Weather event

The first half of the framework answers a deceptively simple question: what, exactly, is the weather doing, and where?

Getting from a broad climate scenario to something a transport model can actually use takes three steps.

The three steps

1. Choose a climate scenario

A climate scenario sets the backdrop. It might be the current climate, or a future climate under a specific set of climate-change assumptions.

2. Derive the weather variables

From that scenario, the platform estimates the variables that define a concrete weather event, such as convective rainfall, an extreme flood or an acute heatwave.

This is the step that turns a climate (decades of statistics) into an event (something with a start, an end and an intensity).

3. Run the hazard model

The event is then applied to a specific UK region and passed through a hazard model: hydrological models for flooding, heat models for temperature.

The output is the useful part. It shows which areas are affected, and how severely, as the event unfolds. Hazard models resolve this finely in time, typically at 20-minute intervals, because a road that floods for twenty minutes during the morning peak is a very different problem from one that floods overnight.

Why the timing matters

It is tempting to treat a hazard as a yes/no state: this road is flooded, that one is not. The IMP deliberately does not.

Because the hazard output varies over time, it can be intersected with a transport simulation that also varies over time. That lets the platform distinguish a closure that catches the commute from one that nobody was travelling through anyway.

Covered in detail