Model

The “Impact based flood forecast and early warnings” is a paradigm shift from the conventional flood forecast approach to forefront the needs of the affected communities, stakeholders and decision makers, recommended by the World Meteorological Organisation (WMO, 2015). UN initiated Early Warning For All is a groundbreaking initiative to ensure that everyone on Earth is protected from hazardous weather, water, or climate events through life-saving early warning systems by the end of 2027.

The approach provides a platform for interaction and sharing of real time information across a river basin to understand and foresee the impacts of an imminent extreme weather event and (or) excessive dam releases/ failures under multiple user defined scenarios. This improves the efficiency of disaster preparedness and management mechanism and equips decision makers to foresee and issue necessary prior warnings and prepare (evacuation, rescue and relief) under multiple disaster scenarios to reduce the impacts.

Approach

There is a lack of monitoring systems, impact based forecasts and large gaps in last mile connectivity that jeopardise the effectiveness of early warning systems. Deficient high resolution flood inundation maps for various weather and river conditions and inadequate local flood action plans limit its further use not only to the affected communities but also to the Disaster Management Authorities. The issue becomes more complex in the small humid tropical rivers where groundwater contribution in floods is significant but less understood. Most of the presently used hydrological models therefore underestimate flood volumes and fail to accommodate the higher time lags for peak runoff.

Our Community-Sourced Impact-based Flood Forecast and Early Warning System (CoS-it-FloWS), bridges the last mile connectivity and data gaps and provides innovative,inclusive solutions to address climate risks through:

• empowering local communities to build hyper-local monitoring systems
• developing hydrological models and impact-based warning systems
• co-creating science-based solutions and decision support systems for flood risk reduction with communities

Main Features

1. Hyper Local Community Sourced Data
2. High resolution Flood Inundation Mapping with AI tools
3. Hybrid flood forecast model leveraging hydrological and machine learning models
4. Impact Based Warning System

Components

• Hydrological Model: Variable Infilteration Capacity Model
• Machine Learning Model
• Django framework for backend
• ReactJS frontend
• MapLibre: Mapping Library

Model Inputs

• Community Sourced hyper local weather, surface and groundwater data collected through the app "Gather".
• High Resolution Digital Elevation Models (DEMs).
• Soil and vegetation/crop data.
• Reservoir and irrigation data.