Thesis
Cold-dry and cold-wet compound events in the United Kingdom : hazard characterisation, and impact-based forecasting for the Scottish rail network
- Creator
- Rights statement
- Awarding institution
- University of Strathclyde
- Date of award
- 2026
- Thesis identifier
- T18195
- Person Identifier (Local)
- 202274537
- Qualification Level
- Qualification Name
- Department, School or Faculty
- Abstract
- Compound events arise from the interaction of multiple climate hazards and/or drivers, often resulting in impacts that exceed those of individual extremes and driving some of the most severe disruptions to society and infrastructure. While compound events have in recent years become a focal point in climate science, research has disproportionately emphasised heat-related hazards relative to cold weather events. Compound cold events remain comparatively underexplored despite their capacity to generate significant impacts across society and critical infrastructure. Even as climate projections indicate a decline in extreme low temperatures, compound cold events represent a critical low-frequency high-impact hazard that warrants systematic investigation. This thesis addresses this gap by developing an end-to-end framework for understanding, detecting, and forecasting compound cold events in the United Kingdom (UK), demonstrated here through a case study of the Scottish rail network. The research defines two types of compound cold events based on co-occurring meteorological conditions: cold-dry (CD) events, characterised by locally low temperatures in the absence of heavy precipitation, and cold-wet (CW) events, where low temperatures coincide with heavy precipitation. This research first establishes the climatological characteristics of compound cold events in the UK using location-specific percentile thresholds applied to gridded observations. A clear spatial pattern emerges: CD events occur more frequently in the west of the UK (up to 14 days per winter) while CW events are more common in the east of the country (2-3 days per winter). By linking compound cold events to a set of 30 weather patterns, this thesis identifies the large-scale atmospheric drivers responsible for the observed regional differences. CD events are predominantly associated with anticyclonic patterns such as those associated with the negative phase of the North Atlantic Oscillation, whereas CW events are driven by transient conditions associated with cyclonic patterns. The research then examines how these compound cold events translate into real-world consequences. Analysing over 10,500 Scottish rail incidents spanning the winters of 2006 to 2023, the thesis shows that high-impact days (i.e., those causing extreme service disruption or customer delays) are strongly associated with compound cold conditions. Critically, many severe disruption incidents occurred on days without heavy snowfall, highlighting the role of snow persistence and event sequencing. When a CW event deposits snow followed by a CD event with sub-zero temperatures, hazardous conditions can persist for multiple days and prolong periods of disruption. This demonstrates that impact severity does not always align with meteorological extremes alone and highlights the importance of adopting an impact-led approach when identifying hazardous weather risks. The final stage of this research builds on these insights to develop a tiered operational forecasting framework that bridges the gap between atmospheric predictability and infrastructure decision-making. The framework provides early indication of heightened risk at extended lead times using a set of “high-risk” weather patterns and dynamically evolving spatial probabilities of CD and CW events. Within this framework, weather patterns serve as a crucial link connecting large-scale atmospheric circulation, compound hazard occurrence, and ultimately infrastructure impacts. This enables predictability to propagate from synoptic-scale patterns through to sector-specific risk information. The framework’s performance demonstrates that actionable forecasts of compound cold events are achievable at operationally relevant lead times. Weather pattern forecasts successfully identify windows of elevated risk up to 10 days in advance of major disruption events, while the spatial probability forecasts capture the location-specific nature of CD and CW hazards across Scotland’s rail network. This thesis makes three principal contributions. First, it establishes weather patterns as effective diagnostic and prognostic tools for compound cold events, revealing systematic relationships between synoptic-scale circulation and regional hazard occurrence. Second, it demonstrates that compound cold events are the primary driver of high-impact disruption days to the Scottish rail sector, and that the sequencing of these events is critical to understanding how impacts unfold. Third, it provides a methodological pathway for operationalising compound event science within impact-based forecasting systems, with transferability beyond the rail sector to other weather-sensitive infrastructure and services. As climate variability continues to challenge infrastructure resilience, such frameworks offer a template for developing multi-hazard early warning systems that anticipate not just what weather will occur, but what consequences it will bring.
- Advisor / supervisor
- White, Christopher J.
- Clarke, J. A. (Joe A.)
- Bloomfield, Hannah
- Resource Type
- DOI
Relations
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
PDF of thesis T18195 | 2026-10-07 | Public | Download |