Artificial intelligence and anticipatory action
Here at JBA, we are leading the way in utilising Artificial Intelligence (AI) to make a step change in anticipatory action in the humanitarian sector; helping to reduce the impacts of flooding disasters globally.
For the past decade, we have integrated AI across many of our activities, from flood modelling and forecasting to advanced vessel route planning for the offshore wind sector. Our commitment to innovation in AI led to the launch of JBAi, a major investment programme dedicated to accelerating the adoption of AI technologies in a safe and ethical manner.
One of the key areas of focus for JBAi is how AI can advance the field of anticipatory action and support the UN’s “early warnings for all” initiative.
How are we supporting humanitarian anticipatory action?
We have developed systems for designing, triggering and operationalising anticipatory action through impact-based flood forecasting that identifies specific communities and infrastructure at risk before a flood occurs. This information is then used by governments, emergency responders and humanitarian organisations to take the actions needed to protect lives.
For instance, JBA’s Flood Foresight technology has been supporting anticipatory action for 10 years, and is active in countries such as Bangladesh, Pakistan and Democratic Republic of the Congo.
What is Flood Foresight?
Flood Foresight is a unique, globally scalable system that combines hydrological and hydraulic models with real-time data inputs, such as river gauges, weather forecasts, and satellite imagery to create impact-based forecasts used to trigger anticipatory action.
Importantly, Flood Foresight computes flood forecasts quickly and objectively; both essential components for effective anticipatory action, allowing for the timely dissemination of warnings and mobilisation of resources.
The future of anticipatory action
But looking to the future of anticipatory action, even faster is better. This is where AI comes in. Over the last two years we have been developing a new AI-powered flood model called JFlowAI, trained from the outputs of our powerful hydrodynamic flood model JFlow. JFlowAI will allow country scale forecasts to be computed in seconds to minutes.
The benefits of this for anticipatory action will be profound. This hyper fast modelling will enable multi-scenario forecasting, uncertainty analysis, and rapid “what-if” forecasting, proving the comprehensive information needed to make risk and data informed decisions that save lives.
However, the integration of AI in flood forecasting is not without challenges. Ensuring the validation and quality assurance of AI models, communicating complex methodologies to non-technical users, and designing schemes that are both effective and understandable are crucial hurdles to overcome.
We believe in the transformative potential of AI for anticipatory action but advocate for its thoughtful application. By designing tailored schemes and utilising expert consultancy when designing trigger mechanisms for anticipatory action, we can fully realise AI's capabilities, bringing significant advancements in flood risk management and beyond.