Exploring GloFAS: AI-Driven Flood Forecasting
GloFAS is the global flood forecasting system of Copernicus Emergency Management Service (CEMS), managed by the European Commission’s Joint Research Centre.
Since 2011 it has provided daily global flood forecasts up to 30 days in advance, aiding in emergency preparedness and response.
Image source: Copernicus
GloFAS complements national and regional data and supports decision making and preparatory measures before major flood events (particularly in large international river basins.
Activity: Play the video introducing the viewer, then register and login to the GloFAS service.
Activity: Explore the different functions of the GloFAS system.
Zoom in a location of interest with a forecast station.
What are the positives and negatives of such AI-driven product in terms of geographical precision, temporal scale, availability of data?
Which of the following data types might be most affected by human biases – rainfall data, river level, weather forecast, river gauges, RADAR systems, satellite imagery?
Image source: GloFAS
The accuracy of AI-based forecasting with GloFAS depends on calibrating the flood frecasting model with river gauge data. .
click image to enlarge
Image source
Accuracy issues with AI flood forecasting concern the:
– quality of data input , such as meteorological and hydrological observations
– spatial resolution, there are limitations modelling small-scale hydrological processes
– type of flood event, GloFAS has not been designed to model events such as flash floods, coastal floods, or urban floods.
Another major challenge exists with basins that do not have river gauges, as it is not possible to gather enough data to assess the reliability of the model.
Activity: Watch the video on the challenges of flood prediction
Utilising artificial intelligence (AI) in flood forecasting offers significant advancements but also presents several ethical challenges including:
– Data Bias: biased or incomplete data can produce inaccurate forecasts, potentially reinforcing existing inequalities.
– Transparency and Accountability: Complex AI algorithms may lack transparency, making it difficult for stakeholders to understand and trust predictions.
– Privacy: AI systems often process vast amounts of personal and geospatial data.
– Equitable Access: lack of technological infrastructure can lead to unequal access to AI-driven flood warnings.
– Misuse of Models: Over-reliance on AI predictions without human oversight can result in misguided decisions.
Activity: Research and consider how these challenges might be addressed.