Floods and Reservoir Safety [4 ed.] 9780727760340, 9780727757715, 9780727757692, 9780727760067

Citation preview

Floods and Reservoir Safety

Floods and Reservoir Safety Fourth edition

Institution of Civil Engineers

Published by ICE Publishing, One Great George Street, Westminster, London SW1P 3AA Full details of ICE Publishing sales representatives and distributors can be found at: www.icevirtuallibrary.com/printbooksales First edition published 1978 Second edition 1989 Third edition 1996 Reprinted 1998 This fourth edition 2015 Other titles by ICE Publishing: Maintaining the Safety of our Dams and Reservoirs (Proceedings of the 18th Biennial Conference of the British Dam Society). British Dam Society. ISBN 978-0-7277-6034-0 Prescribed Form of Record for a High-risk Reservoir, 2nd edition. Institution of Civil Engineers. ISBN 978-0-7277-5771-5 A Guide to the Reservoirs Act 1975, 2nd edition. Institution of Civil Engineers. ISBN 978-0-7277-5769-2 www.icevirtuallibrary.com A catalogue record for this book is available from the British Library ISBN 978-0-7277-6006-7 # Institution of Civil Engineers 2015 ICE Publishing is a division of Thomas Telford Ltd, a wholly-owned subsidiary of the Institution of Civil Engineers (ICE). All rights, including translation, reserved. Except as permitted by the Copyright, Designs and Patents Act 1988, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the publisher, ICE Publishing, One Great George Street, Westminster, London SW1P 3AA. This book is published on the understanding that the author is solely responsible for the statements made and opinions expressed in it and that its publication does not necessarily imply that such statements and/or opinions are or reflect the views or opinions of the publishers. While every effort has been made to ensure that the statements made and the opinions expressed in this publication provide a safe and accurate guide, no liability or responsibility can be accepted in this respect by the author or publishers. While every reasonable effort has been undertaken by the author and the publisher to acknowledge copyright on material reproduced, if there has been an oversight please contact the publisher and we will endeavour to correct this in a reprint. Cover image: Clywedog Reservoir, River Severn, Wales Commissioning Editor: Gavin Jamieson Production Editor: Richard Willis Market Development Executive: Elizabeth Hobson

Typeset by Academic + Technical, Bristol Printed and bound in Great Britain by TJ International Ltd, Padstow

Contents

Preface

vii

01 . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction Background Changes Scope of the guide References

1 1 1 2 2

02 . . . . . . . . . . . . . . . . . . . . . . . . . .

Floods and waves protection standards General The main factors Standards-based approach Risk-based approach Other criteria Other aspects References

3 3 4 4 8 9 11 13

03 . . . . . . . . . . . . . . . . . . . . . . . . . .

Derivation of reservoir flood inflow Objective Flood Studies Report and Flood Estimation Handbook Revitalised flood hydrograph rainfall–runoff method Clarification of appropriate models and design rainfall inputs Summary of the FSR/FEH rainfall–runoff method for the T-year event Summary of the methodology for estimating the PMF event Summary of the rainfall DDF and the rainfall–runoff models Use of local data The contribution of extreme snowmelt Climate change Rapid method of assessment based on FSR procedures Software References

15 15 15 16 16 16 17 17 17 17 19 19 20 20

04 . . . . . . . . . . . . . . . . . . . . . . . . . .

Reservoir flood routing Objective Recommended stages in routing calculation Gated spillways Siphon spillways, automatic toppling weirs and fuse plugs Auxiliary spillways Temporary upstream storage References

21 21 21 23 23 24 24 24

05 . . . . . . . . . . . . . . . . . . . . . . . . . .

Waves, wave overtopping and dam freeboard Scope Water surface and fetches Wind speed, duration and direction Wave height and period prediction Wave overtopping References

25 25 25 25 28 30 33

06 . . . . . . . . . . . . . . . . . . . . . . . . . .

The overflowing and overtopping of embankment dams Introduction Classification Physical factors affecting erodibility Overflowing: assessment Overflowing: remedial measures Overtopping: assessment Overtopping: remedial measures References

35 35 35 35 36 38 40 41 41 v

07 . . . . . . . . . . . . . . . . . . . . . . . . . .

Floods during dam construction and dam improvement works Flood risks during new dam construction Diversion structures Flood risks during improvement and/or dam removal works Health and safety requirements References

43 43 44 44 44 45

Appendix 1 . . . . . . . . . . . . . . . . . . .

Rapid assessment of flood capacity and freeboard at existing dams Purpose Procedure References

47 47 47 47

Appendix 2 . . . . . . . . . . . . . . . . . . .

Detailed advice on applying the FSR/FEH rainfall–runoff method to reservoir safety Introduction FEH rainfall DDF model FEH13 rainfall DDF model ReFH method Advice on detailed aspects of the FSR/FEH rainfall–runoff method Linking the flood frequency curve to the PMF estimate References

55 55 55 55 56 56 60 60

Appendix 3 . . . . . . . . . . . . . . . . . . .

Process diagrams Embankment dam Concrete/masonry dam

61 61 62

Glossary

63

Bibliography General Hydrological Overtopping of embankment, concrete and masonry dams Dam breach analysis Risk assessment

67 67 68 70 71 71

vi

Preface

The Floods and Reservoir Safety Working Group was established in March 2013 by the Institution of Civil Engineers (ICE) in order to update the third edition of Floods and Reservoir Safety. This followed from a request to ICE from the Department for Environment, Food and Rural Affairs that a fourth edition of the guide be prepared to take account of legislative change, research, and guidance on aspects such as hydrology and risk produced subsequent to the publication of the third edition in 1996. The working group met seven times under the chairmanship of A Macdonald. The membership consisted of A Macdonald T M Hewitt Professor A K Hughes E J Stewart Dr P J Mason Professor N W H Allsop C A Goff M Chrimes A Wescott G Jamieson

Atkins Ltd Centre for Ecology & Hydrology Damsolve Ltd HR Wallingford HR Wallingford Institution of Civil Engineers Institution of Civil Engineers ICE Publishing

The working group is pleased to acknowledge the assistance given during the preparation of this edition from a number of other organisations and individuals who have helped in reviewing the guide, undertaking trials of the methodology recommended, or in allowing reservoirs that they own to be used for the purpose of benchmarking against the third edition. These include the Reservoir Safety Consultative Group, the Reservoir Safety Advisory Group of ICE, Bristol Water plc, Dwr Cymru Cyf, Scottish Water, Southern Water Services Ltd, SSE plc, United Utilities Group plc, MWH Global Inc., Jacobs UK Ltd, Mott MacDonald Ltd, and the Environment Agency. In addition, there have been valuable contributions from many people at technical meetings where the guide has been discussed. The authors of the guide are grateful for this input together with the practical support provided by ICE.

vii

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.001

Chapter 1

Introduction Background

UK reservoir safety legislation places an obligation on the owners of certain reservoirs, dependent on the capacity of water above the level of the adjacent ground and the hazard posed by that water, to provide for their inspection in the interests of public safety. A guide relating to flood protection standards, flood magnitude and freeboard for the benefit of engineers exercising personal judgment under UK reservoir safety legislation was first published in its current form in 1978. This was revised in 1989 (minor corrections and additional references) and in 1996 (update taking account of best current research and experience of using the guide). In the period since 1996 there have been changes to extreme flood estimation in the form of rainfall depth–duration–frequency and rainfall–runoff models, reservoir safety legislation, and adoption of a risk-based approach to that legislation. In view of this, the UK Department for Environment, Food and Rural Affairs asked the Institution of Civil Engineers (ICE) to bring the publication up to date by preparing a fourth edition of the guide. In order to undertake this task, a small working group was established to operate under the following terms of reference. The purpose of the working group is to undertake a review of the third edition of Floods and Reservoir Safety and, with the assistance of ICE Publishing, to amend the document such that a fourth edition can be published during 2014. The revised edition should build on the text in the original, only amending those elements where legislative or technical aspects or feedback from the profession have resulted in the need for revision. In particular this will include – legislative changes in England, Wales, Scotland and Northern Ireland – the Flood Estimation Handbook (FEH) (IH, 1999) – recent research into extreme rainfall – guidance on risk assessment for UK reservoirs – research on wave overtopping of embankments. At appropriate events, the working group should also seek the views of owners and panel engineers on the proposed revisions and invite a selection of owners and panel engineers to assess the impact of the proposed changes on a small number of reservoirs prior to publication.

Changes

The flood protection standard required for each dam has historically been based on four categories dependent on the potential hazard to life and damage downstream. There is a general consensus that the existing system is uncomplicated and has been applied in a broadly consistent manner by panel engineers. The categories in Table 2.1 in this guide have therefore been retained from the previous edition, partly because all dams retaining reservoirs that come under existing UK reservoir safety legislation will have been categorised and assessed against those standards. It is recognised that a small change in hazard can result in a change in category and a step change in protection requirements. Changes in legislation could also result in a large number of existing reservoirs being subject to assessment for the first time. Rather than introducing additional categories to smooth out such impacts, the approach advocated in this guide where an existing reservoir fails to meet the standards required in Table 2.1 is to adopt a risk-based assessment, and to review the benefits that would be gained in meeting the 1

Floods and Reservoir Safety

Table 2.1 standard against the cost of doing so, and also the tolerability of the residual risk. Such an assessment should be viewed as part of the decision-making process when assessing the safety of the dam against floods and reflects the principles outlined in the Guide to Risk Assessment for Reservoir Safety Management (EA, 2013) and Reducing Risks, Protecting People: HSE’s Decision-making Process (HSE, 2001). Procedures for deriving reservoir flood inflows have been amended since the last edition by the FEH, which introduced new estimates for rainfall depth–duration–frequency for inflow floods of certain return periods. Chapter 3 has been updated to reflect this change and to clarify the appropriate model and also where the new FEH rainfall model should be used, once it has been released. In scope and general layout the updated guide follows that of its predecessors with the exception that the chapter on the overflowing and overtopping of dams now follows that describing wave height and the prediction of overtopping rates. To avoid confusion, the guide now uses two distinctive terms: overflowing (relatively steady rates of flow) and overtopping (intermittent flows from waves). The approach in previous editions of establishing required flood freeboard margins by estimating the wave surcharge height has been replaced by a method for predicting the rate of wave overtopping. This is better able to take account of upstream slope and crest profiles. Benchmark testing has shown that in general the tolerable limit of overtopping rates suggested in the guide results in a slightly less onerous freeboard requirement than the previous edition.

Scope of the guide

This guide is intended to assist those individuals who bear the personal responsibility that comes from being appointed to the statutory panel of engineers qualified to design and also inspect reservoirs. A desire for brevity and clarity of principle has led to this document being relatively concise. It should be read in conjunction with the latest revision of the FEH, in particular those sections that apply to reservoir safety flood inflow estimation. The working group has updated Chapter 2 on the protection standards to cover all major points of principle, including the use of a ‘standards-based’ approach and a ‘risk-based’ approach. It has also sought to give further clarification and guidance to avoid excessive discrepancies in the conclusions reached by different engineers, when reviewing the required level of protection to a dam against the threat of a flood event. A flowchart outlining the process is included in Appendix 3. The guidance here is not mandatory; however, it is recommended that where an engineer feels it is right to depart from the general principles provided by this guide, or the circumstances are not covered by the principles in this guide, the fact should be recorded, for example in the inspection report or annex to a certificate, with the reasons why. A glossary of terms used in this guide is included after the appendices, together with a schematic diagram, showing the relationship between the principal terms in the guide. This edition of the engineering guide supersedes and replaces the 1978 original and the 1989 and 1996 revised guides. REFERENCES

EA (Environment Agency) (2013) Guide to Risk Assessment for Reservoir Safety Management. EA, London, UK. HSE (Health and Safety Executive) (2001) Reducing Risks, Protecting People: HSE’s Decision-making Process. HSE, Bootle, UK. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK.

2

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.003

Chapter 2

Floods and waves protection standards General

This guide categorises dams in terms of potential hazard to life and damage downstream. Protection standards must resolve acceptably the conflicting claims of safety and economy by giving guidance on the risks associated with these hazards. Recent research published in the Guide to Risk Assessment for Reservoir Safety Management (EA, 2013) (RARS) provides guidance on the methodologies that can be adopted to determine the probability of failure due to a number of identified threats, including floods. Using that guide allows an assessor to determine the risk that a reservoir poses to human life, and the tolerability of that risk, together with a means by which the value in terms of cost to save a life can be judged against the reduction in risk that can be achieved by undertaking works. This relates closely to legislative change in the UK, which is moving towards a risk-based approach for reservoir safety. It is therefore appropriate that this edition of the guide acknowledges these changes and makes provision for them to be considered when undertaking an assessment of a reservoir’s suitability for passing flood flows safely. Although it is now considered possible to design a spillway for the almost total protection of a dam against overtopping by waves or overflowing by stillwater, there is the clear possibility that a dam built with a smaller spillway at less expense would survive several generations without any disaster or damage occurring. Equally, risk can never be totally removed, and while it can be reduced to tolerable levels, the risk of failure can never be reduced to zero. Some dams, even if subject to overflowing, may not breach, and increasing experience of erosion resistance makes it possible to introduce this additional factor into present-day assessments. However, it is not simply a matter of economic judgement. As the Institution of Civil Engineers’ (ICE) Rules for Professional Conduct state, all members shall have full regard for the public interest, particularly in relation to matters of public safety, and in relation to the well-being of future generations and shall show due regard for the environment and for the sustainable management of natural resources. Thus, there are now two methods of approach to reservoir safety with regard to floods: a ‘standards-based’ approach, where the required level of protection is arrived at based on a broad categorisation of downstream damage, including the potential to endanger life, for example as detailed in the previous version of this guide (third edition, 1996), and a ‘risk-based’ approach, where the risk of failure of the dam due to floods is assessed together with downstream damage, including the likely loss of life, and the tolerability of that risk evaluated to arrive at the required level of protection, for example as described in the RARS and Reducing Risks, Protecting People: HSE’s Decision-making Process (HSE, 2001). A key difference is acceptance criteria. The third edition of this guide required a dam to be able to pass the inflow flood with sufficient freeboard as defined in Table 1 of that edition – with an explicit statement regarding the need to correct any deficiency. In the risk-based approach, the dam is assessed with regard to reducing the risk of failure due to floods to ‘as low as reasonably practicable’ (ALARP). The ALARP principle is met when it is deemed grossly disproportionate in terms of expending resources to gain any further reduction in risk. 3

Floods and Reservoir Safety

The approach adopted in this guide is to retain the previous ‘standards-based’ approach and categorisation but where an existing reservoir fails to meet these standards it is recommended that an engineer carry out a ‘risk-based’ assessment to review the benefit that would be gained to reduce risk to life when compared with the costs incurred in meeting the ‘standards’ in this guide. Once more experience in the use of the risk assessment guide has been achieved, a further revision of this guide may occur. For new reservoirs, it is recommended that provision for floods is reviewed against both a ‘standards-based’ and ‘risk-based’ approach.

The main factors

A crucial question when considering flood protection is the combination of circumstances that may arise in progressively rarer events. Three main factors have to be defined: g g g

the initial reservoir level the flood inflow the concurrent wind speed.

Despite continually improving techniques for defining flood hydrographs, wave overtopping and flood routing, the currently available methodologies permit only the independent assessment of each factor and then their combination to estimate maximum reservoir flood levels. This traditional approach ignores the dependence amongst hydrometeorological variables and the complex ‘joint probability’ problem; that is, choosing a coherent set of inputs to yield a maximum reservoir water level of stated rarity. In the early 1990s, prior to the third edition of this guide, the UK Department for Environment funded joint probability studies for reservoir flood safety which explored the application of multivariate extreme value analysis techniques to the problem of choosing design inputs for reservoir flood safety appraisal. A weak dependence between extreme wind and extreme rainfall events was found. A consultative process has been followed in this and previous versions of this guide in order to ensure that the values recommended for the assumptions adopted in this guide are generally acceptable within the profession. Although the general framework may have relevance abroad, the values published in Table 2.1 refer strictly to dams in the UK. Panel engineers must satisfy themselves that the application of values published is valid in particular cases.

Standards-based approach Initial reservoir level

When investigating reservoir safety, it is important to consider at what level the stored water could be when the flood inflow commences. It is recommended that the flood routing calculation should commence from a stable storage situation as shown in Table 2.1 prior to the commencement of flood inflow. The alternative of postulating an antecedent flood that creates a starting water level would only be appropriate for individual complex cases, for example where additional flood storage is created by a narrow slot within an overflow. Further comment is given in Chapter 4. As discussed later in this chapter, flood attenuation embankments are a special case requiring individual consideration, and the ‘just full’ criteria shown in Table 2.1 at the start of the flood inflow may not be appropriate.

Reservoir flood inflow

It is necessary to determine the flood, in combination with wave action, that the dam must be capable of withstanding. The passage of this flood through the reservoir should cause no fundamental structural damage to the dam. However, it is normally uneconomic to provide a waterway below the dam that is sufficiently large to contain all flood outflow within its banks. Damage associated with rare overbank flows below or alongside spillway stilling basins may well be tolerated without risk to the integrity of the dam, but this needs to be confirmed. Similarly, there are situations where it is essential that the spillway channel should be hydraulically designed to carry all the dam flood outflow, for example where the spillway follows the line of the mitre of an embankment dam and out-of-channel flow may lead to failure of the downstream slope. Some engineers prefer a clear distinction between the flood inflow for which the reservoir is designed and the spillway design flood, but by definition only the former is essential to reservoir safety. The standards-based approach in this guide concentrates on the requirements for the reservoir flood inflow, with two criteria referred to as the ‘design flood’ and the ‘safety check flood’ in Table 2.1, for each dam category:

4

Floods and waves protection standards

Table 2.1 Flood, wind and wave protection standards by dam category Dam category

Potential effect of a dam breach

Initial reservoir condition standard

Safety check flood conditions

Design flood conditions

Reservoir flood inflow

Reservoir flood inflow

Concurrent wind speed for assessing the freeboard required to contain wave overtopping, or minimum flood freeboard provision (whichever is the greater)

A

Where a breach could endanger lives in a community

Just full (i.e. no spill)

Probable Maximum Flood (PMF)

10 000-year flood

Mean annual maximum hourly wind speed, with minimum flood freeboard of 0.6 m

B

Where a breach (i) could endanger lives not in a community or (ii) could result in extensive damage

Just full (i.e. no spill)

10 000-year flood

1000-year flood

Mean annual maximum hourly wind speed, with minimum flood freeboard of 0.6 m

C

Where a breach would pose negligible risk to life and cause limited damage

Just full (i.e. no spill)

1000-year flood

150-year flood

Mean annual maximum hourly wind speed, with minimum flood freeboard of 0.4 m

D

Special cases where no loss of life can be foreseen as a result of a breach and very limited additional flood damage would be caused

Just full (i.e. no spill)

150-year flood

150-year flood

Mean annual maximum hourly wind speed, with minimum flood freeboard of 0.3 m

g g

Concurrent wind speed for assessing wave overtopping

                                        

Mean annual maximum hourly wind speed

design flood – the inflow that must be discharged under normal conditions with a safety margin provided by an accepted freeboard limit safety check flood – the inflow beyond which the safety of the dam cannot be assured (i.e. key components exhibit marginally safe performance for this flood condition). (After ICOLD, 1992)

However, it is acknowledged that at the ‘safety check’ level of flood, some damage may occur to the dam due, for example, to overtopping by wave action, or by overflowing. For the case of embankment dams not specifically intended and designed to withstand overflowing or overtopping, the design flood in Table 2.1 is the flood at which no significant wave overtopping of the crest or wave wall should be allowed to occur. An overtopping flow of 0.001 l/s/m, or less, is taken as being no overtopping. For the case of a concrete or masonry dam, wave overtopping would be acceptable at the design flood in Table 2.1, assuming that it is founded on rock, although stillwater level should be below the dam crest or wave wall level unless a rigorous assessment of the founding 5

Floods and Reservoir Safety

conditions at the toe of the dam is undertaken to demonstrate resistance to erosion by overtopping flows. Concurrent wind speed

In the British climate, major rainfalls of several hours’ duration can result from very different storm systems. Intense depressions can be expected to have relatively steady high winds over large areas, whereas thunderstorm clusters can exhibit very different wind patterns within a few kilometres. Sometimes, storms may become stationary over an area, with low-level air rushing in; other storm cells can move in a line, possibly up or down a valley, with consequent wind veering. Waves generated on lakes during a storm, and more particularly at the time of peak flood surcharge at the dam, vary widely as a result. There is some evidence that steep-sided valleys often direct winds close to the valley axis even when winds at higher levels are at significant obliquity to the valley axis. The wind-generated waves on reservoirs and lakes should not be underestimated on the basis that the direction of the valley axis does not align with the predominant wind direction. Chapters 5 and 6 provide guidance on the methodology for assessing waves, wave overtopping, dam freeboard and allowable rates of wave overtopping. Records of past extremes suggest that a wave allowance should always be made and that it should be a more cautious one, particularly where there is a potential risk to human life. This allowance will provide a further margin of safety in many (but not all) rare floods. There is no clear evidence that rare floods will be accompanied by outstandingly rare winds. Balancing a precautionary approach with the need for the justification of capital spending on additional safety, it seems only necessary to test each dam against overtopping from the waves generated by the highest hour’s wind in a year (taken as an average from a series of annual peak hourly wind statistics). Even with valley wind steering there is frequently as much chance that the wind will blow waves away from the dam rather than to it.

Dam categories

The accidental, uncontrolled escape of water from an impounding or other reservoir can threaten life and property. Greater security is required against dam failure where there is a severe threat of loss of life and extensive damage and a lower security where the threat is less severe. All dams should be assessed for the consequences of failure during a flood event, and the categories shown in Table 2.1 indicate the degree of security required of a dam and the likely effects of the failure of the main dam (or if applicable, any other dam) by which the reservoir is retained. It is possible in some circumstances for a reservoir to be retained by dams of different categories. In assessing the consequence of failure, it is the additional damage that would occur if the dam failed under flood conditions compared with the damage caused by the flood were the dam not to fail. The potential effects of dam failure for each of the categories A–D in Table 2.1 are given below. It is recognised, however, that some dams may lie between the stated categories. For instance, a periodically used camp below a dam may not justify a full category A rating: in such cases the judgement of the panel engineer is required to determine the appropriate reservoir flood inflow. In assessing the damage that may be caused as a result of a breach, the panel engineer should take account not only of normal property damage but also damage that may be caused to critical infrastructure, scheduled monuments, protected environmental habitats and the like, where such information is readily available. Published dam breach inundation maps can assist the panel engineer in making an assessment of the risk to life and the extent of the damage that may result from a dam failure. However, for some maps, the limitations of the underlying assumptions and accuracy of the methods used to produce them should be carefully considered.

Category A dams

6

Category A relates to endangering the lives of inhabitants of communities. A community in this context is considered to be not less than about 10 persons who could be affected; it is considered that inspection of any valley will soon reveal whether the presence of a hamlet, school or other social group means that a dam at its head should be in category A. Road and rail traffic caught in a valley flood would only accidentally be involved and would not by itself justify category A. A more difficult situation exists where an occasional camp site exists in the holiday season alongside a reservoired river. If, for example, this is in regular use by school parties it could well

Floods and waves protection standards

justify a community rating, but if it is frequented by a few unrelated short-stay individuals, it need not. Category B dams

Category B(i) is intended to refer to inhabitants of isolated houses and, for example, to operatives in treatment works immediately below a dam and in other places of work in the flood path. (These situations lend themselves to taking measures to buy out the property or to arrange flood escape routes where appropriate.) Category B(ii) refers to extensive damage, including erosion of agricultural soils and the severing of main road or rail communications or other critical infrastructure such as gas mains or transformers.

Category C dams

Category C covers situations with negligible risk to human life and so includes flood-threatened areas that are ‘inhabited’ only spasmodically, such as footpaths across the flood plain and playing fields. In addition, this category also covers damage to scheduled monuments and loss of livestock and crops and protected natural habitats.

Category D dams

Many small reservoirs with low earth dams may cause no real problem, except that of replacement, if they wash out. These special cases, many of which are ornamental lakes kept full for aesthetic reasons, are given a separate category where they pose no significant threat to life or property. A flood intense enough to cause failure of a dam would create some damage even if the valley were still in its natural state; the additional damage caused by the release of stored water may well be insignificant if the lake is small. So, where the amount stored would add no more than 10% to the volume or peak of the flood, it is recommended that the reservoir flood inflow should be that from the 150-year flood. The point of reference for assessing whether the damage is significant or not can be taken as the first site below the dam at which some feature of value exists (e.g. a mill or road bridge). The 1000-year flood hydrograph applicable to the catchment prior to dam construction can be used for making this 10% sensitivity test.

Recommended standards Table 2.1 sets out the standards that are appropriate for the wide variety and scale of dams in the UK. To apply them, it is necessary to route the appropriate reservoir flood inflow using the corresponding initial reservoir condition to obtain the flood surcharge level. The wind speed given in Table 2.1 is then used with the flood surcharge level to determine the wave overtopping discharge, if any, taking account of the geometry and protection standards of the upstream face and crest. For the safety check scenario, the wave overtopping discharge must be less than the allowable wave overtopping discharge as indicated in Chapter 6. These allowable wave overtopping discharges will be sufficient to prevent quantities of flow arising from overtopping causing unacceptable damage to the top of the dam or downstream face, which would otherwise place the dam at risk of a breach. In addition, any wave wall must be able to withstand the loading from the waves generated by the concurrent wind speed indicated in Table 2.1. For fill embankment dams, the elevation of the top of the dam will be governed by one of three conditions, the first being that the safety check flood surcharge level should not exceed the top of the dam, normally the crest level. If the flood peak is particularly prolonged, the safety check flood surcharge level may have to be lower still to avoid harmful leakage through the crest materials above the dam core. The second condition is that the safety check flood surcharge combined with wave action must not lead to a wave overtopping discharge in excess of the allowable wave overtopping discharge indicated in Chapter 6 (unless the dam has been designed to be capable of withstanding overtopping in excess of the recommended standards set out in Chapter 6). The third condition is that the design event flood surcharge combined with wave action must not lead to any overtopping discharge unless the dam has been specifically designed to permit wave overtopping at this level of flood. (A mean wave overtopping discharge rate of 0.001 l/s/m may be taken as zero, when calculated using the method described in Chapter 5.) Where an existing dam is checked against the recommended standards given in Table 2.1, and is found to be lacking in spillway capacity or freeboard, it will be necessary to provide for the safe passage of the safety check flood, unless it can be shown using a risk-based approach that the probability of failure has already been reduced to ALARP. It will also be necessary to check the stability of the dam, and wave wall, particularly if the improvement involves a raising of the flood stillwater level. 7

Floods and Reservoir Safety

For concrete or masonry dams that typically have a spillway over most of their crest length and no fill embankments at the abutments, the concept of freeboard may not be relevant provided that they are constructed on non-erodible foundations, and that there is no chance of erosion caused by overflowing or flow around the abutments. For these dams, it is necessary to be certain that the loading derived from the Table 2.1 safety check flood does not go beyond structural design limits. Where concrete or masonry dams have fill embankments incorporated in the composite structure, the recommended standards for embankment dams will normally govern, as defined by Table 2.1. The above points relating to the application of Table 2.1 are summarised by way of flowcharts in Appendix 3. Table 2.1 is designed to take account of those factors that are weighed together by panel engineers both for the design of new dams and when undertaking inspections of existing dams. Its main intentions are to ensure that, where a community could be endangered by the breach of a dam, the risk of any breach caused by a flood is virtually eliminated. However, expenditure on safety works should be kept to a scale justified by the risk. In this respect, judgement and the rules associated with disproportionality as defined by the UK Health and Safety Executive can be used as an aid to making such assessments. Reference should be made to the publication Reducing Risks, Protecting People: HSE’s Decision-making Process (HSE, 2001).

Risk-based approach The risk-based approach using appropriate tools and methods seeks to provide an approach that General

allows an owner and their advisors to better understand and evaluate reservoir safety risks in a structured way. This then allows for risk-based decisions to be made to reduce risks to people, the environment and the economy but still maintain an important reference to accepted best practice. Where expenditure on remedial works will be significant to meet the standards-based approach to dealing with floods as set out in the previous sections, a risk-based approach could be adopted to assessing the value (cost versus reduction in risk) of undertaking remedial works. This can form part of an overall options appraisal. The process of examining and assessing the significance of estimated risks is tiered risk evaluation, and the UK Health and Safety Executive has a well-established framework for risk evaluations called the tolerability of risk, which splits risk into areas of unacceptable risk, tolerable risk, and broadly acceptable risks. The RARS guide uses a tiered approach to enable a user to move from essentially a qualitative assessment to a detailed quantitative assessment, depending on the levels of uncertainty, the complexity of the problem and the importance of the decisions to be made. In each tier, the ‘loads’ or events which could lead to failure and flooding are considered as well as the consequence of these dam failure scenarios. In each case, failure mode identification forms an important element. In terms of floods, the analyses provide a numerical estimate of failure probability and risk. In assessing the vulnerability of a dam to floods, judgement will need to be exercised. The RARS guide encourages the use of the two-staged phased approach to floods whereby if a reservoir did not meet the requirements of the standards-based approach then the use of the ALARP test can be used to help determine the extent of the works required. In looking at the criteria for Category A and B dams as defined by Table 2.1, which are loosely based on the population at risk (PAR), it can be seen by plotting the values indicated in Table 2.2 on an f–N chart – which relates the estimated annual probability of failure f against the average societal loss of life, N (LLoL) (see RARS Vol. 2, Figure 9.2, for an example of a simple f–N chart) – that most existing Category A and B dams lie within the ‘broadly acceptable’ range (see page 12 for definitions of PAR and LLoL). This illustrates that the standards-based approach can, at least in the short term, sit alongside a risk-based approach. At this time, it is considered that the Probable Maximum Flood (PMF) should be retained as the most onerous inflow flood for UK dams.

8

Floods and waves protection standards

Table 2.2 Relationship between ‘risk’ parameters and category Category A

Category B

Population at risk (PAR)

.10

,10

Likely loss of life (LLOL)

.1

1 to ,10

(1) Safety check flood Annual probability of occurrence

PMF 2.5 × 10 − 6

10 000 year 1 × 10 − 4

(2) Likelihood of failure during a safety check flood very low, say 1% (0.01) chance

0.01

0.01

Annual probability of failure from a safety check flood (1) × (2)

,2.5 × 10

−8

,1 × 10 − 6

Until further research is available, it is recommended that the PMF is assigned an annual exceedance probability of 1 in 400 000 (2.5 × 10 − 6 per year).

Other criteria Dam break wave

Economic considerations

Assessment of the physical effect of a potential dam failure and the consequent flood wave is far from straightforward. However, Figure 2.1 indicates that the flow immediately downstream is influenced mainly by the dam height. Differences between dam types are not strongly marked where storage is substantial. Extrapolation of the curve for a lower dam demands caution because individual circumstances at the site are likely to be more significant. The RARS guide includes methods for estimating peak breach flow, taking account of dam type, height and reservoir volume. Computer programs are available to estimate flood levels as the wave passes downstream, and the models have improved dramatically over the years. Either depth or the velocity of flow or their combination may pose a threat. Although results cannot be precise, such a calculation can help in the assessment of the hazard posed by the possible failure of a dam, and hence its category. Inundation maps generated by others should be treated with caution unless the input parameters are clearly understood. In addition, breach modelling software should be used with care, and analysts should make themselves aware of the inconsistencies, uncertainties and assumptions made in the analyses, whether undertaken by themselves or by others. In general, determining the capacity of a spillway for a new dam is an economic balance between the cost of the provision of the spillway (or the raising of the dam crest) and the protection it affords to life and property downstream. The effect of the recommendations in this guide is to impose safety criteria that effectively constrain the range over which an economic solution is applicable. However, in certain cases the use of a risk-based approach may assist in analysing the additional benefit to be gained. For example, would a spillway capable of passing the design flood without a breach of the dam mean that there would be an improved situation in respect of likely loss when compared with a smaller spillway that would increase the probability of failure of the structure? Normally, a dam breach would be expected to result in a higher loss of life when compared with the natural flood event, but that might not always be the case. For existing dams that do not meet the standards based approach in Table 2.1, it is recommended that the RARS guide is used to assess the value to be gained in undertaking works in order to reduce the probability of failure and downstream impact. A quantitative (Tier 2 risk assessment) approach is likely to be required. A risk-based approach can also be used to support a panel engineer’s judgement where dams are assessed as being between the categories in Table 2.1.

Spillway systems

Due consideration should be given to the performance of a spillway as a system and its ability to cater for the safety check flood outflow, to ensure that the passage of such a flow will not threaten the safety of the dam. The behaviour of the spillway throughout its full extent should be reviewed.

Reservoir cascades

Different categories can exist where reservoirs are in a cascade. In determining the flood category, it is important to consider the effect of a flood-induced failure of an upper reservoir on a lower one, taking account of the likely concurrent water level, as well as the impact of the dam-break flood wave on the intermediate downstream valley. 9

Floods and Reservoir Safety

Figure 2.1 Dam failure flood flow versus dam height. (Source: Kirkpatrick, 1977.) 100 000

13

50 000

Peak flow: m3/s

1

10 000

4

2

5 4

5000 6 9 11

12

7 8 14

1000

10

500 0

10

20

30

40

50

60

70

80

90

Dam height if dam overtopped, or depth of water at time of failure if dam not overtopped; H: m Estimated flood peaks from dam failures. The numbers indicate the name of the dam, its location, type of dam where known, and the year of failure. 1. St. Francis, California, concrete gravity, 1928 2. Swift, Montana, rock fill, 1960 3. Oros, Brazil, earth and rock fill, 1960 4. Apishapa, Colorado, earth fill, 1923 5. Hell Hole, California, rock fill, 1964 6. Schaeffer, Colorado, earth fill, 1921 7. Granite Greek, Alaska, 1971, discharge at 8 km downstream 8. Little Deer Creek, Utah, earth fill, 1963 9. Castlewood, Colorado, rock fill, 1933 10. Baldwin Hills, California, earth fill, 1963 11. Hatchwood, Utah, earth fill, 1914 12. Lower Two Medicine, Montana, 1964 13. Teton dam, Idaho, earth, 1976 14. Dale Dyke, Sheffield, England, earth fill, 1864, discharge at 10 km downstream

Failure to meet recommended standards

10

Early action is required where the dam freeboard is inadequate to contain the stillwater flood surcharge of the appropriate standard: this normally takes the form of a temporary lowering of the top water level prior to remedial works being carried out. However, if a dam is found to be adequate to contain the flood surcharge but unable to cope with the associated waves required by Table 2.1, the action to be taken and the timing thereof will depend on local circumstances. Remedial measures such as a wall or rip-rap protection may be feasible. In some cases, the observation of wave patterns and movements during a storm may show that the dam is sheltered, and enable the panel engineer to review the assessment of wave action on the structure.

Floods and waves protection standards

Where this results in a relaxation, that decision should be recorded. Where the costs of upgrade works are significant, the value of undertaking them should be tested in accordance with the risk-based approach discussed earlier.

Other aspects Gated spillways

Where a gated spillway is employed, high standards of maintenance are required and regular operational testing is essential. This is particularly so where gates form the sole outlet for floodwaters and where the dam is formed by an erodible embankment. The provision of gates should not be encouraged unless the dam owner’s organisation is capable of ensuring the required standards of maintenance and of operation even under emergency conditions. Because the chance of human error or mechanical failure cannot be discounted, it is considered that the following additional safeguards should be provided at dams where overtopping would be likely to cause a breach: g g g

a minimum of two gates the ability of the remaining gates to pass at least the 150-year flood if one gate is out of action if the dam falls into Category A, automatic operating equipment with the necessary standbys and options for feasible manual operation.

Where flood provision on a major dam is dominated by reliance on gated releases, the statistical reliability of the whole operational regime can be assessed by means of a structured systems integrity level analysis. Minimum spillway capacity

A few reservoirs are so large relative to their direct catchment that the routed flood outflow is very small. In such cases, however, the spillway should have a prescribed minimum capacity and be constructed so that it will not tend to block with debris or ice. The capacity, at a head that leaves the required wave freeboard or minimum wave freeboard provision, should be not less than 0.885(SAAR0.83/1000) m3/s per km2 of catchment where SAAR is the standard period average annual catchment rainfall in millimetres (1941– 1970). The equation is equivalent to a daily runoff of 13–65 mm, depending on the location. It should be noted that snowmelt can be a controlling factor in some circumstances (refer to Chapter 3).

Reservoirs without direct For reservoirs with minimal or no natural catchment, no spillway is necessary, provided that catchments g adequate fail-safe provision exists to ensure that the inflow cannot continue to the point of overflowing the embankment (unless overflowing is allowed for in the design) g for open reservoirs, the dam freeboard is adequate to contain direct rain or snowfall into the reservoir and wave overtopping discharges resulting from a 200-year wind speed at the location (see Chapters 5 and 6). Grass-covered embankments

There are recorded cases of existing embankment dams with a reasonably level crest and a surfaced roadway having resisted some overflowing successfully. However, there are many other embankment dams where even a small amount of overflowing could lead to catastrophic failure. Guidance on the tolerance of embankment dams to overflowing is given in Chapter 6.

Fill dams in deep valleys

When fill dams are sited in deep valleys there may be substantial catchment areas that in a severe storm drain towards one or both flanks of the dam. This can result in considerable damage to the embankment. In addition to drains at the junction of the downstream slope with the valley sides, (downstream mitre), additional drainage measures (such as catchwaters) may be necessary to direct storm water clear of the embankment.

Flood attenuation embankments

Some flood attenuation embankments impound little or no water for most of the time. They are usually grass covered, and rarely more than 5 m high. The impoundments are designed to completely fill during storm events. These embankments are usually located upstream of highly urbanised areas, and therefore may fall into Category A. The recommended standards in 11

Floods and Reservoir Safety

Table 2.1 may not necessarily all apply, and each impoundment needs to be judged independently on its own merits. When considering such structures, it should be borne in mind that it is the incremental impact of the failure flood wave that is important. Discharge alterations

Where alterations are about to be made to an uncontrolled spillway or to the operational rules governing flood discharge gates, particularly where this increases the discharge for a particular flood event, the undertaking concerned should draw the facts to the attention of the owners of any impoundments that lie lower down the river; the authority responsible for flood plain management should also be notified.

Rapid assessment of existing dams

In some situations it may be useful to make a rapid assessment of the ability of an existing dam to cope with the combination of flood, waves and initial reservoir level according to the criteria set out in this guide. Appendix 1 describes such a method with a worked example. Although it cannot be precise, it does provide an initial screening tool, particularly for smaller reservoirs, and should reveal consistently those dams that merit fuller investigation of their safety in times of flood. The rapid assessment method retains some elements of the methods presented in the Flood Studies Report (NERC, 1975), which have now been largely replaced by the methods in the Flood Estimation Handbook (IH, 1999). However, it requires only an Ordnance Survey map and a knowledge of catchment average rainfall. The rapid assessment method can also be used to determine whether a full assessment using the methods described in the Flood Estimation Handbook is actually required. Where recommendations are made in the interests of reservoir safety, these should not be based on the results of the rapid assessment method. Also, it should be noted that the rapid assessment method may not be applicable to some types of catchment, such as urban catchments, catchments with a low mainstream slope (where S1085 , 10 m/km), irregular-shaped catchments (see also Chapter 3) and catchments where infiltration rates may be significantly different from the norm.

Population at risk (PAR) and the likely loss of life (LLOL)

It must be recognised and remembered that there is a difference between the PAR and the LLOL. The PAR is defined as ‘the population within the flooded area when the dam fails’ – all those directly exposed to the floodwaters if they took no action to evacuate. The distribution will vary across the flood plain in terms of those subject to deep water and those in shallow water. The portion of the population at risk will vary across those within buildings and those out in the open. Their level of preparedness will also vary. The LLOL is that part of the population at risk who could lose their lives in the event of a dam break. There are many ways of assessing the LLOL, and results can vary by an order of magnitude because it is such a complex issue. The results will depend on the severity of the flood, warning times (if there are any facilities to warn people), the time of failure, and the depth and velocity of the water. Methods include those by the Department for Environment, Food and Rural Affairs in Flood Risks to People (Defra, 2003, 2006), Graham (1999) and Jonkman (2008), the SUFRI (Jo¨bstl et al., 2011) and LIFESim (Aboelata and Bowles, 2005) models and many others. When trying to assess the PAR and the LLOL, the Environment Agency, Natural Resources Wales and the Scottish Environment Protection Agency reservoir flood maps can be used, but it must be recognised that the methodology adopted has a high degree of uncertainty and conservatism built within it. If figures are ‘critical’ to the decision-making process, then a more sophisticated form of breach flow estimate and inundation mapping should be adopted.

Incremental damage

The incremental effect of dam failure on the community downstream should be considered when compared with the no-dam failure scenario. This may allow situations to be accepted where the cost of increasing the size of spillways to meet a larger proportion of the design flood is disproportionately high. The approach recommended at this time is to ‘test’ the dam, adopting Table 2.1 standards, and if the dam meets the standard, it can be considered acceptable. If not, then the risk-based approach should be used to support the justification for works, both in terms of need and societal risk.

12

Floods and waves protection standards

Other structures

It is considered that the standards and assessment methods proposed in this guide be applied to all mine, quarry, power station fly ash and sewage lagoons. Reference to this guide will be included in the soon-to-be-published EU handbook on the hydraulic placement of soils and mineral wastes.

Warning times

The Water Act of 2003 amended the Reservoirs Act 1975 to give the Secretary of State powers to direct reservoir owners to prepare flood plans. Although no known directions have been given at the time of writing, it is considered best practice to prepare on-site plans. The Reservoirs (Scotland) Act 2011 contains similar provisions. The Civil Contingencies Act 2004 established Local Resilience Forums to cover local arrangements for civil protection against emergencies, including floods from the failure of a dam. If people who might be affected by the flood from a dam failure can be warned, then lives can be saved. However, the ability to warn people depends on a number of factors, including the preparedness and vigilance of the undertaker, the type of failure, whether it is a sunny day failure or a rainy day failure, the period of time between the detection of the problem and the breach, the time from the onset of failure to the peak discharge, the number of dams affected, the time of day and day of the week, and the nature of the catchment, the likelihood of the defect being found and access to the site. The RARS guide states that for ‘the base case highest individual risk and average societal life loss it should be assumed that there is generally no warning’. It goes on to state that ‘the exception is where the population at risk is well downstream of the dam with an intervening community where it may be reasonable to assume the alarm would be raised once the flood wave had passed the first community and the population downstream would be warned’. It is likely that significant time would be required to receive the alarm and then to issue a warning downstream.

Legislative categorisation Legislation in the UK that has been introduced are the Flood and Water Management Act 2010, of reservoirs which amends the Reservoirs Act 1975, and the Reservoirs (Scotland) Act 2011. Legislation is also likely to be brought in to cover reservoirs in Northern Ireland. All of these include some form of reservoir risk classification (e.g. high, medium, low, not high). It must be remembered that these are legal definitions and legislative risk designations that are different from the designation of a flood category as determined by an inspecting engineer. RARS – acceptance of different levels of risk

The RARS guide provides a three-tiered approach to assessing reservoir safety, with progressively increasing levels of sophistication in terms of analysis if needed. In some instances, where it is proposed to adopt a lower standard of protection than in Table 2.1, it would be appropriate for the panel engineer to discuss the matter with the owner. Some owners may not wish to accept the possibility of failure of even their smallest or lowest consequence of failure dams because of the resulting reputational loss. They may then opt for more stringent criteria whatever guidance is given. Conversely, if a risk-based analysis shows, in terms of the likely consequence of failure, that it is reasonable to reduce the reservoir flood inflow and not fully meet the standards of Table 2.1 and accept the increased risk of failure that ensues, this may be a choice that an owner would wish to take. REFERENCES

Aboelata MA and Bowles DS (2005) LIFESim: A Model for Estimating Dam Failure Life Loss. Institute for Dam Safety Risk Management, Utah State University, Utah, United States of America. DEFRA/EA (Environment Agency) (2003) Flood Risks to People, Phase 1, R&D Technical Report FD2317/TR. DEFRA/EA, London, UK. DEFRA/EA (2006) Flood Risks to People, Phase 2, R&D FD2321/TR1, The Flood Risks to People Methodology. DEFRA/EA, London, UK. EA (Environment Agency) (2013) Guide to Risk Assessment for Reservoir Safety Management. EA, London, UK. 13

Floods and Reservoir Safety

Graham WJ (1999) A Procedure for Estimating Loss of Life Caused by Dam Failure. U.S. Department of Interior. Bureau of Reclamation, DSO-99-06. United States of America. HSE (Health and Safety Executive) (2001) Reducing Risks, Protecting People: HSE’s Decision-making Process. HSE, Bootle, UK. ICOLD (International Commission on Large Dams) (1992) Bulletin 82. Selection of Design Flood: Current Methods. ICOLD, Paris, France. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. Institute of Hydrology, Wallingford, UK. Jonkman SN and Vrijling JK (2008) Life loss due to floods. Journal of Flood Risk Management 1(1): 43–56. Jo¨bstl et al. (2011) SUFRI – Sustainable Strategies of Urban Flood Risk Management with Non-Structural Measures to Cope with the Residual Risks. Europe. Kirkpatrick GW (1977) Guideline for evaluating spillway capacity. International Water Power and Dam Construction 8(31): Fig. 4. NERC (Natural Environment Research Council) (1975) Flood Studies Report, vols I–V. NERC, Wallingford, UK.

14

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.015

Chapter 3

Derivation of reservoir flood inflow Objective

This chapter explains a procedure for the derivation of the reservoir flood inflow for the range covered in Table 2.1. The flood inflow recommended in Table 2.1 comprises the sum of all inflows to the reservoir. This will include natural catchment runoff (for impounding reservoirs only), direct rainfall on the open reservoir surface, and gravity or pumped inflows from other catchments or reservoir systems separate from the reservoir under consideration. The flood inflow must be specified by its complete hydrograph because the attenuation is a function of the full temporal distribution of the inflow, not simply of the peak flow. The inflow hydrograph at any UK reservoir site can be derived from a rainfall–runoff method, in which an appropriate design storm and associated antecedent conditions are applied to a hydrological model of the catchment. The design model has been calibrated to produce a hydrograph of a specified rarity. This chapter guides the engineer on the application of the relevant methodologies. In particular it g g g g g g

summarises the development of design hydrograph methods clarifies guidance on the design rainfall inputs to rainfall–runoff methods summarises the steps in the Flood Studies Report (FSR)/Flood Estimation Handbook (FEH ) rainfall–runoff method for estimating the inflow flood hydrograph explains how the procedure for the estimation of the return period flood is adjusted to derive the Probable Maximum Flood (PMF) introduces the revitalised flood hydrograph (ReFH) rainfall–runoff method provides a rapid method for deriving a preliminary estimate of the design flood peak inflow.

Research on hydrological modelling techniques appropriate to reservoir safety has a long history, and the methods continue to evolve. Appendix 2 incorporates advice already published and indicates how the conclusions of ongoing research may influence the currently recommended methods.

Flood Studies Report Until the publication of the FEH in 1999 (IH, 1999), UK practice in flood estimation was based and Flood on the 1975 FSR (NERC, 1975) and refined in a number of supplementary reports over the next Estimation decade. The FEH introduced a new set of statistical procedures for estimating rainfall and flood Handbook frequency in the UK, including a new model of UK rainfall depth–duration–frequency (DDF). However, although the FSR rainfall–runoff method was updated to use digital catchment descriptors to estimate model parameters at ungauged sites, and the FEH DDF model was recommended to provide design rainfall inputs, the core of the design hydrograph method remained unchanged. The FEH DDF model was developed to provide rainfall estimates for a range of durations and return periods up to 1000 years. When the model was used to provide estimates of rainfall depths for return periods in excess of 1000 years, concern was expressed about large differences between the FEH DDF model and its predecessor, the FSR rainfall model. MacDonald and Scott (2001) reported that in some cases the 10 000-year return period rainfall exceeded the estimate of probable maximum precipitation (PMP) derived from the FSR. As a result of this concern, a new DDF model has been developed that is designed to provide rainfall estimates for return periods ranging from 2 years to over 10 000 years, and it is proposed 15

Floods and Reservoir Safety

that this should eventually replace the two existing (FEH and FSR) rainfall models for UK hydrological analysis and design. Details of the research leading to the model development are given by Stewart et al. (2013). The new DDF model has now been refined and generalised to provide rainfall estimates for any location or catchment in the UK, and a new web service to deliver the model results is under development. Further details are given in Appendix 2.

Revitalised flood Further research on rainfall–runoff modelling resulted in the development of the ReFH design hydrograph rainfall– method (Kjeldsen et al., 2005). The ReFH method introduces improvements to the key comrunoff method ponents of the FSR/FEH rainfall–runoff method, having a more flexible model structure and being based on more up-to-date flood event data. The ReFH method can be used for hydrological applications requiring hydrographs for return periods of up to 150 years, the current upper limit of the model calibration. This limit means that the current form of the ReFH design method can be used only for estimating the design flood inflow for Category C dams, and the safety check and design flood inflow for Category D dams. It is recommended for the time being that both ReFH and FSR/FEH methods are used to calculate flood inflows for such events. The ReFH method in its current form is not suitable for T-year flood estimation at longer return periods, or indeed for PMF estimation, although research to extend the range of applicability of the method is ongoing. It should be noted that, unlike the FSR/FEH rainfall–runoff method, the ReFH method uses equal return periods for the input rainfall and the output flow hydrograph. Further information about the ReFH model is given in Appendix 2.

Clarification of appropriate models and design rainfall inputs

Before research commenced to develop the new rainfall DDF model, the UK Department for Environment, Food and Rural Affairs (Defra) issued guidance prepared by the Reservoir Safety Working Group of the Institution of Civil Engineers to give clearer direction to panel engineers when considering floods and reservoir safety (Defra, 2004). The guidance states that g

g

g

The FEH DDF model should not be used for the assessment of the 10 000-year return period rainfall. The design rainfall values provided by Volume II of the FSR should continue to be adopted until the results of the new DDF model are available. For 1000-year return period rainfall, assessments should be undertaken for both the FEH and FSR methodologies, with the more extreme value of the two being used for flood assessment. The FEH should be used for the assessment of the 193-year return period rainfall (suitable for the estimation of the 150-year return period flood event using the FSR/FEH rainfall–runoff method).

Until the new DDF model is released for general use, it is considered that this guidance is still generally appropriate. However, the guidance preceded the release of the ReFH model, and so recommendations on its use in reservoir flood studies were not included. It is now considered that the ReFH method is suitable for use with the existing FEH DDF model in the assessment of the 150-year flood (using the 150-year rainfall as the corresponding design input). In future, it is intended that the ReFH method will be used with the new DDF model, when available, but will require calibration of the model and revised software. Attention is drawn to potential problems with regard to the application of the ReFH method to urbanised and permeable catchments.

Summary of the FSR/ FEH rainfall–runoff method for the T-year event

Volume 4 of the FEH gives a comprehensive review of reservoir flood estimation using the FSR/ FEH rainfall–runoff method. It must be emphasised that the methodology, designed primarily to ensure that a complete flood hydrograph and peak of the required rarity are obtained, has to be used in its entirety. In summary, the procedural steps to derive the flood with a recurrence interval of T years are as follows: (a) estimate the critical design storm duration for the reservoir catchment, making an estimated allowance for the reservoir lag (b) estimate the storm rainfall depth averaged over the catchment for the critical duration and of the severity appropriate for the design flood (c) distribute the design rainfall in time according to the winter or summer storm profile (d) derive the time distribution of the excess rainfall by deducting estimated losses

16

Derivation of reservoir flood inflow

(e) estimate the unit hydrograph and the response of the catchment to unit rainfall input (f) multiply the increments of the storm rainfall excess by the ordinates of the unit hydrograph and cumulatively sum to produce the flood response hydrograph (g) add base inflow to obtain the complete flood hydrograph.

Summary of the methodology for estimating the PMF event

The overall FSR/FEH procedure outlined for the T-year event applies also to the derivation of the PMF, with the following adjustments: g g

g

A different storm profile is used for the PMF. For the PMF, the time-base of the unit hydrograph is reduced by one-third, raising its peak by one-half. This adjustment ensures that the PMF is based on a severe condition rather than an average representation of the catchment response to storm rainfall. The PMP is partitioned into winter and summer amounts, as shown in Appendix 2. Winter PMP is assumed to fall on frozen ground and to be augmented at a uniform rate by snowmelt. Design hydrographs for both winter and summer PMF conditions should be estimated, as it is not generally obvious at the outset which condition will generate the critical design condition.

It is recommended that the FSR/FEH graphs and equations that use catchment descriptors should not be applied solely on the basis of an examination of Ordnance Survey maps. Site inspections should be carried out to g g

g

Summary of rainfall DDF and rainfall–runoff models

Confirm the catchment physiography and evaluate the impact on the flood response of catchment developments such as urbanisation, drainage and transport infrastructure. Establish whether drainage paths that convey frequent floods will be drowned in an extreme event or blocked by debris or fallen trees, when floodwater may find a more direct path to the reservoir. Floodwater may even enter the reservoir catchment from an adjacent watershed, effectively extending the reservoir catchment boundary while the event persists. Equally, overland flow may in some circumstances be diverted away from entering the reservoir along roads or gullies. Determine the flow contribution of any catchwaters which should be assumed to lie within the extent of the design storm. There may be a maximum rate of contribution set by the capacity of the catchwater or sewered diversion.

Table 3.1 presents a summary of the appropriate rainfall DDF model and rainfall–runoff model for use in deriving floods of particular return periods.

Table 3.1 Summary of the rainfall DDF and rainfall–runoff models to be used to derive the flood events associated with the dam categories given in Table 2.1 Flood event

Rainfall depth–duration–frequency model

Rainfall–runoff model

150-year return period

FEH until the new DDF is issued

FSR/FEH and/or ReFH (except in highly permeable or urbanised catchments)

1000-year return period

Largest rainfall depth of the FSR or FEH until the new DDF is issued

FSR/FEH until ReFH is extended

10 000-year return period

FSR until the new DDF is issued

FSR/FEH until ReFH is extended

PMF

FSR (PMP)

FSR/FEH until ReFH is extended

Use of local data

Flood estimates can be improved with information from records of the flow and rainfall in the catchment and of the reservoir level. Guidance on how such information can be incorporated into the FSR/FEH procedure is given in Appendix 2.

The contribution of extreme snowmelt

Snowmelt can contribute significantly to large winter floods in the UK, and the FSR recommended that it should be considered in the estimation of a winter PMF by being added to the 17

Floods and Reservoir Safety

Figure 3.1 Sketch map guide to the 24 h snowmelt with a 5-year return period

Contours of the M5 melt (mm/day) 42 Uplands where snow melt rates may exceed 42 mm/day

Inverness

Dundee Oban

Perth Glasgow Edinburgh

Londonderry

Newcastle

Carlisle Belfast

York Preston Leeds

Manchester Liverpool

Kingston upon Hull

Sheffield Nottingham

Stoke Leicester

Norwich

Birmingham

Aberystwyth

Northampton

Ipswich Bedford

Oxford

Swansea

Bristol

Cardiff

Bournemouth 0 Scale

80

LONDON

Brighton

160 m Plymouth

PMP. The value proposed for the 100-year melt rate over a 24 h period was 42 mm (equivalent to 1.75 mm/h). It was believed that this was a suitably rare occurrence for design purposes, particularly when combined with a maximum rainstorm. No geographical variation in the snowmelt rates was given, and it was suggested that this maximum melt rate could continue for 2–3 days. However, commenting on research by Archer (1981, 1983, 1984) and their own analyses, Reed and Field (1992) stated that there is little doubt that melt rates as high as 5 mm/h can occur 18

Derivation of reservoir flood inflow

in UK conditions, although catchment-wide rates are likely to be substantially smaller. They concluded that the FSR specifies a melt rate that is clearly much less frequent in one district (e.g. central southern England) than another (e.g. the Pennines). Hough and Hollis (1997) undertook an investigation of point snowmelt rates at 25 sites in the UK, and provided estimates of the maximum snowmelt rates for a range of durations and return periods. They found that there was a general increase in the maximum snowmelt rates from the south-west to the north-east and with increasing altitude. The results were used to derive Figure 3.1, which indicates areas where 5-year snowmelt rates higher than 42 mm/day (1.75 mm/h) might be expected. In such areas, it would be appropriate to check for sensitivity of the stillwater floodrise to a higher 100-year snowmelt rate. Using the equations and information in Hough and Hollis, an estimate of the 100-year return period 24 h snowmelt can be made at any reservoir site by using geographical variables to estimate the 5-year 24 h rate, and factoring this by the ratio of the 5-year and 100-year return period 24 h rates, which are given, or can be calculated, for the nearest relevant climate station used to calculate the snowmelt rate listed in Hough and Hollis. In such areas it would also be appropriate to consider any available data on local snowmelt rates in the derivation of reservoir flood inflow.

Climate change

The FEH methods, including the new DDF model, have been developed under the assumption of a stationary climate. Simple analyses of annual maximum rainfalls did not detect evidence of an upward trend (Stewart et al., 2013). The latest UK climate projections tool (UKCP09) suggests that, while the annual average rainfall may not change much over the 21st century, there may be changes in seasonal rainfall, with more winter rainfall falling in heavy events (LWEC, 2013). Research suggests that high seasonal rainfall extremes will increase as a result of climate change, with a corresponding increase in the biggest floods. Based on research using UKCP09, the Environment Agency has published factors quantifying the potential impacts that climate change may have on ‘extreme’ rainfall and river flood flows over different time periods differentiated by region (EA, 2011). It should be noted that ‘extreme’ in the context of fluvial flooding relates to much more frequent events than those considered for most reservoir safety categories. These factors could, however, be used in a sensitivity analysis in situations where works were being considered to address deficiencies in a reservoir’s flood-handling ability.

Rapid method of assessment based on FSR procedures

Previous editions of this guide have included a rapid method to assess the ability of an existing dam to withstand the T-year event or the PMF. This was developed to provide a quick and easyto-use preliminary screening method at a time when flood estimation software was not generally available. It was never intended to be used as an alternative to the full rainfall–runoff method. In particular, the rapid method provides only the inflow peak, and does not take into account important effects caused by the presence of the reservoir. Although Volume 4 of the FEH describes it as redundant owing to the accessibility and ease of use of modern flood estimation software, the rapid method has been retained in this edition primarily to facilitate the Tier 1 risk assessment procedure of the Guide to Risk Assessment for Reservoir Safety Management (EA, 2013). Appendix 1 sets out the steps in the rapid procedure, which includes the recommended methods for the estimation of flood surcharge and wave overtopping. For this exploratory procedure the design flood inflow may be computed as a fraction of the PMF, as indicated in Table 3.2.

Table 3.2 Design flood inflows as fractions of PMF for use only with the rapid assessment method Dam category

General standard design flood inflow

Equivalent fraction of PMF for rapid assessment only ( f )

A B C D

PMF 10 000-year 1000-year 150-year

1.0 0.5 0.3 0.2

19

Floods and Reservoir Safety

It is emphasised that the use of a fraction of PMF instead of the appropriate category T-year flood inflow is acceptable for the rapid method only. For complex or unusual reservoirs or catchment configurations such as reservoir cascade systems or reservoirs fed by urban drainage systems, the fractional PMF flood inflow is not recommended, even for screening evaluations.

Software

A number of software packages for applying the design rainfall–runoff methods described here are available. REFERENCES

Archer DR (1981) Severe snowmelt in the North East of England. Proceedings of the Institution of Civil Engineers, Part 2 71(2): 1047–1060. Archer DR (1983) Computer modelling of snowmelt flood runoff in N E England. Proceedings of the Institution of Civil Engineers, Part 2 75(2): 155-173. Archer DR (1984) The estimation of the seasonal PMF. British National Committee on Large Dams Conference, Cardiff, UK, pp. 1–20. Defra (Department for Environment Food and Rural Affairs) (2004) Floods and Reservoir Safety: Revised Guidance for Panel Engineers. Defra, London, UK. http://www.britishdams. org/reservoir_safety/defra-reports/200403Floods%20and%20Reservoir%20Safety%20-% 20Revised%20Guidance%20for%20Panel%20Engineers%20.pdf (accessed 8/10/2014). EA (Environment Agency) (2011) Adapting to Climate Change: Advice for Flood and Coastal Erosion Risk Management Authorities. EA, London, UK. https://www.gov.uk/government/ publications/adapting-to-climate-change-for-risk-management-authorities (accessed 8/10/2014). EA (2013) Guide to Risk Assessment for Reservoir Safety Management. EA, London, UK. Hough MN and Hollis D (2006) Rare snowmelt estimation in the United Kingdom. Meteorological Applications 5(2): 127–138. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK. Kjeldsen TR, Stewart EJ and Packman JC, Folwell SS and Baylis AC (2005) Revitalisation of the FSR/FEH Rainfall–Runoff Method. Department for Environment, Food and Rural Affairs and Environment Agency, London, UK, Report FD1913/TR. LWEC (Living With Environmental Change) (2013) Water Climate Change Impacts Report Card. http://www.lwec.org.uk/resources/report-cards/water (accessed 8/10/2014). MacDonald D E and Scott CW (2001) FEH versus FSR rainfall estimates: an explanation for the discrepancies identified for very rare events. Dams and Reservoirs 11(2): 28–31. NERC (Natural Environment Research Council) (1975) Flood Studies Report, vols I–V. NERC, Wallingford, UK. Reed DW and Field EK (1992) Reservoir Flood Estimation – Another Look. Department of the Environment, London, UK, IH Report 114. Stewart EJ, Jones DA, Svensson C et al. (2013) Reservoir Safety – Long Return Period Rainfall, vols 1 and 2. Flood Management Division, Department for Environment, Food and Rural Affairs, London, UK, Project FD2613 WS 194/2/39.

20

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.021

Chapter 4

Reservoir flood routing Objective

The purpose of this chapter is to comment on the decisions and assumptions that are required at the outset of a flood routing calculation, or when reviewing initial results. Routing calculations generally follow a standard procedure and produce g g g

the flood level attained in the reservoir the time lag between inflow and outflow peaks the spillway discharge hydrograph resulting from the inflow hydrograph.

Computations may be performed manually, but are more usually done by spreadsheet or by subroutines incorporated in specific and dedicated hydraulic modelling programs. Where valid approximations can be made, graphical solutions shorten routing computations. One of these was developed specifically to assist with the rapid assessment of floods at existing dams, and is given in Appendix 1. However, full calculations should be adopted when either a new or an enlarged spillway is to be constructed. In some cases, routing may be immaterial due to the small reservoir storage above the normal top water level compared with the expected flood magnitude. It is easy to check whether this will be so by reference to Figure 4.1, which assumes that the spillway discharge is of the form Q = CBH1.5, where Q is the flow in cubic metres per second, C is the weir coefficient, B the effective length of the weir in metres and H the head over the weir in metres. Typical values of C for a broad-crested weir are around 1.7, and for an ogee-crested weir can be as high as 2.0–2.2. S is the ‘storage ratio’ which can be obtained using the guidance in Appendix 1. Reference should be made to standard hydraulic design manuals and textbooks for more precise guidance.

Recommended stages in routing calculation

Recommendations covering choices in routing calculations are summarised below, and justified subsequently. (a) Initial reservoir level. These are given in Table 2.1 or, where appropriate, calculate antecedent conditions for more complex studies as discussed later in this chapter. (b) Draw-off, indirect catchments and releases during floods. Ignore draw-off and releases. For reservoirs with indirect catchments interrupted by aqueducts or tunnels it is assumed that these add inflow at the lesser of the full conveyance capacity or the appropriate indirect catchment’s runoff. For hydroelectric reservoirs, assume generators are unable to operate. (c) Reservoir level/discharge relationship. Incorporate any restriction on spillway discharge from the reservoir due to backing up from limited downstream culvert, bridge, channel, drop-shaft or tunnel capacities. Derive the relationship up to the level at which dam overtopping would begin (where relevant). At higher water levels, ratings with and without the overtopped section deemed to be passing flow are normally required. Special cases of gated, siphon and auxiliary spillways are discussed later in this chapter. The influence of trash or ice on open-crested spillways, gates or toppling weirs may need to be considered according to the nature of the catchment. Some owners install screens upstream of spillways, and consideration should be given to the possibility of these being blocked unless they can be shown to collapse clear on the differential head without the risk of partial or complete choking of spillways and channels. (d) Reservoir level/capacity relationship. Compute from a graph of the reservoir level against the area flooded (i.e. a depth/storage chart). (e) Reservoirs in cascade. The principles to be used for estimating and routing floods at reservoirs in cascade are summarised in the Flood Estimation Handbook (FEH ), 21

Floods and Reservoir Safety

Figure 4.1 Flood routing graph. (Source: Colombi and Hall, 1977) 1.0

0.9 Standard Average Annual Rainfall

Attenuation ratio R

0.8

3000 mm

0.7 2500 mm

2000 mm

0.6 1500

mm

1000

mm

0.5

0.4 500

mm

0.3 0

2

4

6 Storage ratio S

8

10

12

Volume 4, Section 8.3.2, ‘Multiple reservoir systems’ (IH, 1999). The procedure involves estimating the direct inflow into each reservoir and its routing and superposition with the direct inflow to the reservoir downstream, taking care to preserve the timing of successive contributions. In carrying out such calculations each reservoir should be checked by a tailored analysis using a design storm event appropriate to its entire catchment, also floods from different sub-catchments should only be combined when they have been derived from the same design storm. Reference should be made to the FEH for further details and also examples. Experience has shown that in most cases there is little difference in the results of routing of reservoir flood inflow whether the reservoir starts at the ‘just full’ level or spilling the long-term average flow. For this reason, it has been decided in this fourth edition to simplify matters and adopt the ‘just full’ condition for all categories in Table 2.1 as a general approach. However, where reservoir control procedures require, and discharge capacities permit, operation at or below specified levels defined throughout the year, these specified initial levels may be adopted for the safety check flood providing they are stated in the statutory certificates and/or reports for the dam. However, where a lower initial reservoir level is adopted because of specified drawdown rules, it is recommended that the dam be checked for compliance against the design flood with the reservoir ‘just full’. In general, it is not recommended that initial levels lower than ‘just full’ be adopted just because they have occurred regularly in previous years and not as a result of specific operational rules. They may not be a reliable indication of future conditions under major flood. 22

Reservoir flood routing

Antecedent flow from a preceding flood is important for large catchments where the design storm duration is 1 day or longer. For such large catchments, where insufficient is known about the spacing in time between storms, care must be taken to examine local records of long wet spells. Routing only individual, short, intense floods may be unsafe, as the main impact of antecedent conditions is on an initial level. If a flood routing calculation produces a very small outflow that is maintained for a considerable period, there is a risk that an alternative and more critical sequence of flood-producing events has not been identified. For this reason, a minimum spillway capacity is recommended in Chapter 2 as a safeguard against floods produced by wet periods of many days’ duration. The principal floods used in a standardised approach to reservoir safety are given in Table 2.1. They are the Probable Maximum Flood (PMF), 1 in 10 000-year, 1 in 1000-year and 1 in 150-year return period events. Where a risk-based assessment is undertaken, for example a Tier 2 or Tier 3 analysis as described in the Guide to Risk Assessment for Reservoir Safety Management (EA, 2013), other intermediate flood events are likely to require analysis. This may be to estimate the dam-critical flood at which failure would occur, or back-analysis of the flood required to achieve a given standard of downstream protection for the predicted and likely loss of life. Some guides are now also being produced where alternative reservoir levels and floods are combined to give predetermined risk values for different degrees of reservoir safety and dam stability. For example, the US Army Corps of Engineers’ (2005) Engineering Manual Stability Analysis of Concrete Structures covers the static stability analysis of all forms of hydraulic structures. In the case of dams, stability checks are required for extreme, abnormal and normal load cases and combinations of various conditions and floods including the PMF and the 1 in 300-year and 1 in 10-year flood events. In addition the ‘co-incident pool level’ is used, representing the ‘temporal averaged pool condition’ or the pool elevation that is historically equalled or exceeded 50% of the time (i.e. 182 days per year). Such approaches and alternative reservoir level and flood event combinations may be used in assessing reservoir safety and dam stability where they are taken from a structured approach to reservoir risk and provided they are from a respected, national dam safety body or international code, and provide the level of protection required in the UK, for example the Health and Safety Executive’s Reducing Risks, Protecting People: HSE’s Decision-making Process (HSE, 2001).

Gated spillways

Within the UK, the use of gated spillways has been limited mainly to Scotland. The long experience of SSE plc (previously known as Scottish & Southern Energy) covers as severe a climatic regime as any other UK reservoir owner may anticipate. Its gate reliability is high due to careful maintenance. Its target criterion is of individual flood gates being out of service for up to 6-month overhaul periods at intervals of approximately 30 years, timed to be in the months of probable low flood severity. At these times the other gates would remain available. Gatecontrolled hydropower dams frequently have a complex operating regime that is particularly important in basins subject to prolonged floods and containing several interconnected reservoirs. This, and the recommendation that checks should be made on potential flood levels with one gate immobilised, can require extensive routing studies, but these raise no new issue of principle. It is important that any reservoir flood-routing checks follow the actual gate-operating regime used by operators at the dam in question. Proactive control of discharges from dams during floods will limit the volume of floodwater that will be impounded. No reliance should be placed on such control unless deliberate and effective plans exist to ensure operational reliability under the already extreme circumstances of a rare storm. Site access and the availability of power or personnel for gate or valve operations can all be affected by such storms.

Siphon spillways, Siphon spillways can produce high discharges for relatively small initial changes in the reservoir automatic toppling level. This can be controlled by air regulation and moderated by placing several siphons at differweirs and fuse plugs ing levels to prime in sequence. Toppling (or tipping) weirs also have a suddenly increasing discharge characteristic, and likewise may be designed to tilt in sequence. Siphon spillways are often designed not to exceed the natural rates of flood discharge rise prior to dam construction, so as to add no additional hazard to downstream river users. 23

Floods and Reservoir Safety

The debris of automatic toppling weirs may accumulate to alter the discharge characteristics of the downstream river channel, and this may influence access or conceivably the reservoir discharge over the remaining fixed crest. The full activation of an erodible fuse plug spillway should be demonstrably fast in relation to the inflow hydrograph, and its operation should be reliable even under extremes such as ice cover or packed snow.

Auxiliary spillways

Where auxiliary spillways have been added to reservoirs, it has been to meet the recommended flood standards for the appropriate category. Experience shows that the nature of the site normally dictates the few available options, each of which can readily be tested by an appropriate routing calculation. Where a reservoir has been provided with an auxiliary spillway, which may have its crest at a higher elevation than the main spillway, this will need to be specifically checked to withstand wave overtopping, and then stillwater overflowing, for flood events up to and including the safety check flood.

Temporary upstream An increasing number of reservoired catchments are no longer natural but contain various constorage straints on the occurrence of intense flood runoff. Examples are engineered flood detention schemes and substantial railway and motorway embankments. Where these would temporarily impound significant quantities of water during a flood, and where they are assessed as being stable, they may be treated as if they were additional reservoirs to be incorporated in the routing. Where an existing spillway would be inadequate but for the protection afforded by upstream storage, this fact should be recorded in the inspecting engineer’s report in order to forewarn the supervising engineer should the storage be removed. The closure of railways and the removal of their embankments is a case in point. It should be borne in mind that railway embankments, which are relatively narrow and unsurfaced, may breach in a safety check or design flood event, in which case they can be more of a danger than a protection. REFERENCES

Colombi JS and Hall MJ (1977) A quick screening method for estimating the routing effect of a reservoir. Proceedings of the Institution of Civil Engineers, Part 2 63(4): 935–941. EA (Environment Agency) (2013) Guide to Risk Assessment for Reservoir Safety Management. EA, London, UK. HSE (Health and Safety Executive) (2001) Reducing Risks, Protecting People: HSE’s Decision-making Process. HSE, Bootle, UK. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK. US Army Corps of Engineers (2005) Stability Analysis of Concrete Structures. US Army Corps of Engineers, Washington, DC, USA, EM 1110-2-2100.

24

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.025

Chapter 5

Waves, wave overtopping and dam freeboard Scope

This chapter outlines a method for predicting waves on reservoirs, and (for typical dam crest configurations) to calculate wave overtopping discharges over the dam. The approach outlined here requires establishing fetch distances or areas over which wind can act on the reservoir surface to generate waves; deriving representative wind speeds, directions and durations; and then calculating the appropriate wave conditions. These waves are then combined with parameters describing geometry/properties of the upstream slope of the dam, including the wave wall (if present) to calculate mean wave overtopping discharges. The judgement of appropriate dam freeboard is then derived using this overtopping discharge, and knowledge of the condition and composition of the dam crest and downstream slope to assess the tolerable wave overtopping discharge, discussed further in Chapter 6.

Water surface and fetches

The first steps in making any simple prediction of waves on a given water area are to calculate fetch lengths or areas over which wind can act, and predict extreme wind speeds and directions; and thus calculate wave heights and periods. For the purposes of wave height prediction, the fetch is the maximum straight line over-water distance for a particular wind direction noting that waves can be generated at up to 458 either side of the main wind direction. Fetch lengths are assessed from the point of interest across the open water area to the far shore. For reservoirs that are relatively circular, fetch rays may be drawn at 308 increments to give a realistic resolution for predicted wind directions. For elongated reservoirs, where waves at the dam originate over a narrow range of fetches, or those where shoreline features require a smaller increment, the spacing of fetch rays may be narrowed. For embankment dams, the fetch is usually assessed from the lowest point of the dam crest, but for reservoirs with long embankments, or with wave-sensitive structures at other positions around the reservoir, fetches may be derived for more than one prediction point. On some reservoirs, it may also be necessary to calculate wave heights from a number of different (principal) wind directions. On reservoirs with complex shorelines, promontories or islands may interrupt theoretical fetch lines, apparently making them significantly shorter. But, in some cases, waves may relatively diffract around those promontories, so the theoretical interruption to fetch length should be ignored. For reservoirs that are curved or cranked, there is anecdotal evidence that winds may follow the valley/reservoir axis, so wave heights would be underestimated by using only straight-line fetch rays. For ‘banana-shaped’ reservoirs (Figure 5.1), the fetch rays might therefore be bent to follow the axis of the reservoir into the upper reaches. Note: While simple methods of wave estimation may use a single fetch distance, more complete methods identified by Herbert et al. (1995) may assess the contributions of a spread of fetches either side of the principal direction being considered.

Wind speed, duration and direction

A wind speed map taken from BS EN 1991-1-4 (BSI, 2005) and presented in Figure 5.2 shows the 1 in 50-year maximum hourly wind speed (U50 ) in metres per second, reduced to sea level. A number of adjustment factors are required to adjust these wind speeds for return period, altitude, direction and duration, and for over-water increases. 25

Floods and Reservoir Safety

Figure 5.1 Bent or ‘banana’ fetches Wind direction

Traditional method of measuring fetches

Wind direction

Longest extended fetch Modified method of measuring fetches

Nominal direction change line

Return period adjustment

Where the required return period for wave conditions is shorter (or perhaps longer) than 50 years, as may be required in Table 2.1, then the wind speed derived from Figure 5.2 should be adjusted using factors ( fT ) from Table 5.1.

Altitude adjustment

The required adjustment for altitude ( fA ) is obtained from the equation fA = 1.0 + (0.001 × alt) where ‘alt’ is the reservoir altitude in metres above sea level (source: BS EN 1991-1-4.).

Over-water adjustment

A further adjustment is also required that reflects the increased wind speed over open water as opposed to over land. This adjustment ( fW ) is obtained from Table 5.2, as a function of fetch length.

Duration adjustment

A typical UK reservoir will develop waves fully (wave heights reach equilibrium with the wind speed) in approximately 10–20 min, so the mean hourly wind speeds discussed above must be scaled up by a duration factor. For typical UK reservoir lengths (≤2 km), the duration factor fD may be estimated as fD = 1.05 Factors for other durations are given by Herbert et al. (1995).

Direction adjustment

In the UK, the strongest wind speeds (at a given return period) given in Figure 5.2 are generally from 240–2708N. In the absence of any local information, estimating wind speeds from other directions should use a reduction factor fN , for which generic factors are suggested in Table 5.3. These adjustments are intended for general use in the UK, but regional data on wind direction should be used for specific sites. Particular care should be taken for dams at the east end of a reservoir where winds off the 240–2708N direction might be funnelled along the reservoir/valley axis by steep-sided hills.

26

Waves, wave overtopping and dam freeboard

Figure 5.2 The 1 in 50-year hourly wind speeds (m/s) reduced to sea level. (After BSI (2005) National Annex (informative) to BS EN 1991-1-4)

31

30 29 28 27 26 25 24 Inverness

23

Oban

Dundee

Perth

Glasgow Edinburgh

26 Newcastle

25 Londonderry 24

Carlisle

Belfast

23 24

York Leeds

Preston

Kingston upon Hull

Liverpool

23

Manchester

22

Sheffield Nottingham

Stoke

Aberystwyth

Norwich

Leicester 21 Birmingham

20

Northampton

Ipswich

Bedford Swansea Cardiff

Oxford LONDON Bristol

Bournemouth

Brighton

Plymouth

Table 5.1 Wind speed ratios (fT) for selected return periods Return period: year

Mean annual

5

10

20

50

100

200

fT

0.79

0.85

0.90

0.95

1.0

1.04

1.08

Based on BSI (2005) National Annex (informative) to BS EN 1991-1-4

27

Floods and Reservoir Safety

Table 5.2 Wind speed adjustments (fW) over water Fetch length: m

1000

2000

4000

8000

12 000

fW

1.10

1.16

1.23

1.29

1.31

Source: Saville (1962) in Herbert et al. (1995)

Table 5.3 Values of the wind direction factor fN Direction: 8N clockwise

fN

0 (north) 30 60 90 (east) 120 150 180 (south) 210 240 270 (west) 300 330 360 (north) 0 (north)

0.78 0.73 0.73 0.74 0.73 0.80 0.85 0.93 1.00 0.99 0.91 0.82 0.78 0.78

Adapted from BSI (2005) National Annex (informative) to BS EN 1991-1-4

Required wind speed

Taking into account each of the adjustments above, the wind speed (U ) required to predict wave conditions should be calculated by multiplying the 1 in 50-year hourly wind speed derived from Figure 5.2 by each of the wind speed factors: U = U50 fT fA fW fD fN

Wave height and period prediction

Wind-generated waves on any open area of water contain a range of heights and periods. Such (random or irregular) waves may most usefully be described by the significant wave height Hs (average of the highest one-third of wave heights) and a mean wave period Tm . The recommended method to estimate wave heights is the simplified Donelan/JONSWAP method, for which Hs = 0.00178U(F/g)0.5 where Hs is the significant wave height in metres, U is the required wind speed in metres per second and F is the fetch in metres (not kilometres). (It should be noted that in coastal/ocean engineering there are at least three different definitions of and notations for significant wave height, including Hs , H1/3 and Hm0 . In inland reservoirs, however, these definitions are effectively the same, so no distinction is made within this guide, although alternative versions may be used in, for example, the EurOtop manual (Pullen et al., 2007).) For wave action on UK reservoirs, the peak wave period (Tp ) can be estimated from Tp = 0.0712F 0.3U 0.4 and the mean wave period (Tm-1,0 ) by Tm-1,0 = Tp/1.1 (The more complete method is described by Herbert et al. (1995).)

28

Waves, wave overtopping and dam freeboard

Figure 5.3 Simplified relationship between fetch length, wind speed and significant wave height 30 1.

6

1.

4

1.

2

25 0 1. 0. 8 7 0.

s:

m

0.

ve

he

5

a

w

fic

ni

4

Sig

t an

H ht

ig

6

15

0.

Required wind speed U: m/s

9 0. 0.

20

3

0.

200

2

1

0.

0.

10 100

300

400 500

700

1000 Fetch F: m

2000

3000 4000 5000

7000

10 000

15 000

A plot of the simple relationship between fetch length, wind speed and Hs is presented in Figure 5.3. The methods described for calculating wave heights in the previous section are illustrated for the reservoir in Figure 5.4, where the altitude is + 250 m ODN, and the representative wind speed is U50 = 22 m/s. Three fetch directions have been analysed, for which fetch distances have been measured from a suitable map. The various wind adjustment factors are then used to calculate significant wave height in Table 5.4. Figure 5.4 Example calculations of fetch length, wind speed and significant wave height N

0

F355 F3

Example calculation

355˚

70˚ 30˚

Reservoir F 70

Dam

29

Floods and Reservoir Safety

Table 5.4 Example wind and wave calculations for reservoir in Figure 5.4 Wind direction: 8N

Fetch length: m

fT

fA

fD

fW

fN

U: m/s

Hs: m

355 30 70

1000 700 1500

0.79 0.79 0.79

1.25 1.25 1.25

1.05 1.05 1.05

1.10 1.10 1.13

1.05 1.05 1.05

19.8 18.3 18.8

0.36 0.28 0.41

The critical fetch direction here is 708N, for which the significant wave height Hs = 0.41 m

Wave overtopping

One of the effects of wave action on the dam face (or on other structures attacked by waves) is that waves run up on the structure face, and only a proportion of the waves will overtop the structure. As waves are random in height and period, then the overtopping discharge over any chosen length of dam face will be variable in time, but the mean overtopping discharge (averaged over, say, 1000 waves) can be reliably predicted for many structure configurations using the overtopping prediction methods of the EurOtop manual (Pullen et al., 2007). The mean overtopping discharge so calculated will be substantially lower than any peak discharge (perhaps by 10–1000 times), and must never be taken as an estimate of an equivalent steady state discharge. Limits to the mean overtopping discharge suggested in Chapter 6 already include the natural difference between mean and peak discharges, so should not be further corrected. (So, for instance, if 10% of waves overtop, giving a discharge over all of the waves that averages at (say) 2 l/s/m, then the average discharge over only those waves that do overtop is 20 l/s/m. But, as the peak discharge may reach around 10 times the average, then the individual peak value might be 200 l/s/m. If the overtopping duration is (say) a quarter of a wave period, then a 4 s wave period gives a peak volume of 200 litres per metre run. The overtopping velocities might be of the order of 2–5 times the wave celerity, so for a 4 s wave might be 20 m/s. So, in summary, a mean overtopping discharge of 2 l/s/m may give a peak wave overtopping volume of 0.2 m3/m at 20 m/s.) Wave overtopping discharges depend on the dam structure configuration, the upstream face roughness, the local freeboard and the wave condition. For all structure crest levels, and chosen water levels, the freeboard (Rc ) is simply the vertical height difference between the structure crest level and the water level. The empirical prediction methods for overtopping are described in the EurOtop manual (Pullen et al., 2007), and online software is available for assessing mean overtopping discharges at http://www.overtopping-manual.com/calculation_tool.html, covering a wide range of structural variations. Two example methods are given in the following sections, presenting the basic empirical formulae for two idealised structure types: an idealised slope (with or without a small wave wall), and a tall vertical wall perhaps with a small berm or slope at its toe. Where the structure becomes more complicated and/or differs from either of these simplifications, then more complicated prediction methods may be needed. Empirical and neural network methods are discussed further in the EurOtop manual and at http://www.overtopping-manual.com/ calculation_tool.html. It should be noted that wave overtopping discharges vary exponentially with wave height and/or inversely with freeboard. Small changes in input variables will therefore often give apparently large changes, but, in reality, a change in discharge is not significant until it exceeds a factor of 1.5–2. At low overtopping, predicted discharges less than 0.001 l/s/m may be taken as no overtopping, and will often be reported as such in empirical calculations.

Vertical and steep walls (including large wave walls on submerged slopes)

For dams where the main control on overtopping is given by a vertical or very steep face, the mean overtopping discharge (q, in m3/s per m along the crest) may be predicted by the following:   q R 




= 0.04 exp −2.6 c Hs gHs3 valid for 0.1 , Rc/Hs , 3.5.

30

Waves, wave overtopping and dam freeboard

Embankment slopes (including small wave wall at top of slope)

Predicting mean overtopping discharges on embankment slopes requires the use of a wave breaking parameter jm − 1,0 = tan a/(Hs/Lm − 1,0 )0.5, where a is the slope of the upstream face of the structure in degrees from the horizontal and Lm − 1,0 is the (nominal) deep-water wave length Lm − 1,0 = gT 2m − 1,0/2p. Then, the mean overtopping discharge (again in m3/s per m along the crest) is given by q in the following:   q 0.067 Rc 




= √





gb jm−1,0 exp −4.75 jm−1,0 Hs gb gf gb gv tan a gHs3 with a maximum of   q Rc 




= 0.2 exp −2.6 Hs gf gb gHs3 where gb is the influence factor for a berm (dimensionless), taken as 1.0 on a simple slope; gf is the roughness factor, which depends on the composition of the slope/armour (Table 5.5); and gb is the influence factor for wave obliquity, taken as 1.0 for normal wave attack. The influence factor for a small wall on a slope gv may be calculated by

gv = 1.35 − 0.0078awall where awall is the angle of the wall or upper slope in degrees (between 458 for a 1 : 1 slope and 908 for a vertical wall). For the case of a simple uniform slope with no wave wall gv = 1.00. For the case of a vertical wall gv = 0.648. Some dams use plain or recurved wave walls to reduce wave overtopping. Recurved walls can be highly efficient relative to a simple vertical wall, provided that the water level is below the entry point to the recurve. For higher water levels, however, their efficiency reduces rapidly, and overtopping analysis should treat them as a simple vertical wall on a slope. Methods to predict the effect of wave walls are described in the EurOtop manual (Pullen et al., 2007). Simplified methods use the single additional factor gv defined in the previous paragraph. In the case of the water level being coincident with the top of the slope/base of the wave wall (and pending the publication of further studies on this scenario), an approximate result can be obtained by using the two simplified methods. More reliable results can be obtained by using the more complete methods described in the EurOtop manual. Guidance on acceptable (mean) wave overtopping discharges as a function of the crest and downstream protection are given in Chapter 6. In design work, the overtopping calculations described in this chapter may be reversed using an acceptable discharge as the target overtopping (q), and thus calculating the limiting freeboard (Rc ) from which a structure crest level may be derived. In some instances, it may be practicable to increase the roughness of the upstream face of the dam, perhaps by placing open rock armour (giving a lower value of gf ) in order to reduce the crest level needed for a given target discharge. Guidance on appropriate roughness factors for grassed slopes can be obtained from Equation 5.17/Figure 5.18 of the EurOtop Manual.

31

Floods and Reservoir Safety

Table 5.5 Roughness factor gf for wave overtopping calculations Description of upstream face protection

gf

Concrete (shown in the photograph), asphalt, closed concrete blocks, grass

1.0

Fitted basalt blocks

0.9

Open stone asphalt

0.9

Open concrete revetment blocks

0.9

One-quarter of stone setting 10 cm higher

0.9

Small blocks over 1/25 of surface

0.85

Small blocks over 1/9 of surface

0.8

32

Typical example

Waves, wave overtopping and dam freeboard

Table 5.5 Continued Description of upstream face protection

gf

Rock armour in one layer on a low permeability base

0.6

Rock armour in two layers on a low permeability base

0.55

Typical example

Adapted from the EurOtop manual (Pullen et al., 2007)

REFERENCES

BSI (British Standards Institution) (2005) National Annex (informative) to BS EN 1991-1-4: 2005, Eurocode 1 Actions on structures – Part 1-4: General actions – Wind actions. BSI, London, UK. Herbert DM, Lovenbury HTL, Allsop NWH and Reader RA (1995) Performance of Blockwork and Slabbing Protection for Dam Faces. HR Wallingford and Construction Industry Research and Information Association, Wallingford, UK, Report SR345. Pullen T, Allsop NWH, Bruce T, Kortenhaus A, Schuttru¨mpf H and van der Meer JW (2007) EurOtop: Wave Overtopping of Sea Defences and Related Structures: Assessment Manual. http://www.overtopping-manual.com (accessed 05/09/2014).

33

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.035

Chapter 6

The overflowing and overtopping of embankment dams Introduction

This chapter identifies available methods and gives guidance for assessing the potential for erosion of embankment dams due to overflowing and/or wave overtopping. If flood routing indicates that a flood would overflow the dam, or wave height determination indicates that waves are likely to overtop the dam, the implications of these flows should be assessed to determine whether the duration, depth, velocity and volume of flows are sufficient to be a cause for concern. Exceeding the expected erosion threshold might be tolerated for a short time if the embankment were built of or protected by erosion-resistant material. However, caution is required, and the occurrence of significant damage should be assumed unless there is evidence to the contrary. Masonry and concrete dams are considered to be erosion resistant providing that they are on non-erodible foundations, so are therefore not discussed further, other than to note that limiting criteria may be the hydrostatic loading and the effect of uplift on the overall stability of the structure. Where a concrete or masonry dam includes galleries or chambers within the crest, designers should consider how overflowing and overtopping flows could affect the stability of the dam. The difference between the threshold of erosion for a protected soil surface and an unprotected surface is significant. Once the unprotected soil is exposed to flowing water, its performance is, in relative terms, very much more unpredictable than a grassed, armoured or revetted surface. With cohesive soils, erosion usually proceeds by flow concentration and downcutting at weak spots. A gully then progresses back up the slope by head cutting. Erosion can be rapid due to local concentration of flow, caused by localized features (e.g. a post or stone protruding from the general surface profile, or a side wall). For this reason, the conditions in which soil is directly exposed to erosive forces above the threshold of erosion for any significant duration is generally considered to be the onset of failure, and therefore unacceptable.

Classification

To avoid confusion of the two different mechanisms, and to bring reservoir wave terminology in line with coastal wave terminology, the following distinction is made: g g

‘overflowing’ refers to (relatively) steady flows from flood rise ‘overtopping’ refers to intermittent flows from wave overtopping.

These may occur either separately or in combination. Stillwater overflowing is the residual flow of floodwater inflow that is above that accommodated by the spillway. Wave overtopping is the rise of water created solely by the run-up of waves not contained by the available freeboard Wave conditions can also impose large (dynamic) forces on the upstream face and crest. Guidance on the design of protection to the upstream face is given by Herbert et al. (1995) and McConnell (1998) and in the Rock Manual (CIRIA et al., 2007). Wave overtopping can also impose loadings on the downstream face of the dam, and can have the potential for causing more damage than by stillwater overflowing.

Physical factors affecting erodibility

Physical factors have a strong influence on the initiation and rate of erosion of embankment dams. Factors affecting the downstream face include 35

Floods and Reservoir Safety

g g g

g

g

g

g

g

The existing condition of the embankment and any protection. Grass cover should be regular, without localized patches or tussocks, which could initiate erosion. The embankment configuration, changes in slope angle, types and homogeneity of fill material. The maximum velocities of overflowing flows, which relate to the height of the dam, the downstream slope angle, the discharge rate and the hydraulic roughness of the slope. Discontinuities, cracks or voids on the downstream slope and the effect of appurtenant structures, particularly at the toe. Discontinuities such as berms, roads, structures, trees, shrubs, animal burrows and existing erosion features such as incised paths create local changes in the flow pattern, producing turbulence and erosion. The presence and depth of tailwater on the downstream slope. The energy of flow down the slope can be dissipated within the tailwater, thus reducing the erosion of the underlying toe foundation. The flow concentration at low points along the top of the embankment and the downstream slope, or in the mitres. The uniform distribution of flow over the top and down the embankment slope will discourage localized erosion. The erosion of toe drains formed by granular material, blanket drains and other areas of the dam that may be drained, and which may lead to undercutting of the more cohesive fill material. Excessive cutting or grazing of the grass, which may lead to a loss of cover, and vegetation maintenance including weed control may reduce resistance to erosion. Sustained animal grazing leading to the formation of tracks across the dam face can seriously reduce the erosion resistance of the structure.

As well as considering the potential for failure by surface erosion, if may be necessary to check the geotechnical stability of the embankment subject to overtopping or overflowing, under either saturated or drained conditions. In order to make qualitative assessments of the stability potential for the erosion of an embankment, knowledge of some of the relevant soil parameters can be helpful. These include the natural moisture content, the liquid and plastic limits, particle size gradings, and the shear strength and general nature of the fill. Erosion of an embankment with a granular downstream shoulder and an impermeable element such as a central core or upstream blanket may start to occur once the flood level exceeds the level of the core or blanket. Thus, the ability of the material above the top of the core or blanket needs to be assessed for its in situ permeability and erosion resistance. Once overtopping or overflowing occurs, seepage flow exiting on the downstream slope could initiate erosion. Cohesive embankments have generally proved to be more resistant to erosion than those constructed from granular fill for a given geometry and hydraulic condition. Infiltration into the subsoil of the slope is extremely variable, and in the case of a clayey material is unlikely to prove significant unless the period of overtopping is extensive or there is the potential for shearing off of a clay layer (D’Eliso, 2007). The depth to which the wetted surface advances within the fill will depend on the time for which overtopping/overflowing occurs, the moisture condition of the subsoil and the voids ratio. Often, erosion starts at the toe or where there is a marked transition in the embankment slope as a result of turbulent flow conditions, and then propagates upstream. This undercutting of the slope would cause surface material to collapse under shear failure conditions on the over-steepened slope. On the other hand, there are many incidences of erosion commencing near the top of the dam, with failure occurring before significant undercutting of the downstream toe.

Overflowing: assessment

36

It is important to recognise the range of failure mechanisms and related design criteria that can determine whether or not an embankment dam is able to withstand overflowing. Embankment dams constructed of highly erodible soils often have only minimal erosion resistance once any vegetation cover or other less erodible protection on the face has been lost. However, embankments constructed of rockfill or homogeneous cohesive material are likely to be more resistant to erosion from overflowing.

The overflowing and overtopping of embankment dams

Table 6.1 Overflowing flow and erosion zones Zone

Flow

Erosion

1

From a stillwater upstream energy head to a combination of static and dynamic head, proceeding from the still reservoir to a subcritical velocity state over the upstream portion of the dam crest. In this zone the flow velocities are small, and the small slope of the energy line imposes low tractive force on the surface of the dam crest even at deep flows

Only when the dam crest consists of highly erodible material will erosion occur. Such dams could, for example, comprise an unprotected crest of soil or gravel with poor grass cover

2

Through critical depth control on the crest and supercritical flow across the remainder of the dam crest to the downstream slope. Although the energy level remains similar to zone 1, the slope of the energy line increases, leading to a significant increase in the tractive force

Potential for erosion higher than zone 1, particularly at the downstream edge of the crest. An evaluation of the stability of any crest paving may also be required, as exposed edges are liable to lift off the top of embankments

3

Flow accelerates to high-velocity turbulent supercritical conditions on the steep downstream slope. Energy levels and velocities increase significantly down the slope until, if the slope is long enough, terminal velocity is achieved

The tractive forces are very large, and result in high erosion potential. If erosion is initiated, surface discontinuities may occur, resulting in a cascade flow

4

Supercritical flow conditions generally terminate in a hydraulic jump at the toe, with the downstream water level determined by flow control further downstream

High disturbance and complex flow patterns occur at, and downstream of, the hydraulic jump. Any protection or other physical features exposed above the mean surface level are particularly prone to localized attack

An understanding of embankment performance and erosive forces during overflowing is necessary in order to assess the potential for erosion. In the absence of waves, flow over an embankment with low or no tailwater can be expected to go through the four flow and erosion zones given in Table 6.1 and illustrated in Figure 6.1. Discharge over a uniformly level dam crest can be determined by use of the weir flow equation given in Chapter 4. If the crest has low areas, these will concentrate flow and lead to higher discharge per unit length. If the downstream face is of sufficient length to allow terminal velocity to be established, the velocity of the flow on the downstream face can be determined by iterative use of Manning’s formula: q = vd where q is the discharge per unit width (m3/s/m), v = d 2/3s1/2/n is the terminal velocity (m/s), d is the flow depth (m), s is the surface slope (equal to the energy line slope) and n is Manning’s roughness coefficient.

Figure 6.1 Hydraulic flow regime and erosion Zones. (Source: Hewlett et al., 1987) Flow regime

Supercritial

Subcritial

Erosion zones

1

2

Subcritial 4

3 Critical depth control

Upstream head

Normal depth

Crest acc Flow eler ate s

Hydraulic jump Un nor iform term mal d flow, ina epth l ve loci , ty

Downstream water level

Toe

37

Floods and Reservoir Safety

Figure 6.2 Frictional resistance of grass: recommended Manning’s roughness coefficients for grass slopes steeper than 1 : 10 0

n1

1i

0.030

.5

n7

Manning’s roughnes coefficient n

1i

Flatter than 1 in 10 use VR method

0.025

n5

1i

n4

(see Section 4.2.2 of CIRIA Report 116)

1i

n

n3

1i

er ep e t S in 3 1

tha

0.020 0

0.1

0.2 Embankment slope

0.3

0.4

For slopes steeper than 1 : 10, Manning’s n is independent of hydraulic loading and the grass length, and varies with the face slope, as given in Figure 6.2. Once the overflowing flows and velocities have been computed, it is possible to determine the hydraulic loading by the method presented in CIRIA Report 116 (Hewlett et al., 1987). It should be noted that this report contains embedded safety factors and is therefore more suited to initial design applications than to the reliability assessment of existing installations. The erosion resistance of grass only and composite surfaces with varying qualities of cover is shown in Figure 6.3. An alternative approach is to use a predictive dam breach numerical model such as HR Breach (Mohamed et al., 2002; Morris et al., 2008), MIKE11, AREBA (Van Damme et al., 2012) or EMBREA (Davison et al., 2013) that is able to calculate the overflow water surface velocity and the average shear stress, and hence erosion on the downstream face. While these methods are useful for embankment sections where the fill conditions are the predominant or limiting criteria and there is a need to estimate embankment damage likely to occur, they are less useful in determining the acceptable duration of overflowing.

Overflowing: remedial measures

In the event that an assessment of the stillwater overflowing flow reveals a potential for unacceptable erosion, preventative measures will be required (assuming that the alternatives of raising the dam crest or increasing the spillway capacity are not feasible). Depending upon the degree and period of overflowing anticipated, these measures may range from minor maintenance works to the installation of erosion-resistant surface protection. The following is a range of measures that could be considered: g g g g

38

the selected removal of trees and shrubs from the slope to avoid areas of turbulence and allow even grass cover to establish the establishment of an even grass sward filling of surface fissures or animal burrows, to reduce infiltration the filling of low areas or the removal of high areas on the crest or slope, to achieve a uniform flow depth across the face of the embankment

The overflowing and overtopping of embankment dams

Figure 6.3 Recommended limiting values for erosion resistance of plain and reinforced grass. (Source: Hewlett et al., 1987) 9 Concrete systems, good interblock restraint1 8

7 Other concrete block systems1

Limiting velocity: m/s

6 Open

mats 2

5 Filled

mats, 2

frabric

s3

Mes

4

hes

Plai

n gr

3

Plain Plain

2

ass

– go

od c

gras

ove

s–a

vera

gras

s–p

r

ge c

oor

over

cove

r

1

0 1

2

5

10 Time: h

20

50

Notes 1. Minimum superficial mass 135 kg/m2 2. Minimum nominal thickness 20 mm 3. Installed within 20 mm of soil surface, or in conjunction with a surface mesh 4. These graphs should only be used for erosion resistance to unidirectional flow 5. All reinforced grass values assume well-established good grass cover

g g g g g

maintenance of the grass cover formation of gentle transitions at changes in slope such as berms or toe, to reduce turbulent erosion protection of erodible gravel drains, to avoid the undercutting of fill material relocation of appurtenant structures on the crest or slope, to avoid turbulence around them installation of geotextiles, gabion or concrete armour layers on the crest, slope or mitres, to provide erosion resistance (CIRIA Report 116 (Hewlett et al., 1987) gives guidance on design methods and details to be adopted for the reinforcing of slopes, and includes comparative costs for geotextile and concrete armour systems).

Good management of the embankment slope can be expected to improve significantly the ability of grass cover to resist erosion. An even grass sward of height between 50 and 150 mm well anchored to the soil can be expected to provide satisfactory conditions for erosion resistance. To encourage deeper grass roots, the use of extreme amounts of fertiliser, pasturing with cattle/horses, and burning or very infrequent cutting without clipping removal should all be avoided. Regular mowing will help to inhibit tussock formation and the associated turbulent flow. If sheep pasturing is to be used, this should be carefully managed with the correct ratio of animals to area to achieve the desired grass height. 39

Floods and Reservoir Safety

Overtopping: assessment

Wave overtopping discharges can lead to the erosion of the crest, downstream face or abutments, and may cause a safety hazard to people or property close to the dam structure. These hazards can be important even if wave overtopping occurs from only a few per cent of the waves. It is vital to note that there are significant differences between the mean overtopping discharges that are most commonly calculated by prediction formulae and the peak intermittent flows that may cause danger to people, or loads on wave walls or downstream faces (see Chapter 5). In the past, wave overtopping discharges have been dealt with indirectly for UK reservoir design by a notional ‘wave surcharge allowance’ used to increase the freeboard. Overtopping prediction methods for a wide range of structures were substantially improved in 2007 with the publication of the EurOtop wave overtopping manual (Pullen et al., 2007). This incorporates the latest techniques and data from European and UK research on wave overtopping predictions, covering a wide range of sea and shoreline defence structures, and replaces the empirically based ‘Owen method’ as discussed by Yarde et al. (1996). The EurOtop manual is supported by a web-based calculation tool (www.overtopping-manual. com) that guides the user through the calculation of the mean overtopping discharge, by essentially the same methods described in Chapter 5. When using the online tool for dam assessment purposes, the ‘Probabilistic’ function and ‘Empirical Methods’ tabs should be selected, along with the geometry most suitable for the dam and values of the stillwater flood level, significant wave height (Hs ) and mean wave period (Tm ) as derived from the methodology in Chapter 5. Following the publication of the EurOtop manual in 2007, tests on the tolerance of sea dikes of wave overtopping were performed in 2007 and 2008 on real flood embankments in the Netherlands (van der Meer, 2008a,b, 2009). The first dike tested had a 1 : 3 downstream slope of fairly good clay with grass cover. The overtopping simulator was used to test the erosion resistance of this slope, simulating a 6 h storm at fixed mean discharges, each with appropriate variations wave by wave. Test conditions steadily varied the mean discharge from 0.1 l/s/m up to 50 l/s/m. After all of these simulated storms, the slope was in good condition and showed little erosion. Another test was performed on bare clay by removing the grass cover over the full slope to a depth of 0.2 m. Overtopping (mean) discharges of 0.1 l/s/m up to 10 l/s/m were performed for 6 h periods. Erosion damage started for the initial condition (0.1 l/s/m) and increased during the other simulations. After 6 h at 10 l/s/m there were two large erosion holes about 1 m deep, 1 m wide and 4 m long. This situation was considered as close to breaching. The overall conclusion of these tests was that clay with grass cover can be a highly effective erosion protection system for wave overtopping flows. Even without grass, good-quality clay can survive fairly extensive wave overtopping flows. There is currently limited published guidance available on safe overtopping discharges at UK dams, but based on guidance given in the EurOtop Manual for embankment seawalls, combined with recent physical testing results (van der Meer 2008a,b, 2009), the suggested limits for allowable mean overtopping discharges are summarised in Table 6.2.

Table 6.2 Suggested limits for allowable mean wave overtopping discharge on embankment dams Protection level provided

Allowable mean overtopping discharge q: l/s/m

Dam crest and downstream face are of good grass-covereda clay fill

1

Dam crest and downstream face are of bare clay fill or grass covered erodible fill, or poor grass coverb

0.1

Source: Hewlett et al. (1987) a Good grass cover is assumed to be a dense, tightly knit turf established for at least two growing seasons b Poor grass cover consists of uneven tussocky grass growth with bare ground exposed or a significant proportion of non-grass weed species

40

The overflowing and overtopping of embankment dams

Table 6.3 Hazardous levels of wave overtopping to individuals Hazard type and reason

Hazardous mean overtopping discharge q: l/s/m

Trained staff, well shod and protected expecting to get wet, overtopping flows at lower levels only, no falling jet, low danger of fall from walkway

1–10

Aware pedestrian, clear view of the reservoir, not easily upset or frightened, able to tolerate getting wet, wider walkway

0.1

Source: EurOtop manual (Pullen et al., 2007)

The physical factors described in the ‘Factors affecting erodibility’ section should also be assessed when considering resistance to wave overtopping. For exceptional cases, where the crest and rear slope are strongly protected, with very robust detailing at all interfaces where flows can cross or run along the junction, higher discharges might be allowed. For dams with the crest and rear slope well-protected, then the allowable mean overtopping discharge might be increased to 50 l/s/m. (NB – This implies good condition tarmac or concrete crest road and well established erosion protection system or rockfill on the downstream face, together with careful detailing, especially at the interfaces.) This is likely to be very rare. For those dams that have a well-protected crest, good crest detailing, and good grass cover on the downstream face, then the allowable mean overtopping discharge might be increased to 10 l/s/m. This is likely to be relatively rare. The safety of people accessing the dam crest during periods of wave overtopping should also be considered by dam owners. Table 6.3 shows the result of work in this area undertaken during the development of the EurOtop manual (Pullen et al., 2007).

Overtopping: remedial measures

Many of the strengthening measures recommended for overflowing may also be appropriate for wave overtopping, but it must be remembered that the overall dynamic and erosive effects of wave overtopping may be more severe than those resulting from stillwater overflowing when individual overtopping volumes are large and/or concentrated at low points along the top of the embankment. The extent of any damage will depend largely upon local factors, such as those described earlier. Wind-blown spray from a poorly designed or low wave wall (or other features that concentrate local overtopping flows) has caused severe damage at some dams. This particular aspect needs to be considered, especially where the mean overtopping rate approaches the limit or the top of the dam or the downstream shoulder material has poor erosion resistance (e.g. gravels or soils with low cohesion and poor grass cover). The effects of heavy local rainfall can sometimes cause damage to the downstream face of dams with poor erosion resistance, and this would be a pointer to the need for improvement works if wave overtopping is considered a potential problem. REFERENCES

CIRIA (Construction Industry Research and Information Association), Centre for Civil Engineering Research and Codes and Centre d’Etudes Techniques Maritimes et Fluviales (2007) The Rock Manual: The Use of Rock in Hydraulic Engineering, 2nd edn. CIRIA, London, UK. Davison M, Hassan M, Gimeno O, Van Damme M and Goff C (2013) A benchmark study on dam breach and consequence estimation using EMBREA and life safety model. 12th International Benchmark Workshop on Numerical Analysis of Dams, Graz, Austria. D’Eliso C (2007) Breaching of Sea Dikes Initiated by Wave Overtopping. A Tiered and Modular Modelling Approach. PhD thesis. University of Florence, Italy. Herbert DM, Lovenbury HTL, Allsop NWH and Reader RA (1995) Performance of Blockwork and Slabbing Protection for Dam Faces. HR Wallingford and CIRIA, Wallingford, UK, Report SR345. 41

Floods and Reservoir Safety

Hewlett HWM, Boorman LA and Bramley ME (1987) Design of Reinforced Grass Waterways. Construction Industry Research and Information Association, London, UK, Report 116. McConnell KJ (1998) Revetment Systems Against Wave Attack: A Design Manual. Thomas Telford, London, UK. Mohamed MAA, Samuels PG, Morris MW and Ghataora GS (2002) Improving the accuracy of prediction of breach formation through embankment dams and flood embankments. River Flow 2002. International Conference on Fluvial Hydraulics, Louvain-la-Neuve, Belgium. Morris MW, Hassan M, Kortenhaus A, Geisenhainer G, Visser PJ and Zhu Y (2008) Modelling breach initiation and growth. Flood Risk Management – Research and Practice. FLOODrisk 2008, Oxford, UK. Pullen T, Allsop NWH, Bruce T, Kortenhaus A, Schuttru¨mpf H and van der Meer JW (2007) EurOtop: Wave Overtopping of Sea Defences and Related Structures: Assessment Manual. http://www.overtopping-manual.com (accessed 05/09/2014). Van Damme M, Morris MW, Borthwick AGL and Hassan MAAM (2012) A rapid method for predicting embankment breach hydrographs. FLOODrisk 2012, Rotterdam, Netherlands. Van der Meer JW (2008a) Erosion Strength of Inner Slopes of Dikes against Wave Overtopping – Preliminary Conclusions after Two Years of Testing with the Wave Overtopping Simulator, V1.1 ComCoast Report to Rijkswaterstaat. Van der Meer JW (2008b) Coastal flooding: a view from a practical Dutchman on present and future strategies. Proceedings of the Conference on Flood Risk Management: Research and Practice. Taylor & Francis, London. Van der Meer JW, Schrijver R, Hardeman B et al. (2009) Guidance on erosion resistance of inner slopes of dikes from 3 years of testing with the wave overtopping simulator. Proceedings of ICE Conference on Coasts, Marine Structures and Breakwaters. Thomas Telford, London. Yarde AJ, Banyard LS and Allsop NWH (1996) Reservoir Dams: Wave Conditions, Wave Overtopping and Slab Protection. HR Wallingford, Wallingford, UK, Report SR 459.

42

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.043

Chapter 7

Floods during dam construction and dam improvement works Flood risks during new dam construction

Normally, dam construction requires some form of temporary river diversion to enable construction in the river bed. The choice of a flood against which to design such temporary works is a compromise between economy and safety. Usually during a construction flood the volume of water retained behind any cofferdam is relatively small, and so there is no need to estimate and route the complete hydrograph. Attention may generally be confined to the flood peak discharge alone. A higher risk may be tolerable during a limited construction period where it can be shown that any damage will be restricted to the construction site. However, consideration should be given to the volume that the upstream cofferdam is capable of impounding and whether this exceeds that at which the statutory requirements of reservoir safety legislation apply. The appropriate Flood Estimation Handbook (FEH) procedure for estimating the annual maximum flood peak with recurrence interval T years is to scale up an estimate of the ‘at-site’ median annual flood by a pooling-group ‘growth curve’ of the ratio of the T-year event to the median annual event (IH, 1999). At an ungauged site, the median annual flood is estimated from FEH catchment descriptors. However, this type of relationship provides only a coarse estimate of the median annual peak flood flow. Establishing a suitable compromise between economy and safety will involve calculating the size and effects on the dam site of varying return period floods and for various sizes and types of temporary works. A more formal risk assessment may be appropriate in particular cases. For example, financial risk balances can be made between the costs of protection against given return period events and the net cost multiplied by the probability of damage should those events occur. In assessing risk, concrete structures on foundations are more likely to remain secure during overtopping and be less susceptible to breaching than fill embankment structures. In such cases, overtopping may just lead to a need to pump out and clean up, and to programme delays rather than damage to permanent works. In other cases, overtopping may result in damage to permanent works and to more significant rebuilding costs and programme delays. Beyond a certain flood magnitude the cofferdam could fail, leading to greater damage occurring to the permanent works and the possibility of consequential damage downstream. This case could require dam break analysis to predict the effects of the breach. Potential breach characteristics can be assessed using equations such as those of Froelich for embankments. Downstream routing can be undertaken using standard proprietary river-modelling software. The key to establishing downstream damage limits will depend mostly on establishing realistic breach parameters. Rapid assessments of peak discharge, hydrograph shape and estimations of downstream routing for minor cases such as diversion works can be made using the approximate method outlined in CIRIA Report C542 (Hughes et al., 2000), which was later adapted for use within the Interim Guide to Quantitative Risk Assessment for UK Reservoirs (Brown and Gosden, 2004). Such an approach is probably suitable up to Tier 2-type risk appraisals, and is referred to in the Guide to Risk Assessment for Reservoir Safety Management (EA, 2013). For advanced breach and routing studies, recourse should be made to appropriate proprietary software. Where an arbitrary criterion is appropriate for the design of temporary works, it is sometimes considered acceptable to design for the flood that has only a 10% chance of being exceeded 43

Floods and Reservoir Safety

during the critical period of diversion work. The percentage probability of risk (Pr ) that a flood peak with a return period of T years would occur within the period of risk (r) (i.e. the construction period in years) is given as
 Pr 1 r =1− 1− 100 T The only proviso is that flood peaks are presumed to be independent random events. Evaluating this formula shows, for instance, that the flood corresponding to 10% risk over a 5-year construction period has a 48-year return period. Simplifying for practical purposes, for a 10% risk, the required return interval for a construction period flood peak is T = (period of risk (r) as whole years) × 10

Diversion structures

Temporary diversion works may well involve an upstream cofferdam that provides some storage at the inlet to a tunnel or culvert. However, it is not usual to compute flood peak reduction due to routing. Instead, the introduction of a small safety margin is preferred, since a solution relying on routing may give a false indication of precision. In some circumstances it is possible to design a structure to withstand erosive overflowing forces during construction and so avoid costly diversion works. For cases of moderate overflowing, reference can be made to the recommendations in Chapter 6 of this guide. Planned and significant overflowing of cofferdams can be attractive at narrow dam sites, and has frequently been adopted during the building of concrete dams. The armouring of a rockfill dam near Canberra (Australia) successfully withstood a peak flood of 1.8 m over it in 1976, where the specification required that wire mesh protection should be welded into position at the cessation of each day’s embankment construction. Non-erodible ‘soil cement’ or ‘hardfill’ cofferdams have also been used successfully in recent years.

Flood risks during Improvement works to dams may include modifications to spillways to meet requirements for improvement and/or updated reservoir inflow floods and for embankment raising to restore freeboard or increase stodam removal works rage capacity. The removal of dams may also present flood risks. In general, since the reservoir is unlikely to be completely emptied for most works that may be categorised as improvement, any works that involve a temporary reduction in the dam freeboard may increase the risk of overflowing or failure unless measures are taken such as lowering the retention level during the works. Similarly, repairs or reconstruction of spillways may also increase the risk of dam failure due to the passage of floods. If such works can be made to coincide with the normal drawdown period, the risk of reducing reservoir yield may be minimised. The time available may be assessed by reference to the reservoir operating rules. The risks to the general public depend on the scope of the improvement works, the volume of water remaining in the reservoir during such works and the dam category. The previous edition of this guide suggested designing for a 1% probability flood during critical construction periods; however, this is now regarded as too simplistic and possibly implying too great a downstream risk. It is recommended instead that the choice of flood during improvement works be subject to a failure mode analysis and risk assessment, with the risk of dam failure during critical periods shown to lie within the range of tolerability and reduced to ALARP – ‘as low as reasonably practicable’. During critical periods, both weather and upstream catchment conditions should be monitored, and a contingency plan put in place with appropriate warning levels and emergency actions.

Health and safety requirements

44

Reference should be made to the latest statutory regulations governing the safety of design and implementation of works before decisions are made as to the measures to be adopted to ensure safety. At the time of writing, such regulations may require the contractor to perform a risk assessment for the construction period based on the best available information and the designer may be responsible for supplying this. In all cases reference should be made to the UK Construction (Design & Management) Regulations 2007, or as subsequently amended.

Floods during dam construction and dam improvement works

REFERENCES

Brown AJ and Gosden JD (2004) Interim Guide to Quantitative Risk Assessment for UK Reservoirs. Department for Environment, Food and Rural Affairs. Thomas Telford. London, UK. EA (Environment Agency) (2013) Guide to Risk Assessment for Reservoir Safety Management. EA, London, UK. Hughes AK, Hewlett HWM, Morris M, Sayers P, Moffat I, Harding A et al. (2000) Risk Management for UK Reservoirs. Construction Industry Research and Information Association. London, UK, Report C542. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK.

45

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.047

Appendix 1

Rapid assessment of flood capacity and freeboard at existing dams Purpose

There is a continuing need for a rapid method of assessing the adequacy of an existing dam to withstand flood risk. The method is required to g g g

enable the inspecting engineer to make an initial appraisal indicate the dam’s margin of safety balance the cost of investigation and dam maintenance.

The 1933 report met this objective for a limited range of dams, with graphs giving flood intensity (flow per unit catchment area) and flood reduction factors based on the reservoir level rise and the proportion of the catchment inundated. Some rules were included to cover the minimum freeboard with and without a wave wall. This appendix offers a more comprehensive approach but requires only information that can be obtained from an Ordnance Survey map and the reservoir book (Prescribed Form of Record for a High-Risk Reservoir, ICE Publishing, 2014) and a knowledge of the average annual rainfall (from the Flood Studies Report (FSR) (NERC, 1975) or the FEH CD-ROM (CEH, 2009)).

Procedure

The procedure for rapid assessment is set out in Table A1.1 in the form of an annotated example for an embankment dam with a small vertical wall at the top of the upstream slope. The procedure comprises two components: g g

estimation of the flood surcharge (stillwater) level. estimation of the wave height and the mean wave overtopping discharge for the remaining flood freeboard.

It concludes with two checks: g g

that the flood surcharge (stillwater) level does not exceed the level of the top of the dam or the core that the remaining flood freeboard is sufficient to either contain the wave surcharge, such that no overtopping occurs, or that the overtopping discharge is tolerable given the type of dam and crest and downstream slope details.

REFERENCES

CEH (Centre for Ecology and Hydrology) (2009) FEH CD-ROM 3. Colombi JS and Hall MJ (1977) A quick screening method for estimating the routing effect of a reservoir. Proceedings of the Institution of Civil Engineers, Part 2 63: 935–941. NERC (Natural Environment Research Council) (1975) Flood Studies Report, vols I–V. NERC, Wallingford, UK.

47

Floods and Reservoir Safety

Table A1.1 Rapid assessment of flood capacity and freeboard worked example 1.0 Site information and data

Response

1.1 Site

One more reservoir

1.2 Grid reference

AB 123 456

1.3 Dam category (Table 2.1) Justification: Community at risk Other risk of life Economic losses

A/B/C/D

1.4 Inflow flood (Table 2.1)

Safety check/Design PMF/10 000 year/1 000 year/150 year return period

1.5 Dam type

rockfill/earthfill/concrete/masonry

1.6 Surface of dam crest

grass/tarmacadam/other

1.7 Downstream slope material

grass/reinforced grass/other

1.8 Upstream slope material Roughness factor (Table 5.5)

stone pitching 0.9

1.9 Upstream slope

1(V) : 3(H)

1.10 Top of dam level Nominal Measured low point Measured typical point

250.20 m AOD 250.00 m AOD 250.10 m AOD

Source: Drg 239/6 Source: Topo Survey 25/06/14

1.11 Wave wall top level Nominal Measured Details of gaps, etc. Base of wall level

250.50 m AOD 250.30 m AOD none 249.90 m AOD

Source: Drg 239/7 Source: Topo Survey 25/06/14

249.40 m AOD

Source: Drg 239/6 n/a

1.13 Spillway crest level Nominal Measured

247.68 m AOD 247.65 m AOD

Source: Drg 239/8 Source: Topo Survey 25/06/14

1.14 Spillway crest length, B Nominal (m) Measured (m)

31.00 m 31.00 m

Source: Drg 239/8 Source: Topo Survey 25/06/14

1.12 Core/level Nominal Confirmed

1.15 Spillway discharge coefficient, C Assume that the weir formula Q = CBH1.5 applies Value of C to be appropriate to the high-head situation

48

3

1.7

Source: Topo Survey 25/06/14

Rapid assessment of flood capacity and freeboard at existing dams

2.0 Catchment and rainfall data

Response

2.1 Catchment area, A (km2)

A = 11.76 km2 Source: Flood Estimation Handbook CD-ROM

2.2 Mainstream length, L (km) and slope, S1085 (m/km) Identify mainstream entering reservoir (blue line on 1 : 25000 OS Map) and measure length L (km). Estimate altitude at points 10% and 85% of length from lowest point on mainstream (H10 and H85). Slope is then: H − H10 S1085 = 85 0.75 × L

2.3 Average annual rainfall SAAR (mm) for 1941–70 standard period (can be obtained from FSR V maps or FEH CD-ROM)

3.0 Flood peak inflow 3.1 Peak of probable maximum flood (PMF) inflow, Qm (m3/s) Obtain Qm = 0.454A0.937S10850.328SAAR0.319 in which it is assumed that the catchment soils are impermeable and that there is no urban area in the catchment 3.2 PMF factor, f Select factor from Table 3.2 corresponding to appropriate inflow flood (from 1.4) 3.3 Standard inflow peak, Qi (m3/s) Qi = Qm × f

L = 2.75 km H10 = 270 m H85 = 443 m 443 − 270 S1085 = 0.75 × 2.75 S1085 = 83.88m/km SAAR = 1200 mm

Response Qm = 0.454 × 11.760.937 × 83.880.328 × 12000.319 Qm = 187.6 m3/s f = 1.0 Qi = 188 m3/s (rounded)

4.0 Estimated stillwater flood rise 4.1 Approximate mean reservoir area, a (m2) Estimate area at 0.5 h above weir crest (usually possible directly from the prescribed form of record or the Flood Estimation Handbook CD-ROM) 4.2 Head on spillweir assuming no attenuation, H (m) H = (Qi /CB)2/3 Spillway flow assumed to remain modular 4.3 Time-to-peak of unit hydrograph, TP (s) Obtain from either data or: TP = 0.67KA0.25 (for PMF only) or: TP = KA0.25 (for T-year flood or fraction of PMF) where K depends upon nature of catchment: K Nature 4600 mountainous 5300 hilly 5900 undulating 7900 flat This expression is a simplification of FSR Equation 6.15 in Volume 1 for catchments of typical shape and with area (A) , 100 km2 4.4 Storage ratio, S S = aH/QiTP, where a is obtained in step 4.1 above 4.5 Attenuation ratio, R Obtain from Figure 4.1 for S and SAAR value 4.6 Stillwater flood rise, h (m) h = RH 4.7 Stillwater flood level (m AOD) 1.14 + 4.6

Response a = 220 000 m2 H = [188/(1.7 × 31)]2/3 = 2.345 m

hilly TP = 0.67 × 5300 × 11.760.25 TP = 6576 s

S=

220 000 × 2.345 = 0.42 188 × 6576

R = 0.9 h = 0.9 × 2.345 h = 2.11 m 247.65 + 2.11 249.76 m AOD

49

Floods and Reservoir Safety

5.0 Wave surcharge Wave height 5.1 Fetch, F (m) Determine the longest single-line fetch and/or the longest fetch to the most exposed part of the dam 5.2 Design wind speed, U (m/s) Obtain the ‘at site’ 50-year hourly wind speed U50 from Figure 5.2 (reduced to sea level) Obtain the factor (fT) to reduce the 50-year wind speed to the mean annual maximum value from Table 5.1 Obtain top water level of reservoir from 1.13 (to nearest metre) Obtain factor (fA) to adjust wind speed to altitude of reservoir fA = 1.0 + (0.001 × altitude in metres) Obtain the factor (fW) to adjust wind speed to overwater value given a fetch, from Table 5.2. For all fetches less than 1000 m use fW = 1.10 Obtain the constant fD to adjust hourly wind speed to typical wave set-up time on a reservoir Identify principal axis of reservoir as 8N Obtain wind direction adjustment factor fN given principal axis of reservoir from Table 5.3 Obtain required wind speed U (m/s) U = fT × fA × fW × fD × fN × U50 5.3 Significant wave height, Hs (m) Hs = 0.00178U(F/g)0.5 (or from Figure 5.3)

F = 2000 m

U50 = 23 m/s fT = 0.79

TWL = 248 m AOD

fA = 1.0 + (0.001 × 248) fA = 1.248 fw = 1.16

fD = 1.05 8N = 210 degrees fN = 0.93 U = 0.79 × 1.248 × 1.16 × 1.05 × 0.93 × 23 U = 25.7 m/s Hs = 0.00178 × 25.7 × (2000/9.81)0.5 Hs = 0.65 m

Mean wave overtopping discharge 5.4 Mean wave period (Tm − 1,0) Calculate peak wave period, Tp Tp = 0.0712 F0.3 U0.4 F from 5.1, U from 5.2 Calculate mean wave period, Tm − 1,0 Tm − 1,0 = Tp/1.1 5.5 Calculate wave freeboard (Rc) Top of wavewall level From 1.11 Base of wavewall level From 1.11 Stillwater flood level (SWL) From 4.6 Top water level (TWL) From 1.13 Vertical distance between stillwater flood level (SWL) and dam crest or wavewall crest level

Tp = 0.0712 × 20000.3 × 25.70.4 Tp = 2.55 s Tm − 1,0 = 2.32 s 250.30 mAOD 249.90 mAOD 249.76 mAOD 247.65 mAOD Rc = 0.54 m

5.6 Choose correct mean overtopping flow eqn. Sloping face with small vertical wall 3

50

Sloping face (no wall) n/a

Vertical and steep wall n/a

Rapid assessment of flood capacity and freeboard at existing dams

For sloping upstream face with small vertical wall at top 5.7 Calculate nominal deepwater wavelength, Lm − 1,0 Lm − 1,0 = gT 2m − 1,0 /2p

Lm − 1,0 = 8.4 m

5.8 Calculate wave breaking parameter, jm − 1,0 jm − 1,0 = tan a/(Hs /Lm − 1,0)0.5 NB: cot a = z where z equals upstream slope in the form 1v : zh

jm − 1,0 = 1.20 z = 3 (upstream face slope)

5.9 Select influence factors Berm influence factor, gb Roughness influence factor, gf see Table 5.5 Wave obliquity influence factor, gb Wavewall influence factor, gv (Where: gv = 1.00 − 0.0078awall ) 5.10 Calculate mean overtopping discharge, q   




0.067 Rc q = gHs3 √





gb jm−1,0 exp −4.75 jm−1,0 Hs gb gf gb gv tan a

gb = 1 simple slope gf = 0.9 stone pitching gb = 1 assume normal gv = 0.648 vertical wave wall

q = 8.2E-04 m3/s/m q = 0.82 l/s/m

Check qmax  




3 q = gHs 0.2 exp −2.6

Rc Hs gf gb



5.11 Check q using www.overtopping-manual.com Input parameters used

Output result

qmax = 3E-02 m3/s/m qmax = 30 l/s/m q , qmax therefore OK Select: empirical, probabalistic, composite slope with (small vertical) wall Hs = Hm0 = 0.65 m Tp = 2.55 s Rc = 0.54 m Slope, z = 3.00 gf = 0.90 stone pitching q = 0.76 l/s/m

51

Floods and Reservoir Safety

For sloping upstream face

Not used in example

5.7 Calculate nominal deepwater wavelength, Lm − 1,0 Lm − 1,0 = gT 2m − 1,0 /2p

Lm − 1,0 = m

5.8 Calculate wave breaking parameter, jm − 1,0 jm − 1,0 = tan a/(Hs /Lm − 1,0)0.5 NB: cot a = z where z equals upstream slope in the form 1v : zh

jm − 1,0 = m z = (upstream face slope)

5.9 Select influence factors Berm influence factor, gb Roughness influence factor, gf see Table 5.5 Wave obliquity influence factor, gb Wavewall influence factor, gv (Where: gv = 1.00 − 0.0078awall ) 5.10 Calculate mean overtopping discharge, q   




0.067 Rc q = gHs3 √





gb jm−1,0 exp −4.75 jm−1,0 Hs gb gf gb gv tan a

gb = gf = gb = gv = 1.00 for continuous slope

q = m3/s/m q = l/s/m

Check qmax  




3 q = gHs 0.2 exp −2.6

Rc Hs gf gb



5.11 Check q using www.overtopping-manual.com Input parameters used

Output result

52

q = m3/s/m q = l/s/m Select: empirical, probabalistic, simple slope Hs = Hm0 = m Tp = s Rc = m Slope, z = gf = q = l/s/m

Rapid assessment of flood capacity and freeboard at existing dams

For vertical upstream face

Not used in Example

5.7 Not used 5.8 Not used 5.9 Not used 5.10 Calculate mean overtopping discharge, q   




R q = gHs3 0.04 exp −2.6 c Hs Check validity

5.11 Check q using www.overtopping-manual.com Input parameters used

Output result

q = m3/s/m q = l/s/m 0.1 , Rc /Hs , 3.5 Rc /Hs = Valid/Not Valid

Select: empirical, probabalistic Hs = Hm0 = m Tp = s Rc = m h = m (water depth at toe of wall) gf = q = l/s/m

53

Floods and Reservoir Safety

6.0 Conclusions

Response

6.1 Check provision of adequate dam freeboard Safety check flood conditions Mean wave overtopping discharge rate Compare with Tolerable overtopping rate or

0.8 1.00

l/s/m (item 5.10) ,0.001 l/s/m is considered to be zero l/s/m from Chapter 6/Table 6.2

Conclusion: Waves are/are not expected to overtop the dam in unacceptable quantities. Freeboard is adequate/not adequate

Design flood conditions (i) Check that waves will not overtop dam Mean wave overtopping discharge rate overtopping

l/s/m (item 5.10)

For no overtopping rate must be ,0.001l/s/m (ii) Check that min wave freeboard is met Height available for stillwater floodrise to top of dam/wave wall (highest) (A)

m (item 1.10 – item 1.13, or item 1.11 – item 1.13)

Stillwater floodrise (Item 4.6) (B)

m

Freeboard available to accommodate waves (A)–(B)

m

Min freeboard requirement (From Table 1 and dam category)

m Conclusion: Waves are/are not expected to overtop the dam. The minimum wave freeboard is/is not met. The freeboard is adequate/not adequate

6.2 Check retention of flood surcharge below top of dam level and core Safety check/design flood conditions (i) Height available for stillwater floodrise to top of dam crest (ii) Height available for stillwater floodrise to top of core

2.35 m (item 1.10 – item 1,13) 1.75 m (item 1.12 – item 1.13)

Stillwater floodrise (Item 4.6)

2.11 m Conclusion: Stillwater flood rise is/is not expected to exceed the level of the core Stillwater flood rise is/is not expected to exceed the level of the crest of the dam

Remarks: (a) Check permeability of material above level of core

Date of Asessment 01/07/12014 Assessment by: An Engineer Checked and approved by: Another Engineer

54

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.055

Appendix 2

Detailed advice on applying the FSR/FEH rainfall–runoff method to reservoir safety Introduction

This appendix clarifies the details of current methods to derive design flood hydrographs in the UK and discusses their applicability to reservoir flood inflow assessment. A set of procedures for estimating design flood hydrographs in the UK was originally published in the Flood Studies Report (FSR) in 1975 (NERC, 1975), and was recommended for reservoir flood safety assessment. The methods were subsequently refined in a number of reports in the Flood Studies Supplementary Report (FSSR) and Institute of Hydrology (IH) report series. The Flood Estimation Handbook (FEH), published in 1999, presented a new method for statistical flood frequency estimation but retained the FSR unit hydrograph and losses model, renaming it the FSR/FEH rainfall–runoff method (IH, 1999). Volume 4 of the FEH provides a comprehensive restatement of the basis of the method and its application to flood estimation, including reservoir flood safety assessment. This key reference incorporates much of the previous advice of the FSSRs and Institute of Hydrology reports, which are now superseded, and includes revised parameter estimation equations based on the FEH catchment descriptors. It also provides a number of worked examples to illustrate the recommended methods for estimating the T-year flood hydrograph and the Probable Maximum Flood (PMF).

FEH rainfall DDF model

The release of the FEH included a new model of rainfall depth–duration–frequency (DDF) intended to replace the FSR rainfall model, and this is available on the latest version of the FEH CD-ROM (CEH, 2009). However, the FEH analysis did not include a re-evaluation of probable maximum precipitation (PMP), and, although a number of research projects have considered alternative methods for PMP estimation, the only nationally applicable PMP model is that presented in the FSR. Other aspects of the design storm specification that remain unchanged are the areal reduction factors and symmetrical design profiles of the FSR. The FEH DDF model was developed for return periods of up to 1 in 2000 years only, and, following its release, concern was expressed about the results it produces for longer return periods. In some cases it has been noted that the FEH 1 in 10 000-year return period rainfall exceeds the estimate of PMP derived from the FSR. For this reason, and pending the release of a new FEH rainfall DDF model, the Department of Environment, Food and Rural Affairs (Defra) guidance to revert to the FSR rainfall model for the estimation of 1 in 10 000-year rainfall remains appropriate (see Chapter 3).

FEH13 rainfall DDF model

Research funded by Defra and the CEH has led to the development of a new rainfall DDF model, currently known as the FEH13 model, which provides rainfall frequency estimates for durations ranging from 1 h to 8 days and return periods from 1 in 2 years to over 1 in 10 000 years. The FEH13 model was developed through a complex statistical analysis of an extensive dataset of annual and seasonal maximum rainfall depths from rain gauges throughout the UK. The basic approach taken mirrored that of the FEH rainfall analysis, but with key revisions: (1) the standardisation of the rainfall maxima is now more complex, making the rainfalls at the different sites more similar prior to data pooling; (2) the model of spatial dependence has been revised; and (3) changes were made to the pooling methodology to overcome anomalous behaviour observed across a wide range of test cases. Full details of the statistical analysis are given by Stewart et al. (2013), and subsequent work has refined the structure of the model and generalised and smoothed the results across the UK on a 1 km grid. 55

Floods and Reservoir Safety

From comparisons with the FEH rainfall DDF model across the full range of durations and return periods, several notable features emerge. Firstly, the estimates from the new model are higher over most of Scotland at the shortest durations (,6 h). Secondly, the estimates from the FEH13 model tend to be lower than the FEH at higher return periods (.200 years), and this is thought to be due mainly to the improved model of spatial dependence. At extremely high return periods, estimated rainfalls from the FEH13 model are often considerably lower than the FEH model because the extrapolation of the new model is an approximate straight line on the Gumbel scale whereas the FEH model curves upwards (an exponential extrapolation). Finally, while FEH 1 in 10 000-year rainfall estimates commonly exceeded the FSR PMP, this is less often the case with estimates from the FEH13 model. As in the FEH analysis, the development of the FEH13 model did not include the development of new PMP estimates for the UK. Areal reduction factors and design rainfall profiles also remained outside the scope of the work. A new FEH web service is currently under development, and this will deliver the new FEH13 model results as well as the results of the FSR and FEH models. The web service will also provide access to the FEH catchment descriptors and will replace the existing FEH CD-ROM.

ReFH method

The revitalised flood hydrograph (ReFH) design method was developed to replace the FSR/FEH rainfall–runoff method for the synthesis of design flood hydrographs in the UK for return periods of up to 1 in 150 years. Details of the method are presented by Kjeldsen et al. (2005) and Kjeldsen (2007). The ReFH method uses design storm inputs from the existing FEH DDF model (Faulkner, 1999), although the areal reduction factors and storm profiles applied derive from the original FSR analysis. The method uses a design storm of equal return period to that of the required flood hydrograph. The ReFH model has been found to perform well on a range of different catchments, and therefore can be considered suitable for application to reservoir design inflow estimation for Category C (design flood) and D (safety check and design flood) dams, although limitations are known to exist when the method is applied in urban and permeable catchments. Ongoing research at the CEH is addressing these issues, as well as assessing the performance of the ReFH model in small catchments of less than 25 km2 in area. The method is currently being recalibrated to increase the upper limit of return period to 1 in 200 years. The performance of the ReFH method in estimating flood hydrographs of longer return period has not been tested. However, the structure of the design model includes a correction factor within the soil moisture accounting module, which ensures that the peak flow estimates match those derived from the FEH statistical approach based on pooling-groups. This has been found to cause a conceptual weakness in the model whereby the initial soil moisture parameter decreases as return period increases (Faulkner and Barber, 2009). However, the flood events used in the development of the ReFH model were more extreme than those used in the development of the FSR/FEH model. Further research on this issue is required.

Advice on detailed aspects of the FSR/FEH rainfall– runoff method

The inherent uncertainty in the estimation procedure for ungauged catchments would not normally warrant the subdivision of the catchment. However, the following circumstances are ones in which the normal approach to the representation of the catchment should be modified:

Catchment representation

g

g

If the reservoir area exceeds 5% of the direct catchment, the reservoir area should be excluded from the catchment area, and rainfall incident on the reservoir added explicitly, without losses, during the routing of the inflow hydrograph through the reservoir. If the reservoir extends into the upper reaches of the catchment, the mainstream length and slope should be measured from a stream to the edge of the reservoir and not to the catchment outlet, using a hypothetical stream if none is marked. (Note: this only applies to the rapid assessment method for flood estimation provided in Appendix 1 of this guide. An estimate of the mainstream length and slope, S1085, is no longer required when using FEH catchment descriptors in the full rainfall–runoff method).

The estimation of percentage runoff (PR) is probably the most important part of flood estimation using the FSR/FEH rainfall–runoff method. The PR is made up of a standard term, SPR, representing the normal capacity of the catchment to generate runoff, and dynamic terms representing the variation in runoff depending on the state of the catchment and the storm 56

Detailed advice on applying the FSR/FEH rainfall–runoff method to reservoir safety

magnitude itself. SPR can be obtained from gauged data, or can be estimated (SPRHOST) from the fraction of the catchment falling into each of the 29-class Hydrology of Soil Types (HOST) classification using the methodology of Boorman et al. (1995). The FEH catchment descriptor SPRHOST is available on the FEH CD-ROM (CEH, 2009), and this replaces the FSR procedure based on the WRAP (winter rainfall acceptance potential) maps. The standard and dynamic terms are calculated for a completely rural catchment, and an urban adjustment is subsequently applied. In some PMF calculations, the rural estimate of the PR can exceed the standard 70% value recommended for the PR from impervious surfaces. In such a case, it is advisable to omit the adjustment for urbanisation from the calculation for the PR. Alternatively, the 70% impervious value could be increased subjectively to a value suitably higher than the rural PR. In practice, the role of urban areas in a PMF calculation is very uncertain: urban runoff could actually be held back if the sewers are overloaded. In certain circumstances the SPR indicated by the FEH CD-ROM can indicate too low a value, and will require judgement as to whether it is representative of the catchment. An example of this for HOST soil type 4 that may be found in Pennine catchments is described in Davison (2005). Design storm

The storm duration that gives rise to the critical peak of the routed hydrograph is longer than that which is critical for the unreservoired catchment. The design storm recommended in the FSR is modified as follows for a reservoired catchment: D = (1 + SAAR/1000)(TP + reservoir lag) where D is the storm duration (hours), TP is the time-to-peak of the unit hydrograph (hours) and SAAR is the standard average annual rainfall (mm). Note that g g

g

Seasonal variation of the PMP

The reservoir lag is the time between the peak of the inflow hydrograph and the peak of the routed outflow hydrograph. The reservoir lag cannot be explicitly determined, so the critical storm duration is solved by iteration, starting with an assumed value of reservoir lag. The critical storm duration can change with different reservoir flood controls. TP (time-to-peak) of the PMF hydrograph is taken to be two-thirds of that of a return period hydrograph, to ensure a more severe condition.

Being able to separate summer and winter values of the PMP from the all-year value enables the refinement of adding the snowmelt to the winter PMP alone. All-year limiting values of the PMP are obtained from the FSR: a formal graphical factor method is used for durations between 2 and 24 h (FSR, Volume II, Section 4.3.4), and Table II.4.4 is used for values greater than 24 h. The alternative method of multiplying up the 5-year return period rainfall should not be used. Seasonal factors for PMP are based on estimates of seasonal factors for 100-year rainfall depths in the UK. Table 2.11 in the FSR, Volume II, presents seasonal factors for 5-year rainfall depths, and Table 3.9 indicates how 100-year factors are expected to relate to them. FSR Tables II.2.11 and II.3.9 are combined and scaled to give Table A2.1 in this guide. Further details are given in the FEH, Volume 4, Section 4.3.2. In separating the seasonal values, the all-year value is assigned to one of the seasons: the value for the other season is a fraction of the all-year value. The season in which the all-year value occurs is shown in Table A2.1 by assigning 100%: the percentage fraction is entered for the other season. Snowmelt must be added to the winter season PMP before the critical seasonal PMP can be identified.

Areal precipitation

Point rainfall is transformed to an equivalent reduced areal depth assumed to be uniform over the catchment. The FSR method for deriving catchment area rainfall is in two stages. In the first, estimates at several grid points within the catchment are averaged; in the second, this averaged point estimate is reduced to an areal rainfall based on catchment area and rain duration. Such a method may be appropriate for return period events, but may not hold for distributing a point estimate of the PMP over a catchment. There is agreement that the procedure overestimates the areal PMP, and is accordingly acceptable for use in the assessment of reservoir safety. 57

Floods and Reservoir Safety

Table A2.1 Seasonal variation in the PMP SAAR: mm

Winter PMP as % of all-year 1 h value

500–600 600–800 800–1000 1000–1400 1400–2000 .2000

SAAR: mm

1 min

2 min

5 min

10 min

15 min

30 min

13 15 19 26 30 33

17 19 24 32 38 42

21 24 30 40 47 53

24 27 35 47 55 61

26 30 38 50 59 66

30 33 42 57 67 74

Seasonal PMP as % of all-year value 1h

500–600 600–800 800–1000 1000–1400 1400–2000 .2000

SAAR: mm

2h

Summer

Winter

Summer

Winter

Summer

Winter

100 100 100 100 100 100

33 37 47 63 74 82

100 100 100 100 100 100

38 42 50 69 86 90

100 100 100 100 100 100

45 51 61 79 93 96

Seasonal PMP as % of all-year value 1 day

500–600 600–800 800–1000 1000–1400 1400–2000 .2000

6h

2 day

4 day

8 day

Summer

Winter

Summer

Winter

Summer

Winter

Summer

Winter

100 100 100 100 100 92

55 62 70 79 99 100

100 100 100 100 90 84

63 69 78 85 100 100

100 100 100 100 92 88

64 73 84 92 100 100

100 100 100 100 89 83

67 80 91 96 100 100

Source: FEH, Volume 4, Table 4.2 Note: (1) Snowmelt must be added to winter events; (2) summer = May–October, winter = November–April

Storm profile for large catchments

In catchments that significantly exceed 100 km2, the reservoir level can increase over a period of days as a consequence of a rapid succession of storms. In such cases a symmetrical storm profile of continuous rain is inappropriate. Current advice in the absence of guidance based on research is to adopt the temporal pattern of the most severe sequence of storms of the required duration that has been observed locally. The sequence with the most intense period at the end is generally the most critical case for a reservoir.

Reservoirs in cascade

Reservoirs in cascade are subject to the same design storm, and the inflow to each reservoir is influenced by the cumulative routing effect of all the reservoirs above it. Advice on flood estimation for reservoirs in cascade based on FSSR 10 is given in the FEH, Volume 4. This procedure yields the outflow hydrograph for the bottom reservoir in the cascade, and is applied successively to each reservoir in the cascade, treating each as the bottom reservoir. The procedure involves the estimation of the direct inflow to each reservoir, its routing and superimposition with the direct inflow to the reservoir below, taking care to preserve the timing of successive contributions. An essential feature of the procedure is that all the combined hydrographs are derived from the same design storm. The critical duration of the design storm extends as the procedure is applied to successively downstream reservoirs. When the storm is the PMP, it is possible for an upper reservoir, which was satisfactory when tested as the bottom reservoir, to fail when subject to the longer-duration

58

Detailed advice on applying the FSR/FEH rainfall–runoff method to reservoir safety

PMP storm. The reason is that the shorter storm, on which the reservoir passed, is nested in the longer storm. The reservoir fails because of the greater storm depth in the longer storm. This feature reflects the hypothetical nature of the PMP storm, and does not nullify the stand-alone assessment that should be adopted. Catchment descriptors

The presence of a reservoir, or cascade of reservoirs, can sometimes cause difficulties when determining some digital catchment descriptors. For example, if the reservoir extends well up the catchment, abstracting the mean drainage path length and slope to the dam may lead to a mean length that is too long and a mean slope that is too shallow, which may in turn lead to overestimation of the catchment response time. Similar problems in estimating the catchment response time may occur for the direct subcatchment to a lower reservoir in a cascade. In each case, the recommended guidance is to take appropriate catchment descriptors for the main tributary or a typical tributary to the perimeter of the reservoir, rather than to the dam site, for calculation of the unit hydrograph time-to-peak.

Local data

The preferred method of estimating the parameters of the FSR/FEH unit hydrograph and losses model is to use flood event or hydrometeorological data directly. However, even where the catchment is ungauged, estimates of model parameters from catchment descriptors can often be refined using information from hydrologically similar catchments.

Local rainfall analyses

Local rainfall should not normally be used to modify the design storm inputs derived from results of the rainfall models reported in the FSR and FEH because the length of additional record available is usually so short by comparison with that needed for proper definition of rainfall distributions.

Local unit hydrograph data

The unit hydrograph specific to the catchment may be estimated from observations of the catchment response or of the response of an adjacent catchment. In the absence of sufficient data to derive the unit hydrograph by analytical methods, it may at least be possible to refine the estimate of the time-to-peak, which can then be used to modify the unit hydrograph synthesised in the FSR procedure for ungauged catchments. Detailed advice is given in the FEH, Volume 4.

Local percentage runoff data

Where a site is gauged, the preferred method of deriving estimates of the SPR is by the analysis of observed flood events. The SPR can also be estimated from the baseflow index, from catchment descriptors and by transfer from a similar catchment. Details are given in the FEH, Volume 4.

Other factors Snowmelt

The FSR procedure assumes a melt rate of 42 mm/day (1.75 mm/h) for as long as the maximum snowpack of the 1 in 100-year depth can sustain it. The invariance of the melt rate with geographical location provides a less safe design in those snow-prone regions where the opportunity for a larger melt rate arises more frequently; hence the amended procedure suggested in Chapter 3 of this guide. The presence of lying snow should signal operational procedures that lower reservoir levels in anticipation of flood-producing rainfall. Such prudence would in part balance the uncertainty of any design allowance for snowmelt.

Antecedent flow

The general approach recommended in Chapter 2 is to assume the reservoir is just spilling at the start of the flood flow. Where antecedent flow conditions are judged as requiring assessment, the long-term mean daily flow can be practically assumed to be the catchment baseflow (BF), previously termed the average non-separated flow in the FSR, which more closely represents the higher value of the average winter flow in an upland catchment. On the comparatively few dam catchments exceeding 100 km2, a local study is warranted to check the consequences of adopting the BF rather than the lower mean daily flow. For floods extending over several days in a large basin, this antecedent flow should be allowed to decay to represent the discharge from saturated ground. If left at a constant value, too high a flood volume will be obtained.

Frozen ground

Catchment response in well-drained areas on permeable soils can change dramatically in the very unusual circumstance of heavy rain on frozen ground. When deriving a PMF from a winter 59

Floods and Reservoir Safety

PMP, frozen ground can be represented by assuming that the entire catchment acts as one of the more impermeable soil types. It is recommended that if the original SPR is less than 53%, then the frozen ground SPR is set to be 53%. If the original SPR is already greater than 53%, the value does not need to be amended. Computation time steps

Rainfall increments should be of sufficiently small duration to match the subsequent flood routing procedure. Computer studies will be based normally on time steps rounded down from the definition of 0.2TP in the FSR. Units of 30, 20 or 15 min suit typical UK reservoirs. There should be an odd number of time steps in the storm duration, to maximise the storm depth in the central interval.

Linking the flood frequency curve to the PMF estimate

Although flood estimation may seem to require only a single calculation using criteria obtained from Table 2.1 of this guide, it is common practice to produce estimates for a range of return periods and thereby obtain a flood frequency curve. There is usually a significant difference between the top of this curve (e.g. the 10 000-year flood) and the Probable Maximum Flood. There are two main reasons why it might be helpful to compute floods in this intermediate zone: g g

to provide a check on the 10 000-year flood for use in a risk-based approach, for example to estimate the dam-critical flood at which failure would occur, or for back-analysis of the flood required to achieve a given standard of downstream protection for the predicted and likely loss of life.

Methods of extending the flood frequency curve such that it makes a smooth connection with the PMF estimate are reported by Rowbottom et al. (1986) for Australia and by Lowing and Law (1995) for the UK. The Australian approach is based on geometry, allowing the form of the linkage between flood frequency estimates and the PMF to be influenced by the relative magnitude of the flows concerned and the slope of the frequency curve. The method of Lowing and Law assigns a nominal return period to the PMF and uses an interpolation procedure to reconcile this with the flood frequency curve. An example of the use of both techniques is given in the FEH, Volume 4. REFERENCES

Boorman DB, Hollis JM and Lilly A (1995) Hydrology of Soil Types: A Hydrologically-based Classification of the Soils of the United Kingdom. Institute of Hydrology, Wallingford, UK, Report 126. Davison I (2005) Concern over catchment run-off estimation. Dams and Reservoirs 15(1): 47. Faulkner DS (1999) Flood Estimation Handbook, vol. 2. Rainfall Frequency Estimation. Institute of Hydrology, Wallingford, UK. Faulkner DS and Barber S (2009) Performance of the revitalised flood hydrograph method. Journal of Flood Risk Management 2(4): 254–261. IH (Institute of Hydrology) (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK. Kjeldsen TR (2007) Flood Estimation Handbook. Supplementary Report No. 1. The Revitalised FSR/FEH Rainfall–Runoff Method. Centre for Ecology and Hydrology, Wallingford, UK. Kjeldsen TR, Stewart EJ and Packman JC, Folwell SS and Baylis AC (2005) Revitalisation of the FSR/FEH Rainfall–Runoff Method. Department for Environment, Food and Rural Affairs and Environment Agency, London, UK, Report FD1913/TR. Lowing MJ and Law FM (1995) Reconciling flood frequency curves with the probable maximum flood. British Hydrological Society 5th National Hydrology Symposium, Edinburgh, UK. NERC (Natural Environment Research Council) (1975) Flood Studies Report, vols I–V. NERC, Wallingford, UK. Rowbottom IA, Pilgrim DH and Wright GL (1986) Estimation of rare floods between PMF and the 100-year flood. Civil Engineering Transactions, Australia 28: 92–105. Stewart EJ, Jones DA, Svensson C et al. (2013) Reservoir Safety – Long Return Period Rainfall, vols 1 and 2. Flood Management Division, Department for Environment, Food and Rural Affairs, London, UK, Project FD2613 WS 194/2/39.

60

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.061

Appendix 3

Process diagrams Embankment dam Determine flood category of dam (Table 2.1) Derive the reservoir safety check flood and design flood inflows Route the safety check and design flood through the reservoir and obtain stillwater flood levels for each event Derive significant wave height for the mean annual maximum hourly wind speed Derive the mean wave overtopping discharge for the safety check flood and design flood levels Assess capacity of existing spillway to convey safety check flood outflow Does the reservoir meet FRS4 standards for both the design flood and safety check flood: a) design flood (i) is the mean wave overtopping discharge zero? (ii) does the dam meet the minimum flood freeboard requirement in Table 2.1? and b) safety check flood (i) is the stillwater floodrise below core, or if not, does material above it have sufficiently low permeability to prevent concerns over seepage? (ii) is the stillwater floodrise below dam crest? (iii) is the flood freeboard adequate to contain mean wave overtopping discharge within tolerable limits? (iv) does the spillway have sufficient capacity to prevent unacceptable damage to the embankment? Yes to all

No Consider potential failure modes and identify potential upgrade works up to and including those to meet FRS4 Table 2.1 requirements

Consider adopting a risk-based approach to assess the required level of flood protection (e.g. as described in the RARS guide). Does this indicate that the cost of carrying out works to meet FRS4 requirements is proportionate to the reduction in risk? If not, what works, if any, are required to reduce the risk of failure due to floods to “ALARP”?

Decide, in consultation with the owner, if works to meet FRS4 Table 2.1 are required, or a reduced level of protection is to be adopted, with the likleihood of failure reduced to ALARP. Undertake the appropriate works (if any).

Reservoir has adequate protection against failure from floods.

Note: The above diagram applies to reservoirs with direct catchments. There are specific requirements for reservoirs without direct catchments set out in Chapter 2, concerning fail-safe provision on inflows, direct rainfall on the reservoir, and wind speed for assessing freeboard to accommodate waves. For such structures users should follow the principles of the above but amend the process to suit.

61

Floods and Reservoir Safety

Concrete/masonry dam Determine flood category of dam (Table 2.1)

Derive the reservoir safety check flood and design flood inflows

Route the safety check and design flood through the reservoir and obtain stillwater flood levels for each event

Derive significant wave height for the mean annual maximum hourly wind speed

Derive the mean wave overtopping discharge for the safety check flood and design flood levels

Assess capacity of existing spillway to convey safety check flood outflow

Does the reservoir meet FRS4 standards for the safety check flood: (i) is the loading within structural design limits? (ii) is the stillwater level below dam crest/parapet wall, or if above them, has a rigorous assessment been undertaken to demonstrate resistance to erosion by overflowing, including around abutments? (iii) does the overflow spillway/stilling basin have sufficient capacity to prevent unacceptable damage to the toe area? (iv) is the extent of wave overtopping/stillwater overflow tolerable in terms of accessing and operating equipment critical to flood handling? Yes to all

No

Consider potential failure modes and identify potential upgrade works up to and including those to meet FRS4 Table 2.1 requirements

Consider adopting a risk-based approach to assess the required level of flood protection (e.g. as described in the RARS guide). Does this indicate that the cost of carrying out works to meet FRS4 requirements is proportionate to the reduction in risk? If not, what works, if any, are required to reduce the risk of failure due to floods to “ALARP”?

Decide, in consultation with the owner, if works to meet FRS4 Table 2.1 are required, or a reduced level of protection is to be adopted, with the likleihood of failure reduced to ALARP. Undertake the appropriate works (if any).

Reservoir has adequate protection against failure from floods.

Note: The above diagram applies to reservoirs with direct catchments. There are specific requirements for reservoirs without direct catchments set out in Chapter 2, concerning fail-safe provision on inflows, direct rainfall on the reservoir, and wind speed for assessing freeboard to accommodate waves. For such structures users should follow the principles of the above but amend the process to suit.

62

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.063

Glossary Conformity with the ICOLD Glossary of words and phrases related to dams has not been found practical; the glossary which follows should only be used in the context of this guide as some definitions have been deliberately restricted. Figure G.1 illustrates some of the principal terms. Auxiliary spillway

A secondary spillway designed to operate only during large floods.

Base flow (BF)

Long-term mean daily flow.

Crest of dam

As top of dam.

Design flood

The hydrograph of the flood inflow to the reservoir which produces the maximum stillwater level in the reservoir which the dam is required to accept under normal conditions with no damage and a safety margin provided by an amount of freeboard.

Emergency spillway

Auxiliary spillway.

EurOtop

EurOtop: ‘Wave Overtopping of Sea Defences and Related Structures: Assessment Manual’ published by the Environment Agency in 2008.

FEH

Flood Estimation Handbook published by the Institute of Hydrology in 1999.

Fetch

The maximum uninterrupted straight line over water distance for a particular wind direction: for ‘banana-shaped’ reservoirs the fetch may be bent to follow the axis of the reservoir into the upper reaches. The direction of fetch is always away from the dam.

Flood surcharge

The maximum rise of stillwater level above reservoir top water level (or retention water level) during a flood event. (Flood surcharge water is not retained in the reservoir but is discharged until the top water/retention water level is reached.)

Freeboard, dam

The vertical height from top water level (or retention water level) to the top of the dam. (Freeboard is required to contain flood surcharge, plus wave surcharge plus settlement allowance, plus any other pertinent factors.)

Freeboard, flood or Freeboard, dry or Freeboard, net

The vertical height from the flood surcharge level to the top of the dam. (Preferred term is flood freeboard – denoted Rc in Chapter 5.)

FSR

Flood Studies Report originally published by the Natural Environment Research Council in 1975.

Maximum water level

The maximum stillwater level (flood surcharge) of the safety check flood.

Overflowing flow

(Relatively) steady state flow over the crest or top of the dam, arising from the stillwater level in the reservoir being above the crest or top of the dam, as a result of the flood water inflow being greater than that which can be accommodated by the spillway.

Overtopping flow

Intermittent flows over the crest or top of the dam created solely by the run-up of waves not being contained by the available freeboard.

Percentage runoff (PR)

Made up of the SPR, representing the normal capacity of the catchment to generate runoff, and dynamic terms representing the variation in runoff depending on the state of the catchment and the storm magnitude.

Pitching

A layer of tightly packed irregular-shaped stone, rough dressed rectangular masonry or blocks, sometimes laid in courses, usually placed on a graded filter, providing protection to the upstream face against erosion by waves.

Probable Maximum Precipitation (PMP)

The (theoretical) greatest depth of precipitation for a given duration meteorologically possible for a given basin at a particular time of year. It includes rain, sleet, snow and hail as it occurs, but not snow cover left from previous storms. 63

Floods and Reservoir Safety

Probable Maximum Flood (PMF)

The flood hydrograph resulting from PMP and, where applicable, snowmelt, coupled with the worst flood-producing catchment conditions that can be realistically expected in the prevailing meteorological conditions.

Rainfall excess

Precipitation minus losses, i.e. equivalent to flood runoff volume over and above baseflow.

ReFH

Revitalised Flood Hydrograph Model.

Reservoir Flood Routing

The passage of a flood volume through a reservoir. Generally used to describe the calculation of the attenuation of the hydrograph of the incoming flood as it passes through storage and down the spillway.

Reservoir Lag (RLAG)

The time between the peak of the inflow and the peak of the outflow hydrographs.

Retention water level

As top water level.

Return period

The average expected time (in terms of probability rather than forecasting) between floods equal to or greater than a stated magnitude.

Rip-rap/rock armour

A loose, open layer of graded large random stones, or broken rock, usually placed on a graded filter, providing protection to the upstream face against erosion by waves.

Run-up

The maximum vertical height attained by a wave running up a dam face, relative to the stillwater level.

SAAR

Standard Average Annual Rainfall (Note that this is normally stated for a 30 year period for example 1941–1970, 1961–1990.

Safety check flood

The hydrograph of the flood inflow to the reservoir which produces the maximum stillwater level in the reservoir for which the dam is required to accept beyond which the safety of the dam cannot be assured, i.e. key components exhibit marginally safe performance for this flood condition.

Significant wave height

The average wave height, trough to crest, of the highest one third of waves.

SPR

Standard Percentage Runoff, representing the normal capacity of the catchment to generate runoff.

Stillwater level

The water level in the absence of any wave effects.

Storm losses

That part of precipitation which does not become runoff within the flood period because it has evaporated, infiltrated, been retained in the soil or been temporarily ponded on the catchment surface.

Storm profile

The magnitude and sequence of precipitation in equal time increments during a storm of given duration.

Top of dam

The top level of the dam structure. (Can be the top of the wave wall, if the wall is solid from abutment to abutment, without openings and considered able to withstand potential wave or water loading.)

Top water level

(a) For a reservoir with a fixed overflow sill, the lowest crest level of that sill; (b) for a reservoir from which the overflow is controlled wholly or partly by movable gates, siphons or other means, the maximum level at which water may be stored exclusive of any provision for flood storage. At this level the reservoir is ‘just full’ (see Table 2.1).

Unit Hydrograph (UH)

The runoff hydrograph resulting from unit volume of rainfall excess in a specified duration of time over a given catchment; the rainfall is presumed to fall uniformly or characteristically in both time and space on the catchment in the specified duration.

Wave, dam break

The surge wave of water released down valley by a dam that fails.

Wave overtopping

The intermittent passage of water over the crest of the dam created by waves not contained by the available freeboard, generally expressed as a mean overtopping discharge rate.

Wave surcharge

The rise of water against a dam created solely by the run-up of waves of specified probability.

Wave surcharge allowance

For the particular type and design of dam, the theoretical component of the flood freeboard sufficient to prevent overtopping by waves reaching quantities that could threaten the dam.

Wave wall

A solid wall built along the upstream side of the crest of a dam and designed to withstand or reflect waves.

64

Flood Surcharge

Flood Freeboard (or Dry Freeboard or Net Freeboard)

B: Showing Waves

Flood Surcharge

1/2 Wave Height

Top Water Level

Wave Surcharge (Wave Run-up)

A: Showing Freeboard

Design or Safety Check Flood Stillwater Level

Top Water Level (or Retention Water Level)

Design or Safety Check Flood Stillwater Level

Top of Dam (or Crest of Dam)

Note: Flood freeboard can extend into height of wave wall if wall is solid and considered able to withstand wave action

Dam Freeboard

Figure G.1 Ilustrating some principal terms

Crest of Spillway

Wave Wall

Crest of Spillway

Wave Wall

Glossary

65

Floods and Reservoir Safety ISBN 978-0-7277-6006-7 ICE Publishing: All rights reserved http://dx.doi.org/10.1680/frs.60067.067

Bibliography General

Acreman MC (1989) Extreme historical UK floods and maximum flood estimation. Journal of the Institution of Water and Environmental Management 3: 404–412. ANCOLD (Australian National Committee On Large Dams) (1986) Guidelines on Design Floods for Dams. ANCOLD, Perth, Australia. ASCE (American Society of Civil Engineers) (1974) Inspection, Maintenance and Rehabilitation of Old Dams. ASCE, New York, UK. Binnie and Partners (1989) Estimation of Flood Damage Following Dam Failure. Department of Environment Research Committee on Reservoir Safety, London, UK. Bossman-Aggrey P, Green CH and Parker DJ (1987) Dam Safety Management in the United Kingdom. Geography and Planning Paper No. 21. Middlesex Polytechnic, Hendon, UK. Cantwell B and Murley K (1986) Australian guidelines on design floods for dams. Water Power and Dam Construction 38(4–7): 4–7. Charles JA and Boden JB (1985) The failure of embankment dams in the United Kingdom. Proceedings of the Symposium on Failures in Earthworks. Thomas Telford, London, UK, pp. 181–202. Charles JA and Tedd P (1996) Bibliography of British Dams. Building Research Establishment, Watford, UK. (Revised 2013.) Charles JA, Tedd P and Warren A (2011) Lessons from Historical Dam Incidents. Environment Agency, Bristol, UK. Clarke CL and Phillips JW (1984) Discussion of Floods and Reservoir Safety Guide and the Reservoirs Act, 1975. Proceedings of the Institution of Civil Engineers, Part 1 76: 834–838. Cluckie ID and Pessoa ML (1990) Dam safety: an evaluation of some procedures for design flood estimation. Hydrological Sciences Journal 35(5): 547–569. (Discussion: 36(6): 499– 504.) Cooper GA (1987) The reservoir safety programme in Northern Ireland. Journal of the Institution of Water and Environmental Management 1(1): 39–51. Department for Environment, Food and Rural Affairs (2005) Floods and reservoir safety: revised guidance for panel engineers. Dams and Reservoirs 5(1): 15–16. Department of Environment, Transport and the Regions (2000) Guidance on application of Flood Estimation Handbook. Dams and Reservoirs 10(3): 35–38. FEMA (Federal Emergency Management Agency) (2004) Federal Guidelines for Dam Safety: Selecting and Accommodating Inflow Design Floods for Dams, reprint. FEMA, Washington, DC, USA. Fowler H (2003) Probable maximum precipitation and extreme rainfall estimation. Dams and Reservoirs 13(2): 27–28. Health and Safety Executive (2007) The Construction (Design and Management) Regulations 2007. The Stationery Office, Norwich, UK. ICOLD (International Commission on Large Dams) (1977) Report of Committee on Risks to Third Parties from Large Dams. ICOLD, Paris, France. ICOLD (2012) Bulletin on Safe Passage of Extreme Floods. Bulletin 142. ICOLD, Paris, France. Oosthuizen C and Elges HFWK (1987) Risk-based safety evaluation of dams. The Civil Engineer in South Africa 30: 555–559. Safety and Reliability Directorate (1985) A Feasibility Study into Probabilistic Risk Assessment for Reservoirs. Water Research Centre, Swindon, UK, Report ER188. Wellington NB (1988) Dam safety and risk assessment procedures for hydrologic adequacy reviews. Transactions of Institution of Engineers, Australia 30(5): 318–327. 67

Floods and Reservoir Safety

Hydrological

The Flood Studies Reports (FSR), Flood Studies Supplementary Reports (FSSRs), Flood Estimation Handbook (FEH), and related reports published by Institute of Hydrology and the Centre for Ecology and Hydrology. Boorman DB, Hollis JM and Lilly A (1995) Hydrology of Soil Types: A Hydrologically-Based Classification of the Soils of the United Kingdom. IH, Wallingford, UK, Report 126. Centre for Ecology and Hydrology (2009) Flood Estimation Handbook, CD-ROM 3. Centre for Ecology and Hydrology, Wallingford, UK. Davison I (2005) Concern over catchment run-off estimation. Dams and Reservoirs 15(1): 47. Faulkner DS and Barber S (2009) Performance of the revitalised flood hydrograph method. Journal of Flood Risk Management 2(4): 254–261. IH (Institute of Hydrology) (1983) A Guide to Spillway Flood Calculation for a Cascade of Reservoirs. IH, Wallingord, UK, FSSR 10. IH (1983) Some Suggestions for the Use of Local Data in Flood Estimation. IH, Wallingord, UK, FSSR 13. IH (1985) The FSR Rainfall–Runoff Model Parameter Estimation Equations Updated. IH, Wallingord, UK, FSSR 16. IH (1999) Flood Estimation Handbook, vols 1–5. IH, Wallingford, UK. (Also on CD-ROM.) Kjeldsen TR (2007) Flood Estimation Handbook. Supplementary Report No. 1. The revitalised FSR/FEH Rainfall–Runoff Method. Centre for Ecology and Hydrology, Wallingford, UK. MacDonald D E and Scott CW (2001) FEH versus FSR rainfall estimates: an explanation for the discrepancies identified for very rare events. Dams and Reservoirs 11(2): 28–31. NERC (Natural Environment Research Council) (1975) Flood Studies Report, vols 1–5. NERC, Wallingford, UK. Reed DW and Field EK (1992) Reservoir Flood Estimation – Another Look. IH, Wallingford, UK, Report 114. Winter C, Mason P and Stewart E (2012) The impact of the ‘‘New FEH’’ depth–duration frequency curves on extreme floods for the Nant-y-Moch Reservoir in mid-Wales. In Dams: Engineering in a Social and Environmental Context (Pepper A (ed.)). ICE Publishing, London, UK, pp. 406–412.

Storm rainfall

Bootman AP and Willis (1981) A Discussion on papers 4–6. Institution of Civil Engineers Conference on Flood Studies Report 5 years on, Manchester. Thomas Telford, London, UK, pp. 62–63. Dales NY and Reed DW (1989) Regional Flood and Storm Hazard Assessment. IH, Wallingford, UK, Report 102. Faulkner DS (1999) Flood Estimation Handbook, vol. 2. Rainfall Frequency Estimation. Institute of Hydrology, Wallingford, UK. Kjeldsen TJ, Prudhomm C, Svensson C and Stewart EJ (2006) A shortcut to seasonal design rainfall estimates in the UK. Water and Environment Journal 20(4): 282–286, 10.1111/ j.1747–6593.2006.00028.x. Stewart EJ and Reynard NS (1991) Rainfall profiles for design events of long duration. British Hydrological Society 3rd National Hydrology Symposium, Southampton, UK, pp. 4.27–4.36. Stewart EJ, Jones DA, Svensson C and Morris DG (2010) Reservoir Safety – long return period rainfall. In Managing Dams: Challenges in a Time of Change (Pepper A (ed.)). Thomas Telford, London, pp. 75–86. Stewart EJ, Morris DG, Jones DA and Gibson HS (2012) Frequency analysis of extreme rainfall in Cumbria, 16–20 November 2009. Hydrology Research 43(5): 649–662, 10.2166/ nh.2012.033. Stewart EJ, Jones DA, Svensson C et al. (2013) Reservoir Safety – Long Return Period Rainfall, vols 1 and 2. Flood Management Division, Department for Environment, Food and Rural Affairs, London, UK, Project FD2613 WS 194/2/39. Svensson C and Jones DA (2010) Review of rainfall frequency estimation methods. Journal of Flood Risk Management 3(4): 296–313, 10.1111/j.1753–318X.2010.01079.x. Svensson C and Jones DA (2010) Review of methods for deriving areal reduction factors. Journal of Flood Risk Management 3(3): 232–245, 10.1111/j.1753–318X.2010.01075.x.

Probable Maximum Flood

Archer DR (1984) The estimation of the seasonal PMF. British National Committee on Large Dams Conference, Cardiff, UK, pp. 1–20.

68

Bibliography

Austin BN, Cluckie ID, Collier CG and Hardaker PJ (1995) Radar Based Estimation of Probable Maximum Precipitation and Flood. Meteorological Office, Bracknell, UK. Collier CG, Morris DG and Jones DA (2011) Assessment of the return period of near-PMP point and catchment rainfall for England and Wales. Meteorological Applications 18(2): 155–162, 10.1002/met.191. Lowing MJ and Law FM (1995) Reconciling flood frequency curves with the probable maximum flood. British Hydrological Society 5th National Hydrology Symposium, Edinburgh, UK. Rowbottom IA, Pilgrim DH and Wright GL (1986) Estimation of rare floods between PMF and the 100-year flood. Civil Engineering Transactions, Australia 28: 92–105. US Committee on Techniques for Estimating the Probability of Extreme Floods (1988) Estimating Probabilities of Extreme Floods. National Research Council, Washington, DC, USA. Snowmelt

Archer DR (1981) Severe snowmelt in the North East of England. Proceedings of the Institution of Civil Engineers, Part 2 71(2): 1047–1060. Archer DR (1983) Computer modelling of snowmelt flood runoff in N E England. Proceedings of the Institution of Civil Engineers, Part 2 75(2): 155–173. (Archer DR and Boorman DB (1982) Informal discussion. Proceedings of the Institution of Civil Engineers, Part 1 76(3): 800–803.) Ferguson RI (1984) Magnitude and modelling of snowmelt runoff in the Cairngorm mountains, Scotland. Hydrological Sciences Journal 29(1): 49–62. Folland CK, Kelway PS and Warrilow DA (1981) The application of meteorological information to flood design. ICE Flood Studies Report: Five Years On. Proceedings of a Conference Organized by the Institution of Civil Engineers. Thomas Telford, London, UK. Hough MN and Hollis D (1997) Rare snowmelt estimation in the United Kingdom. Meteorological Applications 5(2): 127–138. Mawdsley JA, Dixon AK and Adamson AC (1991) Extreme snowmelt in the UK. British Hydrological Society 3rd National Symposium, Southampton, UK, pp. 5.17–5.22.

Joint probability studies

Anderson CW and Nadarajahs S (1993) Environmental factors affecting reservoir safety. In Statistics for the Environment (Barneti V and Turkman KE (eds)). Wiley, Chichester, UK, pp. 163–182. Anderson CW, Dwyer IJ, Nadarajah S, Reed DW and Tawn JA (1994) Maximum reservoir water levels. Reservoir Safety and the Environment: British Dam Society 8th Conference, Exeter, UK, pp. 200–213. Reed DW and Anderson CW (1992) A statistical perspective on reservoir flood standards. In Water Resources and Reservoir Engineering (Parr NM, Charles JA and Walker S (eds)). Thomas Telford, London, UK, pp. 229–239. Reed DW (1999) Joint probability problems. In Flood Estimation Handbook, vol. 1. Overview. Institute of Hydrology, Wallingford, UK, appendix B. Reed DW, Anderson CW, Tawn JA, Nadarajah S and Dwyer IJ (1999) Joint Probability Studies for Reservoir Flood Safety. Institute of Hydrology, Warrington, UK.

Flood routing

Colombi JS and Hall MJ (1977) A quick screening method for estimating the routing effect of a reservoir. Proceedings of the Institution of Civil Engineers, Part 2 63(4): 935–941. Houghton-Carr HA (1999) Flood Estimation Handbook, vol. 4. Restatement and Application of the Flood Studies Report Rainfall–Runoff Method. Institute of Hydrology, Wallingford, UK.

Climate change

Environment Agency (2011) Adapting to Climate Change: Advice for Flood and Coastal Erosion Risk Management Authorities. Environment Agency, Bristol, UK. https://www.gov.uk/ government/publications/adapting-to-climate-change-for-risk-management-authorities (accessed 8/10/2014). Kay AL, Crooks SM, Davies HN and Reynard NS (2014) Probabilistic impacts of climate change on flood frequency using response surfaces II: Scotland. Regional Environmental Change 14(3): 1243–1255, 10.1007/s10113–013–0564-x. LWEC (Living with Environmental Change) (2013) Water Climate Change Impacts Report Card. http://www.lwec.org.uk/resources/report-cards/water (accessed 8/10/2014). 69

Floods and Reservoir Safety

Wind, wave surcharge and dam freeboard

Birch KG and Ewing JA (1986) Observations of Wind Waves on a Reservoir. Institute of Oceanographic Sciences, Wormley, UK, Report 234. BSI (1991) BS NA EN 1991-1-4. UK National Annex to Eurocode 1. Actions on structures. General actions. Wind actions. BSI, London, UK. Carlyle WJ (1988) Wave damage to upstream slope protection of reservoirs in the UK. Proceedings of the British National Committee on Large Dams Symposium on Reservoir Renovation, Manchester, UK. CIRIA (Construction Industry Research and Information Association) (1976) Design of RIP-rap Slope Protection Against Wind Waves. CIRIA, London, UK, Report 61. CIRIA and CUR (Centre for Civil Engineering Research and Codes) (1991) Manual on the Use of Rock in Coastal and Shoreline Engineering. CIRIA Special Publication 83. CUR, Delft, Netherlands, Report 154. CIRIA, CUR and Centre d’Etudes Techniques Maritimes et Fluviales (2007) The Rock Manual: The Use of Rock in Hydraulic Engineering, 2nd edn. CIRIA, London, UK. Hasselman K (1973) Measurements of wind wave growth and swell decay during the Joint North Sea Wave Project (JONSWAP). Deutsch Hydrographische Zeitschrift A12: l–95. Herbert DM, Lovenbury HTL, Allsop NWH and Reader RA (1995) Performance of Blockwork and Slabbing Protection for Dam Faces. HR Wallingford and Construction Industry Research and Information Association, Wallingford, UK, Report SR345. Mackey PG (1985) Rehabilitation to Meet Reservoir Safety and Flood Criteria. International Commission on Large Dams Congress, Lausanne, Switzerland. Mackey PG (1988) Derwent Valley And Ladybower Dams. Flood Accommodation Works and Raising to Meet Probable Maximum Flood (PMF). International Commission on Large Dam Congress, San Francisco, CA, USA. Owen MW (1987) Wave Prediction In Reservoirs – A Literature Review. Hydraulics Research, Wallingford, UK, Report EX 1527. Owen MW and Steele AA (1988) Wave Prediction in Reservoirs – Comparison of Available Methods. Hydraulics Research, Wallingford, UK, Report EX 1809.

Overtopping of embankment, concrete and masonry dams

Charles JA (1984) Embankment dams and reservoir safety in Britain: floods, slides and internal erosion. British National Committee on Large Dams Conference, Cardiff, UK, pp. 51–68. Chen YH and Anderson BA (1986) Development of a Methodology for Estimating Embankment Damage Due to Flood Overtopping. Federal Highway Administration and Forest Service, Washington, DC, USA. CIRIA (Construction Industry Research and Information Association) (1987) Protection and Provision for safe Overtopping of Dams and Floodbanks. Report of Overseas Scientific and Technical Expert Mission. CIRIA, London, UK. Hewlett HWM, Boorman LA and Bramley ME (1987) Design of Reinforced Grass Waterways. Construction Industry Research and Information Association, London, UK, Report 116. Hughes AK and Hoskins CG (1994) A practical appraisal of the overtopping of embankment dams. Reservoir Safety and the Environment: British Dam Society 8th Conference, Exeter, UK, pp. 260–270. Johnson TA, Millmore JP, Charles JA and Tedd P (1999) An Engineering Guide to the Safety of Embankment Dams in the UK, 2nd edn. Building Research Establishment, Watford, UK. Kennard ME, Owens CL and Reader RA (1996) Engineering guide to the safety of concrete and masonry dam structures in the UK. Construction Industry Research and Information Association, London, UK. McConnell KJ (1998) Revetment Systems Against Wave Attack: A Design Manual. Thomas Telford, London, UK. Powledge GP, Ralston DC, Miller P, Chen YP, Clopper PE and Temple DM (1989) Mechanics of overflow erosion of embankments. II: Hydraulic and design considerations. ASCE, Journal of Hydraulic Engineering Division 115(8): 1056–1075. Pullen T, Allsop NWH, Bruce T, Kortenhaus A, Schuttru¨mpf H and van der Meer JW (2007) EurOtop: Wave Overtopping of Sea Defences and Related Structures: Assessment Manual. http://www.overtopping-manual.com (accessed 05/09/2014). US Army Corps of Engineers (2005) Stability Analysis of Concrete Structures. US Army Corps of Engineers, Washington, DC, USACE 1110-2-2100.

70

Bibliography

Yarde AJ, Banyard LS and Allsop NWH (1996) Reservoir Dams: Wave Conditions, Wave Overtopping and Slab Protection. HR Wallingford, Wallingford, UK, Report SR 459.

Dam breach analysis Davison M, Hassan M, Gimeno O, Van Damme M and Goff C (2013) A benchmark study on dam breach and consequence estimation using EMBREA and life safety model. 12th International Benchmark Workshop on Numerical Analysis of Dams, Graz, Austria. D’Eliso C (2007) Breaching of Sea Dikes Initiated by Wave Overtopping. A Tiered and Modular Modelling Approach. PhD thesis. University of Florence, Italy. Mohamed MAA, Samuels PG, Morris MW and Ghataora GS (2002) Improving the accuracy of prediction of breach formation through embankment dams and flood embankments. River Flow 2002. International Conference on Fluvial Hydraulics, Louvain-la-Neuve, Belgium. Morris MW, Hassan M, Kortenhaus A, Geisenhainer G, Visser PJ and Zhu Y (2008) Modelling breach initiation and growth. Flood Risk Management – Research and Practice. FLOODrisk 2008, Oxford, UK. Thompson G and Clark PB (1994) Rapid hazard ranking for large dams. Reservoir Safety and the Environment: British Dam Society 8th Conference, Exeter, UK, pp. 306–315. Van Damme M, Morris MW, Borthwick AGL and Hassan MAAM (2012) A rapid method for predicting embankment breach hydrographs. FLOODrisk 2012, Rotterdam, The Netherlands.

Risk assessment

Aboelata MA and Bowles DS (2005) LIFESim: A Model for Estimating Dam Failure Life Loss. Institute for Dam Safety Risk Management and Utah State University, Logan, UT, USA, draft report. Bowles D, Brown A, Hughes A, Morris M, Sayers P, Topple A et al. (2013) Guide to Risk Assessment for Reservoir Safety Management, vols 1 and 2. Environment Agency. Bristol, UK. Brown AJ and Gosden JD (2004) Interim Guide to Quantitative Risk Assessment for UK Reservoirs. Department for Environment, Food and Rural Affairs. Thomas Telford. London, UK. Escuder-Bueno I, Castillo-Rodriguez JT, Perales-Momparler S and Morales-Torres A (2011) SUFRI Methodology for Flood Risk Evaluation in Urban Areas. Decision Guidance for Decision Maker. SUFRI, University Graz, Graz, Austria, WP3. Graham WJ (1999) A Procedure for Estimating Loss of Life Caused by Dam Failure. U.S. Department of Interior. Bureau of Reclamation, DSO-99-06. United States of America. HR Wallingford, Flood Hazard Research Centre of Middlesex University and Risk & Policy Analysts (2006) Flood Risks to People, Phase 2: The Flood Risks to People Methodology. Department for Environment, Food and Rural Affairs and Environment Agency, London, UK, FD2321/TR1. HSE (Health and Safety Executive) (2001) Reducing Risks, Protecting People: HSE’s Decision-making Process. HSE, Bootle, UK. Hughes AK, Hewlett HWM, Morris M, Sayers P, Moffat I, Harding A et al. (2000) Risk Management for UK Reservoirs. Construction Industry Research and Information Association. London, UK, Report C542. Jonkman SN and Vrijling JK (2008) Life loss due to floods. Journal of Flood Risk Management 1(1): 43–56.

71