The solution is a Flood & Multi-Hazard Resilience Platform for cities, built around four core components: LISFLOOD, SUMO, FIWARE, and Cell Broadcast, plus an innovative deep-learning layer.
2.1. Data & Flood Modelling Layer (LISFLOOD)
A dedicated modelling service runs LISFLOOD / LISFLOOD-FP on calibrated catchments covering, its upstream rivers, and coastal/low-lying areas. LISFLOOD ingests numerical weather prediction (rain, snow, wind), sea-level and storm-surge forecasts, snowpack and soil-moisture data to compute grid-based runoff, river stages, and 2D floodplain depth/velocity maps for lead times from hours to days.
Outputs (for each time step) are exposed as geospatial services (raster tiles, vector polygons) and published into FIWARE Orion Context Broker as entities like FloodCell, RiverReach, CoastalZone with attributes such as waterDepth, arrivalTime, maxVelocity.
2.2 Traffic Simulation & Accessibility Layer (SUMO)
A high-resolution Eclipse SUMO model covers Pärnu’s urban network and the regional links that connect to Saaremaa (including key ports, bridges, and ferries).
A “Network Adapter” microservice listens to FIWARE:
When LISFLOOD predicts >X cm water depth on a RoadSegment, it marks the edge as closed or penalised in SUMO.
Real-time traffic from probe vehicles, buses, emergency fleets, and induction loops updates demand and speeds.
SUMO then simulates:
Which roads remain open, realistic travel times, and congestion under evacuation or detour scenarios.
Reachability to hospitals, shelters, power repair depots, and ferry ports for the next 1–6 hours.
Read more from the attachment.
