Resilience tool
In one sentence
In transport research, resilience refers to the capacity of a road network to maintain its operating performance during a disruptive event, such as heavy rainfall, and to recover thereafter.
Resilience is commonly assessed through the degradation and subsequent recovery of a performance measure, such as travel speed or journey time (Mattsson and Jenelius, 2015; Ganin et al., 2017).
What a resilience model does differently
A resilience model characterises the response, not the hazard.
This is the distinction that places this chapter apart from the rest of the handbook. Everything else in the IMP models the hazard and then works out its consequences. A resilience model starts from observed outcomes and asks how the network has actually behaved.
It quantifies two things:
| Quantity | Question answered |
|---|---|
| Magnitude of performance loss | How much worse does the network get? |
| Distribution across the network | Which links bear that loss? |
Empirical, not simulated
The transport model derives disruption by simulating agents on a network under a modelled hazard. The resilience tool derives it by measuring what speeds actually did when it actually rained.
The two are complementary rather than competing. Simulation can explore conditions never observed; empirical estimation is grounded in what has been. The IMP runs both.
The GLM Resilience Model
The GLM Resilience Model provides an empirical measure of this response for the Strategic Road Network.
It builds on established evidence that rainfall systematically reduces traffic speed (Hranač et al., 2006; Becker et al., 2026), and estimates this sensitivity separately for each individual road link, directly from observed data.
Per-link estimation is the point
A single national rainfall-speed relationship would be easy to produce and nearly useless for prioritisation, because it would say every link degrades identically.
Estimating a coefficient per link is what turns the analysis into a map of where the network is weakest.
The acronym is not expanded
The handbook uses "GLM" throughout without defining it. Given that the model estimates per-link statistical sensitivity from observed data, it most likely stands for Generalised Linear Model, but this should be confirmed with the authors before publication.
What the model is used for
Establishing a baseline
The assessment of current conditions establishes a baseline of network vulnerability, and demonstrates the value of related DARe research.
This baseline is particularly relevant to the WP3 work on infrastructure ageing and materials, because the links identified as least resilient are also those where physical deterioration and climate exposure are most likely to interact.
Why that overlap matters
A link that already loses more speed per millimetre of rain than its neighbours is telling you something about its drainage, geometry or surface condition.
Those are the same characteristics that govern how it ages. The least resilient links and the fastest deteriorating links are likely to be the same links, which makes this output a prioritisation shortlist for two work packages at once.
Projecting forward
When applied to projected future demand and to extreme climate rainfall scenarios for 2030, 2045 and 2050, the model estimates how transport conditions across the network are expected to change.
These projections serve as inputs to:
- WP3 risk analysis
- Transport demand projection
- Road resilience assessment
They also support the construction of a combined risk layer that can inform policy advice to local authorities.
Testing interventions
The model enables preliminary policy impact simulation, through which the effect of alternative interventions on transport conditions can be examined under each future scenario.
Where these outputs go
These outputs collectively contribute to the construction of the DSP, the Decision Support Process, which is documented in the companion handbook on this site.
This is one of the clearest points at which the IMP hands evidence to the decision-making side of the framework.
What this chapter covers
| Page | What you will find |
|---|---|
| GLM Resilience Model | The model, its scope and its framework stages |
| Inputs | The four datasets it requires |
| Outputs | Resilience coefficients and projected speeds |
| Further reading | Data sources and methodological literature |