The Scottish Government

Contents |

Listen

APPENDIX B: HYDROLOGY

Catchment Description and Flow Estimation Locations

Each catchment area was identified initially using the FEHCDROM. The FEHCDROM boundaries were checked against OS maps and sometimes after site visits. Where catchments were too small to be located on the CDROM, these were estimated manually.

Each catchment was checked for unusual factors including:-

• high permeability.
• high urbanisation.
• pumped watercourse.
• major lochs, lakes or reservoirs.

Adjustments were then made to the catchment descriptors as appropriate, following methodologies outlined in the FEH.

Baseflow Index

Following SEPA recommendations ²⁸, the Base Flow Index ( BFI) Map of Scotland was also consulted. Where the value for BFI on the paper map differed from that of BFIHOST on the FEHCDROM, then a process of adjustment was followed. BFIHOST was replaced by the mid range BFI extracted from the paper map and the SPR was recalculated using the following equations.

Figure B- 1: Comparison of BFIHOST and BFI

BFIHOST , BFI, SPRHOST and SPR comparison
(note: negative difference equates to BFI or SPR being lower than BFIHOST or SPRHOST)

Out of 20 sites where BFI was assessed, 5 sites saw an increase in BFI, 10 experienced a decrease in BFI and 5 experienced no change.

12.231 Urbanisation

Flood estimates have been calculated for a base date of 2004. The urban extent parameter, URBEXT, on the CD- ROM dates from 1990 and has been updated using OS plans (from which a parameter URBAN has been estimated) and the suggested FEH methodology ²⁹:

URBEXT = URBAN/2.05
where URBAN = urban area/catchment area from OS maps.

It should be noted that the representation of urbanisation in the FEH methods is that URBEXT is used to estimate the net effect of urbanisation on runoff based on historical data. It does not distinguish between catchments with traditional drainage systems and those with restricted runoff as part of Sustainable Urban Drainage ( SUDS) infrastructure.

The adjusted URBEXT2004 values have been used in the flood estimation calculations.

Historical Flooding and Sources of Flood Peak Data

A review of historical flooding and flood peak data was carried out for each of the flood prevention schemes. Data availability varied for each of the sites, but sources of historical data throughout this study included:

• HiFlows- UK - UK website ³⁰
• BHS Hydrochronology ³¹
• Scheme Reports.
• Information supplied by local authorities.
• Alchemy - Database containing scanned CEH gauging station records.

Analogue/ Donor Catchment Analysis

It is good practice not to rely solely on the catchment descriptors for flood flow estimation and it is recommended by the FEH that data from donor or analogue catchments be used to refine these estimates. Donors and analogues were sought initially within a 40 km radius of the flow estimation point. This primarily involved accessing data from the HiFlows- UK website, but in cases where there were no suitable HiFlows- UK gauges within 40km of the site of interest, data from the SEPA gauging network of sites not contained within HiFlows- UK were considered. Annual Maxima data was requested nationally from SEPA for 152 gauging stations. Data for 73 sites was received, the remaining 79 sites were deemed unsuitable for use in peak flow analysis by SEPA.

Gauges within the HiFlows- UK dataset are defined as being either suitable for use during QMED (where QMED is the median flood) adjustment, suitable for use within the pooling group or suitable for both QMED adjustment and pooling.

• A donor catchment is a local catchment with gauged data that are particularly relevant to the subject site. It may be on the same river or on an adjacent and hydrologically similar tributary.
• An analogue catchment is a more distant gauged catchment that is hydrologically similar to the subject site.

The assessment of the hydrological similarity of potential donor or analogue sites was based on analysis of a number of key catchment characteristics (see Table B- 1).

Table B- 1: Donor / analogue catchment descriptors

Adjustment ratios for QMED from catchment descriptors and QMED from flow data were calculated.

FEH Statistical Method

There are two parts to the FEH statistical method: the estimation of an index flood ( QMED) and the estimation of flood growth curves.

QMED was calculated based on catchment descriptors with data transfer from the nearby gauging stations acting as donor or analogues.

Pooling Group

Initial flood growth curves were derived using the WINFAP- FEH software.

The WINFAP- FEH software package was used to derive flood frequency distributions for the downstream limit of each FPS. WINFAP- FEH contains a database of flood data for catchments from across the UK. All FEH Statistical calculations were carried out using the UK HiFlows- UK dataset (up to October 2003). Sites are selected for the pooling group on their hydrological similarity in terms of AREA, BFIHOST and SAAR. A 100 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 to the default pooling groups were made including the inclusion of appropriate HiFlows- UK gauges within 40km of the site in line with recommendations in The Flood Estimation Handbook Statistical Procedures: Automation, Appraisal and Future Development³². This procedure includes the use of geographical proximity measure to adjust and move sites to a new rank within the pooling group.

As long as there were no major changes in catchment characteristics between upstream and downstream limits of the scheme, it was deemed appropriate to use the same pooling group and hence same growth curves as derived for the downstream limits for the upstream limits.

The WINFAP- FEH software was used to determine the suitability of growth curve fittings through the 'goodness-of-fit' test.

Out of the 36 flood prevention schemes analysed, 8 are on rivers with gauging stations. In these cases the growth curves derived using the pooling group were compared to the single site growth curve derived for the individual gauging stations.

Table B- 2: Gauging stations analysed in detail

Variation from the Main Methodology

Galashiels - Incorporation of Historical Data into Flood Estimation for this site

The review of non-gauged historical flood data can lead to a better understanding of the processes that lead to extreme flood events. The use of historical data in flood frequency estimation³³ describes the extension of gauged flow records that exist for the river of interest.

A search of historical data showed the occurrence of at least 5 significant flood events prior to the construction of the flood defence schemes at Galashiels, in 1881, 1891, 1931, 1948 and 1984. The latter being the greatest recorded flood of 195 m ³/s. The 4 earlier flood events pre-date the gauging station record. These 4 historical events have been deemed to be of similar magnitude to that of the 1984 event ³⁴ and hence have been included in the historical analysis.

Using a graphical method, data can be spilt into that which has been recorded (systemic data) and that which is pre-data record (historical data). Using a special plotting position to incorporate the historical data, the historical data is then included in the flood frequency distribution.

On inspecting the flood frequency curves (Figure B- 2) for the three separate methods detailed above, it can be seen that the pooled flood frequency curves result in much lower flood flows at the higher return periods whereas the single site GL- LOM distribution compares favourably with the 'Historical' distribution. Hence the final distribution selected for the estimation of peak flows at the Netherdale and Plumtree FPS's is that of the single site GL- LOM distribution.

Figure B- 2: Flood frequency distributions

Heavily Reservoired Catchments

Two schemes studied lay within catchments heavily influenced by hydroelectric schemes - Conon Bridge (Conon Valley Hydroelectric Scheme) and Fort Augustus (Garry-Moriston Hydroelectric Scheme). As such, both catchments contained a number of storage reservoirs.

Both catchments have gauging stations downstream of the hydro schemes: River Garry @ Craigard (06010) and River Conon at Moy Bridge (4001). In each case, these were used as donor gauges for flow estimation, being taken as reasonably representing the influence of the hydro-electric schemes on the flow regime.

As well as the influence of the raised reservoirs on the natural flow regime, the Conon Valley scheme involves a complex system of aqueducts and tunnels providing flow transfers between the catchments for power generation at the 6 power stations within the scheme. The scheme exerts a considerable effect on flows in the valley downstream and the straightforward application of the FEH statistical method to provide flow estimates is inappropriate, as this would not take into account the effect of the hydro scheme operating regime.

The FEH statistical methodology was partially adopted by estimating the flood frequency curve at Moy Bridge, but deviating from the methodology by using a single site growth curve. The disadvantage of this approach is the uncertainty involved in extrapolating a relatively short period of record to obtain higher return period events. However, it is considered that a pooled growth curve that includes other gauged sites not influenced by the same or indeed any hydro scheme would result in more inaccurate estimates. The final method adopted at Conon Bridge can be summarised as follows:

i) QMED calculation from period of record at Moy Bridge
ii) Growth curve prediction from single site analysis at Moy Bridge
iii) Flood frequency estimates Moy Bridge by combining QMED and growth curve
iv) Deduction of 9.91 m ³/s from predicted flows to account for transfer from Orrin catchment
v) Calculation of area specific runoff (m ³/s/km ²) from catchment to Moy Bridge
vi) Calculation of equivalent runoff from catchment area between Moy Bridge, Orrin Reservoir and Conon Bridge.
vii) Addition of Orrin compensation flow of 1.05 m ³/s
viii) Total flow at Conon Bridge is sum of iii) + vi) + vii)

The period of record at the Garry gauge was insufficient to use annual maximum data or a single site growth curve. In this case, flow estimates for Fort Augustus were estimated by adjusting QMED using POT analysis on the Garry gauge, and constructing a pooling group with a bias to low FARL values.

FEH Rainfall Runoff Method

In some cases flows were calculated using the FEH Rainfall-Runoff method. The Rainfall-Runoff method involves constructing a simple unit hydrograph and losses model of the catchment, with three parameters: Time to Peak of the Unit Hydrograph (Tp) relating to the catchment response to rainfall, Standard Percentage Runoff ( SPR) relating to the proportion of rainfall which directly contributes to flow in the river and baseflow being the quantity of flow in the river prior to the event. These can be best estimated when there are rainfall and river level or flow data for a number of flood events. In the case of ungauged catchments being assessed during this study these parameters have been estimated using the catchment descriptors derived from the FEHCDROM.

The JBA Flood Estimation Software ( FES) was used to generate flood hydrographs and peak flows for the range of annual probabilities.

Small Catchment Hydrology

For a number of schemes, the catchment areas were too small for the FEH methods to apply. Therefore a variety of methods recommended for small catchments were used to estimate flows, and these are detailed below. The method giving what was considered the best estimate (usually the most conservative) was then chosen.

ADAS 345 Method³⁵

The ADAS 345 method was developed for determining the pipe sizes required for field drainage pipes. It is appropriate for use catchments that are predominantly agricultural. The calculations establish a value that is usually taken to equate to the 2 year ( QMED) greenfield runoff rate. The relevant Flood Studies Report regional growth curves are then applied to estimate flows for the range of return periods required for the study.

Institute of Hydrology Report 124 Method

The Institute of Hydrology Report 124 ( IH124) ³⁶ was developed to estimate floods on small ungauged catchments (< 25 km ²). The method uses three parameters: AREA, SAAR and SOIL to determine QBARrural. The area is measured from a 1:25,000 Ordnance Survey map, where it is too small to be identified by the FEHCD- ROM, SOIL is extracted from the maps contained within volume 5 of the Flood Studies Report 1975 ( FSR1975), and SAAR from the FEHCD- ROM.

Rational Method

The Rational Method is a calculation which relies on estimating the average rainfall intensity over the catchment for a given return period and time of concentration. The time of concentration represents the time from the start of the rainfall event until the entire catchment is contributing to estimated flow.

Three different time of concentration estimates were trialled: the Kirpich formula, the Hathaway formula and the Bransby-Williams formula. The value chosen was the shortest of those obtained by these methods, resulting in the highest flows, and the "worst-case" scenario for the catchment.

The average rainfall intensity for the catchment was calculated using the Depth Duration Frequency module of the FEHCD- ROM and the flow estimates were calculated for the range of return periods required.

Contents |