The Scottish Government

« Previous | Contents | Next »

Listen

Appendix C: - Case study

C.1 Menstrie case study

C.1.1 Background

The flood event on the 9 August 2004 caused flooding to approximately 30 properties in Menstrie. Menstrie lies at the foothills of the Ochills in the River Devon catchment and is bisected by Menstrie Burn. Upstream of Menstrie the Burn flows through a heavily wooded gorge with a steep channel slope. In Menstrie the burn gradient gradually slows as it enters the Devon valley. Currently three crossings exist over Menstrie Burn, the Ochill Road Bridge, the A91 road bridge and a vehicular access track. In the August 2004 flood event a footbridge upstream of Burnside Road was washed onto the access bridge immediately downstream of Burnside Road.

Flows in the burn were within channel in the upper part of Menstrie but caused a blockage on a footbridge and access road bridge at Burnside Road. The footbridge was washed away and debris caught on the access road. Flood water left the channel and caused flooding immediately above the footbridge and more substantially along Burnside Road and in the new development at Holly Grove, as shown in Figure C-1.

No records of flood events were found on the Chronology of British Hydrological Events ²¹ for either Menstrie Burn or the River Devon. However, residents who were flooded during this event recall that there had been 2 similar events during the past 20 years. Clackmannanshire Council developed a detailed hydraulic model of the Menstrie Burn and part of the Devon. Cross sections were taken shortly after the event and wrack marks of the flood event were surveyed to allow calibration of the model.

Figure C-1: Flood Extend Menstrie Burn August 2004

C.1.2 Local rainfall

The nearest rainfall station is at Glenochil on the River Devon. The FEHCD- ROM was used to determine the event rarity for both the 2 hour 15 minute and 20 hour durations for over both the Glenochil and Menstrie catchments. Table C-1 displays the results of assessing the event rarity of the recorded rainfall from the Glenochil station. The estimated return periods suggest that the 2.25 hour and 20 hour periods of rainfall could be placed at an approximate 25 year and 9 year return period (4% and 11.1% AEP) respectively.

Table C-1: Design Rainfall Estimates

C.1.3 Design Flood Estimates for Menstrie Burn

As Menstrie Burn is ungauged, initial estimates of QMED were made from FEH catchment descriptors for the site. The initial QMED estimate from catchment descriptors was 4.94 m3/s. As it is not recommended to rely on the catchment descriptors for the QMED estimate, analogue sites were used to adjust the QMED estimate.

A search of the FEH database identified one suitable HiFlow site within 40 km of Menstrie Burn to use as an analogue. Three other stations from the WINFAP- FEH pooling group were added as analogues to estimate the QMED value in Menstrie Burn as 6.85 m3/s. A flood growth curve for Menstrie Burn was then calculated using WINFAP- FEH and the UKHiFLOWs dataset. A 200 year target pooling group was established where the resulting total number of station years required in the pooling group is 5 times the target return period.

Changes were made to the default pooling group for Menstrie Burn and the frequency distribution to give the best fit to the pooled growth curve for Menstrie Burn was found to be the GL distribution. Details of the resulting growth factors and flood frequency curve are shown in Table C-2.

Table C-2: Design Flood Estimates Menstrie Burn

These flows were then modelled in a steady state hydraulic model and an attempt to calibrate the flows against the known flood levels was made. This proved difficult to achieve as the wrack marks were much higher than the modelled flow levels for a 100 year event. In areas around the bridge blockage the anomalies could be readily explained by reduction in channel capacity. On reaches unaffected by blockage the anomaly was not readily understood. Local residents also explained that they had suffered similar flooding twice before within the last 15-20 years. There are no definite dates for these events but if they were of a similar size it would suggest that the flood in August was no more than a 30 year event.

In order to accommodate the uncertainty in the QMED estimate for Menstrie Burn the FEH describes a method to estimate confidence intervals for QMED estimates made from catchment descriptors and adjustment from analogue sites. This is considered to be a reasonable approach to increase the flow estimates whilst still adhering to recommendations within the FEH. For Menstrie Burn the 68% confidence intervals for QMED are estimated to be 4.4 to 10.6 m3/s. Using the upper limit of QMED for data transfer, revised peak flow estimates have been derived.

Table C-3: Revised Flood Estimates

In order to determine the peak flow of the flood event that occurred in Menstrie Burn on the 9 August 2004, rainfall data of the event needs to be translated into run-off. The rainfall data for the event was obtained from the continuous raingauge located at the Glenochil gauging station. Figure C-2 shows the recorded rainfall at the station over a 20 hour period from 23:00 GMT 8 August till 19:00 GMT 9 August 2004. The recorded rainfall depth over the 20 hour period was 55.6 mm. The most intense burst of rainfall during the event occurred in a period of 2 hours and 15 minutes from 15:15 - 17:30 on the 9 August. During this period 28.2 mm of rain was recorded.

There are a number of problems in relating the recorded rainfall at the Glenochil station to generate an estimated flow for the flood event that took place in Menstrie Burn on the 9 August 2004. Firstly assessing the rarity of a storm is a subjective process which involves defining the start and end of the rainfall event and simplifying the spatial and temporal variation in rainfall that occurs over the catchment. The Glenochil station is outside the catchment for the Menstrie Burn and is situated in topography that is not typical of the Menstrie catchment. Therefore there is a degree of doubt over the suitability of using the Glenochil rainfall as representative of the rainfall that fell over the Menstrie catchment during the flood event.

The FEH module of ISIS was used to generate a hydrograph from the recorded Glenochil rainfall for the 20 hour duration. Catchment characteristics for the Menstrie catchment were imported from the FEHCD- ROM. Due to the lack of rainfall and flood data for Menstrie Burn the unit hydrograph was calculated from the FEH catchment characteristics and a time to peak (Tp) of 2.25 hours was calculated for the Menstrie catchment. The standard percentage run-off ( SPR) for the catchment should ideally be calculated from flood event data, however in this case it was again estimated from the FEH catchment descriptors as 44.8. The calculated hydrograph is shown in Figure C-2 and with a calculated flood peak of 14.66 m3/s at 18:45 GMT. The modelled peak of 14.66 m3/s is roughly equivalent to the estimated 100 year peak flow for Menstrie Burn using the initial statistical method adjusting the predictions and using the upper confidence limit revises the flow to one between a 5 and 10 year event.

When considering that the estimated return period for the rainfall event was calculated as 25 years (4% AEP) and three events of similar magnitude have occurred over less than 20 years the lower return seemed more reasonable. The rainfall run off model was tested for sensitivity of the SPR values by adjusting it by +/- 10%. The estimated flood peaks and there equivalent peak flow return periods from the statistical method are shown in Table C-4. It shows that the generated hydrograph is sensitive to the adjustment of the SPR value and that it can have a significant effect on the estimated peak flow and the estimated return period of the flood.

Table C-4: Sensitivity of Rainfall Run-off Model

Figure C-2: ISIS computed hydrograph of the flood event in Menstrie on the 9th August 2004

It was thus concluded that the August 2004 event on the Menstrie Burn was closer to a 10 year fluvial flood.

« Previous | Contents | Next »