Purificación de fragmentos de ADN.

2.6.1

General Flood Risk & Joint Probability Investigations

The River Ouse catchment in East Sussex, UK, suffered severe flooding on the 12th October 2000. Environment Agency (2001a) modelled river flow and sea level in the Ouse catchment as part of the Section 105 study, using flood hydrographs and standard Flood Estimation Handbook rainfall-runoff methods (see Robson and Reed, 1999). No attempt was made to analyse the relationship or joint probability between extreme fluvial flows and high tides or surges.

Following the 12th October 2000 flood in Lewes and Uckfield, Environment Agency (2001c) produced a detailed account of the event. The interaction of sea level and river flow was analysed for the immediate period before and during the peak of the flood. The

causes of the flood were investigated, concluding that the event was primarily fluvially driven, caused by three successive extreme rainfall events in the upper catchment. No probability analysis or long-term flood prediction calculations were undertaken during the study.

Environment Agency (2002, 2004) produced a detailed River Ouse flood management strategy, comprising of probabilities and an iSIS hydraulic flood model. It comments that flooding at Lewes is a complex problem due to the interaction of sea level and river flow, offstream floodplain storage and narrow topography and river channel (including Cliffe Bridge) through the centre of the town. It was concluded that sea level (including surge) alone could not cause flooding in Lewes with the existing defence levels, although the interaction of flow and sea level was not discussed. The impact of fluvial flows was therefore the main focus of the study, and as such, no joint probability analysis was undertaken between sea level and river flow, with sea level was taken as a constant during the modelling exercise.

Environment Agency (2004) commented that sea level rise associated with climate change will increase the importance of extreme sea levels for flood levels in Lewes, particularly downstream of Cliffe Bridge. A brief analysis of the potential impact suggested any increase in the predicted extreme sea levels at Newhaven would require a joint probability analysis of the combined impact of sea level and fluvial events.

MacDonald (2004) derived flood estimates for the Uck sub-catchment (in the upper Ouse catchment), also using standard methods recommended in the Flood Estimation

Handbook. The results updated Environment Agency (2002, 2004), and disseminated general recommendations for extreme river flow analyses. No specific impacts or implications were noted for flood risk calculation in the Ouse catchment.

A recent scoping report for the Ouse catchment flood management plan (Environment Agency, 2006) highlighted the current flood risk in the catchment, using findings from the flood risk management strategy. There was no discussion on the interaction of sea level and river flow at Lewes.

2.6.2

Dependence Studies

The Defra R&D funded joint probability project (e.g. Svensson and Jones, 2003; Hawkes, 2004) calculated dependence values for the Ouse catchment area. Dependence

between two tidal level gauges using a single variable of surge was investigated. The results indicated where surges could occur simultaneously along different stretches of the south coast of England at various tidal gauges, including Newhaven.

Svensson and Jones (2003) calculated that for surge, dependence between the same- variable pairs was weaker in the eastern half of the south coast of England than in the western half. From west to east, similarly distanced station pairs showed decreasing dependence;χ =0.42 for Newlyn and Weymouth,χ =0.25 for Weymouth and

Newhaven andχ =0.08 for Portsmouth and Dover. It was suggested that the decreasing dependence values may be related to the incursion of North Sea surges into the English Channel from the east.

Svensson and Jones (2002, 2004a) also investigated dependence between daily maxima surge and daily mean river flow for station pairs around the UK, including the tidal reach of the River Ouse. Three UK regions displayed significant surge and river flow

dependence which generally exceededχ = 0.1 as the western part of the English south coast, southern Wales, and around the Solway Firth.

Svensson and Jones (2004a) calculated dependence between pairs of daily maximum surge and river flow gauges on the southern coast of Britain. Dependence was often found to be strongest when surge and flow occur on the same day in catchments along the south coast. Dependence between Barcombe Mills river flow and Newhaven surge in the Ouse catchment was calculated asχ =0.05 at the 5% significance level. Higher dependence (i.e.χ >0.1) was generally found in hilly catchments with a southerly to westerly aspect.

It was suggested by the authors that this low value of dependence in the Ouse may be related to the catchments along this part of the coast comprising of a generally permeable (predominantly chalk) underlying geology, which respond slowly to rainfall, and

therefore runoff (and subsequent high river flow) may not form on the same day as a surge occurs. Environment Agency (2002) however categorises the Ouse catchment as being ‘quickly responding’ and ‘flashy’ in nature, raising some doubts about this conclusion.

Seasonal and time-lagged (in days) dependence calculations were carried out for the three UK regions where significance was found to be generally high, but they did not

cover the Ouse catchment as dependence was not found to be significant here for the non-lagged analyses.

With the eastern part of the south coast of England producing low dependence between surge and high river flow, Svensson and Jones (2003) analysed dependence between surge and precipitation in an attempt to avoid any interference from any catchment processes and topography. Precipitation data was used from the Wye precipitation gauge, 40 miles to the east of the Ouse catchment and paired with Newhaven surge. Although the report only draws general conclusions, it found that on the south coast of England, dependence between precipitation and surge was widespread, including a significant level of dependence for the Ouse catchment, although theχ value was not given. Dependence was found to be strongest when high surge and precipitation occur on the same day, but also remains strong for when river flows are lagged one day after the surge. The authors proposed that this was confirmation that the lack of dependence between surges and high river flows was related to catchment processes rather than any other factors.

Svensson and Jones (2002, 2003, 2004a) all utilised the Barcombe Mills river flow dataset for the dependence analyses in the Ouse catchment area, which has been

confirmed as being of poor data quality and reliability (e.g. Environment Agency, 2001b, 2002 and pers comm.). The use of the Barcombe Mills dataset may have had a significant impact on the accurate determination of dependence values, and may explain the

differing findings from the dependence analyses between surges and high river flows, against surges and precipitation.