In 2021, the Emergency Event Database recorded 432 disastrous events related to natural hazards worldwide, with flooding significantly dominating these events with 223 occurrences (EMDAT, 2021). These events affected more than 100 million people and accounted for an economic loss of 74 billion USD. Flooding accounts for approximately 40% of all natural disasters, and as many as 1.47 billion people are directly exposed to flood risks. Flood representation and forecast is essential for the protection of civil and infrastructural assets, either in real-time mode for forecast and crisis management or in hindcast mode for damage assessment, adaptation scenario and flood defense design studies (Mason et al. 2014).
Early warning and emergency management systems such as Copernicus Emergency Management Service, rely on the data driven combination of a dense and reliable Earth Observation (EO) network with hydrodynamics numerical models thus empowered with improved forecasting capabilities. This data assimilation approach inherits from Numerical Weather Prediction and makes the most of the growing volume of EO for continental water surfaces. The observation of inland waters benefits from several altimetry missions that provide high resolution along-track and soon wide swath water surface elevation data. SAR imagery data have also become one of the most efficient ways to map and monitor flood extents in near-real time over large areas, due to their all-weather day-and-night imaging capabilities (Peña-Luque et al. 2021). The Space Climate Observatory coordinates space agencies and international organizations to assess and monitor the consequences of climate change, among which flooding, from remote sensing (RS) observations (Kettig et al., 2021).Data driven hydrodynamic models provide a complete (in space and time) description of a flood event over the past and eventually in forecast mode, acting as a physics-informed interpolator between heterogeneous and complementary observations (Ngyuen et al. 2021). The simulated flood maps offer the opportunity to assess (in hindcast or in real time mode) the socio-economic impact of an event when exploited in concordance with information on industrial, agricultural or urban assets. The financial estimation of flooding impact is of great importance for public institutions and private companies. Indeed, it allows to identify mitigation and adaptation scenarios.  It also drives the strategy for insurance policy and investment fund.

Objectives 

This PhD project focuses on the valorization of flood extent data derived from RS to assess flood risk and estimate the associated socio-economic impact. It relies on the combination of three fields of expertise: Earth Observation, data driven flood modeling and risk assessment. These are respectively the field of expertise of the three partners for this PhD subject (CNES, CERFACS, Université Toulouse Jean Jaurès). CNES will provide a rich environment to work with EO data. With CERFACS, the doctorant will be involved in the development of the system dedicated to the assimilation of RS-derived flood extent data into the 2D hydrodynamic model Telemac. This will lead to the production of water level and flood extent maps that will then be mapped onto a data base of assets in collaboration with QuantCube, generated using IA algorithms for object detection as well as news monitoring technics, thus allowing to estimation the economical of the flood. This will lead to the production of water level and flood extent maps that will then be mapped with QuantCube onto a data base of assets, generated using IA algorithms for object detection as well as news monitoring technics, thus allowing to estimation the economical of the flood. A proof-of-concept will be implemented over a selected number of areas of interest where flooding frequently occurs and where financial stakes are high for industry, agriculture or territory planning actors.

Consortium 

Showing the merits of RS-derived data for flooding is at the core of several projects between CERFACS and CNES. The PhD will be co-advised by teams at CERFACS (Sophie Ricci, Thanh-Huy Nguyen, Andrea Piacentini), CNES (Raquel Rodriguez-Suquet, Jean-Marc Delvit, Santiago Peña Luque, Simon Baillarin) that already collaborate in the context of the SCO-FloodDAM-DT project. Additional inputs on socio-economical aspects that relate to risk assessment and financial stakes will be brought by a collaboration with QuantCube (Alice Froidevaux and Antoine Guiot) and UTJJ (Sébastien Le Corre). Discussions with SWOT-related temas will occur in the context of SWOT-aval and TOSCA-DAHM projects that include CERFACS, INRAE and CS-group (Nicolas Picot, Damien Desroches, Hind Oubanas, Pierre-Olivier Malaterre, Igor Jegadze, Charlotte Emery). On an international level, collaborations with the US partners of the IDEAS project at NASA are planned on the Digital Twin for Hydrology project.