Impacts, Warning and Forecasting in 2005: Lessons for Flood Risk Management

Overview

The January 2005 floods in Cumbria were a major flood event with severe effects on towns, transport, businesses and public services. The Cumbria Floods Technical Report: January 2005 Flooding (2006) shows that the event combined exceptional rainfall, high river levels and widespread flooding across several catchments. In Carlisle, the report states that the 2005 flood at Eden Bridge appeared to be the highest recorded and was more than 1 metre higher than the March 1968 flood. This underlines the scale of the event and why it became an important case for improving flood risk management.

1. Impacts of the 2005 floods

Property impacts

The floods caused extensive damage to homes, businesses and community properties. The report gives examples from different settlements to show both the number of properties affected and the variety of flood mechanisms.

• Cockermouth: 149 properties were affected. According to the Cumbria Floods Technical Report: January 2005 Flooding (2006), about 112 were flooded by the River Derwent and 37 by the River Cocker.
• Rickerby: 14 properties were affected by the River Eden.
• Roe Beck: 17 properties were affected by the River Roe and overland flow, including flooding in the new development at Thistlewood.
• Bullgill: a number of properties flooded on the right bank, and the basement of Main Band House also flooded.

These examples show that impacts were not limited to one town or one river type. Flooding came from main rivers, tributaries, overtopping, overland flow, rising water tables and drainage problems. The report also notes that in Cockermouth the overtopping of existing defences by the Derwent was the main source of flooding, showing that even defended places remained vulnerable when the event exceeded design expectations.

Infrastructure impacts

Infrastructure damage was serious and disrupted movement, emergency response and daily life. The Cumbria Floods Technical Report: January 2005 Flooding (2006) records that in Bullgill the railway was undermined. The report also lists major transport-related losses, including damage and disruption affecting Network Rail and the Stagecoach bus garage.

• Rail: Network Rail recorded losses of £2,500,000.
• Bus operations: the Stagecoach bus garage suffered about £3,000,000 in direct losses and £150,000 in additional losses.
• Emergency infrastructure: Carlisle Fire Station suffered losses including ICT, telecommunications, vehicles, an ATV, furniture and lockers.
• Public services: police staff had to be relocated to Penrith, creating extra travel costs and pressure on resources.

These impacts demonstrate that flood consequences extend far beyond water entering buildings. Damage to transport links, depots, stations and communication systems weakens the operation of the wider region.

Economic impacts

The event had large direct and indirect economic costs. The report provides examples of repair costs, replacement costs, business interruption and longer-term revenue loss.

• Stagecoach bus garage: 85 vehicles were written off. The report also records clean-up, repair and stock losses, replacement of plant and equipment, relocation costs, and loss of revenue from service interruption.
• Network Rail: £2,500,000 of losses linked to cancelled, delayed or incomplete train journeys.
• Businesses: the report notes flooding at Jennings Brewery in Cockermouth, showing impacts on local employers and production.

Economic damage therefore included both physical damage and business disruption. This is important for flood risk management because recovery depends not only on rebuilding structures, but also on restoring transport, services, employment and commercial confidence.

Social impacts

The social impacts can be inferred from the widespread property flooding, disruption to emergency services, relocation of staff, and effects on transport and communications. The report notes the presence of police, fire and rescue, and City Council staff during the event in urban areas, indicating significant pressure on emergency response. Where homes flood, residents face displacement, stress, loss of belongings and interruption to normal life. Damage to transport services and public facilities also reduces access to work, school and support networks.

The scale of flooding in Carlisle and other settlements suggests that social impacts were extensive because they affected households, workers, travellers and service users at the same time. In flood risk management terms, this shows that social vulnerability must be considered alongside hydrology and engineering.

2. Performance of flood warnings in 2005

The source material shows that weather forecasting improved in the days before the flood. In particular, the forecast issued on Wednesday 5 January for Friday 7 January was described in the Cumbria Floods Technical Report: January 2005 Flooding (2006) as “an excellent forecast.” It warned of southwest gales and “rain, heavy and persistent, especially on southwest slopes of the Lake District with some substantial totals likely (120mm at Honister).”

This indicates that the meteorological warning picture was strong by this stage. Forecasts had identified:

• the likelihood of persistent heavy rainfall,
• the importance of the south-westerly airflow,
• the exposure of Lake District slopes, and
• the potential for very large rainfall totals.

However, the report also notes that some forecasts suggested the rain would clear faster than it actually did. For Saturday 8 January, the forecast “Windy with blustery showers” hinted at quicker clearance of the cold frontal rain than occurred in reality. This matters because the timing and duration of rainfall are critical for river response, especially in steep and already saturated catchments.

For flood warnings themselves, the report's statistics show that about 50% of properties in flood warning areas were not registered to receive warnings from the Environment Agency's AVM system. Surface water flooding also began at a number of locations before any warnings were issued. Since 2005, AVM has been replaced nationally by Floodline Warnings Direct, and registration for telephone warnings has risen to up to 75% across Cumbria.

Overall, the warning performance appears mixed but improving:

• Strength: heavy rainfall risk was recognised in advance.
• Limitation: the persistence and exact timing of rainfall clearance were less accurate.
• Implication: even good rainfall forecasts may still leave uncertainty in flood severity and timing.

3. Performance of real-time forecasting models

The Cumbria Floods Technical Report: January 2005 Flooding (2006) explains that the Environment Agency used four types of flood forecasting model of increasing complexity, and its conclusions assess how each performed in January 2005.

What the report concludes about each model type

• Simple trigger models performed well during the event, but they cannot provide forecasts with a long lead time or predict flood peaks.
• Correlation models, which relate an upstream level to a downstream prediction, under-predicted flood levels on the Eden and in Keswick, because flood levels were much higher than those used to calibrate them. Changes were subsequently made to five such models, and a quadratic regression often fitted better than the previous linear one.
• Transfer function models: only the Kendal (Victoria Bridge) model performed well. The Keswick (Low Briery) model did not, and a version calibrated on the 2005 flood over-predicted all other events. The Cummersdale model performed poorly mainly because of the quality of the rainfall radar.
• The ISIS real-time hydraulic model for the Eden at Carlisle, one of the first of its type implemented by the Environment Agency, performed well early in the flood but did not accurately predict the peak, largely because input flows were under-predicted by the then-current stage-discharge relationships at key upstream gauging stations.

Assessment of model performance

The report draws the general lesson that the nature of such extreme floods can be different from more frequent ones, so models calibrated on ordinary events may fail when they are most needed. Simple models were reliable but offered little lead time, while more sophisticated models were only as good as their inputs (radar rainfall, rating curves).

The main lesson supported by the source is that forecasting systems must cope with uncertainty in rainfall timing, persistence and catchment response. A flood risk management strategy should therefore avoid relying on a single model approach and should connect forecasting with warnings, emergency planning and robust defence design.

4. Lessons for flood risk management

The 2005 flood showed several important flood risk management lessons, all supported by the evidence in the Cumbria Floods Technical Report: January 2005 Flooding (2006).

• Extreme events can exceed past experience: Carlisle’s 2005 level was higher than earlier recorded floods.
• Defences reduce but do not remove risk: overtopping occurred in defended locations such as Cockermouth.
• Flood sources are multiple: rivers, overland flow, drainage and groundwater-related effects all contributed.
• Infrastructure vulnerability matters: rail, bus, fire, police and communications systems suffered disruption and loss.
• Warnings need accuracy in timing as well as severity: identifying heavy rain is valuable, but errors in duration and clearance timing affect decisions.
• Economic and social recovery are central: impacts continued after floodwaters fell because transport, businesses and services remained disrupted.

5. Flood risk management improvements made in Cumbria after 2005

Section 7 of the Cumbria Floods Technical Report: January 2005 Flooding (2006) summarises the work carried out after the flood:

• Flood warning: existing Flood Warning Areas were improved and new ones added (for example Armathwaite, Bolton, Eden Hall, Harraby Green, Low Crosby, Rickerby and Warwick Bridge).
• Flood forecasting: further simple trigger forecasts were implemented or improved (for example Appleby/Bolton, Eamont Bridge Farm, Cockermouth, Low Briery and Victoria Bridge), correlations were reviewed, and the ISIS Eden model was improved (ungauged inflows, the floodplain upstream of Carlisle and stage-discharge equations).
• Completed works: piling and raising of flood banks in Carlisle, a new culvert and debris screen at Watson Terrace in Penrith, gravel removal and a new culvert inlet at Cuddy Beck in Keswick, new flood gates in Appleby, and improved Mintsfeet embankments in Kendal.
• Ongoing schemes: the Eden and Petteril scheme for Warwick Road in Carlisle was under construction, with the Caldew and City Centre scheme and several others under investigation.

The report also notes that the flood brought forward planned improvements and provided a wealth of new information for improving models, forecasting and warning. Catchment Flood Management Plans (CFMPs), defined in the report's glossary as large-scale planning documents identifying long-term sustainable policies for managing flood risk across a catchment, support this more strategic, evidence-based approach, which is more aware of residual risk.

Conclusion

The January 2005 Cumbria floods caused major property, infrastructure, economic and social damage. The Cumbria Floods Technical Report: January 2005 Flooding (2006) shows that the event was exceptional in scale, with very high rainfall, record or near-record river levels, overtopping of defences and widespread disruption. Flood warnings benefited from increasingly good meteorological forecasts, especially by 5 January, but uncertainty remained in the exact timing and persistence of rainfall. Among the forecasting models, simple trigger models and the Kendal transfer function model performed well, while correlation models, the other transfer function models and the ISIS model under-predicted, showing the need for robust, multi-layered forecasting and response. The key flood risk management lesson after 2005 was the need for catchment-based planning, better understanding of residual risk, and stronger links between forecasting, warning, defence and recovery.