High Resolution Digital Surface Models Using Drones for Coastal Communities in São Tomé and Príncipe

São Tome aerial image

A collage of drone aerial imagery of a coastal community in São Tome.


São Tomé and Príncipe is highly vulnerable to climate variability and change, coastal and flash flooding, and sea level rise. To help reduce this vulnerability, the Government in increasing its adaptive capacity to improve participatory risk mapping, adaptation activities, and investment design for vulnerable coastal communities.

Risk reduction activities have been piloted in 4 coastal communities, promoting a mix of interventions from grey infrastructures (including a sea wall in Ribeira Afonso and drainage systems in Malanza, Santa Catarina, and Praia das Burras), green infrastructure (including vegetation planting across the 4 communities) and voluntary set back. The project is executed by the Directorate General of Environment, and its components are managed by CONPREC (National Council for Preparation and Responses to Disasters), the National Institute of Meteorology, Capitanery, MARAPA (an NGO for fishermen), and the Directorate General of Fisheries.

A second phase intended to scale-up these pilot activities to an additional 7 communities and to strengthen the institutional capacity building process is under preparation.

GFDRR sponsored activities:

To assist the Government in this project, GFDRR’s team provided the following support:

Coastline exposure projections: the team worked to update and expand the historical and projected coastline trend maps developed in 2013 to high priority coastal settlements, and expand the existing deforestation map to cover the whole of São Tomé and Príncipe. Community buildings were also geo-referenced to assess the change in settlements areas, and specialized training was provided to national experts on hazard and vulnerability mapping for disaster and climate risk assessment.

High resolution Digital Surface Model (DSM): a DSM of the most vulnerable communities was created using drone technology in order to improve hazard assessment models of coastal and river flooding to better inform the design of future investments in coastal protection. This effort also provided reference information to monitor the evolutions of the shoreline due the coastal protection works and environmental changes, and also of the built environment and vegetation.

Drone Activities:

To prepare the activities of the second component, a team from the GFDRR Innovation Lab provided support to define the best strategy to obtain high resolution digital surface models for remote and relatively small-size communities (around 1 km2). The relatively high-density vegetation covering the communities was another constraint which could have prevented the production of high resolution data. However, it was decided to use the photogrammetry using fix wing drones (similar to work in Tanzania) with the support of Drones Adventures.

Another constraint was climate. São Tomé, a small island state located on the equator, has some alternation of rainy and dry seasons. The south of the island receives up to 8m of precipitation per year, nearly all year long. The optimal time window for the imaging was between January and February, with the lower rainy season (gravanita) associated also with some foggy situations beginning end of February, which could reduce the capacity to fly.

Then, Alexis Roze and Alexandre Habersaat from Drones Adventures joined the team in São Tomé from February 26th to March 4th, 2017, to map the 10 most vulnerable coastal communities in the countries, along with the national capital (São Tomé) and the capital of the autonomous region of Príncipe. In each of these communities, and based on the results of the risk assessment performed using the DSM, some adaptation options would be implemented.

Figure 1: Map of projects and images from deployment of drones.

Figure 1: Map of projects and images from deployment of drones.

A team composed by representatives of the Directorate General of Environment, the cabinet of the prime minister, CONPREC, the Observatory of Environment, and the Directorate of Fisheries took part in the drones’ activities in all the communities in São Tomé Island in order to gain experience in planning flight missions, identifying take-off and landing sites, as well as throwing drones. The communities were also involved, with explanationatory sessions and showing of images obtained.

Two drones “Ebee” and one “Ebee plus” have been used during the week, with some simultaneous flights to increase efficiency. The images were processed during the night following their acquisition using Pix4D software to produce high resolution orthomosaic and digital surface models. The average ground sampling distance was around 5cm, to obtain a digital surface model with a vertical resolution between 1 and 3 times this value, suitable for the flood hazard assessment models which would be created before the end of 2017.

Figure 2: Flight plan over Praia das Burras (each bubble corresponds to an image)

Figure 2: Flight plan over Praia das Burras (each bubble corresponds to an image)


In less than a week, all targeted communities were covered. A DSM has been built for each community, and the vegetation was not a major issue (some post-processing might still need to be performed to remove the trees in order to properly model the flooding events). Despite the mission taking place quite late in the season, the weather was favorable, and the few drops received in Praia Abade didn’t prevent the drones from flying. The solution of photogrammetry has proved to be an adequate solution for a small island state like São Tomé, where the remoteness of the country and the relatively small sizes of the targeted areas, compared with the needed resolutions, didn’t justify to use airborne lidar (carried by plane).

In view of the positive results from a technical point of view, but also from the feedbacks of the end-users, similar mapping activities are planned during the second phase of the project to monitor the evolution of the shorelines. The drones will also be used after future disasters to assess the damages.

To ensure re-usability of the data, and that maximum value is derived from it, the team made sure to apply Open By Default principles, and the imagery mosaic are uploaded to OpenAerialMap – a global platform to make remote sensing imagery available as open data (here for download) or to bring directly into GIS software using an image service.

Figure 3: Digital surface model in Praia das Burras

Figure 3: Digital surface model in Praia das Burras.