Search

Are your properties located far enough from rivers, sea shorelines or water bodies? If the answer is yes, this does not mean that they are fully safe from flooding.In an era governed by continuous climate instability and unstoppable expansion of cities, the exacerbation of hydrometeorological events is increasing the occurrence of pluvial floods. Pluvial flooding is induced by the combination of two factors: extreme precipitations and the incapability of the ground/drainage systems to effectively handle excessive rainwater.In an urban environment, the runoff generated by localized and intense rainstorms may quickly inundate streets and buildings undermining the safety of people and assets. The characteristic of being hardly predictable has inspired the definition of pluvial flood as an ‘invisible hazard’ and the related damages and losses are increasingly weighing on the budget of municipalities and private citizens.Looking at the upsetting climate projections, experts are resolute in developing comprehensive methodologies and strategies for flood risk assessment and management.In this work, we present the attempt of accomplishing a high-resolution pluvial flood risk assessment at the city scale. The city of Differdange (Luxembourg's third largest city) is used as case study in which the extreme rainfall-related impacts and hazards are analyzed through the implementation of a fully coupled 1D/2D dual drainage model. This type of hydrodynamic model closely mimics the complexity of an urban landscape allowing to simulate all hydraulic phenomena occurring both on the surface and through the sewer network. Despite the digital accuracy of these models, they are rarely implemented due to the vast amount of detailed information required; which are often unavailable.The implementation of the hydraulic model follows two main steps: the bi-dimensional discretization of the surface and the 1D modelling of the whole drainage network.Nowadays, many countries provide open-access high-resolution digital elevation models of their territories (50 cm for Luxembourg) and up-to-date cadastral planimetries from which essential information for the 2D component are extrapolated. Ground data is enriched by land use/cover and soil maps for the estimation of roughness and infiltration parameters.The drainage network contemplates all pipes carrying rainwater, meaning the newer storm-water system and the old combined sewer network. The geometric specifications required are size, shape, elevation, material of pipes, manholes and tanks. Important infrastructures, such as flooding barriers, have been systematically added to the model. The fully-distributed hydrological engine allows operating the rainfall-runoff transformation on each cell of the domain and the exchange of water between the surface and drainage network occurs through the nodes of the network (storm drains and manholes).The model’s outcomes allow for assessing the level of hazard to which each building is exposed, identifying the critical nodes within the drainage network, and proposing mitigation strategies.Furthermore, these insights may help authorities to improve their warning systems and emergency plans.

Traditionally, flood risk maps used by city officials and water resource managers for urban planning, by engineers for adequate flood defence infrastructure design, or by insurers and re-insurers for estimating financial risk exposure are the result of modelling flood hazard of rivers and their associated floodplain lands at different return periods. Often, any of these stakeholders would use the 1:100 return period of fluvial hazard to plan accordingly. However, with the climate crises signals clearly present during recent flood disasters, and especially with the 2021 Europe floods, water managers, cities and the financial risk sector are now starting to plan differently and are recognizing the need not only for better and more frequently updated flood risk analysis, particularly in urban areas, but also need to consider pluvial and flash floods that can happen in any part of a river basin and oftentimes take place in headwater areas or off the main river floodplains. Flash flooding greatly impacts urban areas where the storm drainage infrastructure is becoming largely insufficient due to the increasing duration and higher frequency of extreme intense rainstorms. Therefore, model simulations of flood hazard that account for these rather unprecedented types of extremely destructive events are required, and those need to be integrating the newest data from all types of sensors. At the same time, we observe that sustainable, nature-based solutions are now sought after because these solutions offer an inviting alternative to ever changing flood risk, particularly under the present and future climate crisis. It is stipulated that increasing healthy urban vegetation cover could reduce this risk and is a form of a nature-based solution for urban areas. Here we combine existing methods from the literature and develop a methodology relating time-series of satellite-based vegetation maps, topography and soil permeability to estimate excess runoff from intense precipitation. The runoff coefficient is mapped through the use of a composite curve number method.. The method of looking at the partition between rainfall and runoff is highly correlated to change in land use, and thus changes in vegetation cover. Relying on the NDVI index for green vegetation mapping, the methodology is able to capture the differences in the hydrological response even for seasonal or canopy integrity changes. Looking at different vegetation cover scenarios therefore allows the creation of different runoff responses, and therefore a possible reduction in flood risk.In this paper, we present initial results of this flood risk analysis, the goal of which is to produce runoff change maps at city, urban neighbourhood or city post code level using different scenarios in rainfall amounts from design storms coupled with existing or planned urban vegetation cover scenarios.

Achieving an efficient management of agricultural fields is crucial for farmers, considering the challenges posed by water resources sustainability. Monitoring tools allow winemakers to keep their vineyards under control and improve the plants' health with targeted actions, such as irrigations scheduling or specific treatments. Knowing the actual number of plants may not be evident, particularly in old plantation, where vines might have been removed or added to the initial planting scheme, or even they have never been counted. However, the knowledge of the number and position of the plants is important to be able to adapt the irrigation network and properly adjust the irrigation schedule. The aim of this project is to build a drone service for precision monitoring of vineyards. Here, an algorithm to detect the position and number of plants in vineyards using drone RGB imagery is presented. First results show a plant detection accuracy of 87%.