NH1.1

This talk reports a new and significantly enhanced analysis of US flood hazard at 30m spatial resolution.  For the first time we consider pluvial, fluvial and coastal flood hazards within the same framework and provide projections for both current (rather than historic average) conditions and for future time periods centred on 2035 and 2050 under the RCP4.5 emissions pathway.  Validation against high quality local models and the entire catalogue of FEMA 1% annual probability flood maps yielded Critical Success Index values in the range 0.69-0.82.  Significant improvements over a previous pluvial/fluvial model version are shown for high frequency events and coastal zones, along with minor improvements in areas where model performance was already good.  The result is the first comprehensive and consistent national scale analysis of flood hazard for the conterminous US for both current and future conditions.  Even though we consider a stabilization emissions scenario and a near future time horizon we project clear patterns of changing flood hazard (3σ changes in 100yr inundated area of -3.8 to +16% at 1° scale), that are significant when considered as a proportion of the land area where human use is possible or in terms of the currently protected land area where the standard of flood defence protection may become compromised by this time.

How to cite: Bates, P., Quinn, N., Sampson, C., Smith, A., Wing, O., Savage, J., Olcese, G., Sosa, J., and Neal, J.: US fluvial, pluvial and coastal flood hazard under current and future climates, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14646, https://doi.org/10.5194/egusphere-egu21-14646, 2021.

Hydrologic models are employed in the flood risk studies to simulate time series of model responses to given inputs. These simulated time series of pseudo-observations can be next statistically analysed and in this way they can extend existing observed records. Simulations of hydrologic models are however associated with modelling uncertainty, often represented through a simulation ensemble with multiple parameter sets. The need of using multiple parameter sets to represent uncertainty is linked however with increased computational costs that may become prohibitive for long-time series and many input scenarios to be analysed. Due to the non-linear input-output relationship in the hydrologic model, a pre-selection of parameter sets is challenging.

This work presents a clustering approach as a tool to learn about the model hydrologic responses in the flood frequency space from the training dataset. Based on this learning process, representative parameter sets are selected that can be directly used in other model applications to derive prediction intervals at much lower computational costs. The study is supported with sensitivity analysis to the number of clusters. Based on results from a small catchment in Switzerland and 10’000 years of streamflow pseudo-observations, it has been found that grouping the full simulation ensemble with 1000 members into 3 to 10 clusters is already suitable to derive commonly applied prediction intervals (90%, 95% or 98%) in the flood frequency space. The proposed clustering approach can be applied in any flood risk analysis to lower the computational costs linked with the use of a hydrologic model.

How to cite: Sikorska-Senoner, A. E.: Clustering model responses in the frequency space for improved flood risk analysis, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1199, https://doi.org/10.5194/egusphere-egu21-1199, 2021.

Every year flood events cause worldwide vast economic losses, as well as heavy social and environmental impacts, which have been steadily increasing for the last five decades due to the complex interaction between climate change and anthropogenic pressure (i.e. land-use and land-cover modifications). As a result, the body of literature on flood risk assessment is constantly and rapidly expanding, aiming at developing faster, computationally lighter and more efficient methods relative to the traditional and resource-intensive hydrodynamic numerical models. Recent and reliable fast-processing techniques for flood hazard assessment and mapping consider binary geomorphic classifiers retrieved from the analysis of Digital Elevation Models (DEMs). These procedures (termed herein “DEM-based methods”) produce binary maps distinguishing between floodable and non-floodable areas based on the comparison between the local value of the considered geomorphic classifier and a threshold, which in turn is calibrated against existing flood hazard maps. Previous studies have shown the reliability of DEM-based methods using a single binary classifier, they also highlighted that different classifiers are associated with different performance, depending on the geomorphological, climatic and hydrological characteristics of the study area. The present study maps flood-prone areas and predicts water depth associated with a given non-exceedance probability by combining several geomorphic classifiers and terrain features through regression trees and random forests. We focus on Northern Italy (c.a. 100000 km², including Po, Adige, Brenta, Bacchiglione and Reno watersheds), and we consider the recently compiled MERIT (Multi-Error Removed Improved-Terrain) DEM, with 3sec-resolution (~90m at the Equator). We select the flood hazard maps provided by (i) the Italian Institute for Environmental Protection and Research (ISPRA), and (ii) the Joint Research Centre (JRC) of the European Commission as reference maps. Our findings (a) confirm the usefulness of machine learning techniques for improving univariate DEM-based flood hazard mapping, (b) enable a discussion on potential and limitations of the approach and (c) suggest promising pathways for further exploring DEM-based approaches for predicting a likely water depth distribution with flood-prone areas. 

How to cite: Magnini, A., Lombardi, M., Persiano, S., Tirri, A., Lo Conti, F., and Castellarin, A.: Combination of geomorphic classifiers through Machine Learning-based techniques for flood hazard assessment, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4723, https://doi.org/10.5194/egusphere-egu21-4723, 2021.

Identification of flood water extent from satellite images has historically relied on either synthetic aperture radar (SAR) or multi-spectral (MS) imagery. But MS sensors may not penetrate cloud cover, whereas SAR is plagued by operational errors such as noise-like speckle challenging their viability to global flood mapping applications. An attractive alternative is to effectively combine MS data and SAR, i.e., two aspects that can be considered complementary with respect to flood mapping tasks. Therefore, in this study, we explore the diverse bands of Sentinel 2 (S2) derived water indices and Sentinel 1 (S1) derived SAR imagery along with their combinations to access their capability in generating accurate flood inundation maps. For this purpose, a fully connected deep convolutional neural network known as U-Net is applied to combinations of S1 and S2 bands to 446 (training: 313, validating: 44, testing: 89) hand labeled flood inundation extents derived from Sen1Floods11 dataset spanning across 11 flood events. The trained U-net was able to achieve a median F1 score of 0.74 when using DEM and S1 bands as input in comparison to 0.63 when using only S1 bands highlighting the active positive role of DEM in mapping floods. Among the, S2 bands, HSV (Hue, Saturation, Value) transformation of Sentinel 2 data has achieved a median F1 score of 0.94 outperforming the commonly used water spectral indices owing to HSV’s transformation’s superior contrast distinguishing abilities. Also, when combined with Sentinel 1 SAR imagery too, HSV achieves a median F1 score 0.95 outperforming all the well-established water indices in detecting floods in majority of test images.

How to cite: Konapala, G. and Kumar, S.: Exploring Sentinel-1 and Sentinel-2 diversity for Flood inundation mapping using deep learning, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10445, https://doi.org/10.5194/egusphere-egu21-10445, 2021.

Flooding seems to be the most widespread and common catastrophe in a tropical country such as India. Efficient rainfall, industrial development, huge population, the effect of the tide, and urban growth are actual reasons for flooding in urban coastal regions. Navsari, the city of Gujarat, located 19 km upstream of the Arabian Sea. The city has experienced a devastating flood on 4rth August 2004. Flash flooding and maximum discharge estimated at the Mahuva gauge station of about 8836 m3/sec were responsible for a disaster that resulted in massive damage to property and lives. A two dimensional (2D) flood simulation model is carried out to assessment of flood inundation in an urban coastal area. HEC-RAS is one of the most popular open-source hydraulic software having 2D capabilities including GIS features. In the present study, the distance between the Mahuva gauge station to the Arabian sea was considered for flood inundation assessment, whereas the SRTM 30 m DEM was used for grid generation for Navsari city. The inflow hydrograph was used as the upstream boundary condition, and normal depth was used as the downstream boundary condition during the 4th August 2004 flood event. The unsteady flow simulation was performed and validated for the year of 2004 flood event. The simulated outcomes show that major areas such as Viraval, Kachiawad, Jalalpore, near Railway station, Kaliawad, Tavdi village, and Near TATA School were flooded with 2-4 m depth. Furthermore, the simulated result demonstrates that, if the discharge exceeds 8836 m3/sec in the area of a floodplain, it may take 11 to 13 hours to make the city inundated. The R² value for the model is 0.9679, which shows that the observed value is the best match with the simulated value. The research study illustrates the accurate flood inundation assessment in the urban coastal area using open-source 2D HEC-RAS model. The present study described the applicability of open-source data and model in flood inundation assessment. The study will fill the gap of flood assessment through 2D HEC-RAS model worldwide areas, which are situated nearby coastal region, accompanied by the benefits of open-source dataset and model.

How to cite: Pathan, A. I., Agnihotri, Dr. P. G., Patel, Dr. D., and Prieto, Dr. C.: Improving assessment of flood inundation of Navsari (India) via open-source data and HEC-RAS model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4187, https://doi.org/10.5194/egusphere-egu21-4187, 2021.

The flood caused by a dam-break event generally contains a large amount of energy, and it can be destructive to the downstream buildings and structures. An experiment-validated three-dimensional numerical model was designed to investigate the impact of dam-break flood on structures with different arrangements. The Eulerian two-phase flow model and the smooth particle dynamics method are applied separately to solve the flow motion, and  the deformation characteristics of buildings under the flood impact are evaluated by fluid-structure interaction model. An experiment is constructed to validate the numerical simulation. The results show that the structure suffers a large instantaneous impact pressure when the flood water first contacts the structure, and the value of this pressure can reach 1.5-3.0 times that of the maximum pressure after the first impact, and the maximum total pressure of the upstream building surface is about 1800N. The deformation near the door and windows is obvious, and the maximum deformation can reach 600μm, which further results in the large deformation of the gable and roof on both sides. Moreover, the arrangement of buildings has different blocking effect on flood. The back-row buildings arranged in alignment along the flow direction still has to bear 20% flood impact, and the front row buildings arranged alternately bear 90% high-speed flow impact. The structural damage is evaluated by the material failure criterion, and the weak position of buildings is identified, providing an optimal design of buildings.

How to cite: Rong, Y., Bates, P., and Neal, J.: Safety evaluation of buildings under flood impact, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6049, https://doi.org/10.5194/egusphere-egu21-6049, 2021.

Wave overtopping on grass-covered dikes results in erosion of the dike cover. Once the dike cover is eroded, the core will be washed away and the dike breaches, leading to flooding of the hinterland. Transitions between grass covers and revetments or geometric transitions are vulnerable for cover erosion and are therefore the most likely locations to initiate dike breach. These transitions affect the overtopping flow and thereby the hydraulic load on the dike cover. For example, bed roughness differences can create additional turbulence and slope changes can result in the formation of a jet that increases the load at the jet impact location. Although it is known that dike cover failure often starts at transitions, the effect of transitions on the hydraulic load remains unknown.

We developed a detailed numerical 2DV model in OpenFOAM for the overtopping flow over the crest and the landward slope of a grass-covered dike. This model is used to study the effects of transitions on the overtopping flow variables including the flow velocity, shear stress, normal stress and pressure. Several types of transitions are studied such as revetment transitions, slope changes and height differences. 

The results show that the shear stress, normal stress and pressure increase significantly at geometric transitions such as the transition from the crest to the slope and at the landward toe. The increase depends on the wave volume and the geometry of the dike such as the steepness and length of the landward slope. Furthermore, the results show that roughness changes at revetment transition on a grass-covered crest has no influence on the maximum shear stress, maximum normal stress and maximum pressure. The flow velocity increases from a rough to a smooth revetment, while the opposite occurs for the transition from a smooth to a rough revetment. The variation in the flow velocity is well described by analytical formulas for the maximum flow velocity along the dike profile. These formulas are also able to describe the variation in flow velocity for a revetment transition on a berm on the landward slope. In this case, the shear stress increases from a smooth to a rough revetment and decreases from a rough to a smooth revetment. This means that a rough revetment can locally reduce the shear stress, however the transitions have no effect on the shear stress downstream.

These model results are used to obtain relations for the increase in the hydraulic variables at transitions. These relations can be used to describe the effect of transitions on the hydraulic load in models for grass cover failure by overtopping waves. Accurate descriptions of the hydraulic load in these models will improve the failure assessment of grass-covered dikes with transitions.

How to cite: van Bergeijk, V., Warmink, J., and Hulscher, S.: How do transitions affect the wave overtopping flow locally as well as downstream?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-679, https://doi.org/10.5194/egusphere-egu21-679, 2021.

Estuarine regions are strategically important from an environmental, economic, and social point of view. To reduce vulnerability and increase resilience, it is crucial to know their dynamics that usually are poorly understood. Numerical models have proven to be an appropriate tool to improve this knowledge and simulate scenarios for future conditions. However, as the modelling results may be inaccurate, the application of the ensembles technique can be very useful in reducing possible uncertainties. In the EsCo-Ensembles project, this technique is proposed to improve hydrodynamic predictions for two Portuguese estuaries: Douro and Minho.

Two already validated numerical models (openTELEMAC-MASCARET and Delft3D), which have demonstrated their ability to accurately describe estuarine hydrodynamic patterns and water elevation for river flow in normal and extreme conditions, were applied. Several scenarios for climate change effects were defined including river flood peak flows for the 100 and 1000 year return periods and sea level extreme values for RCPs 4.5 and 8.5 in 2100.

The results demonstrated a clear difference between the hydrodynamic behaviour of the two estuaries. Model outcomes for the Minho estuary, which is dominated by the tide and therefore by oceanographic conditions, show a pronounced effect of rising sea levels on estuarine hydrodynamics. Whereas, for the Douro estuary, which is heavily dominated by the river flow, the effect of the sea level rise is hardly noticeable during flood events.

These and further results of this ongoing project are expected to (i) provide a complete hydrodynamic characterization of the two estuaries; (ii) evaluate future trends; (iii) estimate the flood risks associated with extreme events and (iv) demonstrate that the combined use of different models reduces their uncertainty and increases the confidence and consistency of the forecasts.

Acknowledgements: To the Strategic Funding UIDB/04423/2020 and UIDP/04423/2020 (FCT and ERDF) and to the project EsCo-Ensembles (PTDC/ECI-EGC/30877/2017, NORTE 2020, Portugal 2020, ERDF and FCT). The authors also want to acknowledge the data provided by EDP, IH and Confederación Hidrográfica Miño-Sil.

How to cite: Iglesias, I., Pinho, J. L., Bio, A., Avilez-Valente, P., Melo, W., Vieira, J., Bastos, L., and Veloso-Gomes, F.: Future estuarine circulation patterns characterization based on a hydrodynamic models ensemble, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6541, https://doi.org/10.5194/egusphere-egu21-6541, 2021.

Land use and delineation of flood-prone areas require valuable and effective tools, such as flood mapping. Local authorities, in order to prevent and mitigate the effects of flood events, need simplified methodologies for the definition of preliminary flooded areas at a large scale. In this work, we focus on the workflow GeoFlood, which can rapidly convert real-time and forecasted river flow conditions into flooding maps. It is built upon two methodologies, GeoNet and the HAND model, making use only of high-resolution DTMs to define the geomorphological and hydraulic information necessary for flood inundation mapping, thus allowing for large-scale simulations at a reasonable economical and computational cost. GeoFlood potential is tested over the mid-lower portion of the river Tiber (Italy), investigating the conditions under which it is able to reproduce successful inundation extent, considering a 200-year return period scenario. Results are compared to authority maps obtained through standard detailed hydrodynamic approaches. In order to analyze the influence of the main parameters involved, such as DTM resolution, channel segmentation length, and roughness coefficient, a sensitivity analysis is performed. GeoFlood proved to produce efficient and robust results, obtaining a slight over-estimation comparable to that provided by standard costly methods. It is a valid and relatively inexpensive framework for inundation mapping over large scales, considering all the uncertainties involved in any mapping procedure. Also, it can be useful for a preliminary delineation of regions where the investigation based on detailed hydrodynamic models is required.

How to cite: D'Angelo, C., Passalacqua, P., Fiori, A., and Volpi, E.: Prediction capabilities of GeoFlood for the delineation of flood-prone areas: the Tiber River case study, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7591, https://doi.org/10.5194/egusphere-egu21-7591, 2021.

In this work, we present a general framework for design and risk assessment of hydraulic structures for water control. The framework relies on a “structure-based approach”, accounting for both the statistical behavior of the hydrological load acting on the river system and the hydraulic response of the structure to the environmental load. This approach allows for the reduction of a multivariate and complex statistical problem to a univariate one, focusing on the damage. The framework is applied to an offline detention basin for flood mitigation based on a general, yet simplified routing model. Furthermore, a real-world case study application is presented, with the specific aim of discussing the role of the design parameters and their effect on the probability distribution of damage. Results show the robustness and the effectiveness of the approach for applications to real cases and provide design guidance for practitioners.

How to cite: Cipollini, S., Fiori, A., and Volpi, E.: Design of offline reservoirs for flood mitigation by using a structure-based risk framework, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9118, https://doi.org/10.5194/egusphere-egu21-9118, 2021.

Earthen levees protect flood-prone areas during severe flood events. In most cases, however, flooding is the result of the collapse of the embankments due to the seepage through and under the levee body. The description of the seepage line is difficult mainly because of the uncertainty on the hydraulic parameters, first of all the soil hydraulic conductivity. Barbetta et al. (2017) proposed a practical method for the seepage analysis based on the Marchi’s equation for the estimation of the probability of occurrence of the levee seepage which provides a vulnerability index under the assumption that the groundwater level coincides with the ground. Recently, the method has been tested also considering the groundwater level below the ground pointing out that such a condition has a high impact on the levee vulnerability to seepage. However, it does not consider the interactions between seepage process in the levee body and in the foundation.

In this context, this work proposes a new approach for the analysis of the infiltration line through the body and the foundation, considering a multilayer soil and assuming a different soil hydraulic conductivity for each layer. The new equation is obtained starting from the continuity equation and the flow equation.

The saturation line estimated through the Marchi’s equation and the one derived through the new multilayer model equation are compared. The analysis is first addressed to identify a threshold of the ratio between water head and water table beyond which the Marchi’s equation is no longer applicable. Indeed, the Marchi’s equation is valid when the river water head is lower than the water table. Different values of these two variables are analyzed and a threshold ratio equal to 0.57 is identified.

Furthermore, the levee vulnerability to seepage estimated with the two approaches is compared and the levee is found more vulnerable when the new approach is applied. The results indicate that the difference between the two vulnerability approaches decreases as the distance between the groundwater table and the ground level tends to zero. The proposed approach is an attempt to quantify the seepage probability with more realistic levees characteristics, hydraulic and soil parameters.

Barbetta, S., Camici, S., Bertuccioli, P., Palladino, M. R., & Moramarco, T. 2017. Refinement of seepage vulnerability assessment for different flood magnitude in national levee database of Italy. Hydrology Research, 48(3), 763–775. https://doi.org/10.2166/nh.2017.101.

How to cite: Bonaccorsi, B., Moramarco, T., Noto, L. V., and Barbetta, S.: A multilayer soil approach for seepage process analysis in earthen levees, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11038, https://doi.org/10.5194/egusphere-egu21-11038, 2021.

During intense periods of drought, the development of cracks is observed in peat and clay dikes. Asset managers of the dikes increase the inspection frequency in times of drought to be able to monitor these cracks. Significant development of the cracks contributes to the development of different failure mechanisms. In this study, the occurrence of the cracks is predicted at a large spatial scale. An inspection database in which the observations from the last three years are stored is used as the basis. The database contains hundreds of observed cracks including the  location and time in which they were observed. The database was extended with attributes such as the precipitation deficit, the peat width at the surface, the orientation of the dike body, the subsidence of the dike body and the soil stiffness. Decision tree algorithms were then used to classify which circumstances will lead to cracks and which circumstances will not. From the resulting decision trees it was deduced that high precipitation deficits, low soil stiffness and the peat width can be used as the main predictors for the occurrence of cracks. Both subsidence of the foundation and the dike body being orientated to the sunny side are also contributors, although less prominent. Time-independent cracking criteria were then used to classify which regions are prone to cracking. Dikes which are rich in peat with a low stiffness were thus highlighted. The Mathews correlation coefficient was used as performance criteria resulting in a 0.3 value for the obtained tree. Application of a random forest increased the coefficient to 0.8. An important conclusion is that proper monitoring of the peat width, soil stiffness and precipitation may result in better asset management.

How to cite: Chotkan, S., Aguilar-López, J. P., van der Meij, R., Vardon, P., Jan Klerk, W., and Chacon Hurtado, J. C.: Predicting drought-induced cracks in dikes with machine learning algorithms, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15435, https://doi.org/10.5194/egusphere-egu21-15435, 2021.

Floods are natural disasters caused mainly due to heavy or excessive rainfall. They induce massive economic losses in Bangladesh every year. Physically-based flood prediction models have been used over the years where simplified forms of physical laws are used to reduce calculations' complexity. It sometimes leads to oversimplification and inaccuracy in the prediction. Moreover, a physically-based model requires intensive monitoring datasets for calibration, accurate soil properties information, and a heavy computational facility, creating an impediment for quick, economical and precise short-term prediction. Researchers have tried different approaches like empirical data-driven models, especially machine learning-based models, to offer an alternative approach to the physically-based models but focused on developing only one machine learning (ML) technique at a time (i.e., ANN, MLP, etc.). There are many other techniques, algorithms, and models in machine learning (ML) technology that have the potential to be effective and efficient in flood forecasting. In this study, five different machine learning algorithms- exponent back propagation neural network (EBPNN), multilayer perceptron (MLP), support vector regression (SVR), DT Regression (DTR), and extreme gradient boosting (XGBoost) were used to develop total 180 independent models based on a different combination of time lags for input data and lead time in forecast. Models were developed for Someshwari-Kangsa sub-watershed of Bangladesh's North Central hydrological region with 5772 km² drainage area. It is also a data-scarce region with only three hydrological and hydro-meteorological stations for the whole sub-watershed. This region mostly suffers extreme meteorological events driven flooding. Therefore, satellite-based precipitation, temperature, relative humidity, wind speed data, and observed water level data from the Bangladesh Water Development Board (BWDB) were used as input and response variables.

For comparison, the accuracy of these models was evaluated using different statistical indices - coefficient of determination, mean square error (MSE), mean absolute error (MAE), mean relative error (MRE), explained variance score and normalized centred root mean square error (NCRMSE). Developed models were ranked based on the coefficient of determination (R²) value. All the models performed well with R² being greater than 0.85 in most cases. Further analysis of the model results showed that most of the models performed well for forecasting 24-hour lead time water level. Models developed using XGBoost algorithm outperformed other models in all metrics. Moreover, each of the algorithms' best-performed models was extended further up to 20 days lead time to generate forecasting horizon. Models demonstrated remarkable consistency in their performance with the coefficient of determination (R²) being greater than 0.70 at 20 days lead-time of forecasting horizon in most cases except the DTR-based model. For 10- and 5-days lead time of forecasting horizon, it was greater than 0.75 and 0.80 respectively, for all the model extended. This study concludes that the machine algorithm-based data-driven model can be a powerful tool for flood forecasting in data-scarce regions with excellent accuracy, quick building and running time, and economic feasibility.

How to cite: Haque, M. H., Sadia, M., and Mustaq, M.: Development of Flood Forecasting System for Someshwari-Kangsa Sub-watershed of Bangladesh-India Using Different Machine Learning Techniques, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15294, https://doi.org/10.5194/egusphere-egu21-15294, 2021.

The low lying and sandy coastal areas of the Emilia-Romagna region are heavily threatened by sea storms, often leading to flooding and coastal erosion events with severe impacts on citizens’ quality of life, damages to the cultural heritage and effects on economic activities (e.g. aquaculture, fisheries, tourism, beach facilities). Climate change projections reinforce the need of strategies and tools to prevent damages and promptly react to extreme events. In this context and in the framework of non-structural mitigation measures, the Hydro-Meteo-Climate Service of Arpae Emilia-Romagna (Arpae-SIMC) developed and operationally manages a Coastal Early Warning System (EWS) for the Emilia-Romagna Region (Northeast Italy).

The EWS was developed during the EU Project FP7-MICORE and it is a state-of-the-art coastal forecasting system that follows a chain of operational numerical models: the meteorological model COSMO, the wave model SWAN-MEDITARE, the ocean model AdriaROMS, and the morphodynamic model XBeach. The latter is currently implemented on a series of cross-shore beach profiles covering eight locations distributed along the Emilia-Romagna shore. Deterministic daily forecasts (72-hours) are generated and Storm Impact Indicators (SIIs) used to assess sea-storm induced coastal risk along the region’s littoral (geo.regione.emilia-romagna.it/schede/ews). 

It is widely known that among the limitations of deterministic approaches, the lack of uncertainty estimation is often problematic as decision-makers might be misled if the only forecast available underestimates (or overestimates) incoming conditions. Hence, following the success of probabilistic forecasting in meteorological applications, storm surge EWSs following ensemble frameworks have been recently developed, allowing for more information available to sustain the decision-making process. Towards the new paradigm change, one of the foreseen outputs of the European Interreg Italy-Croatia CBC Programme project Strategic development of flood management (STREAM) involves the development of a “probabilistic EWS for coastal risk implemented and tested on at least one location along the Emilia-Romagna Coast”. 

The initial implementation of the (semi-)probabilistic framework benefits from the EU ADRION I-STORMS (Integrated Sea Storm Management Strategies) project outcomes, in which wave and sea level multi-model ensembles were developed for the Adriatic Sea giving origin to the Transnational Multi-Model Ensemble (TMES). The TMES was made available as one of the six Integrated Web System (IWS) components, combining five wave and six sea level forecasting systems as means to provide 48-hour forecasts in terms of sea level and wave characteristics (Hs, Tm and Dm). Ensemble mean and standard deviation (SD) are calculated based on different forecasting systems’ results. In the initial approach, four TMES combinations have been tested as XBeach forcing: the TMES mean; the mean minus one SD; the mean plus one SD; the mean plus two SDs. Two months were analyzed together with the already implemented deterministic system for two profiles along the region’s coast.

The methodology followed for the test period will be shown as well as the results. Furthermore, the methodology under development will be also shown as means to enhance the discussion involving storm surge ensemble applications.

How to cite: Germano Biolchi, L., Unguendoli, S., Bressan, L., Giambastiani, B. M. S., and Valentini, A.: A (semi-)probabilistic storm surge EWS implementation for the Emilia-Romagna region (Italy), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15210, https://doi.org/10.5194/egusphere-egu21-15210, 2021.