InfoWorks RS underpins forensic analysis of the Boscastle flash flood

The picturesque small town of Boscastle, on the North Cornwall coast, became a dramatic addition to the UK’s history of extreme flood events on 16 August 2004. On that day, an exceptionally prolonged and intense storm centered on Otterham, near the headwaters of Bodmin Moor, caused levels in its sea-bound exit routes - the river Valency, the Crackington stream, the river Neet and the river Ottery - to rise substantially over a very short period of time.

Unfortunately, because the worst of the storm was concentrated over the headwaters, the townsfolk of Boscastle were unaware of the impending disaster until the normally peaceful river Valency began to rise with startling swiftness. Mercifully no-one was killed, although the dramatic news footage of the village under the full, ferocious onslaught of a massive cascade of water remains one of the most vivid storm images of recent years. Immediately, the obvious questions were posed: What happened? Why there? Could it happen again, there or elsewhere? Was climate change to blame?

The answers were provided in an Environment Agency-commissioned study, through painstaking and inspired detective work from a group of experts led by HR Wallingford - Wallingford Software’s sister company - with Wallingford Software’s InfoWorks RS software solution playing a key role in modeling what turned out to be a fascinating and complex scenario.

The group of experts

HRW was given the role of leading the work and deciding who should be involved in the project. The meteorology of the event was analyzed by the UK’s Meteorological Office; the hydrology by CEH Wallingford - and the hydraulics and geomorphology by HR Wallingford, using InfoWorks RS. Consultants Halcrow and Royal Haskoning undertook post-flood surveys in the Valency/Jordan and Crackington stream catchments respectively, and they and the Environment Agency assembled witness evidence from local interviews.

Certain geographical facts turned out to be key: the catchment area of the river Valency above Boscastle is approximately 20km2, rising to around 300m above ordnance datum (AOD), and the main branch of the river Valency is some 7km long. This means the slope of the river is steep, as is that of a number of it tributaries, some of which are incised into narrow pathways of their own as they approach the main channel. The soils in the area are generally thin over impermeable bedrock; the catchment is predominantly rural and mainly grassland, though there are substantial woods next to the main river and its tributaries.

The event itself was, fortunately, one of the best-documented flood events in the UK. It happened during the day and in the presence of many people, which meant there was an extensive photographic record. The mass of evidence suggested that the river overtopped its bank for around five hours, rising to a peak (from bank-full) in just 1.5 hours. As the floods rose, some witnesses reported extremely rapid short-term water level rises of between 1m and 15m, described as ‘walls of water’, over periods of a minute or less.

Early data

Within two weeks of the event, HR Wallingford had made initial estimates of flood extent and hydraulic roughness from surveys at key locations around the catchment, and Halcrow had surveyed cross-sections of the river and flood plain at key locations, and noted wrack-mark (flood debris) levels along the main rivers. The consultant also made estimates of damage to property and infrastructure in Boscastle within days of the event. From such initial data, the first estimates of the peak Boscastle flood flows were estimated to be in excess of 150m3/sec.

The Met Office data revealed that the extreme rainfall resulted from a sequence of convective storms, the last remnants of Hurricane Alex, that were channeled along the north Cornish coast over several hours.

The team´s Project Director, Dr Colin Fenn of HR Wallingford, explains: “Radar and satellite imagery suggests that a succession of thunderstorms developed one after another

Dr Fenn notes: ‘Factors like the time to peak of the hydrograph, the percentage of the rainfall running off to rivers, the hydraulic roughness of the river channel and the degree of blocking of river bridges were varied in order to match the size and shape of the hydrograph to eyewitness reports.’

In the event, it proved necessary to vary both the standard FEH parameters (the accepted best UK methodology for flood flow estimation is provided in the Institute of Hydrology’s Flood Estimation Handbook) to produce flows of sufficient magnitude, and to model significant blockage of the bridges in order to match the water level predictions from the model to the observed profile of peak water levels.


It is clear that there were many uncertainties associated with the modeling – for instance, it is difficult to reproduce the effect of the destruction of walls and the creation of other obstacles during the storm. However, the results are deemed sufficiently robust to provide valuable information on the magnitude of the event and the probability of its occurrence. The authors of the study hope that the problems experienced in reproducing such an extreme event will guide future model development and research.

The analysis, combined with the work undertaken by the other expert bodies involved in the reconstruction, made it clear that the 2004 flood event was a very extreme event. Its annual probability was estimated at the equivalent of a 1 in 400 year return period – thankfully, an annual probability of just 0.22%.

Customer comments

No comments were found for InfoWorks RS underpins forensic analysis of the Boscastle flash flood. Be the first to comment!