Recommended Practice: Flood Mapping Practice Using Sentinel-1 and Sentinel-2 Imagery

Recommended Practice: Flood Mapping Practice Using Sentinel-1 and Sentinel-2 Imagery

UN-SPIDER has published a new Recommended Practice that leverages both Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 optical imagery to improve flood detection and mapping. Developed by the Regional Centre for Mapping of Resources for Development (RCMRD), this method offers a multispectral and radar-based approach to identifying flood-affected areas with greater precision, especially in regions with persistent cloud cover or challenging terrain.

Strengthening Multi-Hazard Risk Assessment in Kyrgyzstan: UN-SPIDER Leads Hands-On Training in Bishkek

On 15–16 May 2025, the Ministry of Emergency Situations (MES) of Kyrgyzstan hosted a two-day training workshop titled “Multi-hazard Risk Assessment for Risk-Reduction Planning.” Organised by UNOOSA/UN-SPIDER in partnership with ESCAP-APCICT, the Geoinformatics Center (Asian Institute of Technology) and ITC–University of Twente, the course addressed the country’s overlapping threats from earthquakes, landslides, floods and climate-related extremes.

Aims and audience

In Detail: Use of Elevation Data for Storm Surge Coastal Flood Modelling

Worldwide, storm surges pose a threat to coastal communities. Particularly in low-lying African coastal cities, the impact of climate change induced storms and sea level rise is a threat to the vulnerable community  (Nhantumbo et al. 2023).  Especially with climate change, this hazard is likely to intensify (IPCC 2022), calling for urgent action. To identify the exposed groups, assets, and infrastructure, and to implement effective storm surge risk management, an understanding of the hazard, its timing, and geographic patterns is crucial. Hazard maps can serve as a first approximation for estimating the spatial extent of storm surges and sea level rise, as well as the exposed population and assets.

For detailed hazard maps, hydrodynamic models are often used due to the complexity of storm surges. They are influenced by factors such as storm intensity, the storm's track, bathymetry, and more. Hydrodynamic modeling therefore requires high-quality and complex parameters, such as historical data sets or detailed coastal formation data for a specific area, which are not always available or may be costly to obtain. This makes free, detailed modeling of storm surges very challenging.

Other methods, such as the bathtub approach, require fewer datasets and can serve as a first approximation for modeling the potential geographical extent of coastal flooding. This Recommended Practice outlines the steps to visualize the geographical extent of coastal flooding or sea level rise using just two datasets: a digital elevation model (DEM) and a shoreline dataset. The choice of a DEM is crucial; preferably, a high-resolution DEM should be used, as it is more likely to represent the surface accurately. However, high-resolution datasets are not always available for free. In such cases, a mid-resolution DEM, such as the Copernicus 30m DEM, which is globally accessible, should be used when open-access DEMs are the only option.

The flood masks created during this practice can be further used to identify exposed elements in specific regions. The Recommended Practice on Exposure Mapping can be applied for that (Link: https://www.un-spider.org/advisory-support/recommended-practices/recommended-practice-exposure-mapping ). 

Computer Requirements

To perform the step-by-step procedure, it is required to install QGIS (preferred 3.28.15 Firenze or newer version).

Required Datasets

  1. Digital Elevation Model (e.g. Copernicus 30m DEM, WorldDEM neo)
  2. Coastline vector layer (e.g. OpenStreetMap water-body layer, WorldDEM Ocean shoreline a product of Airbus)

You can freely download the Copernicus 30m DEM from the OpenTopography, CODE-DE or the Copernicus Data Space Ecosystem website as described on the step-by-step page.
 

  • Develop scenarios of impacts of storm surges in specific coastal areas;
  • Conduct an initial assessment of risk associated with storm surges;
  • Improve storm surge early warning systems using scenarios of the extent of coastal flooding.

Strengths:

  • Can be used in all coastal areas of the world, the COP30 DEM and other commercial DEMs are available globally
  • Free and open-source software can be used (e.g. Quantum GIS)

COP30 DEM:

  • Resolution:                                       1 arcsec (approx. 30m)
  • Absolute vertical accuracy:          < 4m (90% linear error)
  • Relative vertical accuracy:            < 2m (slope <= 20%) / < 4m (slope <= 40%)

Limitations and Restrictions:

The globally accessible Copernicus DEM has a resolution of 30 m. When using this dataset, finer details may be omitted due to its resolution.

To address this limitation, alternative datasets with higher resolution can be considered. One such dataset is the WorldDEMneo from Airbus. However, for some regions there are restrictions regarding geographic availability. Consult with Airbus to determine whether data can be supplied for your region of interest.

DLR Document: TD-GS-PS-0021 (2016). TanDEM-X Ground Segment - DEM Products Specification Document, Issue 3.1, 05.08.2016. Available at https://tandemx-science.dlr.de/pdfs/TD-GS-PS-0021_DEM-Product-Specification_v3.1.pdf

IPCC (2021): Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. V. Masson-Delmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger et al.United Kingdom and New York, NY, USA.

Nhantumbo, B., Dada, O. A. and f. E.K. Ghomsi (2023): Sea Level rise and Climate Change – Impacts on Africal Coastal Systems and Cities. DOI: 10.5772/intechopen.113083

World Meteorological Organization (2011). Guide to Storm Surge Forecasting. Publication No. 1076. Geneva, Switzerland, 2011. Available at: https://library.wmo.int/pmb_ged/wmo_1076_en.pdf

Step-by-Step: Use of Digital Elevation Data for Storm Surge Coastal Flood Modelling

The following steps were carried out, as described in more details below.

This practice offers two approaches for implementation:

  1. A Step-by-step procedure (Section 4a)
  2. A QGIS-model (Section 4b)

The first approach provides a detailed guide through each step, helping you gain comprehensive QGIS knowledge. In contrast, the second approach allows for faster processing by utilizing an integrated QGIS model that automates the entire project.

Feel free to explore both options. If you are only interested in the second approach, proceed directly to Section 4b.

This Recommended Practice requires the use of an open-accessible digital elevation data set, preferably a digital surface model and a shoreline. Step 1 describes how to access the COP30 DEM and a shoreline. The procedure makes use of the free and open QGIS desktop software.

1. Download datasets

For the Recommended Practice the following datasets are needed and can be downloaded from:

Digital Elevation Model (DEM)

Use a DEM with the highest resolution available. Some countries offer openly accessible high-resolution datasets; however, this is not the case globally. Partners such as Airbus can provide commercial, high-resolution data, such as the WorldDEMneo. This dataset offers global coverage and supports modeling at a global scale. For inquiries about WorldDEMneo, please contact Airbus (see Airbus contact).

An openly available and practical alternative is the Copernicus 30 m Digital Surface Model (DSM), which is derived from Airbus’s WorldDEM data. This DEM is used in this practice to demonstrate the feasibility of modeling storm surges using openly accessible elevation data.

For more details on the COP30 DEM see: https://dataspace.copernicus.eu/explore-data/data-collections/copernicus-contributing-missions/collections-description/COP-DEM

Note: There are two types of Digital Elevation Models:

  • Digital Surface Model (DSM): represents the bare-earth surface with all natural and built features (houses, trees, cars,...). This one is preferred for storm surge modelling because it counts in potential barriers against storm surges.
  • Digital Terrain Model (DTM): represents only the bare-earth surface without natural and built features

 

Fig.1: Difference between DSM and DTM. MartinOver, CC BY-SA 3.0, via Wikimedia Commons

 

Please note that if a local, more accurate DSM (e.g., WorldDEM Neo) is available, take advantage of this resource.  

Shoreline dataset

The shoreline dataset should preferably be with global coverage. If a local, more accurate shoreline vector file is available, take advantage of this resource.

DatasetDownload
Copernicus DEM

Option 1: OpenTopography (quick access, no registration required, see full description below)

Option 2: CODE-DE (requires registration)

Option 3: Copernicus Data Space Ecosystem (requires registration and approval for CCM DEM/COP DEM data, approval might take a few days)

Shoreline from OpenStreetMap

https://osmdata.openstreetmap.de/data/water-polygons.html

Download > Format: Shapefile, Projection: WGS84 (large polygons are split) > Download button (see detailed description below)

Download DEM data from OpenTopography 

Go to the website: https://opentopography.org/ 

Click on Data > Data Catalog

Select Global & Regional DEM and then click on Copernicus Global Digital Elevation Models.

Fig.2: OpenTopography website
Fig.3: Download from OpenTopography

Select the Copernicus 30m DEM for the highest and most feasible resolution available as openly accessible data. The following Figure 4 provides a detailed overview of the available Copernicus DEMs.

Figure 4: Download from OpenTopography

Scroll down to ‘1. Select area of data to process’ and select a region of interest. For our example we use the AOI around Accra, Ghana. Click “select a region” and draw a rectangle.

Fig.5: Download from OpenTopography, Selecting a region of interest

 Scroll down and keep the default settings. Under 'Job Description,' define a title for the area and add a description to identify the dataset's content. Enter your email address and submit (Fig. 6). 

Fig.6: Download from OpenTopography

Check your email for a link to the download section. This process may take a few minutes depending on the dataset's size.

Download the dataset.

(Alternative) Download DEM data from CODE-DE

Alternatively, you can download the DEM data from the CODE-DE.

Go to website: https://explore.code-de.org

Login or register a new account. Download the data, by selecting the COP DEM and time frame, draw a polygon in the area of interest, click on search and download the desired tiles (Fig. 7).

Fig.7: Settings for CODE-DE

(Alternative) Download DEM data from Copernicus Data Space Ecosystem (CDSE)

Alternatively, you can download the DEM data from the CDSE.

Go to website: https://browser.dataspace.copernicus.eu

Login or register a new account. Then request access to CCM data. After a few business days, you should be able to download the data, by selecting the COP DEM and the time frame, draw a polygon in the area of interest, click on search and download the desired tiles (Fig. 8).

Fig.8: Settings for CDSE

Download "shoreline" data from OpenStreetMap

Go to the following website: https://osmdata.openstreetmap.de/data/water-polygons.html

On this website you find global water polygons that are used for creating our shoreline.

Download the water polygons shapefile dataset with the WGS84 projection (Fig. 9).

Fig.9: Download from Osmdata

 Save the DEM and the shoreline dataset in one folder and open QGIS. 

2. Load data in QGIS

Load the DEM raster file and shoreline vector file in QGIS (Fig. 10).

Fig.10: Loading data in QGIS

In the background, the loaded Digital Elevation Dataset (greyscale image) is visible (layer name: output_hh); here it is the Copernicus DEM with a 30 m resolution (Fig. 10). The shoreline vector file (layer name: water_polygons) is displayed in transparent blue. When sourced from OpenStreetMap (OSM), this file provides global coverage and appears in separate tiles, encompassing the entire world in the QGIS project. In the next step, the shoreline dataset will be preprocessed to enable further analysis for the chosen region of interest.

3. Preprocessing

3.1 Shoreline

Search for the 'Clip Vector by Extent' function in the QGIS Processing Toolbox. Use the worldwide water polygon layer from OSM as the input layer and select the raster layer as the clipping extent (Fig. 11). 

Fig. 11: Preprocessing Shoreline

Under 'Clipped (extent),' select a folder and click “save to file” and assign it a name (e.g., Ghana_coastline). You can save it as shapefile (.shp) or as Geopackage (.gpkg). Here it was saved as shapefile. Then click 'Run' (Fig. 12).

Fig. 12: Preprocessing Shoreline

A new layer has been created in your QGIS project, displaying only the extent of the selected raster DEM. To maintain a clearer overview, remove the 'water_polygon' layer by right-clicking on the layer and selecting 'Remove Layer'.

The clipped vector coastline layer is still divided into several tiles. To merge them into a single connected tile, search for 'Dissolve' in the Processing Toolbox. Fill in the parameters as shown in the screenshot (Fig. 13). Remember to save the layer in the respective folder, naming it for instance 'Ghana_coastline_final'. After selecting the folder, click 'Run' to create the final coastline.

Fig. 13: Preprocessing Shoreline

You can remove the 'Ghana_coastline' layer to keep only the necessary layers in your QGIS project. The layer will now appear as shown in Figure 14.

Fig. 14: Preprocessing Shoreline

3.2 DEM - Vertical and horizontal reference system adjustments

Note: If you aim to compare different digital elevation models, it is important to adjust both, the vertical and horizontal reference systems. The horizontal Coordinate Reference System can be adjusted using the 'Warp (Reproject)' tool, which can be found in the Processing Toolbox. For the vertical adjustment, a reference model is required.

The COP30 DEM has the vertical reference system EGM08. Since this is suitable for the planned storm surge modelling, no further vertical adjustment will be made to the DEM.

4. a) Processing with step-by-step procedure

4.1 Load data in QGIS

If you followed the preprocessing steps you already have all the necessary data in your QGIS project (the Copernicus 30 m DEM and the preprocessed shoreline layer, see Fig. 14). If you have your own datasets, please load a DEM and shoreline layer in your QGIS project.

4.2 Raster calculation

In this step, all elevation pixels below a certain elevation threshold will be detected. Open the Raster Calculator by selecting Raster > Raster Calculator (Fig. 15).

First, select your DEM layer and click on the ‘less or equal’ (<=) math operator to create a Raster Calculator Expression. Then set the threshold for the flood level you would like to simulate. In this sample case, a value of 5m has been set. In the equation you can also adjust the m to any height you like. Save your layer under ‘Output layer’ (here named: COP30_Ghana_5m) and set the ‘output format’ to GeoTIFF (Fig. 15).

Fig. 15: Raster calculation

The result is a binary raster layer that classifies all potentially flooded areas with a value of 1 and non-flooded areas with a value of 0 (Fig. 16). The results of step 4.2 are depicted in Fig. 16, where all pixels at or below 5m are represented in white, and pixels above 5m are represented in black with a value of 0.

Fig. 16: Raster calculation. Binary layer. White stands for flooded area and black stands for non-flooded area.

4.3 Define NoData value

In this step, all non-flooded areas in the classification mask will be set to NoData using the ‘Translate (Convert Format)’ tool. The tool is accessed via Raster > Conversion > Translate (Convert Format). Set the ‘Input layer’ to the 5m raster layer, produced in section 4.2 and save the new layer under ‘Converted’ (here named: COP30_Ghana_5m_0_set_nodata). The NoData value has to be assigned to 0 (Fig. 17). Click Run.

 

Fig. 17: Defining a non-data value

While the black area previously had a value of 0, it no longer has any value. The white area still retains the value of 1, indicating the flooded area and other water bodies.

4.4 Polygonize

In this step the flooded area (area with the value 1) is going to be converted to a vector file using the Polygonize tool. Open the tool through Raster > Conversion > Polygonize (Raster to Vector) and set the input file (here: COP30_Ghana_5m_0_set_nodata). Save the polygonized layer under ‘Vectorized’ (here: COP30_Ghana_5m_polygonized).  Tick the ‘Use 8-connectedness’ box (Fig. 18). Click ‘Run’.

Fig. 18: Polygonize

4.5 Filter inland water patches

In this step, inland water patches that meet the elevation threshold (<= 5m) but are not connected to the ocean are filtered out. The connectivity between the ocean shoreline and the inland flooded areas is used to reduce the overestimation of flooded areas.

To identify only the flooded areas connected to the shoreline, the polygonized potential flood extent must intersect with the shoreline vector layer. This can be done using the 'Select by Location' tool, which is located under Vector > Research Tools > Select by Location.

Select the polygonized layer from the previous step as the 'Select features from' option. Tick the 'Intersect' box under the geometric predicate. For 'By comparing to the features from,' select the preprocessed coastline layer (from Step 3, here named 'Ghana_coastline_final'). Click 'Run' to proceed (Fig. 19).

Fig. 19: Filter inland water patches

Note that you have now only created a selection of the polygonized layer (here: COP30_Ghana_5m_polygonized). The selected areas are highlighted in bright yellow in your QGIS project. Figure 20 shows an example, where the water patches connected to the shoreline are displayed in yellow, and the water patches that have not been selected (due to no direct connection to the shoreline) are shown in blue.

Fig. 20: Filtered inland water patches. The area with connectivity between the shoreline and modelled flooded area is depicted in yellow.

For saving the selected features in a separate layer, right-click on the polygonized layer then select Export > Save Selected Feature As (Fig. 21). 

Fig. 21: Saving the filtered inland water patches.

Change the format to ESRI shapefile and select a folder to save your 5m flood extent layer (here: Waterlevel_Ghana_5m): Make sure that the box ‘Save only selected features’ is selected. Click OK to save the new layer (Fig. 22).

Fig. 22: Saving the filtered inland water patches.

You now successfully created the flood extent layer for a 5m flood height event. You can now proceed to step 5 and create a map with your layers. If you wish to create flood extent layers for different heights, you can repeat the steps 1 to 4 and adjust the water level height in step 4.2.

 4. b) Processing with QGIS Model

 Another way to fulfill the steps (Processing 4.1 to 4.5) is by using the QGIS model provided at the following link (https://github.com/josibregulla/SPEAR---Storm-Surge-Modelling). In this GitHub project, you will find models that simulate surge heights ranging from 1 to 6 meters (Fig. 23). The 1-meter model includes the creation of a hillshade, which can serve as a visual enhancement for the final map. Download the models of your choice by clicking on the model and then on the download button on the top right (Fig. 24). 

Fig. 23: GitHub data for storm surge modelling

 

Fig. 24: Download model from GitHub.

We begin again with step 4.1, 'Load data in QGIS', ensuring that both the DEM and coastline are loaded. Next, load the QGIS model (which you downloaded in the previous step) by clicking on Processing > Model Designer. The Model Designer will open. To load the downloaded model from GitHub, click Model > Open Model. Navigate to the model file (here: ModelDEM_waterlevel_5m), which you downloaded from GitHub, and open the model. Your screen should then display the following (Fig. 25):

Fig. 25: QGIS storm surge model for 5m water level

On the right side, you will see the QGIS model, which will automatically carry out steps 4.1 to 4.5. This approach offers the advantage of faster processing, particularly when assessing multiple areas worldwide and different storm surge heights.

Click the green 'Play' button. Select the DEM, the preprocessed coastline, and save the layer. (Note: If you don't select a folder to save the layer in this step, a temporary layer will be created, automatically named 'water_level_5m') (Fig. 26).
 

Fig. 26: Processing with the QGIS model

Click ‘Run’. After a few seconds, you should get the following result (colors may differ; Fig. 27).

Fig. 27: Results of the QGIS model

You have now created a flood height layer (here displayed in orange; adjust to the color of your choice in the next step) and can proceed to step 5 to create a thematic map.

5. Create a map

The processed layers, including the final layer from Step 4a) or 4b), the coastline layer, and the DEM, can now be used to create a map showing the extent of the flood. You may add additional base maps or create hillshades to enhance the visualization. An example map is provided in Figure 28.

Fig. 28: Example map of two areas in Ghana for storm surge modelling in coastal areas.
https://un-spider.org/sites/default/files/AIRBUS_Logo.png
https://un-spider.org/sites/default/files/UniBonn_ZFL_Logo_Resc.png

June Webinar: Tackling Barriers to the Use of Earth Data in Flood and Drought Response

This is event is available for participation on an ongoing basis

0
0
No certificate issued or not known
0.00
USD
06/18/2025, 12:00am
Visible
1

Wuhan University (WHU) in China

China
0
0

Wuhan University (WHU) stands as a global leader in remote sensing science and technology. Since 2017, its remote sensing discipline has consistently ranked No. 1 in the world according to the Shanghai Ranking (ARWU), a testament to its academic excellence and research prowess. The university hosts a multidisciplinary team of over 100 experts specializing in forest, agricultural, urban, and geological remote sensing, driving innovation across diverse sectors. Since 2015, WHU has successfully designed and deployed six remote sensing satellites, encompassing both optical and radar imaging technologies. This achievement underscores its commitment to advancing space-based observation capabilities. In 2024, Prof. Deren Li—a distinguished authority in the field—was awarded China’s highest scientific honor, the National Top Science and Technology Prize, for his pivotal contributions to the development of high-resolution satellite systems.

On 22 April 2025, WHU further solidified its global leadership by signing a Memorandum of Understanding (MoU) with the United Nations Office for Outer Space Affairs (UNOOSA) to establish the China Regional Support Office (RSO) of the UN Platform for Space-based Information for Disaster Management and Emergency Response (UN-SPIDER). Leveraging WHU’s cutting-edge satellite constellation and remote sensing expertise, the RSO will actively support the United Nations Sustainable Development Goals (SDGs), the Sendai Framework for Disaster Risk Reduction, and the Global Development Initiative (GDI). Through this partnership, WHU aims to enhance international collaboration in space-based solutions for disaster resilience, environmental monitoring, and sustainable development worldwide.

Fields:

  • Disaster rapid mapping
  • Night-time light imagery analysis
  • High resolution imagery interpretation
  • SAR imagery analysis
  • Remote sensing of economy

Experts: 

  • Prof. Deren Li, LIESMARS, Wuhan University
  • Prof. Xi Li, LIESMARS, Wuhan University
  • Prof. Mian Yang, Economics and Management School, Wuhan University
  • Prof. Ailong Ma, LIESMARS, Wuhan University

Currently, the China Regional Support Office (China RSO) can provide remote sensing imagery from the Luojia-4 Satellite and Qimingxing Satellite. The Luojia-4 Satellite offers 20-meter resolution hyperspectral imagery and night-time light data, while the Qimingxing Satellite provides 20-meter resolution hyperspectral imagery. These data can support detailed observations in fields such as environmental monitoring, disaster risk assessment, and urban sustainable development planning.

The China Regional Support Office (RSO) possesses a robust portfolio of training resources in remote sensing and its applications for disaster risk reduction. Since 2011, the International GeoInformatics Summer School (IGSS) has been hosted by the State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS) at Wuhan University. Held annually each July, IGSS delivers a one-week intensive program comprising theoretical and hands-on courses in Geoinformatics, attracting over 100 participants worldwide each year.

In addition to IGSS, the RSO can leverage other institutional resources, such as the foreign aid training programs organized by Wuhan University on behalf of China’s Ministry of Commerce. These programs have a proven track record in delivering high-quality capacity-building initiatives. Building on Wuhan University’s extensive training expertise, the RSO is well-positioned to design tailored training curricula specifically aimed at cultivating remote sensing professionals in disaster reduction. These courses will integrate cutting-edge technologies and global best practices to address the urgent needs of disaster management in the region and beyond.

Dr. Xi Li
Professor, State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing , Wuhan University

Deputy Director, Office of International Affairs, Wuhan University
Research team Webpage: http://nightlight.whu.edu.cn/en 

Head, China Regional Support Office, UN-SPIDER

Lead of GEO Initiative - GEO Night Light, Group on Earth Observations

E-mail: li_rs [at] 163.com (li_rs[at]163[dot]com), lixi [at] whu.edu.cn (lixi[at]whu[dot]edu[dot]cn)  

Address:

Room 403A, Xinghu Building, Wuhan University, 129 Luoyu Road, Wuhan, China