At Fluvialycs, we are committed to pushing the boundaries of flood prediction technology. As part of our ongoing efforts, we conducted rigorous testing of our AI model during Storms Ciara and Dennis in 2020. Our objective was to assess the efficacy of Fluvialycs in predicting flood events under real-world conditions.

To simulate the real scenario of Fluvialycs in action, we utilized specialized models that were not trained with data from Storms Ciara and Dennis. This approach allowed us to evaluate how Fluvialycs would perform in forecasting flood events without prior knowledge of these specific storms. The test was undertaken simulating to be on February 13th 2020 at 18:00. The below graph illustrates river Trent levels at Colwick, Nottingham, for the past 5 days and the predicted levels for the next 5 days.

To validate Fluvialycs’ predictions, we have overlaid the actual observed Trent River levels for the future 5 days. This comparison allows us to assess the accuracy and reliability of Fluvialycs’ forecasting capabilities in real-world scenarios. By comparing Fluvialycs’ predictions with the observed level, we can ascertain the model’s effectiveness in anticipating flood events. Notably, the graph reveals that Fluvialycs predicts the peak level several days before it occurs. This early warning capability is a key tool for enhancing preparedness and response efforts, enabling stakeholders to take proactive measures to mitigate the impact of flooding.