Hyperion is an operational intelligence layer for compound urban flooding.

It reads rainfall, tidal loading, river response, terrain, drainage, and the built environment in one regional frame. Instead of treating those signals as separate alerts, it makes their interaction inspectable for the people deciding what to watch and where to prepare.

Flood decisions are often assembled from fragments: a rainfall warning, a tide chart, a river gauge, a static susceptibility map, and a forecast that does not make its limits clear. That fragmentation can create both delay and false precision at the moment an operator needs a coherent regional picture.

Hyperion combines a shared decision architecture with regional adapters rather than assuming that one map or model transfers unchanged everywhere.

The work joins common data layers, susceptibility and event modelling, compound nowcasting, calibrated scores, visible uncertainty, event-specific validation, and transfer testing. The interface keeps evidence and model limits close to the decision, not hidden behind a single risk label.

Hyperion gives Genesis contact with a bounded physical world in which several causal systems interact at once.

It can contribute spatial relations, temporal dynamics, incomplete observations, uncertainty under real consequences, and transfer between environments — while preserving the evidence that the same schema can require different regional interpretations.