:
As users gain more experience using QGIS, using updated versions will become easier.
The following toolbars are particularly useful, and they should be enabled:
• File - provides quick access to creating, opening, saving QGIS projects, and creating and
managing print composers.
• Manage Layers - contains tools to add vector, raster, database, web service, text layers,
or create new layers.
• Map Navigation - contains tools useful for panning, zooming, and refreshing the map
display.
• Attributes - provides access to information, selection, field calculator, measuring,
bookmarking, and annotation tools.
• The administrative boundaries data used are shape file data from HDX data source
(Humanitarian Data Exchange). Users can download administrative boundaries data from
various other sources as shown in the below table.
This table outlines sources for the administrative boundaries data that are most frequently used.
• The Census data used in this guide is obtained from HDX data source (Humanitarian Data
Exchange). Users can download population data from various other sources as shown in
the below table.
This table outlines sources for the most frequently used Global population data sets.
• The global land cover data used in this guide is obtained from Copernicus Global Land
Service. Users can download land cover data from various other sources as shown in the
below table.
1999-
Landsat 7 30M https://earthexplorer.usgs.gov/
Present
Landsat 8 2013 30M https://earthexplorer.usgs.gov/
Aster 1999 30M https://earthexplorer.usgs.gov/
https://earthexplorer.usgs.gov
Sentinel 2015 10M
250M,
2003- https://search.earthdata.nasa.gov
Modis 500M, 1Km,
Present
5Km
https://cds.climate.copernicus.eu/cdsapp
1992-
ESA 300 M #!/dataset/satellite-land-
Present
cover?tab=overview
Copernicus
2015 https://land.copernicus.eu/global/product
Global Land 100M
Present s/lc
Service
• The flood hazard data used in this guide is downloaded from Global Risk Data Platform.
Users can download flood hazard data from various other sources as shown in the table
below.