RISC-KIT: EWS-DSS Hotspot Tool

RISC-KIT: EWS-DSS Hotspot Tool Robert McCall Ap van Dongeren Deltares www.risckit.eu This project has received funding from the European Union s Seventh Programme for Research, Technological Development and Demostration under Grant Agreement No. 603458. This presentation reflects the views only of the authors, and the European Union cannot be considered liable for any use that may be made of the information contained therein.

The RISC-KIT Toolkit 1. Storm Impact Database of present and historic socio-economic and physical data. 2.Coastal Risk Assessment Framework (CRAF) to identify - at the regional scale (100 s km) - present and future hot spot areas of coastal risk 3. Web-based Management Guide offering innovative, cost-effective, ecosystem-based DRR measures; 4. Quantitative, high-resolution Hotspot Tool (EWS/DSS) to evaluate the effectiveness of DRR measures in hot spots (with a scale of 10 s of km) and 5. MCA Guide to assess strategic alternatives with stakeholders

Hotspot tool Detailed model for one hotspot location (following from identification in CRAF) Hotspot tool Computes consequences (socio-economic, ecological, cultural) of storm events by coupling Computed onshore hazards (erosion, overtopping, inundation) Receptor information (infrastructure, land use / ecosystem, population, etc.) Vulnerability relations

Hotspot tool Hotspot tool can be used as an ex-ante planning tool: investigate the potential impact of storm events (when, where, how often) help design, assess and optimise DRR measures in the hotspot location (e.g., seawall versus flood-proofing) Hotspot tool can be used as part of an Early Warning System provide predictions of coastal impacts including uncertainty ranges designed to fit in with current flood EWSs

Computing onshore hazards Physics-based numerical model to compute multi-hazards (e.g., erosion, overtopping, inundation) In RISC-KIT many partners using XBeach (natural and urbanised coasts, exposed to waves) Choice of model can be tailored to each site (e.g., LISFLOOD- FP, XBeach-G, etc.) Storm-induced dune erosion, overwash and breaching around a coastal apartment building in Ocean County, New Jersey, using XBeach.

Computing consequences Combine local hazards (e.g., flood depth), receptor information (buildings) and vulnerability (depth-damage curves) to compute consequences Receptor (housing) data and vulnerability relations at the RISC- KIT Porto Garibaldi site

Assessing DRR measures Compute effect of DRR measures on consequences of storms Measures that affect the hazard pathway (flood defences) Measures that affect exposure and vulnerability (evacuation) DRR measure: dune nourishment Flood extent without and with DRR measure (RISC-KIT site Kristianstad)

Bayesian Network During planning/assessment phase 100s of model simulations can be run, describing Changes in storm conditions (small-large return periods, climate change and sea level rise scenarios) Changes in the physical setting Varying DRR measures Results of all model simulations and measured data are stored in a Bayesian Network

Bayesian Network Bayesian Network identifies probabilistic relations between storm characteristics and DRR measures, and local hazards and consequences Interactive, fast and efficient communication tool

Bayesian Network RISC-KIT Kristianstad BN Storm characteristics DRR Measures Local hazards Receptor info Consequences

Bayesian Network Large storm Storm characteristics DRR Measures Local hazards Receptor info Consequences

Bayesian Network Area-specific Storm characteristics DRR Measures Local hazards Receptor info Consequences

Bayesian Network Effect of DRR measure Storm characteristics DRR Measures Local hazards Receptor info Consequences

Early Warning System In RISC-KIT we use the free Delft FEWS software package Delft FEWS currently used by EA as part of the National Flood Forecasting System Delft FEWS is an open data handling platform, allows connections between any forecasting model

Early Warning System RISC-KIT project has extended the use of Delft FEWS to coastal flooding data types model adaptors data visualisation Application of Delft FEWS as a coastal EWS is currently being tested at 10 RISC- KIT case-study sites around Europe

Early Warning System Example FEWS for the Ria Formosa, Algarve Wind forcing

Early Warning System Example FEWS for the Ria Formosa, Algarve Wave model

Early Warning System Example FEWS for the Ria Formosa, Algarve Local tide model

Early Warning System Example FEWS for the Ria Formosa, Algarve Local hazards model

Early Warning System - BN In RISC-KIT we have coupled FEWS with the Bayesian Network developed at the RISC-KIT case-study sites Many advantages: EWS provides one prediction, BN provides a probability distribution The BN can be used in place of the physics-based numerical hotspot model is the computational demand is too great for an operational EWS The BN can continue learning from the hotspot model results if it is part of an operational EWS

Early Warning System - BN Example of BN in place of a computationally-expensive physic-based model

Visualisation on the web RISC-KIT developing web-based viewer of results of hotspot tool Bayesian Network data can be used during planning phase of DRR measures data sharing with stakeholders, building consensus public outreach

Progress Meeting, Berlin, 2015-06-18

Visualisation on the web RISC-KIT developing web-based viewer of results of hotspot tool Can be used as part of an EWS easy data access BN data to help decision-making in preparedness and response actions

Visualisation on the web Presentation of flow discharge on the RISC-KIT web viewer

Take home: Hotspot tool is Model to predict local hazards and consequences Bayesian Network to identify probabilistic relations and provide uncertainty ranges Delft FEWS used to run operational coastal flood forecasting Web viewer for dissemination and data sharing Hotspot tool used for Ex-ante planning and assessment of DRR measures Coastal operational flood early warning

Thank you www.risckit.eu