Supporting Disaster Risk Financing in Morocco Through EO-Based Flood and Exposure Mapping

Morocco is regularly affected by natural hazards, particularly floods and earthquakes, creating significant challenges for disaster response and financial risk management. For institutions such as the Fonds de Solidarité contre les Événements Catastrophiques (FSEC), estimating potential losses in advance is essential. In practice, however, this has long been constrained by the limited availability of consistent and spatially detailed information on both hazards and exposed assets.

Within ESA’s GDA Disaster Resilience activity, an EO-based effort was carried out to strengthen risk assessment capabilities in Morocco. The World Bank DRFI Program and FSEC helped define user requirements and operational workflows.

The activity combined two complementary components. The first focused on exposure, generating datasets that describe where built-up assets are located, how dense and tall they are, and what types of buildings are present. This included settlement extent, imperviousness, built-up height, and building type classification. The second focused on hazard, generating flood datasets from satellite observations to identify the spatial extent and recurrence of flooding events.

Together, these datasets improved national-scale hazard and exposure assessment. Their value became particularly evident after the 2023 Marrakesh–Safi earthquake, when the exposure information was used to rapidly assess the distribution and types of affected buildings, supporting early damage appraisal and response prioritisation. The case shows how EO can strengthen disaster risk financing and support preparedness, response, and recovery.

Morocco has established institutional mechanisms to manage the financial consequences of natural disasters, with FSEC playing a central role in compensation and recovery. For these mechanisms to work effectively, risk needs to be assessed in advance by combining information on where hazards occur with a realistic representation of the assets that may be affected.

In practice, this has often been difficult. Information on buildings and infrastructure is frequently fragmented, uneven in detail, or not regularly updated. In particular, datasets describing not only where built-up assets are located, but also how dense, how tall, and what type of buildings are present, are often missing or incomplete. Hazard information, especially for floods, can also be inconsistent across the country or not readily usable within financial risk models. As a result, estimating potential losses and designing financial protection mechanisms becomes more uncertain.

As part of ESA’s GDA Disaster Resilience activity, Morocco served as a practical case through which EO-based approaches could be tested against real disaster risk financing needs. The World Bank DRFI Program and FSEC were closely involved in framing the use case, articulating the priority information gaps, and providing feedback so that the resulting datasets could be aligned with operational workflows and decision-making requirements.

The work was therefore developed in close interaction with users, allowing the products to be refined iteratively based on feedback. This helped ensure that the final outputs were not only technically robust, but also practical to use within existing institutional workflows.

To address the identified gaps in exposure and hazard information, the activity produced a set of complementary EO-derived datasets designed to provide a spatially detailed and nationally consistent representation of risk components across Morocco.

On the exposure side, the work was built on DLR’s World Settlement Footprint (WSF) framework, which provides satellite-derived information on the spatial distribution of human settlements. For Morocco, this included settlement extent information together with dedicated layers describing imperviousness and built-up height. The imperviousness layer estimates sealed surfaces within built-up areas, supporting assessment of urban density and flood-related runoff. The built-up height layer captures the vertical dimension of settlements.

In addition to these layers, a dedicated Building Type Classification product was developed specifically for this Morocco activity. It distinguishes between classes such as villas, apartments, Moroccan houses, rural housing, informal settlements, and a broader non-residential category including industrial, commercial, military sites, and greenhouses. This added an important dimension to the exposure analysis, because different building types are associated with different vulnerabilities and economic values, which in turn influence potential loss estimates and recovery needs. Examples from Casablanca (Figures 1, 2, 3) show how the dedicated Building Type Classification moves beyond a simple built-up mask, distinguishing between settlement types..

On the hazard side, flood extent datasets were generated from satellite observations by combining radar and optical information to improve detection under different environmental conditions (Figures 4, 5). These products support flood extent and frequency analysis.

Together, the datasets provided integrated hazard and exposure information suitable for risk analysis and financial modelling.

The activity combined several EO and geospatial data sources to characterise exposure and flooding.

Sentinel-2 optical imagery played a central role in the exposure component. It supported the delineation of settlement patterns, the estimation of imperviousness, and building type classification. Sentinel-1 SAR data was mainly used for the hazard component, as it enables flood detection even under cloud cover and in rapidly changing conditions. It also contributed to the broader characterisation of built-up areas through settlement-related analyses.

To better represent the vertical dimension of exposure, global elevation datasets were integrated, notably the Copernicus DEM and ALOS World 3D. These datasets were used both to estimate built-up height and to support flood analysis. In addition, terrain information derived from FABDEM was used to support the building type classification.

The activity also incorporated VIIRS nighttime light observations, which provided additional information on the functional character of built-up areas. These data were particularly useful in helping distinguish between residential and non-residential zones.

Complementary geospatial datasets also played an important role. OpenStreetMap and Facebook road data provided contextual information supporting building type classification and neighbourhood delineation.

A combination of EO and auxiliary geospatial data enabled richer representation of exposure and flood-prone context for risk analysis.

The activity strengthened the ability of Moroccan institutions to assess disaster risk using more consistent and spatially detailed information on both hazard and exposure. A key benefit lies in the richer characterisation of exposed assets. Exposure analyses can now consider settlement density, height, and building type. This makes it possible to represent exposure more realistically within financial risk models and to support more robust estimation of potential losses.

The dedicated Building Type Classification product is particularly valuable in this respect. Different building types imply different vulnerabilities and economic values. Introducing this level of differentiation helps move from generic exposure estimates towards assessments that are more meaningful for financial planning, emergency appraisal, and post-disaster recovery analysis.

The practical value of the datasets became especially evident after the 2023 Marrakesh–Safi earthquake. In that context, the exposure layers were used to rapidly assess the likely distribution and typology of affected buildings, providing an early indication of the types and distribution of impacted assets. This supported initial damage appraisal and helped prioritise response efforts at a time when rapid information was essential.

Beyond individual events, the integration of hazard and exposure information supports a more structured approach to risk analysis. Stakeholders can identify where built-up assets are concentrated in hazard-prone zones and better understand the characteristics of those exposed assets, helping to inform preparedness planning, prioritisation, and longer-term risk reduction strategies.

At the same time, the activity highlighted important limitations. In particular, the rapid and short-lived nature of many flood events in Morocco makes them difficult to capture using satellite observations alone. Some exposure products, especially the Building Type Classification, also depend on the availability and representativeness of reference and ancillary data, which may affect generalisation in some parts of the country. These limitations highlight areas for further refinement.

Overall, the activity shows how EO-based information can strengthen disaster risk financing by providing consistent, scalable, and operational inputs for decision-making. Beneficiaries include disaster risk management and financial planning institutions, as well as affected communities. This is particularly important in contexts where faster and better targeted early assessments can influence how recovery support is prioritised after a disaster.