Flood Modelling Guidance for

Responsible Authorities
Version 1.1
1 Executive Summary
Developing our knowledge of flooding and how we can effectively assess potential impacts
on our communities is essential in implementing effective and sustainable flood risk
management plans. Flooding is a complex natural process arising from a range of sources
(e.g. rivers, sea, surface water/pluvial) and from a range of mechanisms. Using information
on the characteristics of floods we are able to assess the potential impact of floods using
computer simulations. Computer flood models are key tools in assessing, testing and
informing the delivery of flood risk management actions; while there will remain inherent
uncertainty in representing natural systems, quality models enable the production of flood
maps and data that support communication with the public and enable key policy and
investment decisions.

Utilising a consistent framework to guide model development will support a common

understanding and effective communication of flood study needs within and between
organisations. Ensuring a clear description of model needs enables models to be developed
appropriately to deliver study objectives. Thus, with an improved knowledge base, common
language and improved communication, model quality can be raised to improve confidence
in information that empowers decision-makers to act in support of flood risk management.

This modelling guidance document is therefore intended to provide a common technical

basis to support Responsible Authorities in the planning, development and use of flood
models to inform flood risk management decisions. As flood modelling science is constantly
evolving, this guidance will be revised as our understanding improves. SEPA’s strategic
flood risk team (strategic.floodrisk@sepa.org.uk) welcome any questions, comments or
suggestions for improvement regarding this guidance.

Flood Modelling Guidance for Responsible Authorities v1.1 1

Contents

1 Executive Summary ....................................................................................................... 1

2 Introduction .................................................................................................................... 6
2.1 Purpose of the Flood Modelling Guidance ............................................................. 6
2.2 How to use this guidance ....................................................................................... 8
2.3 SEPA input into modelling studies ......................................................................... 9
3 Scoping ........................................................................................................................ 10
3.1 Introduction.......................................................................................................... 10
3.2 Define Modelling Objective .................................................................................. 10
3.3 Developing a Conceptual Model .......................................................................... 10
3.4 Spatial Extent ...................................................................................................... 12
3.5 Joint Studies ........................................................................................................ 14
3.6 Level of Assessment............................................................................................ 17
3.7 Modelling Approach ............................................................................................. 17
3.7.1 Selection of Modelling Software ..................................................................... 20
4 Commissioning a study ................................................................................................ 22
4.1 Introduction.......................................................................................................... 22
4.2 Statement of Requirements (SoR) ....................................................................... 22
4.3 Quality Control ..................................................................................................... 22
4.4 Resources ........................................................................................................... 22
4.5 Timescale/Milestones .......................................................................................... 23
4.6 Risk Register ....................................................................................................... 23
4.7 Meetings and Progress Reports .......................................................................... 23
4.8 Intellectual Property Rights .................................................................................. 23
5 Data ............................................................................................................................. 24
5.1 Introduction.......................................................................................................... 24
5.2 General Considerations ....................................................................................... 24
5.3 Data collection ..................................................................................................... 24
5.3.1 Data registers ................................................................................................ 24
5.3.2 Licensing ....................................................................................................... 25
5.4 Historic Flooding Information ............................................................................... 25
5.4.1 Use of data .................................................................................................... 25
5.4.2 Available Datasets ......................................................................................... 25
5.5 Topographic and Bathymetric Data...................................................................... 27
5.5.1 Reuse of existing data ................................................................................... 27
5.5.2 Commissioning a New Survey ....................................................................... 28
5.5.3 Data Format ................................................................................................... 28
5.5.4 Channel Survey ............................................................................................. 28
5.5.4.1 Commissioning a New Channel Survey ................................................... 28
5.5.4.2 Cross-Sections......................................................................................... 29
5.5.4.3 Hydraulic Structures ................................................................................. 30
5.5.4.4 Flood Defence Survey .............................................................................. 30
5.5.5 Digital Elevation Models (DEMs) .................................................................... 34
5.5.5.1 Existing Data Sets .................................................................................... 34
5.5.5.2 New Data Collection................................................................................. 35
5.5.6 Bathymetric Data ........................................................................................... 38
5.5.6.1 Commissioning New Bathymetric Surveys ............................................... 38
5.5.6.2 Coastal Defence Survey .......................................................................... 38
5.6 Operating Information .......................................................................................... 38
5.7 Existing Studies ................................................................................................... 39
5.8 Hydrometric Data................................................................................................. 40
5.8.1 River Levels and Flows .................................................................................. 40
5.8.2 Rainfall Data .................................................................................................. 42
5.8.3 Soil Moisture and Evaporation Data ............................................................... 43

Flood Modelling Guidance for Responsible Authorities v1.1 2

 5.8.4 Reservoir Data ............................................................................................... 43
5.8.5 Sea Level Data .............................................................................................. 43
5.8.5.1 Gauge Data.............................................................................................. 43
5.8.5.2 Tide Tables .............................................................................................. 44
5.8.5.3 Sea level boundary conditions ................................................................. 44
5.8.6 Wave Data ..................................................................................................... 44
5.8.7 Wind data ...................................................................................................... 45
5.9 Other Data ........................................................................................................... 45
5.9.1 Sewer Network Data ...................................................................................... 45
5.9.2 National Coastal Change Assessment ........................................................... 45
5.9.3 Water Framework Directive Status/Pressures ................................................ 45
5.9.4 Mapping Data ................................................................................................ 45
6 Boundary Conditions .................................................................................................... 47
6.1 Introduction.......................................................................................................... 47
6.2 Terminology......................................................................................................... 47
6.3 Fluvial .................................................................................................................. 48
6.3.1 Review of Hydrometric Data .......................................................................... 50
6.3.2 Design Flows ................................................................................................. 50
6.3.2.1 Statistical Method (single site and pooled analysis).................................. 51
6.3.2.2 Design Event Method (Rainfall-Runoff Method) ....................................... 51
6.3.3 Hydrograph Shapes ....................................................................................... 52
6.3.4 Catchment Schematisation/Boundary Locations ............................................ 53
6.3.5 Flow Reconciliation ........................................................................................ 53
6.3.6 Climate Change ............................................................................................. 54
6.4 Pluvial .................................................................................................................. 54
6.4.1 Rainfall Models .............................................................................................. 54
6.4.2 Losses ........................................................................................................... 55
6.4.3 Climate Change ............................................................................................. 56
6.5 Coastal ................................................................................................................ 58
6.5.1 Still Water Boundary Conditions..................................................................... 59
6.5.2 Waves............................................................................................................ 62
6.5.2.1 Offshore design wave conditions.............................................................. 62
6.5.2.2 Joint probability of extreme still water levels and extreme wave conditions
63
6.5.2.3 Wave transformation modelling ................................................................ 63
6.5.2.4 Wave Overtopping ................................................................................... 64
6.5.3 Climate Change ............................................................................................. 64
6.6 Joint Probability ................................................................................................... 64
6.6.1 Extreme Sea Level and Fluvial Flows ............................................................ 64
6.7 Groundwater ........................................................................................................ 65
6.8 Uncertainty .......................................................................................................... 65
7 Model Schematisation .................................................................................................. 66
7.1 Introduction.......................................................................................................... 66
7.2 1D fluvial Models ................................................................................................. 66
7.2.1 Channel representation.................................................................................. 67
7.2.1.1 Cross Sections ......................................................................................... 67
7.2.1.2 Roughness ............................................................................................... 68
7.2.1.3 Hydraulic Structures ................................................................................. 68
7.2.2 Floodplain Representation ............................................................................. 70
7.2.3 Common Problems with 1D Models ............................................................... 72
7.3 2D models ........................................................................................................... 73
7.3.1 Model Resolution ........................................................................................... 74
7.3.2 DTM............................................................................................................... 75
7.3.2.1 Building representation ............................................................................ 75
7.3.2.2 Hydraulic Structures and Linear Features ................................................ 76

Flood Modelling Guidance for Responsible Authorities v1.1 3

 7.3.2.3 Roughness ............................................................................................... 77
7.3.3 Numerical precision ....................................................................................... 77
7.3.4 Common Problems with 2D Models ............................................................... 77
7.4 1D-2D models...................................................................................................... 78
7.4.1 Common Problems with 1D- 2D Models ........................................................ 78
7.5 Boundary Types and Locations ........................................................................... 79
7.5.1 Fluvial Models (1D and 2D)............................................................................ 79
7.5.2 Pluvial Models (2D) ........................................................................................ 79
7.5.3 Coastal Models (2D) ...................................................................................... 79
7.6 Initial Conditions .................................................................................................. 79
8 Calibration, Validation & Sensitivity Analysis ................................................................ 81
8.1 Introduction.......................................................................................................... 81
8.2 Calibration ........................................................................................................... 81
8.2.1 Fluvial ............................................................................................................ 83
8.2.2 Coastal .......................................................................................................... 85
8.2.3 Pluvial ............................................................................................................ 86
8.3 Common Issues with Calibration ......................................................................... 86
8.4 Sensitivity Analyses ............................................................................................. 87
8.5 Confidence .......................................................................................................... 89
9 Scenarios ..................................................................................................................... 92
9.1 Introduction.......................................................................................................... 92
9.2 Fluvial .................................................................................................................. 93
9.3 Coastal ................................................................................................................ 93
9.4 Pluvial .................................................................................................................. 93
9.5 Integrated Catchment Studies ............................................................................. 94
10 Climate Change........................................................................................................ 95
10.1 Introduction.......................................................................................................... 95
10.2 UKCP09 Projections ............................................................................................ 95
10.3 Fluvial .................................................................................................................. 96
10.4 Pluvial ................................................................................................................ 100
10.5 Coastal .............................................................................................................. 101
10.5.1 Extreme Still Water Level Rise ..................................................................... 101
10.5.2 Waves.......................................................................................................... 102
11 Defences ................................................................................................................ 103
12 Natural Flood Management .................................................................................... 104
12.1 Introduction........................................................................................................ 104
12.2 Modelling Approaches for NFM Studies ............................................................. 104
12.2.1 Short Listing of Measures ............................................................................ 104
12.2.2 Options Appraisal ........................................................................................ 106
12.2.3 Detailed Design ........................................................................................... 106
12.3 Consideration of NFM within Flood Studies ....................................................... 106
12.4 Hydraulic Modelling as part of an NFM Study .................................................... 107
12.5 Hydrological Modelling as part of an NFM Study ............................................... 109
13 Flood Mapping........................................................................................................ 112
13.1 Introduction........................................................................................................ 112
13.2 Fluvial ................................................................................................................ 112
13.2.1 1D flood Mapping ......................................................................................... 112
13.2.2 2D Flood Mapping ....................................................................................... 113
13.3 Coastal Flood Mapping ...................................................................................... 113
13.3.1 Horizontal Projection Method ....................................................................... 113
13.3.2 2D Coastal Flood Mapping........................................................................... 113
13.4 Pluvial Mapping ................................................................................................. 114
13.5 Post-Processing ................................................................................................ 114
13.6 Quality Checking ............................................................................................... 114
14 Quality Assurance and Quality Control ................................................................... 115

Flood Modelling Guidance for Responsible Authorities v1.1 4

 14.1 Introduction........................................................................................................ 115
14.2 Scoping and Commissioning a Study ................................................................ 115
14.2.1 Conceptual Model ........................................................................................ 117
14.3 Data Collection .................................................................................................. 117
14.3.1 Model Build .................................................................................................. 117
14.3.2 Model Audit .................................................................................................. 117
14.3.3 Responsible Authority Review of Results ..................................................... 118
14.4 Reporting ........................................................................................................... 119
15 Deliverables ........................................................................................................... 120
15.1 Introduction........................................................................................................ 120
15.2 Reporting Requirements .................................................................................... 120
15.2.1 Technical Report.......................................................................................... 120
15.2.2 Non-Technical Summary ............................................................................. 120
15.2.3 Model hand over report ................................................................................ 120
15.2.4 Audit report .................................................................................................. 120
15.2.5 Format ......................................................................................................... 121
15.3 Results Files ...................................................................................................... 121
15.3.1 Gridded output ............................................................................................. 121
15.3.1.1 Depth and Elevation ............................................................................... 121
15.3.1.2 Hazard ................................................................................................... 121
15.3.1.3 Velocity .................................................................................................. 122
15.3.1.4 Other Gridded Outputs ........................................................................... 122
15.3.2 Flood Extents and Area of Benefit................................................................ 122
15.3.3 Miscellaneous Outputs................................................................................. 122
15.3.4 Animations ................................................................................................... 123
15.3.5 Other formats ............................................................................................... 123
15.3.6 Tables of results and long sections .............................................................. 123
15.4 Model Files ........................................................................................................ 123
15.5 Survey ............................................................................................................... 124
15.6 Photographs ...................................................................................................... 124
15.7 Other deliverables ............................................................................................. 124

Appendices

A Template Flood Study SoR ........................................................................................ 126

B Example Survey SoR ................................................................................................. 127
C Quality Control ........................................................................................................... 138
C.1 Example technical review certificate .................................................................. 138
C.2 Fluvial Flood Estimate Review ........................................................................... 146
D Report Template ........................................................................................................ 151
E Model Deliverables .................................................................................................... 152
E.1 General.............................................................................................................. 152
E.2 InfoWorks(CS and ICM) ..................................................................................... 152
E.3 MIKE Flood ........................................................................................................ 152
E.4 HECRAS ........................................................................................................... 152
E.5 FloodModeller 1D .............................................................................................. 153
E.6 FloodModeller 2D .............................................................................................. 154
E.7 TUFLOW ........................................................................................................... 155
E.8 FloodModeller-TUFLOW .................................................................................... 155
E.9 FloodModeller 1D-2D......................................................................................... 155
F Model Node Naming Structure ................................................................................... 157
G Glossary .................................................................................................................... 159
H Bibliography ............................................................................................................... 162

Flood Modelling Guidance for Responsible Authorities v1.1 5

2 Introduction

2.1 Purpose of the Flood Modelling Guidance

This technical flood modelling guidance document is intended to support those in
Responsible Authorities who are responsible for developing and commissioning flood studies
in respect of flood risk management planning. This guidance provides technical guidance for
the modelling aspect of flood studies and should be used in conjunction with the Local
Authority Flood Study Checklist 1, which provides additional information on non modelling
aspects of a flood study. The Flood Risk Management (Scotland) Act 2009 (FRM Act) has
established a collaborative approach to the development of new information in support of
flood risk management. As part of this approach the adoption of consistent principles at each
step of the process will ensure that we develop nationally comparable risk-based
information. This will allow Responsible Authorities to develop, share and understand flood
risk on a common basis. Establishing a common understanding through guidance will
enable strategic and local needs to be linked by common values, approaches and definitions
of quality for appropriate use.

The flood hazard and flood risk information which underpins flood risk management
decisions is often derived from computer flood models. Flood models use simplifications and
assumptions to represent complex natural systems and this leads to inherent uncertainty,
which must be acknowledged when making decisions based on model results. This
document therefore seeks to provide guidance for Responsible Authorities on where
uncertainty may arise in flood modelling and how it may be managed through the modelling
process so that it can inform appropriate decisions. An important component of this is that
contractors work to a common set of best practice guidelines in building models and in
documenting the modelling process; Responsible Authorities are encouraged to refer their
contractors to this guidance document to promote compliance with best practice. For site-
specific Flood Risk Assessments (FRAs) to inform land use planning, guidance is provided
in SEPA’s Technical Flood Risk Guidance for Stakeholders (SEPA, 2015). Table 2-1 shows
where this guidance and the Technical Flood Risk Guidance for Stakeholders should be
followed.

Who? Developer Responsible Authority

What? FRA Flood Study NFM Study Integrated catchment

study
Which Technical Guidance This Guidance Scottish Water and
Document? for Stakeholders UDG guidance for
(SEPA, 2015) modelling sewer
systems in addition to
this Guidance.
Table 2-1: Which document do I need?

This guidance document is intended to provide a technical overview of the development of

flood models to a standard that allows them to inform flood risk management decisions.
Project-specific requirements will need to still be considered as each model is unique to the
location and scale of study as well as to budget, time and resource. This document,
however, seeks to identify the main elements to consider when establishing a flood
modelling project and offers guidance on planning, input data, model set-up, model
calibration, model quality checks and reporting requirements.
1
Local Authorities who do not have a copy of the Flood Study Checklist can obtain this from
frmplanning@sepa.org.uk or from the SCOTS – Flood Risk Management group web portal (Knowledge
Hub) https://khub.net.

Flood Modelling Guidance for Responsible Authorities v1.1 6

Flood Modelling Guidance for Responsible Authorities v1.1 7
 Usually carried out by a
Responsible Authority
Scoping Data
SoR
The guidance in this (Chapter 3) Collection
section is targeted at
Responsible Authorities.
Key points for Responsible
Commissioning a
Authorities are given in
study (Chapter 4)
green boxes.

Usually carried out by a

contractor Data Collection

Quality Assurance and Quality Control

(Chapter 5)
These sections should be
read by Responsible
Authorities and Hydrology Model
contractors. technical Schematisation
note (Chapters 6 and 7)

(Chapter 14)
The discussion in these
sections aims to give Calibration Calibration and
Responsible Authorities technical sensitivity testing
sufficient background to note (Chapter 8)
critically assess contractor
modelling.
Design runs
Key points for contractors (Chapter 9)
to ensure a consistent
approach, and quality of
flood modelling in Flood Mapping
Flood Maps
Scotland are given in red (Chapter 13)
boxes.

Flood study Reporting

report (Chapter 15)

Figure 2-1: Layout of this document.

The guidance covers fluvial, pluvial and coastal flood studies. Fluvial and coastal studies
are covered in detail; however, for pluvial studies where a detailed representation of the
surface water drainage network is required, Scottish Water and CIWEM Urban Drainage
Group2 (UDG) guidance for modelling sewer systems should be followed. For studies
covering combined pluvial and fluvial flooding or combined pluvial and coastal flooding this
guidance should be used for the fluvial and coastal components of the studies.

2.2 How to use this guidance

Figure 2-1 shows the structure of this guidance document, and how the different chapters
relate to different phases of a flood modelling project. For studies to be carried out later in a
Flood Risk Management Planning cycle, Responsible Authorities may wish to carry out the

2
Formerly WaPUG

Flood Modelling Guidance for Responsible Authorities v1.1 8

scoping stage earlier in the cycle, in order to identify the need for any additional data
collection.

Responsible Authorities often appoint specialist modelling contractors to carry out some or
all aspects of a flood study. Where they consider it appropriate Responsible Authorities may
pass this guidance onto their contractors and ask them to consider the relevant aspects in
their work. To assist with this, key points for contractors are highlighted in red boxes while
key points for Responsible Authorities are highlighted in green boxes.

2.3 SEPA input into modelling studies

SEPA is able to support flood studies via the provision of data, technical advice and a review
of outputs. Early notification of a study by a Responsible Authority will enable SEPA to plan
its resources and consider how best it can provide support as well as helping to ensure that
study outputs can be incorporated into future hazard map updates and inform future flood
risk management strategies. In the first instance ensure that the Regional Planning Manager
is up to date with plans to progress a study via the local partnership and that contact is made
with SEPA’s Strategic Flood Risk team (strategic.floodrisk@sepa.org.uk). Where appropriate
a named contact within Strategic Flood Risk will be identified who will assist in developing
links with other SEPA teams.

SEPA is able to provide support for modelling studies at the following stages:

Scoping stage – SEPA can advise on any known linkages with other studies and the
suggested study area.

Developing a Statement of Requirements (SoR) – a template SoR can be provided on

request and SEPA may also be able to review SoRs if required. SEPA can provide details of
the available hydrometric data, and other data where appropriate for inclusion in the SoR.

Data Collection – Details of the data held by SEPA which can be supplied for use in a flood
study are given in Chapter 5. Data held by other organisations which may be required for a
flood study is also given in Chapter 5.

Review of draft outputs – SEPA may be able to assist with the review of the following draft
outputs where included in the project scope; technical notes on hydrology tidal/coastal
boundary conditions, calibration results, draft design maps and flood levels, draft models and
the draft model report. SEPA may consider an independent review or audit of hydraulic
models to support consistent quality in Responsible Authority studies.

Due to the number of studies identified, in the first Flood Risk Management cycle, SEPA’s
resource requirements for supporting modelling studies are likely to be significant. To
enable SEPA to plan and prioritise input to studies it would appreciate being informed of
planned delivery dates for key project outputs and notification of any significant changes to
these timescales

Flood Modelling Guidance for Responsible Authorities v1.1 9

3 Scoping

3.1 Introduction
The first task is to define the scope of the study. This will establish the purpose of the
assessment, the level of assessment and the data requirements to inform decisions.

Notifying SEPA that a study is planned for a particular area will allow SEPA to provide
advice to feed into the scoping phases. In particular SEPA can advise on known linkages
with other studies and can provide details of hydrometric and other data (held by SEPA) for
the study area. It is recommended that SEPA’s Strategic Flood Risk team
(strategic.floodrisk@sepa.org.uk) and the appropriate Regional Flood Risk Planning
Manager are contacted to advise that the study is taking place.

This chapter provides guidance for Responsible Authorities in:

• Developing a conceptual model of a catchment;

• Identifying opportunities for joint studies;
• Identifying the appropriate level of complexity for a flood study;
• Selecting suitable modelling software to meet the study objectives.

Further information on scoping Natural Flood Management (NFM) Studies or studies with an
NFM component is given in Chapter 12.

3.2 Define Modelling Objective

It is important to develop a clear statement of the purpose of the proposed modelling study
as this will determine the level of assessment carried out. The purpose may be for flood
mapping to support development planning and management, options assessment for a flood
prevention scheme, detailed design etc. Consideration should be given to the required
accuracy and level of quality control given the purpose of the model. This should be stated in
the Statement of Requirements (SoR). The required study outputs should be listed as part
of the modelling objective and further guidance is given in Chapter 15.

Potential future uses of the model and outputs should also be considered, as this may
enable the model to be built in such a way as to maximise reuse, and will ensure that
necessary outputs are supplied.

3.3 Developing a Conceptual Model

Key points for Responsible Authorities
• Develop a conceptual model of flood risk using the source-pathway-receptor-
impacts approach.
• Carry out a catchment walk-over.
• Review available data to determine the need for any additional data collection.
• Set out knowledge of the catchment in the SoR.

Key points for Contractors

• Develop and revise the conceptual model of the catchment throughout the study.
• Describe the conceptual model in the modelling report.

Flood Modelling Guidance for Responsible Authorities v1.1 10

Before commissioning a study, it is important to understand flooding mechanisms at the
study location. A good understanding of the links between the sources and impacts of
flooding can help identify the most appropriate modelling approach or whether modelling is
the correct approach. A summary of the flooding mechanisms should be provided in the
SoR to enable contractors to propose an appropriate methodology for the study.

To help understand the interaction of different actions across catchments and coastlines, the
Responsible Authority should use the source–pathway-receptor–impact approach to build a
conceptual model of the key processes which need to be considered in the study (Figure
3-1). This approach is a well-established framework in flood risk management. It provides a
basis for understanding the causal links between the source of flooding, the route by which it
is transmitted and the receptor, which suffers some impact:

• Sources are the weather events or conditions that result in flooding (e.g. heavy
rainfall, rising sea level, waves etc.);
• Pathways are routes between the source of flood waters and the receptor. These
include surface and subsurface flow across the landscape, urban drainage systems,
wave overtopping;
• Receptors are the people, industries and built and natural environments that can be
impacted upon by flooding;
• Impacts are the effects on exposed receptors. The severity of any impact will vary
depending on the vulnerability of the receptor.

For any area there may be multiple sources, pathways and receptors which interact with
each other.

Source Pathway Receptor Impacts

•e.g. rainfall, •e.g. •e.g. property, •e.g los of life,

wind, waves overtopping, people, stress, material
overflow environment damage,
environmental
degradation

Figure 3-1: The Source-pathway-receptor-impacts model, based on the DEFRA/Environment

Agency R&D Technical Report FD2302/TR1 (Sayers, Gouldby, Simm, Meadowcroft, & Hall,
2003).

In developing the conceptual model historic flood information for the area, including any
anecdotal evidence, should be examined. A catchment walk-over in conjunction with a
desk top review of Ordnance Survey maps and aerial photography should be used to identify
physical features which may affect flood pathways and possible receptors. Previous studies
and SEPA’s national flood hazard maps http://map.sepa.org.uk/floodmap/map.htm can also
be used to identify possible flooding mechanisms and whether they are adequately captured
by the previous modelling approach. For coastal areas the National Coastal Change
Assessment http://www.dynamiccoast.com/ should be used to identify areas where coastal
change may be a factor influencing flooding. The historic flood datasets which may be
available to help with developing a conceptual model are discussed in Section 5.4.

Available data should be assessed during development of the conceptual model in order to
determine the need for any additional data collection. Responsible Authorities may consider
assessing available data for studies later in a Flood Risk Management cycle so that data
gaps can be identified and filled prior to commencement of the study. Information on
relevant data is given in Chapter 5. A list of available data, with a brief description, should

Flood Modelling Guidance for Responsible Authorities v1.1 11

be set out in the SoR, together with any survey requirements etc. SEPA can assist in
determining the availability of hydrometric and other data for inclusion in the SoR.

Knowledge of the catchment should be set out briefly in the SoR, including any key areas
and known flooding mechanisms which need to be considered. These key features should
also be marked on a location plan to be included with the statement of requirements.

A catchment walkover with an appointed contractor can help develop a shared

understanding of the study area. Contractors should develop the conceptual model of the
catchment, as understanding of the study area increases throughout the study, and describe
the conceptual model in the modelling report.

3.4 Spatial Extent

Key points for Responsible Authorities

• Use the conceptual model in setting the study extent.

• A single study covering a larger area is likely to be significantly cheaper than
several smaller studies and may also be more accurate and robust
• Studies for inclusion in SEPA’s hazard maps should have a study area consistent
with the strategic nature of the maps.

The conceptual model described in section 3.2 should be used in setting the study extent.
Consideration should also be given to the area of interest, availability of calibration data,
future use of the model and cost and time for a study

The study extent should be sufficient to represent the assumed flooding mechanism (i.e. it
should cover the flood source, flood pathway and any receptors). Boundaries of the study
area should be sufficiently far away from the area of interest, considering flow controls, to
have no impact on the results. Good places to set study boundaries are areas where the
flood extents are relatively constrained for large events or where there is a hydraulic control
such as a weir or tidal boundary. SEPA’s flood hazard maps
http://map.sepa.org.uk/floodmap/map.htm can be used with large scale mapping to identify
constrained sections, which should be reviewed at a site visit.

Data available for calibration should be considered in setting the study extent, as extending
the study area to cover calibration data may significantly improve confidence in the study
output. For fluvial studies covering gauged rivers it is strongly recommended that the study
extent covers at least one and preferably two or more gauges to assist in calibration. This is
discussed further in section 5.8 and Chapter 8.

The availability of topographic data should not be used to constrain the study area where
other considerations suggest that a larger area would be more appropriate. The preference
should be for additional data collection rather than a reduced study extent.

Flood Modelling Guidance for Responsible Authorities v1.1 12

Figure 3-2: Considerations in setting study extents for a hypothetical catchment level study.

Flood Modelling Guidance for Responsible Authorities v1.1 13

The cost and time required for a modelling study will increase with the study extent.
However, potential future uses of the model should be weighed against any increase in cost
as a single study covering a larger area is likely to be significantly cheaper than several
smaller studies and may also be more accurate and robust.

SEPA will consider the study area in deciding whether Responsible Authority studies can be
used to update the national hazard maps. For inclusion in SEPA’s hazard maps, studies
should cover reaches so that they can be tied in smoothly with the national mapping. As
inconsistencies between different modelling approaches may be particularly evident in urban
areas, studies should not normally have boundaries within continuous urban areas as shown
by the Scottish Government Urban Rural Classification 3. SEPA’s hazard maps are
strategic level and small study areas inconsistent with this level of mapping will not
be considered for inclusion in SEPA’s national flood hazard maps, although they may
be suitable for site specific FRAs submitted in support of planning applications. In this case
the guidance in SEPA’s Technical Guidance for Stakeholders on Flood Risk Assessment
should be followed (SEPA, 2015). Figure 3-2 shows considerations in setting study extents
for a hypothetical catchment.

The study extent should be set out in a location plan included with the SoR.

3.5 Joint Studies

Key Point for Responsible Authorities

Overall knowledge and understanding of flooding might be improved by combining your

study with those of other organisations. This may also provide efficiencies in cost, time
and quality.

There may be cost efficiencies and quality improvements resulting from a joint study either
with partner organisations covering nearby areas or with organisations with different
objectives in the same study area.

Benefits may include:

• Sharing costs in model development across multiple organisations.
• Reduction in mobilisation costs for survey.
• Reduction in project management time.
• Reduction in modelling costs and time as modellers will become familiar with the
study area and model set up.
• Improved calibration if a study area covers several gauges or extends to an area for
which historical data is available.
• Ability to investigate the impact of flood risk management measures and catchment
changes upstream of the area of immediate interest.
• Ability to investigate the impact of flood risk management measures on the
downstream area.

In order to identify these opportunities, effort should be made to speak to the following
organisations during the scoping phase:
• Local authorities upstream or downstream of a study area along the same
watercourse, or along the same stretch of coastline.

3
http://www.gov.scot/Topics/Statistics/About/Methodology/UrbanRuralClassification

Flood Modelling Guidance for Responsible Authorities v1.1 14

 • Scottish Water where surface water flooding or combined flooding is considered to
be an issue.
• Commercial organisations or government agencies such as Transport Scotland and
the Forestry Commission where it is known that large planning applications or
developments are proposed within an area.
• Major land owners.
• SEPA which may be aware of other studies or work within the area, including studies
being carried out by SEPA’s Flood Forecasting and Warning and River Basin
Management Planning teams. The Strategic Flood Risk team will act as a single
point for this type of enquiry at scoping stage.

Case Study – River Kelvin Flood Mapping Study

Glasgow City Council, East Dunbartonshire Council and SEPA worked together to
commission the River Kelvin study. This enabled the model to be calibrated against two
gauges in the river Kelvin catchment, increasing confidence in the results and ensuring
flood maps were consistent across the local authority boundary.

Flood Modelling Guidance for Responsible Authorities v1.1 15

Case Study – Metropolitan Glasgow Strategic Drainage Programme

The Metropolitan Glasgow Strategic Drainage Partnership (MGSDP) was created in

response to the significant flooding that occurred in Glasgow in July 2002. The extreme
rainfall event led to flooding from multiple sources requiring an integrated, catchment
based, partnership approach to reduce the risks and impacts of uncontrolled flooding,
improve water quality and also as a result enable sustainable urban regeneration and
growth. The key partners involved are Glasgow City Council, South Lanarkshire Council,
North Lanarkshire Council, Renfrewshire Council, East Dunbartonshire Council, SEPA,
Scottish Water, Scottish Canals, Network Rail, the Scottish Government and Clyde
Gateway as organisations responsible for, or with an interest in, flood risk management in
the area. Further information about the MGSDP Vision, Objectives and Guiding Principles
is available on the website – www.mgsdp.org

A dedicated Programme Management Office was established to provides overall co-

ordination of the partnership and facilitate collaborative working to review and improve
knowledge and understanding of flooding issues and integrated solution delivery. A
phased approach was undertaken with an initial information gathering phase identifying
historical flooding issues, data shortfalls and existing studies leading onto pluvial
modelling. Given the complex nature of flooding the need for integrated catchment
modelling was identified and now been progressed based on catchment areas served by
the several waste water treatment works within the Metropolitan Glasgow area. The
modelling is currently being used to help develop and deliver a range of appropriate flood
management and water quality interventions.

Based on these studies the benefits of integrated working include:

• Service delivery improvements
• Economies of scale
• Efficiency savings through improved systems and practices
• Sharing of knowledge and good practice
• Streamlining of communication
• Development of a strategic and co-ordinated approached to project delivery
• Creating cost and time savings
• Development of collaborative and integrated flood risk strategies and projects in
line with Flood Risk Management (Scotland) Act 2009 duties
• Enhancement of public confidence.

Further information and contact details are available here:

http://www.mgsdp.org/index.aspx?articleid=2017.

Flood Modelling Guidance for Responsible Authorities v1.1 16

3.6 Level of Assessment
The level of assessment should be considered Point to Note:
at the scoping stage as this has implications for
the amount and quality of data required, the It should be noted that, in some cases,
modelling effort and ultimately the cost of the limitations in scientific understanding
study. The level of assessment should be may mean that a more detailed and
appropriate to meet the modelling objective, for complex approach may not provide
example, a high level scoping study to additional confidence in the
understand areas for further work is only likely conclusions which can be drawn from
to require a simplified modelling approach or the modelling. This particularly applies
desktop study. Generally, as a study tends to some aspects of Natural Flood
towards the detailed assessment and design of Management (NFM); see chapter 12.
flood risk management options, the model
complexity will increase.

SEPA’s modelling framework sets out a hierarchical approach with 3 levels of assessment
national or strategic, catchment or feasibility and local or design, Figure 3-3 and Table 3-1.

National •Aim: Meet the

(Strategic) requirement of the
EU Floods
Directive.

Catchment •Aim: Inform options

(Feasibility) appraisal, inform NFRA,
support LLA.

Local •Aim: Support LLA in LFRMP

(Design) development

Figure 3-3: SEPA's modelling hierarchy.

3.7 Modelling Approach

There are a range of possible modelling approaches that can be adopted. Responsible
Authorities should understand which approaches will be suitable to meet the modelling
objectives at scoping stage as this will affect time and cost of a study. The appropriate
approach depends on the required study outputs, the flooding mechanisms, and the level of
assessment required. Models can by categorised by whether they are steady or unsteady
and by the dimension of the modelling. Table 3-2 provides an overview of model dimensions
which might be considered and the circumstances where they may be applied. Table 3-3
provides an overview of steady and unsteady analyses and where they may be appropriately
applied.

Flood Modelling Guidance for Responsible Authorities v1.1 17

 National (Strategic) Catchment (Feasibility)
Local (Design)
Producing national hazard maps to meet Options appraisal. Detailed design of flood defences or NFM measures.
Example Objectives requirements of EU Floods Directive. NFM studies. Site specific FRAs for developments.
Screening maps for LUP in more complex areas.
Broad scale assessment; national overview; Catchment scale assessment; improved understanding of the local focus; consider interactions of sources; detailed
based on national datasets; strategic direction catchment; inform flood risk management option feasibility; representation of local features
General description setting; e.g. FRM planning -> national flood warning area development; emergency planning
comparison; awareness raising; improved
understanding
National, large or multi- catchment, or long Catchment scale, major firths or long reaches of coastline with Small catchment or section of larger catchment, short lengths of
Scale reaches of coastline. similar orientation. coastline e.g. town frontage.
Use of historic flood extents, 1D or 2D hydraulic 1D, 2D or 1D-2D hydraulic modelling. Hydrology using FEH Detailed hydraulic modelling (1D or 1D-2D); range of modelling
Fluvial modelling. Flows from national datasets e.g. methods and distributed inflows. Range of scenarios. scenarios; consideration of climate change. Local-scale
CEH flow grid. assessment of design flows; explicit use of local information.
Use of historic flood extents, rapid flood 2D hydraulic modelling. FEH DDF model. Range of scenarios.
Approaches Pluvial spreading, 2D hydraulic modelling. FEH DDF
model. Limited number of durations
Use of historic flood extents, Level projection, Hydrodynamic and wave transformation modelling. Detailed hydrodynamic and wave transformation modelling,
rapid flood spreading, 2D hydrodynamic 2D hydrodynamic model for flood inundation. Range of overtopping modelling. Consideration of climate change.
Coastal
modelling. scenarios. Consideration of coastal processes.
Consideration of climate change 2D hydrodynamic model for flood inundation.
Typically blockages caused by structures Definition of key structures using basic information to represent Explicit and refined definition of key structures and channel
Fluvial removed from DTM. height, length, etc. Structures such as foot bridges may be features.
Structures Buildings likely to be removed from the DTM. omitted.
Pluvial
Raised defences added to DTM. Overtopping analysis for Overtopping analysis for defences.
Coastal
defences.
Check of selected locations against historic Check against historic records; sensitivity testing. Calibration at gauges for multiple events.
Calibration/ records. Reality checks Calibration at gauges. Installation of additional gauging.
Sensitivity Testing Comparison with historic records.
Sensitivity testing.
Not suitable for site specific information. Assumptions on Hydrology and limited assessment of local Can be time consuming and expensive.
Large spatial range -> relatively lower spatial influences
Limitations
accuracy

DTM DTM DTM

Fluvial
Channel cross section survey Channel cross section survey
DTM DTM
Data Pluvial
Drainage network information
Requirements
DTM DTM
Coastal Bathymetry Bathymetry
Defence and beach profiles
Review of input data. Review of input data. Review of input data.
Review of method. Review of models and hydrology by independent internal Detailed review of models and hydrology by independent internal
Quality Control Consistency checks. reviewer. reviewer or external reviewer.
Screening for physically unrealistic results.
Table 3-1: Overview of different levels of assessment.

Flood Modelling Guidance for Responsible Authorities v1.1 18

Model Type Description Uses
0-D / Rapid 0D models use simple rules to spread water over the floodplain between adjacent depressions. To provide a general overview of pluvial flooding when the
Flood Advantages: Short run times. Limited data requirements. detailed data required by most 1-D or 2-D models are
Spreading Limitations: Only produce final state of inundation, do not capture flow pathways. Flow velocity not available. unavailable.
Input data requirements: Digital Elevation Model (DEM) or Digital Terrain Model (DTM) This may be appropriate for strategic levels studies covering
large areas; however for local flood studies the 1D and 2D
approaches below are more likely to be appropriate.
1D - 1D routing models use a storage equation to model changes in the shape of the flood hydrograph along a river reach. Parameters To provide boundary conditions for more detailed river models.
Hydrological in the equation are either calculated from the channel geometry or estimated from observed data. Frequently used in flood warning where short run times are
Routing Advantages: Short run times. Limited data requirements (sparse cross sections). required.
Limitations: Loss of accuracy. Levels are not calculated.
Input data requirements: Cross sections or a DTM from which cross sections can be extracted
1D In 1D models flow is averaged over depth and across defined cross sections. Models can be steady or unsteady. Fluvial: Used for modelling In bank flows, in channel hydraulic
Advantages: Accurate representation of flow and level within channel. Good representation of hydraulic structures. Usually short structures and narrow well defined floodplains.
run times allowing multiple scenarios to be run. Pluvial: Used for modelling surface water drainage networks.
Limitations: Complex floodplain flows cannot be represented. Floodplain velocity not available.
Input data requirements: Channel sections and structures.
2D In 2D models flow is averaged over the flow depth and horizontally over a model grid cell or element. 2D models may be divided Fluvial: Used for modelling floodplain flows where the channel
into those which solve the full shallow water equations and those which solve a simplified version of the equations to reduce capacity and transition between in bank and out of bank flow is
computational effort. Either a regular grid or irregular mesh may be used. not important (e.g. larger flood events).
Advantages: Accurate representation of floodplain flow. Simple model build. Pluvial: Used for modelling overland flow where interaction with
Limitations: Cannot accurately represent structures or in channel flow particularly for narrow watercourses. Higher resolution the surface water drainage network is not important.
models may have long run times. Coastal: Used coastal inundation modelling.
Input data requirements: Digital Elevation Model (DEM) or Digital Terrain Model (DTM).
1D-2D The channel is modelled in 1D and linked to a 2D model of the floodplain so that they can exchange flow. Fluvial: Used where there is a need to understand both the
Advantages: Accurate representation of flow and level within channel and floodplain flow routes. channel and floodplain processes.
Limitations: Large time inputs required to set up complex models. Long run times. Potentially costly. Pluvial: Used where there is need to understand how the
Input data requirements: channel cross sections and DEM. surface water drainage network interacts with overland flows e.g.
an Integrated Catchment Model (ICM).
Coastal: May be used where there is a well-defined drainage
network along which flood waters can propagate.
3-D or 3d and physical models allow vertical flow to be modelled. Not commonly used in flood studies. They may be necessary for
Physical Input data requirements: 3d structure information, DEM, bathymetric survey. design in some complex cases for example, understanding of
Models Advantages: Can represent vertical variations in flow. flow over, through and around structures.
Limitations: Costly and complex.
Nearshore Simulate the processes such as energy dissipation due to bottom friction, white capping, or depth induced breaking, growth due to Used in coastal flood studies to transform wave conditions from
wave wind, refraction and shoaling. offshore wave models inshore to produce boundaries for
models Advantages: Allow nearshore wave conditions to be assessed in the absence of long series of observations. overtopping models.
Limitations: Large time inputs required to set up complex models. Long run times. Potentially costly.
Input data requirements: Bathymetric survey and a DTM of the intertidal area. Water levels. Offshore wave conditions.
Overtopping Empirical and physically based models are available. Used in coastal flood studies to predict the rate of water
Models Advantages: Allow overtopping rates to be assessed. overtopping flood defences, and provide boundaries for 2D
Limitations: Depend on the type of model. Expert judgement may be required to determine the appropriate model in each case. inundation models.
Input data requirements: Nearshore wave conditions and water levels. Crest and toe levels, and defence profile, including
information on roughness.
Table 3-2: Modelling approaches

Flood Modelling Guidance for Responsible Authorities v1.1 19

 Model Type Description and Uses
In steady state modelling it is assumed that flow does not vary with
Steady Flow
time at given location. This assumption can be made for many
applications where floodplain flow is limited, for instance in narrow
valleys.
Unsteady Flow Unsteady models allow flow to change with time. Unsteady models are
required for problems where flood wave propagation or flood storage or
attenuation within the system is of interest.

The areas where unsteady flow should be considered include where:

• there is a tidal boundary;
• there is flood storage within the system;
• hydrograph timing is of interest;
• floodplain flow needs to be taken into account;
• there are pumps or gates to which control rules are applied.

Table 3-3: Overview of basic model types

3.7.1 Selection of Modelling Software

A variety of hydraulic modelling software is available for use in flood model development
although there is no universal package which can be recommended for all applications. The
Environment Agency has carried out benchmarking tests of 1D and 2D hydraulic modelling
packages commonly used in Flood Risk Management in the UK (Crowder, et al., 2004)
(Neelz & Pender, 2013). These provide a summary of the key features of the modelling
packages and the time taken to set up models using different packages. It is not advised
that software should be used unless it has been thoroughly peer reviewed or extensively
tested through an extended period of use by several different organisations. Advice should
be sought from SEPA if the suggested software is not covered by the Environment Agency
benchmarking, including where the use or development of novel tools is proposed in
developing areas of flood risk science such as Natural Flood Management or Climate
change.

In any tender the contractor should set out which modelling software will be used and why it
is appropriate to meet the project objectives.

In addition to the ability of the selected software to meet the project objective the
Responsible Authority should consider:

• If there is a sufficient pool of people in the industry experienced in using the selected
software to enable the model to be reviewed and audited
• If use of the proposed software will restrict future use and development of the model
to specific contractors, either because the software is in-house to a particular
contractor or because there are limited skills.
• If there is support and training available for use of the software to allow any bugs or
issues to be addressed and for expert advice to be sought for difficult or unusual
schematisations.
• Licensing options and cost if the Responsible Authority either wish to rerun or update
the model themselves or to view the model schematisation and results. However it
should be noted that higher license costs may add functionality necessary for
meeting the project objectives or be offset by improved workflows, customer support,

Flood Modelling Guidance for Responsible Authorities v1.1 20

 improved model stability or reduced run times which may reduce time costs for the
modelling study.

SEPA has licenses and skills in a number of modelling packages, Table 3-4, and may be
able to review models or assist local authorities in viewing models and results and rerunning
models if necessary. Some modelling packages either have free versions with limited
functionality, or viewers available at a reduced cost when compared to the full license.
Information on this is available from the relevant software suppliers.

Coastal

Modelling

1D Sewer
Software Supplier

1D River

2D Land
Software

Waves
Surge
FloodModeller 4 https://www.floodmodeller.com/en-gb/     
Delft 3d / SWAN https://www.deltares.nl/en/software/delft3d-     
4-suite/
HECRAS http://www.hec.usace.army.mil/software/he     
c-ras/
Infoworks ICM http://www.innovyze.com/     
Infoworks RS 2D     
JFLOW http://www.jbaconsulting.com     
MIKE FLOOD 5 http://www.dhigroup.com/     
TUFLOW/ ESTRY http://www.tuflow.com/     
Table 3-4: Flood modelling software currently used by SEPA.

4
Formerly ISIS.
5
The MIKE suite is capable of modelling waves and hydrodynamic flows with estuaries, but SEPA’s flood risk
teams do not currently have a license for this functionality.

Flood Modelling Guidance for Responsible Authorities v1.1 21

4 Commissioning a study

4.1 Introduction
Responsible Authorities have commissioned and managed flood studies successfully over a
number of years. The guidance in this chapter does not aim to replace Responsible
Authority expertise or procurement and project management procedures, however it
highlights some factors which Responsible Authorities should consider in commissioning a
study to ensure:

• Flood modelling is at a consistently high standard across Scotland

• Data from flood studies can be shared with other Responsible Authorities
• SEPA has sight of timescales and project milestones, so it can provide timely input if
required.

Key points for Responsible Authorities

• SEPA may be able to assist with reviewing SoRs.
• Ask contractors to provide evidence of quality control.
• The Intellectual Property Rights (IPR) for any data collected explicitly for the
project or generated by the project should be held by the Responsible Authority.

4.2 Statement of Requirements (SoR)

A SoR for the study area should be developed following the scoping exercise set out in
Chapter 3. A template SoR suitable for a range of flood studies at different scales and levels
of detail is available and further details are given in Appendix A. Responsible Authorities
may use this template where they consider it appropriate for their study; use of template
should produce modelling and mapping outputs which meet the requirements for inclusion in
the national hazard maps. Developing an appropriate SoR is one of the most important
parts of any modelling study and, where required, SEPA may be able to assist with their
review.

4.3 Quality Control

The contractor should provide evidence of how they will carry out quality control and quality
assurance through the project, including setting out internal review procedures for modelling
and hydrology. Evidence of quality control should be included as one of the deliverables
within the SoR. Further detail on this is given in Chapter 14.

4.4 Resources
The project must be adequately resourced from both the contactor and Responsible
Authority sides.

The contractor should demonstrate that sufficient staff are available with the correct level of
experience to deliver the project in their tender. This should include identification of an
internal reviewer with sufficient experience who is not directly involved in the project.
Contractors should also be able to demonstrate that they have sufficient computational
resources and software available to deliver the project.

The Responsible Authority should ensure that appropriate staff will be available at key points
in the project to enable timeous supply of data and information, and so that outputs can be
reviewed within the stated timescale.

Flood Modelling Guidance for Responsible Authorities v1.1 22

In exceptional cases there may be additional computational resource requirements for large
projects, including purchase of data storage, computer processing power or additional
software licenses, and contractors may include costs for purchasing these in their tender.

4.5 Timescale/Milestones
A proposed programme should be included with all tenders. The programme should show
times when input from client and other external stakeholders will be required. A minimum of
10 working days should be allowed for the commissioning body and other stakeholders to
review draft outputs.

Where requested by the Responsible Authority, SEPA will seek to support Responsible
Authority studies within its resources. Due to the number of studies identified, SEPA’s
resource requirements for supporting modelling studies are likely to be significant over the
first flood risk management planning cycle. To enable SEPA to plan and prioritise input,
SEPA should be informed of planned delivery dates for key project outputs and notified of
any significant changes to these timescales.

4.6 Risk Register

A risk register should be included in any tender identifying the risk, the likelihood of the risk
occurring, the consequence of the risk if it did occur, any mitigation to be taken, and whether
the risk is owned by the Responsible Authority of contractor. This register should be
reviewed and updated as the project progresses.

4.7 Meetings and Progress Reports

Regular meetings between the contractor and Responsible Authority are advised to ensure
good communication links are established.

An inception meeting can be combined with a walk-over survey or site meeting with an
appointed contractor. This is advised to ensure that there is a shared understanding of the
study area at the outset of the project.

The use of progress reports alongside an agreed programme of works is key to effective
project management. These should be at a suitable frequency commensurate to the scale
and complexity of the project, however fortnightly to monthly reporting in any agreed format
are typical frequencies.

4.8 Intellectual Property Rights

It is important that the Intellectual Property Rights (IPR) for any data collected explicitly for
the project or generated by the project are held by the body commissioning the study are not
retrained by the consultant. This includes survey data, any models and associated model
outputs produced as part of the study and photographs collected on site visits. This
requirement should be included within any contract.

Where data is provided by 3rd parties for use in the study, appropriate licensing agreements
should be in place to ensure that use of the data does not affect future use of the model. If
this is not possible the benefits of collecting new data rather than using the 3rd party data
should be assessed.

Flood Modelling Guidance for Responsible Authorities v1.1 23

5 Data

5.1 Introduction
A wide range of data is used in flood modelling Point to Note:
studies. Data requirements depend on the
project objective including any requirements for Data underpins a modelling study.
quality, the level of detail and adopted modelling Data requirements will be informed by
approach. At the scoping stage the Responsible study objective, time and budget.
Authority should seek to understand the data However, good quality data provides a
available and whether any new data is likely to significant step towards a good quality
be required as this can significantly affect the model.
cost, quality and timescales for a modelling
project. General data requirements for different
modelling approaches are set out in Table 3-2

This chapter aims to:

• Describe the different types of data required for a flood modelling study and their
usage and limitations.
• Describe how to obtain datasets held by SEPA and other organisations.
• Provide technical guidance for Responsible Authorities specifying new data collection
for flood modelling.

5.2 General Considerations

5.3 Data collection

Key points for Responsible Authorities

• Start data collection early to avoid project delays. SEPA may be able to prepare
data requests prior to a contract being awarded.
• Data registers should be kept by the Responsible Authority and the contractor.
• Data licencing agreements should be in place for all data used within the
modelling study.

For most modelling activities some data will already exist, while other data will need to be
generated.

It is necessary to remember that the data collection process can take some time and may be
seasonally dependent; this needs to be incorporated into the project plan. Delays in data
collection or in providing data to contractors can cause significant project delays.

Where SEPA holds the data being requested, it may be able to respond to data requests
prior to the contract being awarded.

5.3.1 Data registers

A clear record of the data used for a project is important for understanding if there are any
licensing restrictions which may impact future use and subsequent implications on the
project output if quality issues are identified with input datasets.

Flood Modelling Guidance for Responsible Authorities v1.1 24

It is recommended that a data register is used to record when information is issued or
received, where it is located, the date the data was collected, and any relevant licence
terms. Where a contractor is used both the Responsible Authority and contractor should
keep data registers, and the contractor’s data register should be included in the project
deliverables.

5.3.2 Licensing
Data licensing agreements should be in place for all datasets used within the modelling
study. This for example should stretch to include site surveys, the digital terrain model, the
hydrology and all model outputs. Where possible, licences should not restrict the future use
of any models or any derived data by SEPA or Responsible Authorities. SEPA should be
consulted for advice about how to proceed if a comprehensive licence cannot be achieved.

5.4 Historic Flooding Information

Key points for Responsible Authorities
• A flood study should include an assessment of the local flood history.

5.4.1 Use of data

Historic flooding information should be used at the scoping Point to Note:
stage when developing a conceptual model of the flooding
mechanisms (as discussed in Section 3.3), in setting the Information on ‘real’ flood
study extent (as discussed in section 3.4) and in identifying
events provides a clear
possible calibration events for inclusion in a SoR (as
benchmark for
discussed in Section 3). Historic flooding information should establishing good quality
be used at the calibration and verification stage to increase
flood models. A range of
confidence in model results, as discussed in Chapter 8. information is available
Historic flood information may also be used in development
and should be considered
of flood frequency curves, and further guidance on this is
in developing and
given in Bayliss & Reed (2001). Observed depths and
finalising models.
extents may also be used for strategic level studies instead
of a modelled outlines.

5.4.2 Available Datasets

To ensure that the local flood history and mechanisms are fully understood the local flood
history evidence should be compiled and assessed as part of a flood study. All available
information should be compiled including photographs, flood outlines, anecdotal descriptions
of onset of flooding and receptors affected and summarised in the report. The outline SoR in
Appendix A includes compiling and assessing the local flood history.

SEPA maintains an extensive flood event database. The database is subject to

development and ongoing quality control of older records but it is a useful source of
information in many areas. A large amount of this data was supplied by local authorities, so
may duplicate Local Authority records. Depending on the data source, flood records in the
SEPA database are available for a range of spatial scales. At present the database does
not contain any level information, however SEPA may hold post flood event survey with
levels or photographs from which levels can be derived. Other potential sources of flood
information are given in Table 5-1.

Flood Modelling Guidance for Responsible Authorities v1.1 25

Source Sub-source Type Comments
Responsible Flood Prevention 1) Biennial Flood 1) Often available on Council
Authority Authority Reports / flood photos. website.
Planning 2) Flood Prevention 2) Feasibility studies are
Authority Scheme studies often undertaken for areas
3) Strategic Flood Risk where no formal flood
Assessments prevention measures
4) Flood Risk currently exist.
Assessments for Many councils have an e-
planning planning website.
Scottish Water Flood incident reports.
SEPA Flood Risk Flood photos, post- SEPA’s flood risk team hold
Hydrology flood survey data. information on past flood
events in Scotland in various
formats
mailto:strategic.floodrisk@se
pa.org.uk

National Flood Digitised records of Available via SEPA website

Risk Assessment past flooding from http://map.sepa.org.uk/nfra/ma
multiple sources. Note p.htm
this is less detailed than
the information held in
local teams.
British University of Chronology of British Available on-line at
Hydrological Dundee Hydrological Events http://www.dundee.ac.uk/geo
Society graphy/cbhe/
Media Television and Flood reports/ Material may be found on-
Newspaper photographs. line.
SNIFFER Coastal Flooding Final report and GIS Information on past coastal
in Scotland: A data can be found at flood events in Scotland as
Scoping Study http://www.sniffer.org.u well as the dominant coastal
2008 k/files/7013/4183/7993/ processes.
FRM10_final_030908_
with_security.pdf
Local Flood Local residents Anecdotal accounts of flooding and/or flood photos.
Groups
Library/ Books, journals, Historical flood information & photos.
Archives magazines,
newspapers,
church records.
Internet Web search Accounts of flooding Numerous data sources exist
and photos. on-line.
Buildings/ Can be on a Epigraphic flood data Often levels of past extreme
bridges plaque or floods are marked on buildings
chiselled mark and bridges.
Table 5-1: Potential sources of historic flood information.

Flood Modelling Guidance for Responsible Authorities v1.1 26

5.5 Topographic and Bathymetric Data
Key points for Responsible Authorities

• Existing topographic data should be assessed to ensure that it is appropriate to

use in a new study.
• Use the Environment Agency’s Standard Technical Specifications, Version 3.2 for
any new survey.

Topographic and bathymetric data is used in the construction of hydraulic models and in the
production of flood maps from the subsequent model results.

5.5.1 Reuse of existing data

Existing topographic data should be carefully assessed to ensure that it is appropriate to use
in a new study. In particular it should be clear that:

• There have been no major changes to the study area since the data was collected
including significant erosion or deposition, vegetation growth, construction of new
structures or alteration or removal of existing structures.
• The datum used for the existing survey is the same as for any new survey and any
DTM used in the study.
• That the IPR for the existing data allows reuse in this study
• The original survey is of an appropriate quality for the modelling study.

If in doubt a check survey should be used to ensure that the existing survey is suitable for
further use. The appropriateness of data for reuse depends on the purpose of the modelling
for instance, older channel survey data may be appropriate for strategic level mapping, but is
unlikely to be appropriate for detailed design. Where older topographic data is in paper
format only it may add considerably to the time and cost associated with model building as
well as increasing the possibility of human error. It should be noted however that there may
still be significant value in original engineering drawings from flood defence schemes,
culverts etc. and, where possible, these should be provided for use even if other survey is
considered necessary.

Survey data may be contained within existing models however, where possible, original
survey plans, drawings, photographs, and any other datasets should be supplied together
with the existing hydraulic model as:

• there may be schematisation errors in the existing model;

• different software may require different or additional information for the same
structures;
• the location of cross sections may not be clear, particularly for older hydraulic models
which are not georeferenced;
• the model may contain a combination of detailed survey information, interpolated
sections, and other information such as levels interpolated from Ordnance Survey
contours, and it may not be clear which information is from the detailed survey.

Flood Modelling Guidance for Responsible Authorities v1.1 27

5.5.2 Commissioning a New Survey
The Environment Agency’s Standard Technical Point to Note:
Specifications, Version 3.2 (Environment Agency,
2013) provide a comprehensive technical survey Use of the Environment Agency’s
specification. It is recommended that the Standard Technical Specifications
relevant sections of this specification are used v3.2 should ensure that data is
when specifying survey for hydraulic modelling. supplied in an appropriate format.
This should provide efficiencies as survey and
hydraulic modelling contractors operating within The EA Specifications are referenced
other parts of the UK are already familiar with in this guidance document as the ‘EA
producing and using data produced to this Specification’.
specification. A template SoR for use of these
specifications is provided in Appendix B.

5.5.3 Data Format

It is important that any new survey is delivered in a suitable format to enable efficient entry of
survey into the model as this may significantly reduce model build time and costs as well as
reducing the potential for data entry errors. In particular cross section survey should be
supplied in an electronic data format which can be imported directly into the hydraulic
modelling software used for the project. In most cases this will require specifying survey
to be delivered in the specific proprietary format for the modelling software to be used for the
project; however the recently developed non-proprietary Environment Agency Channel
Survey Data (EACSD) format can now be imported directly into some river modelling
software. Surveying software packages 6 are available which are able to export data in the
formats required by commonly used modelling software and the EACSD format. Supply of
data in Excel spreadsheet format does not generally enable efficient import of data
into hydraulic models.

In addition to requesting data in the correct format for import into modelling software, it is
recommended that data is also requested in the following formats:
• Drawings and plans of the survey in CAD software. This allows key points to be
clearly marked and for measurements to be scaled of drawings,
• GIS layers showing section locations and survey points, this allows cross sections
to be mapped quickly using GIS software.
• Photographs of cross sections and survey locations.

5.5.4 Channel Survey

Channel survey is required where a detailed assessment of the in-channel hydraulics is
required, usually for 1D and 1D-2D fluvial models (discussed further in section 7.2). Survey
is required for open channel sections and for structures.

5.5.4.1 Commissioning a New Channel Survey

If a new channel survey is required, section IV of the Environment Agency’s Standard
Technical Specifications, Version 3.2 (Environment Agency, 2013) (hereafter referred to as
the EA Specifications) provides an appropriate specification. Note that the EACSD format
referred to in section Data Format5.5.3 has not yet been adopted by all software vendors. It
is therefore recommended that the survey is requested in a suitable format for import into the
hydraulic modelling software to be used for the project, as well as the EACSD format.
Structures and natural features which have a similar impact on flow to manmade structures
should be surveyed as in Section IV of the EA Specifications. Culverts should be surveyed

6
Storm Georiver http://storm-georiver.com/ and MBS Survey Software RXS Tools http://surveymbs.com/our-
software/mbs-rxs-tools

Flood Modelling Guidance for Responsible Authorities v1.1 28

as in section XI of the EA Specifications. If detailed economic appraisal is also being carried
out, inclusion of topographical survey associated with this (threshold levels, etc.) should be
considered within the same commission to minimise costs.

A survey specification for channel survey should contain

• start and end points for the survey
• a specified cross-section spacing
• cross-section width
• instructions for any specific locations to be surveyed
• an instruction to survey structures
• a plan showing the area to be surveyed
• the location of any known flood defences
• whether the bed levels to be surveyed are hard or soft or both
• the hydraulic modelling software format to be used for the deliverables

Access to the river should also be considered when commissioning a new survey. While
SEPA and Local Authorities have powers of entry under the FRM Act (section 79(1)) 7 it is
recommended that access is pursued by mutual agreement with landowners, taking
cognisance of any constraints. Introductory letters may be provided to surveyors to facilitate
access

5.5.4.2 Cross-Sections
The appropriate cross-section spacing depends on the physical characteristics of the
channel and the scale and purpose of the study; for instance, cross-sections may be further
apart for a channel with a uniform cross section and slope, and more frequent cross sections
may be required for design of flood defences. It is therefore difficult to provide general
guidance on cross section spacing, however generally:
• For large rural rivers on low slopes the maximum cross-section spacing should be
around 200 m;
• For smaller streams, sleeper slopes, or within urban areas the maximum cross
section spacing should be around 50 m;

Cross-section data is also generally required at the following points:

• All major obstructions to flow, such as bridges and culverts as well as road and rail
embankments across the floodplain;
• Points of significant changes of shape in the channel and/or change in the floodplain
width;
• Significant changes in stream slope, or near control sections (e.g. rapid drops at
weirs and dams;
• Areas where there is a significant change in channel or bank roughness;
• At existing flood protection structures;
• Upstream and downstream of confluences with significant tributaries.
• At gauging stations or other locations where information is available for calibration
(SEPA can provide information on gauging station locations, and recommended
survey requirements).
• Other key areas of interest, e.g. adjacent to proposed development sites.

7
The FRM Act gives powers of entry to persons authorised by SEPA (79.1) and local authorities (79.2) for
carrying out certain functions including the production of flood hazard maps and the preparation of flood risk
management plans.

Flood Modelling Guidance for Responsible Authorities v1.1 29

As an illustration, Figure 5-1 shows the factors
considered by SEPA in specifying cross sections Point to Note:
for a model of the River Garry.
A high quality (LIDAR) DTM means
The floodplain width and the availability and cross-sections need not cover the
appropriateness of other data for representing entire floodplain width.
out of bank flow paths should be considered in
establishing the required cross-section width. A low quality DTM (e.g. NEXTMAP
Digital Terrain Models (DTMs) used to represent data or ‘poor’ quality LIDAR) means
out of bank flow paths are discussed in Section cross-sections should cover the entire
5.5.5; if a sufficient quality DTM, typically LIDAR, floodplain width, or a new DTM
is available over the entire width of the should be collected for the floodplain.
floodplain, and considered to be a good
representation of the actual ground surface, cross sections should be extended as in Section
IV of the EA Specifications (e.g. 5 m beyond the top of bank or for vegetated banks 5 m
beyond the vegetation line but no more than 50 m).

Where the DTM is not of sufficient quality, typically areas where only NEXTMap data is
available (see section 5.5.5.1), there are quality issues with the LIDAR or significant ground
changes have occurred since the LIDAR was flown, extending cross sections across the
entire floodplain width may be desirable. An initial estimation of floodplain width can be
made from SEPA’s flood maps and checked during a site visit.

5.5.4.3 Hydraulic Structures

Information is required for all hydraulic structures which may have an impact on flow at the
scale of the study. This includes natural drops which may have the same effect as man-
made weirs. The information which is required and the structures which are important
depend on the level of detail of the modelling. Table 5-2 provides a guide on whether
structures may need to be considered at the different levels of assessment described in
section 3.6 although it should also be considered where structures are likely to be blocked.
In general data is not required for temporary structures. Details on the information ideally
required for modelling different types of hydraulic structures for different levels of study is
given in Table 5-3. It should be noted that it may be more cost effective to survey all
structures at the level required for a detailed study rather than remobilise surveyors to
collection additional data for a more detailed study at a later date.

5.5.4.4 Flood Defence Survey

Fluvial flood defences should be included within a channel survey. In general,
embankment and wall crest levels (ideally as a georeferenced string across the top of
the embankment), the size and location of any gates, outfalls, weirs etc. should be
supplied. If undefended modelling is required, information on embankment and wall
toe levels will be necessary to remove flood defences from the base model.

The Scottish Flood Defence Asset Database (SFDAD), available at

http://www.scottishflooddefences.gov.uk/Site/SE_Splash.asp, contains some survey data for
some flood defence schemes constructed under the Flood Prevention (Scotland) Act 1961.
However, due to the age of some of the survey data, checks should be carried out as in
Section 5.5.1.

Flood Modelling Guidance for Responsible Authorities v1.1 30

 Figure 5-1: Consideration of cross section spacing.

Flood Modelling Guidance for Responsible Authorities v1.1 31

Level
Compound Structures
of Bridges and short culverts Weirs and Gates Long Culverts Dams/Reservoirs and Lochs
and Mills
study
Not necessary, to include Not necessary, to Simplified representation Not necessary, to include Embankments and spillways
explicitly, but there must be a include explicitly. If likely to be necessary. explicitly but there must be need not be represented
continuous flow path through gates are usually open a continuous flow path explicitly.
the structure. there must be a Information on the path of the through the mill.
Simplified

continuous flow path culvert, survey of inlet and Flow attenuation should be
through the gate. outlet structures and walk Short lades may generally accounted for in the hydrology.
through or CCTV survey be omitted.
identifying constrictions may
be sufficient.

Should generally be Should generally be Should generally be included The main flow path and Embankments and the main
included. Footbridges and included, unless from cross section survey. opening should be controls should be represented
Intermediate

pipe bridges may be omitted drowned at low flows. included based on survey explicitly.
if they are considered to Manholes and sewer data. It may be appropriate
have a negligible impact on connections may be omitted. to omit minor flow paths. Flow attenuation should be
flow e.g. wooden plank accounted for in the hydrology.
footbridges and some pipe
bridges.
Should generally be included Should generally be Should be included from All flow paths through the Embankments and spillways
unless clear span. included. cross section survey. Mill should generally be should be represented
Detailed

Manholes and sewer included. explicitly.

connections should generally
be included. Flow attenuation should be
modelled.
Table 5-2: Guide for including different types of hydraulic structures at the different levels of assessment, described in Section 3.6 This is a guide
only and the appropriate level of detail will depend on the local circumstances.

Flood Modelling Guidance for Responsible Authorities v1.1 32

 Weirs and natural drops or
Bridges Culverts Mills Dams and reservoirs Gates
constraining features
Size and shape of opening Section along weir crest Size and shape of the opening As for bridges culverts Top of dam elevation; As for
including soffit and (crest elevation may suffice and gates. weirs
springing levels; for strategic study) Entrance type, shape of Normal depth elevation;
wingwalls etc. Number
Length of bridge tunnel Weir length Spillway type and type
Bar size and spacing for any of gates
Upstream (and Weir long profile trash screens Inlet and outlet elevations
downstream channel and dimensions Sill levels
sections if different to Skew angle or length Any bends, cross section
upstream section); perpendicular to river for changes, obstructions, or Depth area/volume Gate
labyrinth weirs changes in bed slope along relationship. heights
Size and shape of opening, culvert.
Sections immediately Maximum
Number and width of piers upstream and downstream of Soffit and bed elevation at opening
weir entrance and exit and along
Bridge parapet elevation culvert.
and type. Sections upstream and
downstream where channel Upstream and downstream
Skew angle. returns to normal cross channel sections
section.
Similar information is Barrel roughness and condition
required for any relief Similar information is also
arches or culverts. required for any side weirs. Any manhole locations

Table 5-3: Information required for modelling different types of hydraulic structures.

Flood Modelling Guidance for Responsible Authorities v1.1 33

5.5.5 Digital Elevation Models (DEMs)
A DEM is a 3D representation of a ground surface created from elevation data. A model of
the ground surface within the floodplain is required for almost all flood modelling studies. For
2D or 1D-2D modelling approaches the ground model is used to route flows over the
floodplain. For 1D modelling approaches a ground model is required for developing the out
of bank schematisation, and for mapping the results.

Ground models may be bare earth where features such as buildings have been removed,
usually referred to as a Digital Terrain Model (DTM), or may contain the elevations of
surface features (e.g. buildings, vegetation), commonly referred to as a Digital Surface
Models (DSM). DTMs are more commonly used in flood modelling.

DTMs may be constructed from ground-based topographic survey, from remote sensing data
(e.g. LIDAR) or from a combination of these. The required resolution and accuracy for a
DTM depends on the modelling objective and approach, and the study area. The DTM
resolution determines the finest possible 2D model resolution, as the resolution cannot
practically be increased beyond that in the available DTM. Further discussion on the
required resolution for 2D models is given in Section 7.3.1.

DTM information may also be required at a Point to Note:

higher resolution than the 2D model resolution
in order to allow the modeller to identify the Recommended DTM accuracy:
elevations of flow paths and obstructions in the
floodplain. Local, detailed flood studies require RURAL floodplains
a more accurate and higher resolution DTM Vertical: 0.5m
than strategic or catchment level studies. For Horizontal; spatial resolution of at least
rural floodplains a general recommendation is 10m
that the DTM has a vertical accuracy of 0.5m
and a spatial resolution of at least 10 m, while URBAN floodplains
for urban areas a vertical accuracy of 0.05 m Vertical: 0.05m
with a spatial resolution of 0.5 m may be Horizontal; spatial resolution of up to
required to resolve gaps between buildings 0.5m in some cases.
(Mason, Schumann, & Bates, 2001).

5.5.5.1 Existing Data Sets

There are several existing DTMs, based on remote sensing data, available for Scotland. In
assessing whether these are sufficient for use or if additional data may need to be collected
it is important that the resolution, accuracy and collection date are considered.

The accuracy depends on the method of data capture, and how well surface features have
been removed from the DSM to create the DTM. In areas with dense vegetation or buildings
the accuracy of the DTM may be reduced due to the need to remove features from the DTM.

Table 5-4 provides a summary of the different datasets available for Scotland together with
comments on their use and limitations. Generally LIDAR collected from aeroplanes is the
best DTM for flood modelling. Figure 5-2 shows current LIDAR coverage for Scotland.

Flood Modelling Guidance for Responsible Authorities v1.1 34

There is no programme to collect a DTM for Scotland at regular intervals so in almost all
cases the best available data represents a single snapshot in time. There may therefore be
issues if there have been significant changes since the DTM was collected, such as infilling
of docks, river realignment, developments involving land raising or erosion or deposition.
For coastal studies the state of the tide when the DTM was collected should also be
considered.

Whether linear features such as flood defences, agricultural embankments, railway and road
embankments and cuttings and small watercourses are picked up in a DTM depends on the
resolution as well as the size of the structure. For this reason a DTM should always be
supplied to modellers at the highest resolution available. It is however unlikely that any of
the available DTMs for Scotland are of sufficient accuracy to determine elevations of flood
defences or resolve local drainage networks.

5.5.5.2 New Data Collection

It may be necessary to collect new data where the existing DTM is insufficient to meet the
modelling objectives. There are 2 principal survey methods which may be used for this –
ground-based topographic survey or LIDAR surveys.

5.5.5.2.1 LIDAR
High mobilisation costs for airborne LIDAR may be prohibitive for surveys of small areas. It is
generally more cost-effective (in terms of the cost per km2 of data collected) to survey larger
areas; working in partnership with other organisations can increase the size of survey areas
and reduce costs.

Ground control points should be compared with the elevations in the LIDAR data and if
possible a data validation exercise should be carried out following data collection.

LIDAR should be flown when there is no dense vegetation cover as this may obstruct the
laser from reaching the ground surface. Generally, data collection should be carried out in
the period following significant autumn leaf fall and before the main spring growing season.
Where data collection is to be carried out during the winter months issues with snow cover
should be considered. Snow cover prevents the laser from reaching the ground surface.
Data collection should also not take place when the ground is flooded, as this will also
prevent the laser from reaching the ground surface. Surveys should be planned with some
contingency time to allow for local conditions. For coastal studies LIDAR should be flown as
close to low tide as possible to allow for a detailed representation of the coastline to be
collected.

Ground-based LIDAR can be used to collect more detailed data, such as kerb heights, which
may be of use for detailed surface water studies. The data collected in these surveys can
also be used to develop 3D visualisations of flooding.

Flood Modelling Guidance for Responsible Authorities v1.1 35

Figure 5-2: LIDAR coverage for Scotland (June 2015).

Flood Modelling Guidance for Responsible Authorities v1.1 36

 Horizontal Vertical Comments/
Data Source Description Coverage Availability and licensing Year
Resolution Accuracy Limitations
Various See Figure 5-2 1m–2m ± 0.15 m Dependant on data source and Varies Best available.
LIDAR proposed use. Generally ties in well
datasets with ground based
available For Local Authority studies LIDAR survey.
LIDAR should be supplied by the Local
Authority. Other Public Bodies may
hold LIDAR which SEPA
Other public bodies should contact is unaware of.
The Scottish Government
Synthetic National 5m Depends Contact Scottish Government 2002- Known issues in
Aperture on area ± 2003 forested and urban
Radar (SAR) 1 m or ± areas. Accuracy means
NEXTMap 0.7 m that there may be a
considerable offset with
other data in some
locations.
Stereo Aerial National 5m ±0.6 m Contact Getmapping Not widely tested by
Photography SEPA. Found to be
Getmapping
better than NEXTMap in
some areas but not all.
Stereo Aerial National 5m ±2m Ordnance Survey Quarterly New dataset not tested
Photography updates, by SEPA.
OS Terrain 5 and ground but may
based survey not cover
all areas.
National 50 m Ordnance Survey – open data Annual Resolution probably too
updates, low for most mapping
OS Terrain
but may studies.
50
not cover
all areas.
Table 5-4: Summary of known DTM data sources for Scotland.

Flood Modelling Guidance for Responsible Authorities v1.1 37

5.5.5.2.2 Ground-Based Topographic Survey
Ground-based survey to develop a DTM can either be on a grid pattern across the study
area at the required resolution or targeted at particular features such as embankments,
ditches or curb or threshold levels.

Where the data is targeted at particular features, modellers may be able to combine the
ground based survey with existing remotely sensed DTM to create a new DTM. However,
there may be issues resolving inconsistencies between the 2 datasets especially in areas
where LIDAR is unavailable.

New topographic survey should be according to Section III of the EA Specifications.

5.5.6 Bathymetric Data

Existing bathymetric data is available from a number of sources.

The European Marine Observation and Data Network EMODnet has produced a 1/8th arc
minute 8 resolution bathymetric dataset of European waters (approximately 130 m east-west
resolution and 230 m north-south resolution at 56º north). The bathymetry uses data from
hydrographic offices, authorities responsible for management and maintenance of harbours,
coastal defences, shipping channels and waterways, and research institutes and industry,
and General Bathymetric Chart of the Oceans (GEBCO) bathymetry where no other data is
available. The bathymetry is freely available from the EMODnet bathymetry portal. This
data is unlikely to be at a sufficient resolution for detailed coastal studies but may be suitable
for regional models. There are some discontinuities in the data at the boundaries between
datasets.

Some gridded and point data from the United Kingdom Hydrographic Office (UKHO) is
available under the Open Government License and can be downloaded from the UKHO
inspire portal http://aws2.caris.com/ukho/mapViewer/map.action. However, note that the
data has not been processed to remove conflicts between datasets at boundaries.

Published Admiralty charts provide bathymetric data at a range of resolutions. Electronic

datasets can either be generated by digitising these charts, subject to license from the
UKHO, or by purchasing existing electronic versions of the datasets from commercial
suppliers (i.e. C-Map, SeaZone). Sea Zone also produce a gridded dataset combining chart
data and bathymetric survey. However, license conditions for the electronic versions from
these suppliers may restrict usage.

Harbour authorities may also have datasets which they are willing to share for flood studies
and should be contacted directly.

5.5.6.1 Commissioning New Bathymetric Surveys

If new bathymetric survey is required, section 7 of the EA Specifications provides an
appropriate specification.

5.5.6.2 Coastal Defence Survey

In addition to toe and crest levels for wave overtopping models a profile through the flood
defence is required, including information on the roughness.

5.6 Operating Information

8 th
An arc minute is 1/60 of degree.

Flood Modelling Guidance for Responsible Authorities v1.1 38

Information on the operation and maintenance of flood defences, river channels and
reservoirs should be provided where possible. There are no set formats for the supply of
this information.

Where there are flood defences, or other structures which may be operated during flood
events, control rules and procedures will be required for any detailed study. Where
structures are manually operated, information on when they were operated during any
events to be used for calibration will be required.

Information on known/frequent blockage locations and dredging and weed cutting regimes
can be used to inform model sensitivity tests.

5.7 Existing Studies

Key points for Responsible Authorities

• Any existing models should be reviewed by an experienced modeller to

determine if they are suitable for use in the new study.

SEPA and the Responsible Authority may be able to identify if any existing models cover the
study area. Where existing models are available it should be considered whether they can
satisfy the new study. However, it should be noted that sometimes the modification of an
existing model to meet the objectives of a new project is as much work as starting from
scratch. Questions which may need to be considered are:

• What was the purpose of the existing modelling? Are there assumptions and
limitations which make it unsuitable for the new purpose?
• Is the study area appropriate or would it need to be extended?
• Is it at a sufficient level of detail at the study location, or is it too detailed? Are there
any areas of the model with lower confidence?
• Have there been any significant catchment changes since the model was
constructed?
• What data was used for the original model? Is better data now available e.g. LIDAR
instead of NEXTMap for the out of bank DTM, or are there known issues with some
of the original data sets?
• Are there any reasons to suspect that the existing model does not provide a good
representation of the system? Is the model calibrated?
• Is the model numerically stable, or are run times excessive?
• What software and version was used for the original study. Will it run in more recent
versions? Are there any significant changes to results between old and new software
versions?
• If the model was built for a flood prevention scheme is it a design model or an as built
model?
• Is the model georeferenced to enable flood mapping?
• Does the model contain sufficient out of bank representation to model the range of
scenarios required for the project?
• Do intellectual property rights allow the model to be used for the new study?
• Can the model be obtained by the Responsible Authority?

Flood Modelling Guidance for Responsible Authorities v1.1 39

 • Is the model sufficiently well documented?

An experienced modeller will be required to answer some of these questions and a

contractor should allow time in a tender for auditing any existing model and reviewing
modelling reports.

Irrespective of whether an existing model is to be reused for the new study, reports from
existing studies should be referenced by the new study. These reports may provide
information on historic flood events, highlight issues with adopting particular approaches for
the study area, and identify flooding mechanisms which require further analysis in the new
study or help in identifying the reasons for any discrepancies between the existing and new
studies.

5.8 Hydrometric Data

Hydrometric data can be used to develop flood frequency curves and generate boundary
conditions for design events and to calibrate a model. These uses are covered in Chapter 6
and in Chapter 8.

In general two types of data series are of interest

Point to Note:
for flood studies: time series data and series of
extremes. Two types of extreme datasets are
Time series data – a more or less
available: Annual Maxima (AMAX) the largest
continuous record normally
event in any given year and Peak Over
considered in relation to specific
Threshold (POT) all events over a given
events.
threshold. For AMAX series years should be
defined so that series are not cut at a flood prone
Extreme series data – dataset of
time of year, to ensure that maxima in
extreme records; two types:
consecutive years are independent. In general
extreme series will be required for the full length
AMAX – the annual maxima
of record, whereas time series data is only
POT – all events over a given
required for particular events. However, it may
threshold
still be necessary to supply entire time series to
enable suitable events to be identified.

Responsible Authorities should consider the need for any additional hydrometric data as
early as possible in a Flood Risk Management Planning cycle, as the greatest benefits are
likely to be obtained from a longer period of monitoring.

Key points for Contractors

• River flow data held in the NRFA for Scottish sites is not up-to-date. Data for any
Scottish gauges used in the study should be requested from SEPA.
• Complete time series should be requested for any sites used in model calibration.
• Rain gauge data should be requested for gauges within, and surrounding the
catchment.
• Any Met Office data (if required) should be requested directly from the Met Office
enquiries@metoffice.gov.uk. Government bodies should include the phrase
Government Enquiry in the subject line to avoid being charged commercial rather
than government rates.

5.8.1 River Levels and Flows

Flood Modelling Guidance for Responsible Authorities v1.1 40

SEPA maintains a network of river flow and level gauges. Information on the location of
these gauges is available from datarequests@sepa.org.uk. SEPA is planning to add a web
link to a database in the near future. A proportion of SEPA gauges, including those
contributing to the Peak Flow Database (formally known as HiFlows-UK), appear on the
Centre for Ecology and Hydrology (CEH) website
(http://www.ceh.ac.uk/data/nrfa/data/search.html), which has a number of search options
including an interactive map option. The data available from these gauges includes 15
minute time series data, daily max, monthly max and annual maximum (AMAX) series,
although not all of this information is required for the majority of flood studies.

AMAX or POT series of flows are required for design flow estimation for fluvial flood studies.
AMAX series are sufficient, unless it is necessary to generate flow estimates for events with
a recurrence interval < 2 years, or the record length is < 14 years in which case POT series
are desirable. Data is required for any gauges within the study area, gauges upstream or
downstream of the study area along the same watercourse, for any sites used in a pooling
group and for any donor sites. AMAX series should be generated based on UK water years
(1st October – 30th September) rather than calendar years.

Peak flow estimates are derived from a stage-flow relationship (rating) that is calibrated
from sample flow measurements (gaugings) through as much as the flow range as possible.
However, the derived flow from the rating can only be guaranteed up to the highest gauging,
which is often well below the highest recorded stage. Derived flows above the highest
gauging are based on a simple extrapolation of the rating and are only credible up to bank
full level or as a rule of thumb, 10% above the highest gauged flow. The exception to this is
where any floodplain flow has been modelled. The upper limit of sensible derived flow from a
rating needs to be understood and factored into how the data is used. The local hydrometric
team can advise on rating limits for each station.

AMAX and POT series for some SEPA gauges is Point to Note:
available from the National River Flow Archive
(NRFA) managed by held CEH. This includes the Early engagement with the local
Peak Flows dataset for use with the Flood SEPA Hydrometry team is
Estimation Handbook (FEH) discussed in Section essential to confirm the availability
6.3. The Peak Flow Database has not been and quality of gauged data.
updated with SEPA data since October 2006,
however, SEPA has recently transferred revised and up-to date data to the CEH for 30
stations and these should appear on the website by the end of 2016. SEPA will work to
update all the SEPA gauging stations in the Peak flows station pool, currently numbering
140 stations, over the new few years. In the meantime, up to date AMAX for the study area
for any Scottish sites used in a pooling group analysis, and for any donor sites, should be
requested directly from SEPA. It is critical to liaise with the local SEPA hydrometric team to
ensure that flows are derived using the most appropriate rating for flood estimation. SEPA is
currently working to ensure that all AMAX are derived from the most appropriate ratings and
until this exercise is complete, a check with the local team is necessary. On occasion it may
be simpler for SEPA to supply level data and a rating rather than derived flow data. The
exercise also aims to produce POT datasets for those sites contributing to the Peak Flow. It
should be noted that SEPA may have revised its rating curves since the last update to
HiFlows-UK. These discrepancies will be resolved in future updates as per the exercise
described above. Updates on this project can be obtained from SEPA.

Time series of both flow and level are required for model calibration. Calibration events may
be identified based on a number of factors including data availability at several gauges and
in, most cases, it is easier to request complete datasets rather than submit multiple requests
for chunks of the same dataset. Time series of flow are required for any gauges within the
study area and gauges upstream or downstream of the study area along the same

Flood Modelling Guidance for Responsible Authorities v1.1 41

watercourse. Time series of level are required for any gauges along the modelled reach.
Time series of flow may also be used in deriving design hydrograph shapes. If spot
gaugings are available these can also be used to assist with model calibration, however this
should be discussed with the local hydrometric team who can advise on rogue or
problematic gaugings excluded during rating development.

SEPA’s Hydrometry team is able to provide information about the reliability of particular
gauges and suitability for measuring high flows, including a history of the site. Rating curves
are also useful for investigating any discrepancies between the flow record and model
output.

Other organisations such as Local Authorities or water companies may also operate flow or
level gauges.

On large costly projects consideration should be given to installing additional flow monitoring
equipment to collect data. This would be of particular advantage on ungauged watercourses
and could prevent over-design with resulting cost savings.

5.8.2 Rainfall Data

SEPA maintains a network of rain gauges. Information on the location of these gauges is
available from datarequests@sepa.org.uk. SEPA is planning to add a web link to a database
in the near future. Other organisations such as water companies may also install temporary
rain gauging. Data from Environment Agency gauges should be requested from the
Environment Agency.

Radar data is available from the Met Office and may also be useful in model calibration. To
avoid being charged commercial rather than government rates, government bodies
(including local authorities) should request radar data directly from the Met Office
enquiries@metoffice.gov.uk and include the phrase Government Enquiry in the subject line.
Contractors should not request data directly from the Met Office.

Time series of rainfall may be used for calibrating fluvial and pluvial models. For calibrating
models data will be required for gauges within and surrounding the study catchment as the
most representative gauge may not be within the catchment and rainfall applied to a model
may be from area weighting rainfall from the a number of different gauges. If it is necessary
to generate antecedent conditions for the calibration events rainfall data may be required for
a long period prior to the calibration event. Calibration events may be identified based on a
number of factors, including data availability at several gauges, and in most cases it is easier
to request complete datasets rather than submit multiple requests for chunks of the same
dataset.

The FEH Depth Duration Frequency (DDF) model is generally used for generating design
rainfall so annual maximum and POT series are not usually required. The FEH DDF model
gives rainfall depth as a function of return period and storm duration for all catchments > 0.5
km2 and on a 1 km grid across the UK. Example output from the FEH DDF model is shown
in Figure 5-3. A new version of the DDF model (FEH 2013) was released in 2015, and
replaced the existing DDF model (FEH 1999) for a complete range of return periods and
durations. Both the FEH 1999 and FEH 2013 DDF models are available through the FEH
web service at https://fehweb.ceh.ac.uk/. The FEH 1999 DDF model is also available
through the FEH CD-ROM.

FEH 2013 incorporates a significant amount of additional data to FEH 1999 and uses an
improved statistical model. In Scotland FEH 2013 includes 176 hourly rain gauges
compared to 58 in FEH 1999 which is a significant improvement in data coverage for the

Flood Modelling Guidance for Responsible Authorities v1.1 42

rainfall durations likely to be of most interest to flood risk management. As a result of the
improved data coverage and statistical methods, FEH 2013 has higher depths than FEH
1999 for short duration rainfall (< 6 hours) for most locations in Scotland up to the 0.1% AEP
event. Further information on the development of the FEH 2013 model is found in the Defra
technical report Reservoir Safety – Long Return Period Rainfall (Stewart, et al., 2010).

Figure 5-3: Example output from the FEH Depth Duration Frequency Model (DDF).

5.8.3 Soil Moisture and Evaporation Data

Soil moisture and evaporation data are used to assess catchment antecedent conditions for
flood event analysis, and to calibrate rainfall runoff models. The Met Office Rainfall and
Evaporation Calculation System (MORECS) provides an assessment of soil moisture on a
40 km square grid for the UK. To avoid being charged commercial rather than government
rates, government bodies (including local authorities) should request MORECS data directly
from the Met Office enquiries@metoffice.gov.uk and include the phrase “Government
Enquiry” in the subject line. Contractors should not request data directly from the Met Office.

5.8.4 Reservoir Data

Where reservoirs are considered to have a significant impact on flow regimes within a
catchment, time series information on levels and releases from the reservoirs and depth
volume curves are likely to be required for model calibration, and in developing flood
frequency curves. This data should be obtained from the reservoir operator. For reservoirs
with a retained volume of 25,000 m3 or greater details of the reservoir operator can be
obtained from the public register.

5.8.5 Sea Level Data

5.8.5.1 Gauge Data

Tide gauge data is available from SEPA, the British Oceanographic Data Centre (BODC)
and port/harbour authorities.

Data from the national tide gauge network can be obtained free of charge from BODC
https://www.bodc.ac.uk/data/online_delivery/ntslf/. There are 43 gauges within this network
of which Leith, Aberdeen, Wick, Lerwick, Kinlochbervie, Ullapool, Stornoway, Tobermory,
Port Ellen (Islay), Millport and Portpatrick are in Scotland. Both the measured sea level and
the residual or surge (difference between the measured sea level and the astronomical tide)
are supplied. Time series and monthly extremes are available.

Information on the location of SEPA gauges is available from datarequests@sepa.org.uk.

Flood Modelling Guidance for Responsible Authorities v1.1 43

Data from harbour or port authority gauges should be requested directly from the relevant
port authority.

Time series of measured sea level are required for calibration of coastal models and for the
tidal areas of fluvial models. Extreme series of sea level are used to develop design water
levels, however in most cases design sea levels in the Environment Agency Report Coastal
Flood Boundary Conditions for UK mainland and islands (McMillan, et al., 2011), described
in section 5.8.5.3 are used. Extreme sea level series are therefore only needed if it is
wished to extend the analysis to sub-annual events or areas within estuaries or sea lochs
which are not covered by the study.

5.8.5.2 Tide Tables

Astronomical tide curves and harmonic constants are available from the Admiralty Tide
Tables, Admiralty Total Tide and other suppliers such as C-Map.

5.8.5.3 Sea level boundary conditions

Design sea levels are given in Environment Agency Report Coastal Flood Boundary
Conditions for UK mainland and islands (McMillan, et al., 2011). The design sea levels are
developed from a statistical analysis of extremes from the class A tide gauge network
together with data from a small number of other sites. The study gives extreme still water
levels and representative surge shapes for 1, 2, 5, 10, 20, 25, 50, 75, 100, 150, 200, 300,
500, 1000 and 10 000 year return period events. The study applies to the open coast only,
and the further analysis required to develop design sea levels within estuaries is discussed
in section 6.5.1. Levels from the Coastal Flood Boundary Project have already been
supplied to Local Authorities by SEPA. Surge shapes from the Coastal Flood Boundary
Project should be requested from SEPA.

The SEPA Coastal Hazard Mapping Project (Royal Haskoning DHV and JBA, 2013)
extended the extreme sea level analysis from the CFB study to sea lochs and estuaries
within Scotland but did not derive representative surge shapes for these locations. Where
surge shapes are required within sea lochs and estuaries, a hydrodynamic model may be
required to model how the shape of the surge changes within the estuary, section 6.5.1.

5.8.6 Wave Data

Wave observations for some sites are free to download from websites including those
referenced in Table 5-5, however, there are limited gauge sites and, for the majority, only a
short record is available, limiting use of the data to model calibration and validation.

In the absence of long duration wave observations time series data at wave model offshore
boundaries are typically taken from wave model hindcasts. Wave model hindcasts are also
used to generate AMAX and POT series for design event analysis and are available on
request from the organisations noted in Table 5-6.

Organisation Web Address

Cefas Wave Net http://www.cefas.defra.gov.uk/our-science/observing-and-
modelling/monitoring-programmes/wavenet.aspx
BODC https://www.bodc.ac.uk/data/online_delivery/waves/
Table 5-5: Organisations providing wave observation data.

Organisation Time Period Data Available Contact Address

1980-2014 enquiries@metoffice.gov.uk (Nb. Local Authorities
Met Office
should contact SEPA in the first instance)
NOAA 1979-2007 http://polar.ncep.noaa.gov/waves/CFSR_hindcast.s
html

Flood Modelling Guidance for Responsible Authorities v1.1 44

Table 5-6: Organisations providing wave hindcast data.

Contractors should not request data directly from the Met Office to avoid being charged
commercial rather than government rates. The NOAA model also covers the UK and is
available from the NOAA website although it is not calibrated specifically for the UK.

The available data from these models is typically wind wave, swell and resultant (wave
and swell combined) waves, significant wave height, mean period and mean direction;
the complete frequency spectrum is not usually available.

On large, costly projects consideration should be given to installing additional wave buoys to
collect data, this can be expensive, but will may increase confidence preventing over-design
with resulting cost savings.

5.8.7 Wind data

Time series of wind data may be required for calibrating coastal models and for generating
wind wave boundary conditions. Wind observations are primarily available from the Met
Office although data may also be available from local weather stations. To avoid being
charged commercial rather than government rates, government bodies (including local
authorities) should request radar data directly from the Met Office
enquiries@metoffice.gov.uk and include the phrase “Government Enquiry” in the subject
line. Contractors should not request data directly from the Met Office.

5.9 Other Data

5.9.1 Sewer Network Data

Scottish Water is the organisation responsible for the design and management of Scotland’s
sewerage systems. It can therefore provide information on the sewer network, including the
location of outfalls. Data should be requested from the Scottish Water Area Asset Manager.
Sewer network data can be useful in identifying catchment boundaries in urban areas where
catchments may be significantly different from the natural catchment boundaries due to the
surface water drainage network.

5.9.2 National Coastal Change Assessment

For coastal studies the National Coastal Change assessment http://www.dynamiccoast.com/
provides information on the susceptibility of the coast to erosion. This can provide an
indication of whether the potential for coastal erosion to contribute to flood risk should be
considered in flood studies.

5.9.3 Water Framework Directive Status/Pressures

Information on Water Framework Directive (WFD) Status and Pressures as identified in the
River Basin Management Plans can be found on the SEPA Water Environment Hub
http://www.sepa.org.uk/data-visualisation/water-environment-hub/. This information should
be reviewed at scoping stage to identify potential opportunities for studies and measures
with a joint WFD and Flood Risk Benefit, and to assist in identifying structures (including
informal flood defences) which would need to be included in any channel survey.

5.9.4 Mapping Data

Mapping data has several uses in flood modelling. At the scoping stage, maps are used to
identify structures and other features which may influence flow pathways and possible
receptors. During model build, maps are used to identify building footprints, and areas likely
to have different roughness, or loss of water to the drainage network. Maps are also used
in communicating model results. Sources of mapping data typically used in flood modelling,

Flood Modelling Guidance for Responsible Authorities v1.1 45

and their principle uses are given in Table 5-7. Ordnance Survey data is available to Public
Bodies under the One Scotland Mapping Agreement 9, and the license conditions also allows
public bodies to pass data to their contractors, subject to conditions.

Map Type Location Notes

OS Colour https://www.ordnancesurvey. Display of model outputs
Raster Maps co.uk/ Identification of structures and features.
OS Building footprint identification.
Mastermap Identification of roughness zones.
Historical http://maps.nls.uk/index.html Identifying former paths of watercourses, potential
Maps culvert routes and artificially drained areas. This is
useful when developing a conceptual model.
Land Cover http://www.ceh.ac.uk/landcov Identification of roughness zones. Identification of
Map 2007 ermap2007.html rural and urban areas.
Scottish http://www.gov.scot/Topics/S Identification of rural and urban areas.
Government tatistics/About/Methodology/
Urban Rural UrbanRuralClassification
Aerial Various sources Display of model outputs
photographs Identification of structures and features.
Identification or roughness zones.
Table 5-7: Map data used to support flood modelling studies.

9
https://www.ordnancesurvey.co.uk/business-and-government/public-sector/mapping-agreements/one-
scotland-mapping-agreement.html

Flood Modelling Guidance for Responsible Authorities v1.1 46

6 Boundary Conditions

6.1 Introduction
Hydrological analysis is required to determine design flows and probability of a flood. These
are used as input boundary conditions for hydraulic models.

Estimates of probability are not static but may change over time due to changes in climate or
catchment and due to changes in the data record or best practice flood frequency analysis
techniques. This means that it is important to review the hydrology for any new flood study,
even if a hydrological analysis has been carried out for a previous study.

Hydrometric data in the UK are generally of high quality; however uncertainty is inherently
present when conducting flood frequency estimates due to the length of record compared to
typical design probabilities of interest, the range of different of analysis methods available,
and the coverage of the gauge network.

This chapter aims to:

• Give a brief description of the methods which can be used for hydrological analysis of
fluvial, pluvial and coastal flooding.
• Describe the outputs which would be expected from a hydrological study.
• Describe the circumstances in which SEPA may seek to support a review of
hydrological analysis.

Key points for Responsible Authorities

• All flood frequency estimates are inherently uncertain and subject to change.
• The hydrology should be reviewed for any new flood study.

6.2 Terminology
Key points for Contractors

• Where possible, flood studies should use the annual exceedance probability
terminology rather than return period.

Two terms are commonly used to describe the flood frequency in the UK, the return period
and the annual exceedance probability.

• The return period of an event is the average interval between years containing an
event of the same or greater magnitude. A similar measure is the average
recurrence interval, which is the average period between events of a same or
greater magnitude.
• The annual exceedance probability (AEP) is the probability that an event of the
same or greater magnitude will occur in any one year. This is the reciprocal of the
return period.

Use of the terms return period or average recurrence interval can cause some confusion
amongst non-specialists who can misinterpret it to mean that events occur at fixed intervals.
For this reason use of annual exceedance probability, rather than return period is preferred.

Flood Modelling Guidance for Responsible Authorities v1.1 47

 Return Average Annual
Period Recurrence Exceedance
(years) Interval (years) Probability (%)
2 1.44 50.0%
5 4.48 20.0%
10 9.49 10.0%
30 29.50 3.3%
50 49.50 2.0%
100 99.50 1.0%
200 199.50 0.5%
1000 999.50 0.1%
Table 6-1: Different ways of presenting flood probability.

6.3 Fluvial

Key points for Responsible Authorities

• Design flows should not be finalised or signed off by either the Responsible
Authority or SEPA before model calibration and reconciliation is complete, and
should be reviewed and revised as a modelling study progresses.
• Identify key locations (reconciliation points) where it is important for flows in a
hydraulic model to match hydrological estimates, and agree the locations at the
inception meeting.
• The variation in design flows along a catchment should be physically justifiable
and explained in the modelling report.
• It may be necessary to run multiple model scenarios for the same AEP event in
order to match the design flow at different points in the model.
• Ensure that contractors allow time for appropriately experienced staff to review
any hydrological analysis. Encourage use of the supplied check list in Appendix
C.2 .

Flood Modelling Guidance for Responsible Authorities v1.1 48

 Key points for Contractors

• Determine the robustness of all hydrometric data in consultation with local SEPA
Hydrometric teams.
• Compare flow estimates from statistical (single site and pooling) and rainfall-runoff
methods.
• Consider the method used to derive hydrograph shape and run large catchment
models for multiple storm durations if required.
• The modelling report should include sufficient details of the analysis to enable an
experienced hydrologist to reproduce the flow estimates.
• Further detail on approach to hydrological analysis is available in SEPA’s
Technical Guidance for Stakeholders.
• Refer to specific guidance in the FEH.
• Document the approach used for reconciliation in the modelling report, and give
values for any scaling factors.

For all fluvial flood studies the hydrological analysis will need to produce design flow
estimates. For studies involving unsteady modelling the hydrological analysis will need to
produce design hydrograph shapes. For catchment scale studies the analysis will also need
to consider the distribution of inflows.

In most cases hydrological analysis should be based on the methods in the Flood Estimation
Handbook (FEH) (Institute of Hydrology, 1999) which provides the industry standard
methods and guidance for fluvial flood estimation within the UK. The FEH largely
supersedes the Flood Studies Report (Institute of Hydrology, 1975) and its associated
reports. The FEH provides a framework for flood estimation, however user expertise and
experience is required to judge the most appropriate methods / data to use in any individual
circumstance. No single method is considered superior to others for all situations and in
some cases other flow estimation methods than those in the FEH may be appropriate
depending on the catchment characteristics. However, if FEH methods are not used a
comparison with FEH methods should be made with justification provided as to why the FEH
methods were not considered appropriate.

Responsible Authorities should note that hydrological analysis may be more complicated or
uncertain in the following cases:
• Urbanised catchments;
• Small catchments < 25 km2;
• Catchments containing reservoirs, lochs and hydroschemes;
• Pumped catchments;
• Assessment of natural flood management measures (Chapter 12).

This means that more time may be required to be allowed for the analysis, and/or there may
be lower confidence in the results. Where these factors exist within a catchment the
approach taken to deal with these cases should be discussed in the report.

An appropriate hydrological analysis is one of the most important parts of a flood

study. Responsible Authorities should ensure that contractors allow time for appropriately
experienced staff to review any hydrological analysis and encourage use of the check list in
Appendix C.2 .

Flood Modelling Guidance for Responsible Authorities v1.1 49

6.3.1 Review of Hydrometric Data

All data received to support a flood study should be independently reviewed and checked
against the quality statements and advice of the data provider. This also applies to
hydrometric data. Before undertaking a hydrological analysis any hydrometric data supplied
for the project should be reviewed in full. SEPA makes every effort to ensure that all
hydrometric data supplied by SEPA is accurate however it is possible that some issues and
inconsistencies may only become apparent through use in a detailed flood modelling study.
As a minimum the data review should cover:

- Robustness of the rating for any gauges used in the statistical analysis e.g. variation
of the rating with time, hysteresis, location of any discontinuities
- Catchment changes which may mean that sections of the data series are no longer
valid e.g. construction of reservoirs.
- Gaps in the hydrometric data series.
- Suitability of the data for use in hydrological analysis. This should include a review of
whether the NRFA FEH indicative suitability is correct.

Any apparent anomalies should be discussed with the local hydrometric team.

6.3.2 Design Flows

The relationship between flow and probability is known as the flood frequency curve. There
are two common approaches to estimating the flood frequency curve: (i) statistical analysis
of flood peak data (single site or pooled analysis) and (ii) the design event method using a
rainfall-runoff model. For many catchments, either approach can be applied however they
can produce very different results.

For catchment or local scale models flow estimates using both approaches should be
compared and the adopted method justified. For all methods sufficient details of the analysis
should be given in an appendix to the modelling report to enable an experienced hydrologist
to reproduce the flow estimates.

Design flow estimates are required at the upstream boundary of any modelled watercourses
and at reconciliation points where it is important for the flow in a model to match the
hydrological estimates. Reconciliation points are typically chosen at:
• All gauging stations in the model domain
• Downstream of major confluences
• Key receptors such as urban areas
• Any points where the SOP of a flood defences is required.

Reconciliation points should be agreed between the Responsible Authority and the
contractor at he inception meeting.

Design flow estimates should be reviewed and revised as a modelling study progresses and
should not be accepted or signed off by either SEPA or the Responsible Authority
commissioning the study until calibration and reconciliation of any models is complete.
Where possible a combined hydrological- hydraulic approach should be should adopted for
model calibration, as the following could indicate both problems with design flow estimates
or missing processes and errors in a hydraulic model:

Flood Modelling Guidance for Responsible Authorities v1.1 50

 • Problems representing the observed flood history e.g. annual flooding is observed
and the 50% AEP event is in bank, or 50% AEP shows flooding in an areas with no
flood history.
• Inconsistencies in design flows along the catchment, e.g. increases/decreases in
design flow estimates downstream which cannot be explained by increase in
catchment area/ attenuation.

6.3.2.1 Statistical Method (single site and pooled analysis)

The statistical method is generally the first choice method where there is a long record of
gauged flood event data available. The method involves the construction of a flood
frequency curve based on the estimation of QMED (the median of the set of annual
maximum flood data with an annual exceedance probability of 50%) and a growth curve
which gives design flows for other return periods as a function of QMED.

For gauged locations QMED is typically estimated from the AMAX or POT series at the
gauge, and these data series are discussed in section 5.8. For ungauged locations QMED
is estimated using another, usually nearby, catchment with similar catchment characteristics,
a process known as donor transfer, or from catchment descriptors. Other less common
approaches include the use of channel dimension data or continuous simulation.

The FEH describes 6 methods to estimate QMED. In addition, a revised method of

estimating QMED based on catchment descriptors is described in the study by CEH
(Kjeldsen, et al., 2003). While the choice of method(s) used in the study should still be
justified, SEPA recommends that the revised QMED estimation method is considered and
that a precautionary approach is taken when estimating QMED and the subsequent
derivation of the design flood.

Where only short gauged records are available these may be used to support estimates of
QMED.

The method for estimating the growth curve depends on the length of the gauged record for
which data is available and the required return period, there are two methods a single site
analysis based on local gauged data, or a pooling group analysis which estimates QMED
and the growth curve from a group of gauging stations on other similar catchments. For
gauged catchments with a sufficient length of record a comparison of the single site and
pooling group analysis should be made. For ungauged catchments only the pooling group
method will be appropriate.

Where level only gauges are available, or where no high flow rating is available, it may be
possible to extract a rating from a hydraulic model provided that the gauge is surveyed and
modelled in sufficient detail. In this instance SEPA should be contacted to discuss survey
requirements, gauge history and available information for calibration.

6.3.2.2 Design Event Method (Rainfall-Runoff Method)

The FEH Rainfall-Runoff method uses a design storm based on the FEH DDF model
discussed in section 5.8.2. This is then run through a simple catchment model (based on
the Unit Hydrograph and a loss model) to produce a design flow estimate Figure 6-1. This
approach produces a full flow hydrograph as opposed to a peak flow produced by the
statistical method. As a result, this method is usually applied when flood volumes or
durations are important such as in the design of flood storage areas or reservoir spillways.
This method is potentially most applicable to small catchments especially where they
are ungauged.

Flood Modelling Guidance for Responsible Authorities v1.1 51

Figure 6-1: FEH Rainfall Run-off model. The model consists of a losses model which
describes the proportion of rainfall falling on the catchment which runs off rapidly (net rainfall)
and a unit hydrograph which describes the response to the net rainfall.

A new conceptual rainfall-runoff model called the Revitalised Flood Hydrograph Model
version 2 (ReFH2) has been developed which supersedes the FSR/FEH Rainfall-Runoff
Method. Recent improvements to the method have now rendered it applicable for design
flow estimation within Scotland however the methodology has been calibrated to catchments
without significant storage (i.e. lochs and reservoirs). The methodology is still being
assessed and, like any other flood estimation methodology, it should only be used in
combination with others for comparison. Note that ReFH version 1 is not considered
suitable for use in Scotland due to the limited number of Scottish gauges and lack of Scottish
specific calibration.

As discussed in Section 5.8.2 a new version of the FEH DDF model FEH 2013 has recently
been published and should generally be used over the previous version, FEH 1999 due to
the improvements in data and techniques. The differences between FEH 2013 and FEH
1999 may also have implications for previous design flood estimates carried out using
rainfall-runoff models.

Both the FEH and ReFH2 rainfall models are lumped hydrological models (i.e. they use a
single unit to represent a catchment and model parameter values are averaged across the
catchment). More complicated distributed hydrological models allow factors such as soil
type or rainfall to vary across a catchment. Distributed hydrological models are not
commonly used for flood estimation in Scotland and would not usually be accepted by
SEPA for design flow estimates; however they are potentially useful for some NFM
studies. This is discussed in chapter 12.

6.3.3 Hydrograph Shapes

The shape and duration of a hydrograph is determined by a variety of factors such as
drainage efficiency, catchment shape, channel characteristics, vegetation cover, land use,

Flood Modelling Guidance for Responsible Authorities v1.1 52

soil type and storm patterns. Hydrograph shapes are required for unsteady modelling (Table
3-3).

For ungauged catchments the hydrograph shape should be generated using a rainfall-runoff
model, as in section 6.3.2.2

Where there is gauged data available close to the model boundary a hydrograph shape can
either be derived by either:

• standardising hydrographs by their peaks and averaging (Archer, Foster, Faulkner, &
Mawsdley, 2000),
• using a hydrograph from a large observed flood event or
• using a rainfall-runoff model

In order to identify a representative large observed event or suitable events for the averaging
approach, 15 minute times series data will be required as discussed in Section 5.8.1.

In all cases the hydrograph is forced to fit the design flow estimate, either by adjusting the
rainfall runoff model parameters or by scaling the derived hydrograph shape.

The method used to derive the hydrograph shape and why it was chosen should be
discussed in the hydrology report.

6.3.4 Catchment Schematisation/Boundary Locations

For local assessments, a single upstream boundary with a flow estimate and hydrograph
based on the catchment upstream of the site may be appropriate provided there are no
significant inflows within the study extent. However, for catchment or strategic scale studies
where the flow may be reasonably expected to increase throughout the model, inflow
hydrographs are required at several points throughout the study area. In this case the
catchment is split into sub-catchments and the hydrograph for each sub-catchment is routed
through the model. A map of the catchment schematisation should be included in the
hydrology report together with the FEH catchment descriptors for each sub-catchment in an
appendix.

It is not physically realistic for different duration critical storms to occur at different points in
the same catchment at the same time. Where a catchment has been split into sub-
catchments the same design storm is applied to each sub-catchment, and the model is run
for a range of storm durations to find the critical storm duration which produces the worst
case flow or level at the site of interest.

6.3.5 Flow Reconciliation

For large catchments the AEP of a flood event may vary throughout the catchment. This
means it is not always possible to apply a consistent design storm across the catchment so
that the flow in a hydraulic model matches the design flow estimates at the reconciliation
points. In this case it is necessary to adjust the boundary conditions so that the flow in the
model matches the design flow at the reconciliation points. This process is termed flow
reconciliation, and should be undertaken after model calibration, chapter 8. There is no
single method to achieve this but typical approaches include:

Adjusting the storm duration: Different critical storm durations may be identified for
different parts of the catchment. Generally the critical storm duration will increase
downstream; any decreases in critical storm duration downstream should be justified in the
model report (e.g. confluence with major tributary with steeper/more urbanised catchment).

Flood Modelling Guidance for Responsible Authorities v1.1 53

Adjusting the hydrograph shape: Different hydrograph shapes may be required to match
design flows at different points in the catchment, or for different AEPs. If this approach is
taken it would be expected that a higher volume hydrograph would be used for larger flood
events or for reconciliation points lower in the catchment. Where this is not the case this
should be justified in the modelling report.

Scaling tributary inflows: Different scaling factors may be required for different
reconciliation points and for different storm durations. Typically the same scaling factor is
applied to all tributary inflows for the same model run for simplicity. In most cases scaling
factors will reduce the flood peak from the design flow estimates on the tributaries, as
downstream of a confluence the flood peak is likely to result from a combination of more
frequent events on the tributaries. Although in some instances design flows may decrease
downstream due to attenuation, and the AEP of a particular observed flood event may
increase downstream, it is not possible to require a lower probability flood event to occur on
a tributary in order to generate a higher probability flood event downstream.

Adjusting the phasing of tributary inflows: The start time of tributary inflows may be
adjusted to make them more or less coincident. This would typically be used where a
combination of gauged data and rainfall-runoff boundaries have been used for model
inflows.

It may be necessary to run multiple model scenarios for the same AEP event in order to
match the design flow at different reconciliation points. The confidence in hydrological flow
estimates and the ability of the hydraulic model to reproduce any attenuation in the design
flows should be considered in determining to what extent the flow in the hydraulic model
should be constrained to match the design flow. The approach used for reconciliation
should be described and justified in the modelling report.

6.3.6 Climate Change

A discussion of climate change allowances for fluvial modelling is given in Chapter 10.

6.4 Pluvial
Key points for Contractors

• If necessary model several storm durations to identify the worst case.

• Describe and justify the method used to determine infiltration and drainage losses.
• Undertake sensitivity analysis if the approach adopted is not demonstrably
conservative.
• The modelling report should include sufficient details of the analysis to enable an
experienced hydrologist to reproduce the rainfall estimates.

Boundary conditions for pluvial models need to consider how much rainfall falls on the
surface and how much of this rainfall is lost due to infiltration into the ground or is carried
away by the surface water drainage network.

6.4.1 Rainfall Models

It is suggested that representative rainfall applied over a model domain is constructed based
on the FEH 2013 Depth-Duration-Frequency model (section 5.8.2). FEH 2013 should
generally be used over FEH 1999 due to the improvements in data and techniques between

Flood Modelling Guidance for Responsible Authorities v1.1 54

FEH 2013 and FEH 1999. As in many cases depths in FEH 2013 are higher than in FEH
2013 than in FEH 1999, use of FEH 1999 could lead to an underestimation in rainfall depths.
The DDF model gives total rainfall depth for a given probability of occurrence and flood
storm duration. The storm duration which causes most flooding for a given probability of
occurrence will depend on the catchment being modelled, and it may be necessary to model
several storm durations to identify the worst case; SEPA’s pluvial hazard maps used 1 hour
and 3 hour duration storms.

Standard rainfall profile shapes are given in the FEH Point to Note:
and used to generate rainfall profiles from the DDF
model. The FEH provides two standard profiles; winter A summer storm profile
and summer (Figure 6-2), which do not vary with presents a shorter duration but
duration or location. The summer profile has a more higher intensity storm and is
pronounced peak, representative of the convective generally recommended for
storms more common in summer, and is generally application to urban
recommended for application to urban catchments catchments.
where a shorter period of high intensity rainfall is
generally more critical. The summer storm profile was used in SEPA’s pluvial hazard maps.
The choice of rainfall profile should be justified in the modelling report.

Figure 6-2: Standard FEH Hyetograph Profiles, based on Faulkner, (1999).

6.4.2 Losses
Two types of losses need to be
considered in surface water modelling; Point to Note:
infiltration into the ground and loss of
water into the urban drainage system. SEPA’s pluvial hazard maps use a
conservative approach in its assumptions.
Infiltration rates should vary between
urban and rural areas to account for the Runoff rates:
effect of extensive impermeable surfaces • 100% of the urban area is
in built-up regions. The FEH handbook impermeable
advocates the use of 70% runoff for • Urban runoff rate of 70% and rural of
impervious areas (Institute of Hydrology, 55%
1999). Within an urban area there will be
a component of the catchment which acts Drainage loss allowance:
as natural and has lower run off rate. The • National models: 12mm/h
percentage of the urban area which is • Regional models: average 20% AEP
impermeable varies according to the
(1 in 5) rainfall loss

Flood Modelling Guidance for Responsible Authorities v1.1 55

physical characteristics of the area and the scale and accuracy of the mapping used to
define the urban extent. The FEH handbook assumes that 61.5% of the urban area is
impermeable based on the digital land cover maps used to define urban extents in FEH
(URBEXT). Impermeable areas may be defined at higher resolution using aerial
photography or from large scale mapping. SEPA’s pluvial hazard maps use a conservative
approach, assuming that 100% of the urban area is impermeable and then applying a flat
runoff rate of 70% in urban areas and 55% in rural areas 10. Other, more detailed,
approaches are possible, such as using the losses model from the ReFH rainfall runoff
method which allows the percentage runoff to vary with time, or by specifying different
infiltration rates for different surfaces. The method used to determine infiltration losses
should be described and justified in the modelling report and sensitivity analysis may be
advised if the approach adopted is not demonstrably conservative.

In urban catchments it may be appropriate to incorporate a realistic drainage value to

remove a proportion of the rainfall input. This is not required if the drainage network is
represented explicitly in the model (i.e. an integrated catchment study). Research
conducted by the Environment Agency during the creation of The Flood Map for Surface
Water advocates the application of drainage loss rates of 12 mm/hr. If there is sufficient
information on the drainage network it may also be possible to define a local loss value.

An alternative approach is to assume a given service level for the drainage system. SEPA’s
national pluvial mapping assumed a 12 mm/hr loss to the drainage system whilst the
regional pluvial mapping assumed an average 1 in 5 year rainfall event loss (Figure 6-3).
Where a level of service is assumed, the loss applied may be constant or may vary with
time, Figure 6-4.

The approach taken to determine losses to the drainage system should be described and
justified in the modelling report. In general no loss to the drainage system should be
considered for rural catchments.

6.4.3 Climate Change

A discussion of climate change allowances for pluvial modelling is given in Chapter 10.

10
Urban areas in SEPA’s regional pluvial modelling are defined using the Land Cover Map 2007 (LCM2007) and
urban areas in the national pluvial maps are defined using the Scottish Government Urban/Rural Classification

Flood Modelling Guidance for Responsible Authorities v1.1 56

Figure 6-3: Areas covered by SEPA's national and regional pluvial modelling.

Flood Modelling Guidance for Responsible Authorities v1.1 57

Figure 6-4: Example of Different Methods to Remove 5 year Drainage Allowance (JBA
Consulting, 2014).

6.5 Coastal

Key Points for Responsible Authorities

• Coastal flooding is due to a combination of astronomical tides, surge and waves.

Interactions between these elements should also be considered. The conceptual
model should identify which processes are important for coastal flooding at a
study location.
• Responsible Authority coastal flood studies should seek to adopt a more detailed
approach than that used to derive SEPA’s national hazard mapping. This may
involve the use of several models.

Coastal flooding is due to a combination of astronomical tides, surge and waves. The
combination of astronomical tide and surge is referred to as the still water level. Still water
levels and waves are often treated separately however, waves may increase still water
levels at the coast due to a process called wave setup. In turn still water levels influence
where waves break and hence the total amount of overtopping. The key processes leading
to coastal flooding for a particular study location should be identified through the conceptual
model of the study area; see section 3.3.

Section 5.8.5 describes available sea level data and section 5.8.6 describes available wave
data. Information on extreme still water levels is not generally available within lochs and
estuaries and wave data is usually only available offshore. This means that further
modelling is often required to bring boundaries inland and, in complicated areas, generating
boundary conditions for coastal inundation models may require:

• A hydrodynamic model to look at how surge and tide change as they move up sea
lochs and estuaries.
• A wave transformation model which looks at how waves change as they move inland.
• A wave overtopping model to look at the rate of water overtopping flood defences.

Flood Modelling Guidance for Responsible Authorities v1.1 58

Inland flooding due to still water levels may be modelled by applying a level boundary to an
inland flood model or else the hydrodynamic model used to model surge and tide may be
extended in land. Inland flooding due to wave overtopping is usually modelled by applying a
flow boundary to an inland flood model. The different factors which need to be considered in
developing boundary conditions for coastal flood models are given in Figure 6-5.

Figure 6-5: Factors to be considered in developing coastal boundary conditions.

The SEPA Coastal Hazard Mapping Study mapped flooding due to extreme still water levels
only for the entire of the Scottish coast. This strategic level study was based on level
projection of these water levels only and did not consider the duration for which the levels
were high. It is expected that any coastal flood study undertaken by a Responsible Authority
will be more detailed; probably with time-varying still water boundaries and wave boundary
conditions. More information on coastal modelling techniques and their applicability to
Scotland is given in (Stokes, Masselink, & Conley, 2016).

6.5.1 Still Water Boundary Conditions

Key Points for Contractors

• For the open coast use boundary conditions from the Environment Agency Project
Coastal Flood Boundary Conditions (CFB) for UK mainland and islands (McMillan,
et al., 2011).
• For sea lochs and estuaries review the method used to extend to the CFB dataset
inland for the SEPA Coastal Hazard Mapping Project (Royal Haskoning DHV and
JBA, 2013) to determine if it is appropriate for a more detailed study.
• The base year for the study levels should be documented in the report.

Development of still water boundary conditions needs to consider:

• Extreme water levels, surge and tide shapes on the open coast;
• The change in extreme water levels & surge and tide shapes within estuaries and
sea lochs;

Flood Modelling Guidance for Responsible Authorities v1.1 59

 • The effect of wind and wave set up on still water levels.

The conceptual model of the study area should identify which of these factors are important,
and the modelling report should contain a discussion of the reasoning behind including or
excluding these factors from the study.

For locations on the open coast boundary conditions should be taken from the Environment
Agency Project Coastal Flood Boundary Conditions (CFB) for UK mainland and islands
(McMillan, et al., 2011). This gives extreme still water levels for 16 annual exceedance
probabilities at 2 km spacing around the open coast of Scotland, England and Wales and
provides guidance on developing standard storm tide curves to be used with extreme sea
levels at each location. These storm tide curves may be applied directly as a level boundary
for inland flood models for locations on the open coast if wind and wave set up are not
considered important. The report should state the CFB point used, which surge shape was
used to generate the boundary and how the base astronomical tide was derived.

The CFB dataset extends into the outer parts of some estuaries and sea lochs but not the
inner parts of estuaries because local bathymetric effects can significantly affect tide levels
within estuaries. The SEPA Coastal Hazard Mapping Project (Royal Haskoning DHV and
JBA, 2013) extended the analysis from the CFB study to sea lochs and estuaries within
Scotland using a combination of observed data not included in the CFB study, local
modelling studies and relationships between open coast and estuarine locations from
similarly shaped and aligned estuaries. For local flood studies in areas covered by the
SEPA Coastal Hazard Mapping Project a review should be undertaken of the method used
to extend the CFB dataset to determine both if the level of confidence in the levels is
appropriate for a more detailed study and if storm tide curves can be derived. This review
should be documented in the modelling report. The location of points taken directly from the
CFB study and the SEPA study are shown in Figure 6-6.

If the SEPA Coastal Hazard Mapping Project is not suitable for providing boundaries a
hydrodynamic model of the estuary or loch may be required. This will usually be 2D, but in
some instances a 1D model may be sufficient for narrow estuaries and it may also be
possible to include inland flooding within the same model. Detailed bathymetric data is
required for constructing a coastal hydrodynamic model and possible sources of data for this
are described in Section 5.5.6. Where possible the model should be calibrated against
observed tide levels or admiralty tide tables and this is discussed in Section 8.2.2. The
relevant CFB storm tide curve should be used as a boundary for the hydrodynamic model.
Construction and calibration of any coastal hydrodynamic model used to generate boundary
conditions should be documented in the modelling report.

The CFB boundaries and the SEPA Coastal Hazard Mapping levels include the effect of
storm surge but do not take into account wave setup which may increase sea levels on a
downwind coast. In some locations wind set up is also not accounted for. Where this effect
has been identified as important in the conceptual model it may be necessary to carry out a
joint probability analysis for surge and wave and wind setup, and to add wave and/or wind
boundary conditions to any hydrodynamic model used to bring surge inland. In this case
additional modelling may also be required for the open coast. Any analysis used to
determine the effect of wind and wave setup should be documented in the modelling report.

Sea levels in the CFB and SEPA Coastal Hazard Mapping study are referenced to 2008.
The base year for the study and the adjustment made for climate change since 2008 should
be documented in the report.

Flood Modelling Guidance for Responsible Authorities v1.1 60

Figure 6-6: Location of CFB points.

Flood Modelling Guidance for Responsible Authorities v1.1 61

6.5.2 Waves
There are two different types of waves:
• Wind waves (also referred to as sea waves). These and are generated by the local
wind, have a shorter period and are often irregular.
• Swell waves. These waves are generated by more distant weather systems, have a
longer wave period and are more regular.

The sea state is the combination of wind waves and swell. Both components need to be
considered in wave overtopping studies..

A description of available wave data is given in section 5.8.6. There are few long time series
observations available so, in most cases, the best available data are hindcasts from wave
forecast models. Forecasts points from these models are some distance offshore, and need
to be transformed inshore, usually through use of a numerical wave transformation model.
Extreme still water levels and extreme wave conditions may occur independently however,
the worst case situation for flooding is likely to be when large waves occur at high tide or
during a high tide and surge, so a joint probability analysis of extreme waves and
extreme still water levels is required. Wave conditions inshore cannot be used as a direct
input into a hydraulic model, and run up and overtopping models are needed to determine
the rate of flow over the defences, which can then be used as an input to an overtopping
model. Developing wave boundary conditions for input into a flood inundation model
therefore requires:
• Offshore design wave conditions
• Joint probability analysis of still water levels and extreme wave conditions
• Wave transformation modelling
• Overtopping Modelling

Case Study – Eyemouth Flood Study

Wave overtopping studies can be complex. The Scottish Borders Council Eyemouth
Wave Overtopping and Flood Study (Royal HaskoningDHV, 2013) used three models to
generate inflow hydrographs for a flood inundation model from the offshore results from
the Met Office wave model.

Waves were transformed inshore using a regional Firth of Forth wave model, and then
further inshore using a local Eyemouth wave model, both of which were developed for
SEPA’s Firths of Forth and Tay flood warning scheme. The results from the Eyemouth
model were used as boundary conditions for an overtopping model of the Eyemouth sea
wall.

6.5.2.1 Offshore design wave conditions

Offshore design wave conditions are derived from an extreme value analysis of a wave
model hindcast (Section 5.8.6). This is carried out for the range of direction sectors (typically
in 30º increments) which may be incident on the coast. Results for all wave direction sectors
analysed should be tabulated in the report, and sufficient detail of the analysis should be
given to allow the results to be reproduced.

Typically the Met Office wave model hindcasts are used for the UK. As with all models, wave
model hindcasts are not perfect and may systematically over or under estimate the extreme
wave conditions of interest for flood studies. Where possible the performance of any model

Flood Modelling Guidance for Responsible Authorities v1.1 62

used for boundary conditions should be assessed against observed data. In all cases a
literature review should be carried out to ensure that known limitations in the model are
understood, and these should be documented in the modelling report.

6.5.2.2 Joint probability of extreme still water levels and extreme wave conditions
Extreme still water levels are discussed in section 6.5.1. A joint probability analysis of
extreme still water level and wave height should be undertaken. Different approaches for
joint probability analysis are given in DEFRA guidance FD2308 (Hawkes & Svensonn, 2005)
(Hawkes, 2005), including a desk study approach which is simple to apply and will be
sufficient in many cases. However since publication of the DEFRA guidance new
techniques have been developed which may be more appropriate for complex areas or
where full uncertainty analysis is required (Gouldby, Méndez, Guanche, Rueda, & Mínguez,
2014).

Results from the extreme value analysis and joint probability analysis should be tabulated in
the report for the different wave direction sectors considered. Again sufficient details of the
analysis should be provided to allow the results to be reproduced. To obtain the worst case
flood extents it will generally be necessary to consider a number of different wave and SWL
scenarios, each with the same joint probability of occurrence.

6.5.2.3 Wave transformation modelling

Forecast wave model points are usually some distance off shore, while flood modelling
requires wave conditions at the coast. Wave conditions from the wave forecast points must
be therefore transformed inshore in order to provide wave conditions at the coast and this is
carried out using a numerical wave transformation models.

Wave and water level boundary conditions used for the wave transformation modelling come
from the joint probability analysis discussed in Section 6.5.2.2. Wind boundary conditions
may also be required and are usually developed using a simple regression analysis between
wind speed and significant wave height from the wave model hindcast.

Development of a wave transformation model requires good quality bathymetric survey and
shore survey (see section 5.5.6). For complex regions multiple wave transformation models
may be required, with a coarse resolution regional model covering a larger area used to
provide boundary conditions for higher resolution nested models of inshore areas, Figure
6-7. In this case joint studies may be cost effective. SEPA has several wave models
developed for flood warning schemes which may be made available for Responsible
Authority studies.

Where possible the wave transformation model should be calibrated using observed data,
and the set up and calibration of any wave transformation model used should be described
in the modelling report.

Flood Modelling Guidance for Responsible Authorities v1.1 63

Figure 6-7: Wave transformation model domain for the Firth of Forth developed for SEPA’s
flood forecasting scheme. To generate boundary conditions at Eyemouth 2 nested model
domains are used. Figure from Royal HaskoningDHV (2013)

6.5.2.4 Wave Overtopping

Wave overtopping models are used to calculate an overtopping rate from:

• Inshore wave and still water level conditions, and

• Detailed information on the flood defence or beach profile. Typically the crest
height, toe level and profile or type of the defence is required.

Several different overtopping models are available (e.g. Pullen, et al., (2007), Hedges &
Reis, (1998)) and the most suitable model may change according to the defence and beach
type. (Stokes, Masselink, & Conley, 2016) provide a review of different approaches which
may be relevant to different types of coastline around Scotland. The modelling report should
justify the overtopping model used with reference to the available literature.

6.5.3 Climate Change

A discussion of climate change allowances for coastal modelling is given in Chapter 10.

6.6 Joint Probability

6.6.1 Extreme Sea Level and Fluvial Flows

Joint probability analysis to investigate the potential combined effect of extreme sea levels
and high fluvial flows may be required for river reaches which are tidally influenced. The
worst case scenario would involve the concurrence of high tide, a surge and high fluvial
flows. Joint probability analysis typically includes running a model with different combinations
of downstream tidal boundary and fluvial inflow boundary conditions. Consideration should

Flood Modelling Guidance for Responsible Authorities v1.1 64

be given to the backwater influence of the tidal boundary on upstream water levels as this
can have an effect further upstream than the tidal limit. Joint probability analysis should be
according to the methodology in the DEFRA/Environment Agency Project FD2308 Joint
probability: dependence mapping and best practice (Hawkes & Svensonn, 2005) (Hawkes,
2005)

6.7 Groundwater
Currently there are few confirmed instances of groundwater flooding in Scotland and a
recent scoping project suggests that it is not as widespread a problem in comparison to
other parts of the UK. However, groundwater flooding is possibly underrepresented in
Scotland because of the difficulty of differentiating it from other types of flooding.

If groundwater is perceived to be an issue it can be investigated through desk studies or on-

site ground investigations and groundwater level monitoring in conjunction with other
hydrological data.

6.8 Uncertainty
All flood frequency estimates are inherently uncertain due to the length of record compared
to typical design probabilities of interest, the range of different of analysis methods available
and the incomplete data coverage. This can be one of the largest sources of uncertainty in a
modelling study and understanding this can help in making decisions such as the level of
freeboard to apply to a defence based on modelled water levels.

Where the analysis method permits, error bounds should be given on estimates of design
flows or levels; and design flows or levels should not be reported with a higher level of
precision than is justified by the data or analysis. The impact of the effect of uncertainty in
flood frequency estimates should be addressed through sensitivity testing. This is discussed
in chapter 8.

Flood Modelling Guidance for Responsible Authorities v1.1 65

7 Model Schematisation

7.1 Introduction
This chapter provides guidance for Responsible Authorities wishing to critically review
models received from contractors. It describes:

• The options which can be considered when schematising different features within a
hydraulic model and how these may affect the results.
• Common errors and problems with hydraulic models.
• Expected good practice in model building.

7.2 1D fluvial Models

Key Points for Responsible Authorities

• 1D models cannot be used to give flow velocity on the floodplain

• The floodplain schematisation determines the out of bank flow routes possible in
the model.
• There are several common issues with 1d models which need to be checked for
during a model review.

Key Points for Contractors

• Comments should be added to the model giving the source of any data, the
reasons for any structure representation, and the location of structures and cross
sections.
• Sensitivity testing should be carried out for roughness, structure blockage, and
structure representation.
• The modelling report should explain the choice of flood plain representation, the
channel and floodplain roughness values

In a 1D model the channel is represented as a series of cross sections. A single flow and
level value is calculated for each of these sections and velocity is averaged over the depth
and cross section width. As there is only a single level at each cross section, all points
below the calculated water level are wet simultaneously, even if there is no connection
between one area of the cross section and another, Figure 7-2. This also means that 1D
models cannot be used directly to give flow velocity in the channel and on the floodplain, or
variations in level across a section for instance, due to superelevation at a bend. Where flow
is not predominately 1D, typically at structures, different equations have to be used to
calculate energy losses through the structure.

Key factors to consider in schematising a 1D model are the representation of the channel
between hydraulic structures, the representation of any hydraulic structures and the
floodplain representation.

Flood Modelling Guidance for Responsible Authorities v1.1 66

7.2.1 Channel representation

7.2.1.1 Cross Sections

Cross sections are required at all points of interest, upstream and downstream of structures,
where there are significant changes in channel shape or slope and at the model boundaries.
Elsewhere the model cross section spacing must be sufficiently close to capture variations in
the hydraulic properties; and rules of thumb for cross section spacing are given in Figure
7-1. Too sparse cross section spacing can result in numerical instability and unphysical
attenuation of the hydrograph. Where survey sections are too far apart most hydraulic
modelling software has an option to increase the spacing by interpolating between upstream
and downstream survey sections. This is acceptable provided that there are no major
changes in channel shape and slope between the survey sections.

For a cross section spacing ∆x, the following rules of thumb apply

1. ∆𝑥 ≈ 𝑘𝑘, where B is the top width of the channel, and k is a constant with a
recommended range from 10-20.
𝐷
2. ∆𝑥 < 0.2 where D is the bank full depth and s is the slope.
𝑠
𝑐𝑐
3. ∆𝑥 < where c is the speed of the flood wave, T is the period of the flood
𝑁𝑔𝑔
wave and Ngp is a constant between 30-50.

Figure 7-1: Rules of thumb for cross section spacing in hydraulic models (from Castellarin, Di
Baldassarre, Bates, & Brath, 2009).

If cross sections are too close together, this can also cause problems with numerical
stability and lead to long model run times as the model has to be run with a small time step.
This may mean that a modeller has to exclude some surveyed sections if the survey is very
closely spaced.

Braided channels or split flow paths can be modelled either as a single set of cross sections
covering both flow paths, as in Figure 7-2, or separate cross sections can be used for each
branch of the channel with some representation of the flow pathway between the cross
sections. Where a single section is used the 1D methodology means the calculated water
level in both channels is the same and the flow split between the channels is not calculated.
The approach taken will depend on the level of detail required and the local hydraulics. For
strategic or catchment level models use of a single section may be appropriate especially if
there is significant flow between the channels during all flood events of interest. For local or
design models, or where there are receptors between the flow paths, use of separate
sections may be more appropriate.

Comments should be added to the model giving the source of the cross section data
(including surveyed date) and any modifications made during the model build. This can
prevent or help identify potential problems if the model is reused in future, for instance, steps
in the bed due to a different survey datum or loss of channel capacity due to siltation
between surveys. Where it is not clear from looking at the model if a section is from survey
or has been interpolated, this should also be recorded. In addition, useful information such
as nearby street names or location identifiers should be included as comments where
available.

For new models in software which requires cross section names to be entered (e.g.
FloodModeller), names should be logical and based on the chainage from a downstream
confluence with a larger watercourse or the tidal limit. This can help in identifying errors with

Flood Modelling Guidance for Responsible Authorities v1.1 67

cross section locations, and adding any additional survey to a model at a later date. This is
described further in Appendix F.

22
21
20
Elevation (m AD)

19
18
17
16
15
14
13
12
11

0 10 20 30 40 50 60 70 80

x (m)

Figure 7-2: Example of split channel modelled as single section.

7.2.1.2 Roughness
Hydrodynamic models include estimates of the surface roughness for the channel and the
floodplain areas (right and left 11) at every cross section in the model (1D model) and for
every cell grid in the model (2D). The most common representation of roughness is in the
form of Manning’s coefficient (n). Factors that affect roughness include: the nature of the
channel bed material, channel bed forms, channel structure, any obstructions (e.g. debris)
and the time of the year (i.e. vegetation cover). For free open flowing rivers roughness
decreases with increased stage and flow but if the banks of the river are rougher than the
channel bottom then the composite ‘n’ value will increase with increased stage.

Estimation of roughness is generally subjective based on modeller expertise, site visits or

photographs and look up tables for example Chow (1959) or the Conveyance Estimation
System (CES) roughness advisor (Fisher & Dawson, 2003). The subjective estimation of
roughness means that it is one of the most uncertain variables in a hydraulic model.

• Roughness values should be reasonable and defensible and able to withstand

independent review.
• Sensitivity testing to roughness should be carried out as in section 8.4.
• The modelling report should document the values used for roughness and the reason
for selection.

7.2.1.3 Hydraulic Structures

Around structures where flow is 2D or 3D, 1D models have to use different sets of equations
or parameterisations to represent the flow. Depending on the software used and the
structure in question these parameterisations may be purely empirical, based on laboratory
and field tests, or they may be based on physical equations. Most 1D modelling software
has built-in representations for a range of structure types with options to switch between

11
Right and left are defined as viewed downstream

Flood Modelling Guidance for Responsible Authorities v1.1 68

different equivalent representations of some structures. The precise details of how a
structure is represented in a hydraulic model will depend on the software and on the
structure; however, there are some general factors which have to be addressed.

Structure coefficients or parameter values In almost all cases representing a structure in

a 1D model involves estimation of parameter values by the modeller. This estimation is
usually based on user experience, site visits, photographs and survey drawings as it is
unusual for detailed measurements to be available at a structure at during flood flows. This
introduces uncertainty into the model and sensitivity testing should be carried out for
structure coefficients especially where there is high uncertainty in the value to be used or the
output of the modelling is expected to be sensitive to the parameter value.

Inclusion of all flow paths There may be multiple flow paths around a structure, particularly
during flood events. For instance there may be flow over a bridge deck or out of bank
around the side of a weir. All relevant flow paths should be represented. Where a particular
flow path has been excluded from the representation this should also be commented upon
(e.g. if the flow pathway over the bridge deck is not included as it is above the level of the
largest event modelled).

Level of detail The level of detail will determine which hydraulic structures are included in
the modelling. For a strategic or catchment level model it may be appropriate to omit some
structures such as small footbridges, however these may need to be included for a detailed
or design model. Further guidance on this is given in Table 5-2.

Angle to flow Where structures are skewed across the channel this should be accounted for
in the structure representation. If the modelling software cannot effectively account for the
decrease in effective length due to drowning then different models may be required for high
and low flows to ensure that both the drowned and the undrowned states are appropriately
captured.

Blockages Where structures have been identified as being at risk of blocking the effect of
blockages should be investigated through sensitivity testing; further details on this are given
in section 8.4.

Comments should be added to the model giving the location of the structure to aid
identification, the reasons for the structure representation used if several options are
available, the source of the structure data and how any parameter values were chosen or
calculated.

7.2.1.3.1 Bridges
Flow in the vicinity of bridges may be a combination of free surface flow where flow is
below the bridge deck, pressurised or surcharged flow where the flow is in contact with
the deck and weir flow over the bridge deck. The bridge schematisation should be sufficient
to represent all modes of flow which occur in reality. This may involve the use of multiple
model units to represent the bridge (e.g. a bridge unit and a weir or spill unit to represent
flow over the bridge). Most forms of hydraulic modelling software have several available
bridge representations for each mode of flow – a review of representations available in
different software packages is given in Samuels (2004). Typically there is no suitable
calibration data available and the choice of representation used is largely based on user
experience. If necessary, sensitivity tests to different bridge representations can be carried
out.

Flood Modelling Guidance for Responsible Authorities v1.1 69

7.2.1.3.2 Culverts
The hydraulic analysis of flow in culverts is complicated and reference should be made to
industry guidance e.g. CIRIA’s Culvert design guide (Balkham, Fosbeary, Kitchen, &
Rickard, 2010). All culvert models should contain representation of the inlet losses, outlet
losses and losses due to friction along the culvert barrel; where a culvert changes shape,
bends or is obstructed due to service crossings additional losses should be included to
represent these. Losses due to trash screens, where present, should be included in the
representation of the culvert inlet.

For detailed or local models it may be necessary to include representations of manholes and
surface water sewer connections although in most cases these can be omitted from
catchment level models.

Culverts can be particularly prone to blockage and for detailed or local studies sensitivity
testing to sedimentation of the culvert barrel and due to inlet blockage should be carried out
as set out in the Culvert Design Guide (Balkham, et al., 2010).

Short culverts, where the effect of friction along the length of the culvert can be neglected,
can also be modelled as an orifice. This can increase model stability in some cases.

7.2.1.3.3 Weirs and Gates

For structures involving several gates and openings such as mills it may be appropriate for
strategic and catchment level modelling to only model the main flow path through the
structure, rather than including each sluice individually. Where gates are operated during
flood events this can be included in a hydraulic model through use of control rules.

7.2.2 Floodplain Representation

There are three primary methods for representing out of bank flow in 1D models: extended
channel sections, storage areas and parallel channels Table 7-1. The most appropriate
method depends on the floodplain geometry and several methods may be combined to
represent the floodplain in a single 1D model. The floodplain schematisation determines the
out-of-bank flow routes possible in the model so care should be taken to ensure that the
schematisation is appropriate. The choice of floodplain representation should be covered in
the modelling report.

Where the cross section survey does not cover the full width of the floodplain the floodplain
representation is typically based on a DTM (section 5.5.5). Difficulties may arise in all
methods if there is a discontinuity between the cross section survey and the DTM. Where
this occurs it is unlikely that it can be resolved without collection of further out of bank
survey.

Flood Modelling Guidance for Responsible Authorities v1.1 70

 Description Comments
Not suitable for embanked watercourses; as
there is only one value for stage across a cross
section extended sections allow the area
behind the embankment to flood before the
Extended Sections

embankment is overtopped.

Cross sections are extended Cross sections must be perpendicular to the

across the full width of the flow direction and for meandering rivers they
floodplain until they intersect with may be kinked, with different reach lengths
high ground (Figure 7-3). being set for the channel, left and right banks.

Cross sections should not cross each other.

Care must be taken to ensure that there are no

significant floodplain changes between
surveyed cross-sections.
Storage Areas

Depth area relationships are Most suitable for flat areas with negligible out
defined for areas of the of bank flow. Only one elevation value is
floodplain. Flow between these calculated for the storage area so, if the areas
areas and the channel is are large or steeply sloped, other areas of the
controlled by the weir equation storage area may become wet before the
(Figure 7-4). areas next to the channel.
A separate 1D channel is added There may be stability issues associated with
channels
Parallel

parallel to the main river. Flow the parallel channel drying out, particularly at
between the main channel and the start of the modelled event.
the parallel channel is controlled
by the weir equation.
Table 7-1: Methods for floodplain representation in 1D models.

98 0.06

96

94
Elevation (m AD)

Mannings n

92

90

88

86

84 0.04
200 300 400 500 600 700 800 900 1,000 1,100

x (m)

Figure 7-3: Example of extended section in Flood Modeller Pro.

Flood Modelling Guidance for Responsible Authorities v1.1 71

Figure 7-4: Example of storage area configuration in HEC-RAS from Brunner (2010).

7.2.3 Common Problems with 1D Models

Several common issues may arise with 1D models which should be checked at the moel
review stage.

Data entry errors 1D models can require significant amounts of manual data entry
particularly at structures as there is no single survey or input file format which can be used to
read structure data into all hydraulic modelling software 12. This is a potential source of
model error and a detailed check on structure data entry against the original survey data
should be carried out.

Inappropriate choice of structure coefficients Structure coefficients should be checked to

ensure that they are within normal ranges and are physically realistic based on the available
evidence.

Glass walling Where 1D models exceed the maximum level in a cross section the modelling
software will automatically add a vertical wall at the end of the section to allow water to rise
above the section level. This leads to increased water levels within the section as the water
cannot spread over the floodplain. Where this occurs the floodplain representation should
be extended, unless there is a reason for allowing the model to glass wall (e.g. to represent
a very high defence).

Embanked sections If 1D model cross sections contain embankments the area behind the
embankment will become wet in the model before the embankment is overtopped. This may
be acceptable in some circumstances, for instance if the embankments provide a very low
standard of protection or if there is a small area behind the embankment in a catchment or
strategic level model. In HECRAS, for example, levees can be used to prevent areas behind
the levee becoming wet before it is overtopped. Other software may require the cross
sections to be cut back to the top of the embankment and a different representation for the
area behind the embankment.

12
The EACSD file format aims to address this issue, but uptake amongst software providers has been low.

Flood Modelling Guidance for Responsible Authorities v1.1 72

Numerical stability 1D models can be prone to numerical instability. This often manifests
as unphysical spikes or oscillations in the water levels or flows. The model log file should be
examined for any reports of numerical instability, hydrographs should be examined for
unphysical features at each cross section and a long section should be animated to check
for unphysical oscillations.

7.3 2D models
Key Points for Responsible Authorities

• 2D models can give depth-averaged velocity on the floodplain but may not
represent channel flow well if the resolution is insufficient, particularly around
hydraulic structures.
• The model resolution affects the computational effort required to run a model.
Balance the need for increased detail due to higher resolution and understanding
uncertainty in the modelling through more sensitivity tests and scenarios.
• Future use of the model should be considered in choosing the software as this
affects whether the model resolution can vary through a model domain, whether it
can be linked with 1d model at a later date and how structures can be represented
within the model.
• There are several common issues with 2D models which need to be checked for
during a model review.

Key Points for Contractors

• Avoid representing buildings using mesh voids or full height buildings as these
representations cause problems when undertaking depth damage calculations
using the model results.
• The approach taken to represent any hydraulic structures or linear features in the
DTM should be described in the modelling report, together with the data used.
• The roughness values assigned to different land uses should be described in the
modelling report, together with the data used to determine the land use.

2D models calculate water level and depth-averaged velocity over a regular or an irregular
2D grid. Schematisation of 2D models is similar for coastal, fluvial and pluvial flooding, with
the main difference being in the type and location of boundary conditions applied.

Key factors to consider in schematising any 2D model are the grid type and resolution, the
representation of features within the DTM and the roughness. There are also several
considerations regarding schematisation which have to be made in choosing 2D modelling
software before deciding on the schematisation within a particular software package.

Grid Type – Models may have regular or irregular grids. Irregular grids allow resolutions to
vary across the model domain so that higher resolutions can be used only where required,
such as for urban areas, steep slopes and along channels, however regular grids may be
easier to set up.

Flood Modelling Guidance for Responsible Authorities v1.1 73

Shock capturing – Shock capturing models are better at capturing hydraulic jumps and
discontinuities in flow. This is important for dam break modelling and breach analysis.

Computational efficiency and software architecture – Different model codes use different
software architecture, which can affect model run times and the number of concurrent
simulations which can be run. Shorter model run times are beneficial if large numbers of
scenarios or multiple calibration runs are required.

Linkage with 1D models – If the model can be linked with a 1D model at a later date. This
is most likely to be a consideration for fluvial models.

Structure Representation – Different packages have different methods to represent

features smaller than the model grid such as kerb lines, fences or underpasses, or hydraulic
structures such as bridges.

DTM manipulation – The ease with which the DTM can be modified to produce a good
representation of the ground surface for modelling (e.g. to include buildings or walls can vary
significantly between software packages).

7.3.1 Model Resolution

The model resolution depends both on the scale and objectives of the modelling and the
physical characteristics of the study area.

A rule of thumb is that 3-4 grid cells are

required to resolve major flow paths. This Point to Note:
limits the minimum resolution which can be
used for modelling watercourses in 2D. For The available DTM affects the model
instance, modelling a 5 m wide channel in 2D resolution as it cannot usefully be
would require a minimum 2D grid resolution of increased beyond the DTM resolution.
around 1.5 m whereas, for a 30 m wide river, a
10 m resolution may be adequate. A rule of thumb is that 3-4 grid cells are
required to resolve major flow paths.
For modelling the floodplain of large rivers in
The computational effort required to run
rural areas the resolution does not have to be
a model is largely dependent on the
high enough to resolve features such as
drainage ditches, as these tend to have more resolution
influence during low flows than flood flows.
Embankments, where present, do usually affect
flood flows however; these can be incorporated through forcing elevations in the DTM or
incorporation of 1D structures (see section 7.3.2.2). It does not necessarily require a higher
resolution. SEPA’s national hazard maps use a resolution of between 5 m and 20 m in
rural areas. In urban areas this resolution is sufficient to resolve flow paths along roads;
however, a higher resolution may be required to model flow paths between buildings.

Adequate representation of detailed urban flow pathways (e.g. flows between buildings or
even obstructions to flow due to kerb heights) requires a finer spatial resolution. SEPA’s
fluvial hazard maps use a resolution of 5 m in urban areas and SEPA’s pluvial maps use a
resolution of 2 m or 5 m in urban areas.

The computational effort required to run a model is largely dependent on the resolution and
halving the grid cell size typically results in an increase of run times by a factor of 8. If
computing resources are limited it may be necessary to balance the desire for finer grid
resolution with the need to run multiple scenarios or sensitivity tests.

Flood Modelling Guidance for Responsible Authorities v1.1 74

Where higher resolution is only required in part of a model domain, for instance an urban
area within a larger catchment model, this can be achieved either through nested grids in
models with a regular grid or through decreasing the element size in models with an irregular
grid (examples of these are shown in Figure 7-5). Some nested grid approaches do not
conserve momentum at grid boundaries and this can lead to unphysical flow patterns around
the boundary. Where nested grids are used the boundaries between grids should be located
far enough away from the area of interest so that they do not affect the results; to ensure this
is managed appropriately, results near the boundary should be reviewed.

(a) (b)

Figure 7-5: Examples of locally increased resolution; (a) an irregular grid with higher
resolution along a watercourse and (b) an example of a regular nested grid, in this case a
higher resolution grid covers an urban area.

The available DTM also affects the model resolution, as the resolution cannot usefully be
increased beyond the DTM resolution. However, if other considerations identify that a
higher resolution model is required new data should be collected to increase the DTM
resolution.

7.3.2 DTM
The basis of the DTM should be stated in any modelling reports, as well as the date it was
collected. An assessment of accuracy should also be provided.

7.3.2.1 Building representation

There are several options for representing buildings within 2D models and a review is given
in Syme (2008). Common representations include:

Voids Buildings are left as void polygons in the mesh where no values are calculated. This
is a conservative approach as storage within the buildings is not accounted for but it may
overestimate the obstruction to flow caused by a building. As the model does not calculate
within buildings, the depth outputs required for damage calculations are not available as a
model output. Instead, depths within buildings have to be interpolated from the water level
adjacent to the building and the threshold level which can be problematic if there are
differences in water level around the building. Due to the difficulties in interpolation to get
values at building centres, this method is not recommended where depth damage
calculations are required.

Full Height Buildings Buildings are represented as blocks at their actual height. This has
similar limitations to voids and the method is not recommended where depth damage
calculations are required.

High Roughness Buildings are represented by increasing the roughness over the building
footprint. This allows buildings to store water and gives depths within buildings. There is
limited guidance available on the appropriate roughness value to use however; values

Flood Modelling Guidance for Responsible Authorities v1.1 75

should be significantly higher than the surrounding roughness class. This approach was
used for SEPA’s national fluvial hazard maps.

Stubby Buildings Buildings are represented by raising the DTM within the building footprint
to the threshold level. Typically thresholds are assumed to be 0.3 m above the bare earth
DTM. The roughness may also be increased over the building footprint. The stubby
buildings approach without an increase in roughness was used for SEPA’s regional
pluvial hazard maps.

Porous Walls Buildings are represented as partially porous walls, with the porosity and
height of the walls specified by the modeller.

In most cases OS Mastermap data is used to identify building footprints however, other
datasets such as detailed ground-based survey may also be used. The method used to
represent buildings in the DTM should be stated in the modelling report together with the
data used to identify building footprints.

Consideration should be given to future use of the model when choosing the building
representation as both the void and full height building methods do not allow depth-damage
calculations to be carried out. In particular depth-damage calculations are required to inform
Flood Risk Management Strategies so models using these building representations will not
be used to inform future Strategies.

7.3.2.2 Hydraulic Structures and Linear Features

Linear features such as raised flood defences or road and rail embankments or drainage
channels may or may not be picked up automatically in a DTM depending on both the model
and the input DTM resolution. To ensure that features are represented in the model at the
correct height most modelling software contains options either to set the height of grid cells
along a line or to add a 1D weir type structure into the model.

Bridges, culverts, underpasses and similar features are likely to appear as false blockages,
or complete obstructions to flow in the DTM. This can result in over estimation of flooding by
the model upstream of the structure but underestimation downstream as not enough flow is
passed forward. Some representation of these structures in the DTM is required. Depending
on the software, different options are available; a review of structure representation in
TUFLOW is given in Syme (2001).

Lowering DTM through the structure This allows flow to pass but does not represent a
constriction to flow due to the soffit of the structure. Depending on the scale of the structure
and grid resolution some changes in velocity due to constriction/expansion at the entrance
and exit may be captured, but constriction due to the soffit is not included. This method can
be implemented quickly with limited structure information and is appropriate for strategic
scale modelling.

Use of a 1D structure Most 2D modelling packages contain the option to include 1D

representations of structures. This has advantages as most 1D software has in-built
representations available for a wide range of structure types. Constriction due to the soffit
and constriction/expansion losses can be accounted for through the choice of loss
coefficients as in a 1D model although they may need to be reduced if some of the changes
in the flow patterns are captured in the 2D model. Due to the 1D-2D link momentum is not
conserved through the structure and this reduces the accuracy with which flow patterns
around the structure are modelled.

Flood Modelling Guidance for Responsible Authorities v1.1 76

Use of 2D structures Some 2D packages have the ability to model structures in 2D by
partially constricting flow along cell sides and at the soffit. Some changes in velocity due to
constriction/expansion at the entrance and exit are by captured by the change in 2D flow
paths but there is usually also an option to add a “form loss” which accounts for changes
which are not modelled either due to the 2D grid resolution or because they involve 3d flow.
There is limited guidance available on determining loss coefficients in 2D models and a
check against other methods is required.

It may also be necessary to modify the DTM to ensure there are flow paths between
buildings or other closely spaced structures and to ensure that forested or heavily vegetated
areas have not been picked up as obstructions in the DTM.

The approach taken to represent any hydraulic structures or linear features in the DTM
should be described in the modelling report together with the data used.

7.3.2.3 Roughness
In 2D models roughness can vary across the model grid. Typically the model domain is split
into different land use classes using information from land cover layers such as OS
Mastermap and a roughness value assigned to each class. The roughness for some land
use types may depend on resolution for example; if features such as buildings and walls are
not resolved in the DTM a higher roughness may be set for general urban areas to account
for this. The roughness values assigned to different land uses should be described in the
modelling report together with the data used to determine the land use.

7.3.3 Numerical precision

Rounding errors can lead to mass balance problems if the numerical precision is not
sufficient. This is most likely to be an issue for pluvial modelling where very small depths of
water are added to each model grid cell.

7.3.4 Common Problems with 2D Models

Several common issues may arise with 2D models, which should be checked at review
stage.

False Blockages Blockages in the DTM which cause an obstruction to flow where in reality
there is a flow path.

Leaking Embankments Embankments or walls which are not picked up properly in the
DTM and allow flow through where, in reality, there is no flow path.

Glass walling Where 2D flows reach the edge of a model domain and no boundary
condition has been defined to allow the water to flow out, the model can insert a vertical wall.
This leads to ponding and increased water levels at the edge of the model domain.

Numerical stability 2D models can be prone to numerical instability, particularly if the

timestep is too long or the domain is very steep. This can show up as oscillations in velocity
and water level. The model log file should be examined for any reports or numerical
instability, and the results animated and examined for nonphysical flow patterns.

Mass Balance Errors The model may gain or lose mass, particularly if there is frequent
wetting and drying, or if the inflow at each grid cell is small. The mass balance files should
be checked to ensure the mass balance is within acceptable limits, usually within ±1%
cumulative error.

Flood Modelling Guidance for Responsible Authorities v1.1 77

7.4 1D-2D models
Key Points for Responsible Authorities

• 1D-2D models can give a depth averaged velocity on the flood plain and are able
to represent channel flow well, but they can take significant time and resources to
set up.
• There are several common issues with 1D-2D models which need to be checked
for during a model review. In particular models can be prone to mass balance
errors and models should not be accepted without evidence of mass balance
checks.

Key Points for Contractors

• The mass balance should be reported in the modelling report.

Coupled 1D-2D models use a 1D model component to represent river channels and/or the
surface water drainage network and a 2D model component to represent the floodplain.
Flow is dynamically passed between the 1D and the 2D components. This approach is used
where there are complex floodplain flow paths which cannot be represented in 1D and where
increasing the grid resolution to resolve the channel in a purely 2D model would be
impractical or where there are hydraulic structures which can be best represented in a 1D
model. Coupled 1D-2D models are most commonly used for fluvial flood modelling or for
detailed surface water drainage network models; however they may also be used within
estuaries. Despite the advantages of 1D-2D models of representing both the channel and
floodplain, they are not recommended if a purely 1D or 2D model will deliver the
desired objective as they are complex to set up and prone to numerical instability which
can take significant time and resources to solve.

The guidance for 1D models in section 7.2 and 2D models in section 7.3 applies to the 1D
and 2D components of coupled models. The other considerations are the features
represented in the 1D and 2D components of the model and the type of link.

It is also possible to have a 1-way link using the output from a 1D model as the inflow into a
2D model. This approach is only suitable where flow is away from the system represented in
1D and is not as prone to numerical instability.

7.4.1 Common Problems with 1D- 2D Models

In addition to the problems which can occur with separate 1D and 2D models there are
potential issues which may arise with coupled 1D-2D models which should be checked at
review stage.

Elevation of links If the wrong link elevation is set for lateral links between 1D and 2D
models this may lead to flow across the link before the bank is overtopped.

Numerical stability can be a problem for 1D-2D models particularly if there is frequent
exchange of water across the link. This can show up as oscillations in velocity and water
level in both the 1D and 2D components. The model log file should be examined for any
reports or numerical instability and the results animated and examined for nonphysical flow
patterns.

Flood Modelling Guidance for Responsible Authorities v1.1 78

Mass Balance Errors can be a particular problem for 1D-2D models. The mass balance
files for the 1D component, 2D component and combined model should be checked to
ensure the mass balance is within acceptable limits, usually within 1% cumulative error. The
mass balance should be reported in the modelling report.

7.5 Boundary Types and Locations

All models only cover a particular area of interest. Boundary conditions are required to
define what happens at the boundary between this area and the area outside. Defining the
location and type of any boundary conditions is an important aspect of model
schematisation. The type of boundary condition depends on the source of flooding and the
dimension of the modelling.

7.5.1 Fluvial Models (1D and 2D)

An upstream boundary condition is required at the upstream end of any modelled river
reaches. These conditions are generally specified as a flow hydrograph for unsteady
models, or a constant flow for steady models.

A downstream boundary condition is required at the downstream end of any river reaches.
These conditions are generally specified as stage hydrograph, flow hydrograph, single
valued rating curve, normal depth or critical depth boundary. If the flow reaches the domain
boundary other than at the defined downstream boundary the model extent should be
examined.

Additional inflows can be added to models between upstream and downstream boundaries
to account for any increase in flow between the upstream and downstream boundary.
Lateral inflows trickle flow in gradually along a reach and are typically used to account for
the increase in catchment area along a reach while point inflows add flow at a specific point
and are typically used to represent tributary inflows.

Upstream and downstream boundary conditions should be located a sufficient distance from
the area of interest so that any errors in the boundary will not significantly affect predicted
water levels at the study area.

• A rule of thumb L=0.7D/S (where D= bank full depth and S= river slope) can be used
when considering the location of downstream boundary from the study site.
• If possible, the downstream boundary should be located where relationship between
level and flow is well defined e.g. weir
• If the downstream boundary is tidal; the downstream boundary should be located
where a tidal curve can be defined.

7.5.2 Pluvial Models (2D)

Typically a depth of rainfall is applied to every grid cell in the modelled area at each time
step. The edges of the model domain are usually set so they are sufficiently far away from
the area of interest so as not to have an impact on results.

7.5.3 Coastal Models (2D)

For coastal inundation models a level boundary is typically defined along the coast for still
water flooding. Flow due to wave overtopping is usually added as a flow hydrograph. The
inland boundary should be sufficiently far inland to be outwith the coastal flood extents.

7.6 Initial Conditions

Initial conditions describing the state of the system before the start of the flood event are
required for all types of model. The initial conditions consist of flow and level information at

Flood Modelling Guidance for Responsible Authorities v1.1 79

each point in the model (1D cross sections and storage areas and at each 2D grid cell);
although typically for 2D models the flow velocity is assumed to be zero at the start of the
simulation.

If the initial conditions are incorrect this may lead to instability at the start of the model run or
storage areas having the wrong volume of water in them at the start of a simulation. The
model initial conditions should be checked during review of the model. Particular care
should be taken in situations where there are significant amounts of storage or artificial
drainage.

Flood Modelling Guidance for Responsible Authorities v1.1 80

8 Calibration, Validation & Sensitivity Analysis

8.1 Introduction
Model calibration and validation is important for determining the degree of confidence which
can be placed on model results. Sensitivity analysis is important for understanding the level
of uncertainty in the modelling. These affect the practical use the model can be put to and
the confidence in decisions which are based on the modelling. This chapter describes:

• What calibration is and why it is important;

• Expected good practice in model calibration;
• What sensitivity analysis is and why it is important;
• Expected good practice in sensitivity analysis.

8.2 Calibration
Key Points for Responsible Authorities

• Specify the number of calibration events and target criteria for model calibration in
the SoR. A minimum of 3 calibration events and one validation event is
recommended.
• Additional sensitivity testing and uncertainty analysis should be carried out if
limited calibration data is available. In some cases installation of additional
gauging to enable a higher level calibration may be appropriate.
• Review model calibration reports critically to ensure common issues with model
calibration are avoided.

Key Points for Contractors

• The calibration process should be fully documented in the modelling report.

Changes to parameters and the rationale for revising must be clearly
documented.
• Models must not be forced to fit the data by varying parameters outside physical
ranges or in ways which are not supported by the available data.
• Consider the possibility of data errors and changes to the study area since the
calibration event.
• Carry out additional sensitivity testing if it possible that several parameter
combinations may give the same fit to the observed data.

Calibration is the process of adjusting model parameters, such as the surface roughness,
within physically defensible ranges until the resulting predictions give the best possible fit to
a selected observed event. A model is said to be validated if it is able to provide accurate
predictions against other observed events (i.e. non-calibration events) within acceptable
limits. Model calibration and validation provides an understanding of the appropriateness of
the model considering observed flow/stage data. The main objective of model calibration and
validation is to provide a demonstration of the quality of model predictions. If calibration is
not carried out, confidence in the model application will be significantly reduced.

Flood Modelling Guidance for Responsible Authorities v1.1 81

Model calibration appropriate to the level of study should be carried out where sufficient data
is available. The quantity, quality and type of observation has a primary influence on model
reliability and confidence in outputs and, as discussed in section 3.4, the data available for
calibration should be considered in setting the study extent. Where stage and flow calibration
data is not available a reality check on the predicted outlines and levels for an event should
be carried out using other historic flooding information such as photographs and anecdotal
descriptions of flooding. It is important that the reliability of the data is checked prior to use.

The level of calibration which would be expected for different levels of study is given in Table
8-1.

Level of Study Gauged Ungauged

Sensitivity testing. Check of design events against historic record at
Strategic
sufficient locations to understand limitations in the method.
Calibration against • Reality check against historic data.
Catchment gauged data • Sensitivity testing

• Reality check against historic data.

• Sensitivity testing.
Local
• New data collection.

Table 8-1: Suggested level of calibration for gauged and ungauged catchments for different
levels of study. Target calibration levels for gauged study areas discussed for different
sources of flooding in section 8.2.1, 8.2.2 and 8.2.3 .

The main information required is recorded flows and/or water levels and flood extent from
observed events; however a range of different types of data may be used in calibration and
reality checks as listed below;

• Observed extents from survey ;

• Aerial photography (particularly if taken at the flood peak of the event);
• Historic flood levels;
• Trash lines;
• Anecdotal reports of which properties/streets flooded, anecdotal reports of depths of
flooding experienced (this may be in the form of descriptive terminology e.g. “waist
deep” etc.);
• Photographic evidence for example of levels at structures during events;
• Information on structure operations and blockages during the event.

The number of calibration events should be specified in the Statement of Requirements

(SoR). It is recommended that a minimum of 3 calibration events and one validation event
are used for local and catchment scale studies. However, model confidence can still be
improved by calibration/validation for fewer events if sufficient calibration data is not
available for more events. After calibration, performance of the model (and adjusted
parameters) should be validated through simulation of at least one separate observed event.
Possible calibration events should be identified in the SoR, but it may be necessary to use
other events if there are issues with the data record or the identified events are not
considered representative (e.g. flooding mechanisms such as ice jams or culvert blockages
or multiple sources of flooding).

In some areas calibration data may only be available for a short period of time, with no
significant events. Calibration and validation of models under these circumstances can still

Flood Modelling Guidance for Responsible Authorities v1.1 82

be useful (e.g. to check the performance of a wave transformation model or flows predicted
by a rainfall runoff model). In some cases installation of additional gauging to enable this
level of calibration may be appropriate, and the costs of installation and maintenance should
be considered against the project aims and required accuracy.

Target calibration acceptance criteria should be defined in the SoR or early in the project.
Typical requirements are set out in the template SoR which can be made available on
request, appendix A.

The calibration process should be fully documented in a report and should include calibration
event dates and measurements and locations of historic floods. Changes to parameters and
the rationale for revising must be clearly documented.

8.2.1 Fluvial
River models should be calibrated for flow and levels at gauging stations. It is strongly
recommended that, where possible, the study extent covers at least one and preferably two
or more gauges to assist in calibration.

The hydraulic parameters which are usually varied during model calibration are the surface
roughness (e.g. Manning’s n) and structure coefficients. Model boundaries, including
parameters in hydrological models, may also be varied, for instance the parameters in the
rainfall run off model. Where possible a combined approach to hydraulic-hydrological model
calibration should be undertaken.

The calibration events should cover both in-bank and out-of-bank scenarios to ensure that
both the channel and the floodplain are modelled correctly. Although inclusion of larger
events is important, not all the events need not have caused extensive flooding as it is also
valid to show the model correctly predicts water not reaching particular locations. Utilising
recent events may minimise the impact of recent changes in hydraulic structures or
catchment characteristics.

Specific data requirements for each calibration event are:

• 15 minute flow and level time series for any gauges within the study reach, including
tributaries. Particular care should be taken in extrapolating rating curves.
• 15 minute rain gauge data for any gauges within or surrounding the catchment.
• Tide gauge data if the downstream boundary of the model is tidal.

If the ReFHv2 rainfall run off model is used, the following are also required:

• MORECS/MOSES evapotranspiration and soil moisture data for 2 years prior to the
event.
• Daily rainfall for any gauges within or surrounding the catchment.

Flood Modelling Guidance for Responsible Authorities v1.1 83

 Case Study – Hydraulic model calibration for Selkirk

Selkirk has a history of flooding, with notable recent events in 1977, 2003, 2004, 2005,
and 2009. Following these events Scottish Borders Council categorised Selkirk as a high
priority for investment in flood protection measures and between 2002 and 2011 a phased
programme of modelling and investigation was undertaken in order to inform the flood
protection measures.

In order to have confidence in the hydraulic model predictions of flood levels, the model
was calibrated using a combination of gauged data and historic flooding information.
Four events were identified where there was suitable recorded data, observed flood levels
and/or anecdotal evidence to be used as calibration events; October 1977, January 2005,
October 2005 and November 2009 (Halcrow, 2011).

Flow records from the 2 gauges within the study reach were used to derive model inflows,
and the hydraulic model was calibrated by adjusting individual model parameters in order
to obtain a good correlation between predicted and observed water levels. The final
phase of the study reported a reasonable fit to observed data for all four events, with
discrepancies of less than 0.2 m recorded for the majority of locations across all
calibration events. In instances where larger discrepancies were noted, further
investigation was undertaken in the form of sensitivity testing. This helped to refine the
conceptual model of the catchment by identifying factors which could increase flood levels
which had not previously been considered. Once such case was at the Riverside
footbridge in the November 2009 event, where sensitivity testing identified that a 40%
blockage of a bridge was required in order to reproduce observed water levels, and whilst
this was initially considered high, this was supported by a review of the observed data for
this event.

The calibration was revisited at different phases of the work following updates and
expansion of the model. This ensured that confidence in the model outputs at each stage
was understood, and provided confidence in the use of the model for the design of the
Selkirk Flood Protection Scheme and alleviation.

The calibration criteria should consider;

• peak water level;
• overall hydrograph shape; and
• timing of the peak level.

Target accuracy in the calibration should be set using tolerances of both peak water level
(e.g. +/- ± X mm) or less and in the timing of peak level (e.g. within X hours or less).

For catchment scale studies it is recommended that tolerances for peak water level at
measured locations are in the order of +/-300 mm or less and that, for local scale or
detailed studies, it is recommended that tolerances for peak water level are in the order of
+/-150 mm or less depending on the application. This can be considered to correspond to
medium and high confidence in the outputs respectively.

Target accuracy in the timing of the peak level will depend on the hydrograph duration; for
most purposes a target accuracy of 30 minutes would be appropriate however a larger
tolerance may be acceptable for catchments with a long time to peak and a 15 minute
tolerance may be required for very quickly responding catchments.

Flood Modelling Guidance for Responsible Authorities v1.1 84

Figure 9-1: Example of a calibration plot comparing modelled and observed flow

Calibration plots can be produced comparing modelled flows with those recorded at the
gauging station for an event. Tables can also be used to present a comparison of observed
and modelled peak flow, time to peak and stage at particular locations.

8.2.2 Coastal
Hydrodynamic models used to bring extreme still water estimates inland should be calibrated
for level at tide gauges. Where tide gauges are not available performance should be
checked against tide tables over a spring neap cycle, while installation of additional
temporary gauging may be considered. The hydraulic parameters which are usually varied
during model calibration are the bed roughness and eddy viscosity. Changes may also
be made to model boundaries and bathymetry.

Wave transformation models should be calibrated for significant wave height at wave buoys.
Where these are not available, installation of temporary gauging should be considered.
Calibration should be over a sufficient period to cover all wave direction sectors. The
hydraulic parameter usually varied during model calibration is the bed roughness, although
boundary conditions may also be adjusted.

It is unlikely that sufficient data will be available for calibration of wave overtopping
and coastal inundation models.

It is recommended that tolerances for coastal hydrodynamic models are that:

• Levels are within ±0.1 m at the mouth of firths, estuaries and sea lochs and ±0.3 m
at the head.
• Directions are within ±22.5 degrees
• Timings of high water are within ±15 minutes

The Environment Agency document Best Practice in Coastal Flood Forecasting (HR
Wallingford, 2004) classifies wave and surge models as having high, medium and low
confidence if predictions of height are within about ±20%, ±30% and ±40% respectively. For

Flood Modelling Guidance for Responsible Authorities v1.1 85

shoreline models of overtopping rate and probability of breaching, the same document
classifies models as having high medium and low confidence if most predictions are
expected to be within factors of about 5, 15 and 50, respectively, due to the much lower
expectation of accuracy in these models. This corresponds to expected accuracy for the
area flooded within about ±30%, ±45% and ±70%, respectively if overtopping or breaching
has occurred.

8.2.3 Pluvial
A detailed pluvial model with explicit representation of the surface water drainage network
should be calibrated according to the UDG guidance (WaPUG, 2009), which is used in the
collaborative Scottish Water and Local Authority Integrated Catchment Studies (ICS). Data
requirements for calibration are also given in this guidance.

There is unlikely to be sufficient data available to calibrate strategic and catchment scale
models. In this case a reality check against observed data should be carried out.

8.3 Common Issues with Calibration

It should be noted that no model will ever Point to Note:
give a perfect fit to data. The differences
between the calibrated model and recorded Despite best efforts, no model will ever
data should be acknowledged in the technical give a perfect fit to data. From model
report and any areas or events where the fit is calibration it should be clear as to the
poor should be explained. Models should quality of the model and the confidence
not be forced to fit the data either by upon which decisions can be undertaken
making unphysical changes to parameter based on its output.
values or by making changes to model
parameters at a spatial scale which cannot be
supported by the data. At best this will lead to an inappropriate level of confidence in the
model and, at worst will hide model and data error leading to the incorrect conclusions being
drawn from the study. This in turn could lead to poor investments and decision making. A
list of common issues is highlighted below.

Model calibration parameters should only be adjusted within published and accepted
ranges For instance accepted Manning’s roughness coefficients for cultivated areas are
between 0.020 and 0.050 depending on the crop condition (Chow, 1959); so changing the
roughness coefficient to 0.1 for areas covered by cropland would not be appropriate.

Data errors should be considered a possibility during model calibration For instance
an incorrect gauge datum would lead to a mismatch between model and data and although it
might be possible to adjust model parameters to improve the fit to data this would lead to an
error in the model.

Changes to parameter values should be appropriate given the available data For
instance if roughness classifications have been based on land use maps it would not be
appropriate to vary roughness between individual fields with the same land use class unless
other information such as a site visit or aerial photography provided evidence for different
roughness. It would however be appropriate to vary roughness for the entire land use class
within published and accepted ranges. Similarly changing bridge coefficients based on a
single wrack mark upstream of a bridge may be inappropriate, as several other parameters
may affect water levels upstream of the bridge.

Missing flood mechanisms should be considered during calibration For instance a

fluvial model may not reproduce observed flood extents if surface water or groundwater

Flood Modelling Guidance for Responsible Authorities v1.1 86

flooding contribute to recorded flooding. In this case it would be incorrect to adjust
parameters of the fluvial model to reproduce the observed flooding.

Catchment changes since the recorded flood event should be considered For instance
if flood defences have been constructed at a location the model would not be expected to
reproduce flood extents prior to the defences being constructed. It may be necessary to
construct a model of the historical condition of the catchment to assist in calibration however,
where changes have been extensive, the resulting calibration may be of limited use for the
present day case. It should also be noted that not all catchment changes will be
documented.

Calibration runs should always be driven using observed data A comparison of a 0.5%
AEP design event against an observed 0.5% AEP event would be a useful reality check but
would not constitute calibration as the design event may differ from the observed event in
several ways (e.g. hydrograph shape, combination of waves and tides, etc.).

The possibility of several parameter combinations giving the same fit to the observed
data should be considered Typically with flood models there are many fewer
measurements available for calibration than there are model variables. For example, in a 1D
fluvial model variables include roughness at each cross section and coefficients at each
hydraulic structure, while in a 2D pluvial model, variables could include roughness and
evaporation loss at each grid cell. Conversely flow measurements may only be available for
a single point within the catchment. Where there are more variables than measurements,
the problem is said to be “underdetermined” and there may be more than one combination of
parameter values which gives the same fit to the calibration data. These combinations of
parameter values may exhibit different sensitivity and give different results for extreme
conditions outwith the range of the calibration data. Where this is considered a possibility,
sensitivity tests should be carried out to assess the impact of choosing different plausible
parameter sets. This is particularly important for some hydrological models which may have
many more parameters than either 1D or 2D flood models.

Calibration data are not always available and, in such circumstances, greater emphasis
should be put on understanding the model sensitivity and model uncertainties.

8.4 Sensitivity Analyses

Key Points for Responsible Authorities
• Model sensitivity tests should be undertaken for all modelling studies in order to
give the modeller, reviewer and users an understanding of what parameters affect
the model and in what ways. The required sensitivity tests should be specified in
the project SoR.

Key Points for Contractors

• Results from the sensitivity analyses should be presented in the modelling report.

Model sensitivity tests should be undertaken in order to give the modeller, reviewer and
users an understanding of what parameters affect the model and in what ways. Sensitivity
testing involves varying an element of the modelling and assessing how this alters the model
results. This helps develop an understanding of the confidence in the model and its outputs.
Sensitivity analysis is particularly important where limited data is available to validate or
calibrate the model or where there is large uncertainty in model parameters or input data.

Flood Modelling Guidance for Responsible Authorities v1.1 87

There are two types of sensitivity analysis that should be carried out; sensitivity to
parameters which may affect the numerical solution (which is discussed in Chapter 14) and
sensitivity to physical parameters, which is discussed below.

Sensitivity to the following are typically tested:

• Boundary conditions e.g. a 20% increase in design flow

• Surface roughness e.g. increasing or decreasing the Manning’s n used in the model.
• Location and type of upstream and downstream boundary conditions to ensure there is
no impact on results within the area of interest.
• Blockage of critical structures such as culverts and other hydraulic structures which may
be prone to blockage during flood events. Models can be run with full and/or partial
blockage to better understand the impact of these processes.

Additional sensitivity testing of the following may be required, depending on the specifics of
the model:

• Model resolution e.g. increasing or decreasing the cell size e.g. 20 m -> 5 m.
• Key structure coefficients e.g. at bridges and weirs.
• Banktop/floodplain spill coefficients
• Initial conditions/initial water levels in storage areas such as ponds and flood storage
reservoirs
• For pluvial modelling, testing the sensitivity of the model to the storm duration used may
be appropriate.
• Wind boundary conditions, particularly for coastal surge models.
• For wave overtopping models beach/defence profile and overtopping model parameters.

The required sensitivity tests should be specified in the project SoR. Where information is
missing or uncertain, additional sensitivity testing may be valuable such as for example
influence of floodplain embankments.

The following sensitivity tests were carried out on SEPA’s strategic level national fluvial
hazard mapping models. More detailed studies may consider a wider range of tests in
addition to these.

• Sensitivity to a 20% increase in flow for the 1 in 10 and 1 in 200 year defended and 1
in 1000 year undefended scenarios.
• Sensitivity to a 40% increase in roughness for the 1 in 10 and 1 in 200 year defended
scenarios.
• Sensitivity to blockages for the 1 in 10 year defended and 1 in 1000 year undefended
scenarios

Results from the sensitivity analyses should be presented in the modelling report. Sensitivity
analyses results can be presented in several ways. For 1D models the analyses are usually
presented by displaying the different sensitivity model run results on a long section plot.
Alternatively plots showing difference in water level against chainage for each of the
sensitivity runs or tables showing the predicted level at key locations or model nodes of
interest for each of the runs can be produced.

Flood Modelling Guidance for Responsible Authorities v1.1 88

8.5 Confidence

Key Points for Responsible Authorities

• Uncertainty is inherent in all models. Understanding the level of uncertainty in a

modelling study helps informs decisions based on the study outputs.
Understanding the sources of uncertainty can reduce uncertainty in future studies
by targeting the largest sources of uncertainty.
• There are a range of existing methods for analysing uncertainty which vary in
complexity and effort required. Use a risk-based approach to uncertainty analysis
where the level of risk informs the level of uncertainty analysis required.
• Consideration of the type and level of uncertainty analysis required should be
made at scoping stage and the level and form of uncertainty analysis required for
the project should be specified in the SoR.
• Further information on uncertainty is available in SEPA’s Uncertainty Framework
which can be made available to Responsible Authorities.

Key Points for Contractors

• Uncertainty should be considered in all flood studies. The modelling report

should contain a description of the uncertainty analysis undertaken together with
identification of potential sources of uncertainty and an indication of the level of
uncertainty. Decisions or judgments made about the uncertainty, including any
assumptions, should be documented.
• Use the FRMRC Framework for addressing uncertainty in Fluvial Flood Risk
Mapping to determine the appropriate level of assessment for catchment or local
scale studies.
• Ensure that the quoted level of precision in any outputs is appropriate given
uncertainty in the modelling, and include error bounds on study outputs where
appropriate.

Uncertainty is inherent in all models. Uncertainty arises at each level or stage in the process
of modelling flood risk and from a range of sources. Figure 8-1 shows examples of potential
sources of uncertainty in flood risk models.

The level of confidence in the output will reflect the uncertainties within each of the stages
of assessment, such as within the input data, parameters, the model and the way the
outputs may be transformed (Walker, et al., 2003).

Guidance from the Scottish Government (Scottish Government, 2011) states that
“Uncertainty should be clearly presented in flood risk assessments showing what
approaches have been used to quantify them and how decisions have been influenced by
uncertainties. Any assumptions made should be clearly set out”.

Flood Modelling Guidance for Responsible Authorities v1.1 89

 13
Figure 8-1: Examples of potential sources of uncertainty in models

The level of uncertainty analysis should be proportional to the costs and potential benefits.
Detailed uncertainty analysis, with associated resource and time implications, may be
justified where the level of confidence in the model predictions would affect the outcome of a
decision or where the product would be used in evaluation of significant investment, such as
construction of a major flood defence scheme (Beven, 2011).

Consideration of confidence should reflect the purpose of the model and the decisions which
it is intended to inform. It should also consider the level of detail and modelling methods
applied at each level of modelling. In this way, a risk-based approach should be followed
(i.e. the level of risk informs the level of modelling which then informs the level of uncertainty
analysis required).

Identification of the relative importance of different sources of uncertainty, related to the

magnitude of the impact on the final output, may be useful since the largest future reductions
in uncertainty could be gained through targeting the largest sources of uncertainty (e.g.
through collecting cross section data to improve a DTM, or additional gauging to improve the
investigation of inflow boundaries).

There are a range of existing methods for analysing uncertainty including both qualitative
and quantitative methods. These range from simpler forms of analysis (e.g. sensitivity
analysis and approaches which qualitatively score uncertainty), to complex approaches (e.g.
formal, expert elicitation where the opinion of several authorities on the subject is used to
inform confidence intervals, Bayesian methods, regressions and approaches involving
defining distributions for propagating the effects of different sources of uncertainty to see
how these influence model output). The ability to conduct detailed evaluation of uncertainty
may be affected by the availability of data required.

13
Sources of uncertainty identified in: (Apel, Thieken, Merz, & Blöschl, 2004); (Apel, Aronica,
Kreibich, & Thieken, 2009) (Maier & Ascough, 2006), (Bales & Wagner, 2009)

Flood Modelling Guidance for Responsible Authorities v1.1 90

Generally, detailed or complex quantitative approaches are expected to be most applicable
at the detailed modelling scale where there is potential for significant investment in
measures, meaning that a detailed indication of the potential uncertainty is of particular of
benefit. Quantitative approaches can provide a fuller insight into how uncertainty in inputs
propagates through models (to the outputs) and, in some cases, detailed information about
uncertainty in different sources can be produced in the form of probabilities of outcomes,
confidence intervals and/or probability distributions. Quantitative uncertainty analysis may
be used to inform calculation of freeboard levels for flood prevention schemes.

Qualitative approaches focus on identification and grouping of the sources of uncertainty.

They then characterise the uncertainty usually using judgements about the level of
uncertainty. Qualitative estimates and/or sensitivity analysis may be the selected way of
assessing uncertainty where the approach is constrained by available data (e.g. where no
historical data are available in order to constrain estimates of uncertainty).

At the catchment to local modelling scales, the decision flow diagrams contained within the
FRMRC Framework for addressing uncertainty in Fluvial Flood Risk Mapping (Beven, 2011)
may assist in determining which methods of uncertainty analysis are appropriate. It is
recommended that at the outset of the project there should be consideration of the form of
uncertainty analysis required for the project and this should be specified in the SoR.

For all studies a description of the uncertainty analysis undertaken should be provided,
together with identification of potential sources of uncertainty and an indication of the level of
uncertainty. Decisions or judgments made about the uncertainty, including any assumptions,
should be documented.

Flood Modelling Guidance for Responsible Authorities v1.1 91

9 Scenarios

Key Points for Responsible Authorities

• To maintain consistency with national maps and appraisal methods, SEPA

recommends a particular set of model scenarios to be provided to enable an
update to national flood hazard maps.
• Running additional scenarios may add to the cost of a flood study however, this is
likely to be cheaper than running additional scenarios at a later date.

9.1 Introduction
The range and number of scenarios run through a model will depend on the use of a flood
study. For example, a flood risk assessment according to SEPA’s Technical Guidance for
Stakeholders only requires the 0.5% annual exceedance probability (1 in 200 year) event
while a minimum of 5 flood events spanning a range from high to low probability are required
for detailed damage calculations (Penning-Rowsell, et al., 2013). Natural flood management
techniques are expected to be most effective for frequent flood events so an NFM study may
require consideration of more frequent flood events than a design for a hard flood defence
scheme.

Running additional scenarios is likely to increase the cost of a flood study but future use of
the model and results should be considered when specifying the required scenarios as the
costs of contractors rerunning the model at a later date to produce additional scenarios is
likely to be greater due to;

• additional project management costs;

• the risk that models may not run, or produce different results in newer versions of
modelling software meaning significant work is required to run the additional
scenarios;
• additional time required for a different modeller to familiarise themselves with the
model if the original modeller is no longer available, or creating the additional
scenarios is awarded to a different contractor.

Extending the range of scenarios to cover more frequent events may improve confidence in
the modelling, as there is more likely to be data available for validation including anecdotal
evidence on the frequency of flooding.

The science on how climate change may affect flooding is still developing and recommended
allowances for climate change may go up or down in future. Estimates of present day
extreme flows and levels may also change as new data is collected or analysis methods are
improved. A wider range of scenarios can provide a measure of future proofing for a study
as new flow estimates may correspond to a scenario which has already been run.

SEPA’s national hazard maps use a consistent set range of scenarios across each source of
flooding which provide a suitable spread for the damage calculations used to inform the
Flood Risk Management Strategies. To maintain consistency, SEPA requires the same
scenarios to be provided for any update to the national hazard maps. This chapter sets out
the minimum scenarios required for an update to SEPA’s hazard maps. Additional scenarios
may be required in some instances, depending on the study area and the purpose of the
study.

Flood Modelling Guidance for Responsible Authorities v1.1 92

9.2 Fluvial
The scenarios used for SEPA’s hazard maps are given in Table 9-1. To maintain
consistency the core scenarios are required for any study used to update SEPA’s fluvial
hazard maps. The methodology used to develop SEPA’s national fluvial hazard maps only
considers out of bank flow so SEPA does not currently publish flood maps for a 50% or 20%
AEP event. However, including these scenarios is strongly recommended as for many UK
rivers the bank full capacity is the 50% AEP event so this can provide a useful sense check
on model results, particularly in the absence of historical flood information. Sensitivity tests
are discussed in section 8.4.

Core Scenarios Sensitivity Tests

Return Period

Manning’s n
Undefended
Exceedance
Probability

+20% Flow
sensitivity
Defended

Blockage
Change
Climate
Annual

+40%
(%)

50 2 
20 5 
10 10    
3.33 30   
2 50 
1 100  
0.5 200     
0.1 1000   
Table 9-1: Scenarios used in SEPA's national fluvial hazard mapping.

9.3 Coastal
The scenarios used for SEPA’s hazard maps are given in Table 9-2. To maintain
consistency these scenarios are required for any study used to update SEPA’s coastal
hazard maps.

SEPA’s coastal flood maps do not include the effect of waves. For wave overtopping studies,
a joint probability analysis of waves and extreme still water level should be undertaken as
there will be multiple combinations of wave and extreme still water level which could
constitute for example a 0.5% AEP event. This may mean that a range of combinations of
extreme water level and waves need to be run for each flood probability.

Annual Exceedance
Return Period Undefended Climate Change
Probability (%)
10 10 
4 25 
2 50 
1 100 
0.5 200  
0.1 1000 
0.01 10000 
Table 9-2: Scenarios used in SEPA's national coastal hazard mapping.

9.4 Pluvial
The scenarios used for SEPA’s pluvial maps are given in Table 9-3. To maintain
consistency these scenarios are required for any study used to update SEPA’s pluvial
hazard maps.

Flood Modelling Guidance for Responsible Authorities v1.1 93

 Annual Exceedance
Return Period Undefended Climate Change
Probability (%)
10 10 
3.33 30 
2 50 
1 100 
0.5 200  
10 10 
3.33 30 
2 50 
1 100 
0.5 200  
Table 9-3: Scenarios used in SEPA's pluvial hazard maps.

9.5 Integrated Catchment Studies

The scenarios used for Integrated Catchment Studies should be agreed with Scottish Water.

Flood Modelling Guidance for Responsible Authorities v1.1 94

10 Climate Change

10.1 Introduction
Key Points for Responsible Authorities

• An investigation of the impact of climate change on flood risk should be included

in all modelling studies.
• Detailed information on climate change, including appropriate allowances for
different purposes, will be produced by SEPA’s flood risk and climate change
group and this chapter updated as appropriate.

The Climate Change (Scotland) Act, 2009 places duties on public bodies regarding climate
change, including acting in the best way calculated to deliver the Scottish Government’s
adaptation programmes, and to act in the way they consider the most sustainable.
Consideration of climate change is also a key part of the FRM Act. An investigation of the
impact of climate change on flood risk should therefore be a component of any modelling
study.

Information now available on potential climate change and its impacts on flows in our rivers
and the sea level at our coasts provides a clearer consideration of the probable range of
change across Scotland’s regions than ever before. The provision of regional climate
impacts in a probabilistic manner represents a significant change from the long-term
approach of considering the impact of climate change as a single figure uplift applied flatly
across the country.

The new information provides greater flexibility to consider climate impacts in a risk-based
framework although it could, initially, appear confusing. The change of approach warrants
further, specific guidance which will be forthcoming. This chapter of this guidance,
however, deals specifically with the consideration of climate change for strategic
modelling issues in support of FRM actions and summarises the latest information on
climate change impacts on flows.

This chapter;

• Summarises available climate change information for changes in peak river flow,
short duration rainfall, and sea level rise
• Discusses the approach used for these variables in SEPA’s national hazard maps
and whether this is still considered appropriate.

Detailed information on climate change, including appropriate allowances for different

purposes will be produced by SEPA’s flood risk and climate change group, and this chapter
will be updated as appropriate following the discussions of the group. The chapter will also
be updated as appropriate to take account of new studies and scientific recommendations.

10.2 UKCP09 Projections

The leading source of climate information for the UK is the UKCP09 climate projections. The
projections are probabilistic, quantifying uncertainty in climate change projections arising
from the representation of climate processes and the effects of natural internal variability in
the climate system. A user interface allows users to easily access information relevant to a
geographical area or for a particular climate variable. The projections and associated

Flood Modelling Guidance for Responsible Authorities v1.1 95

scientific reports are available from the UKCP09 website
http://ukclimateprojections.metoffice.gov.uk. Projections are available for high, medium and
low emissions scenarios for different time horizons until 2100.

Although the UKCP09 projections were published in 2009, the guidance document “Flood
and Coastal Defence Appraisal Guidance FCDPAG3 Economic Appraisal Supplementary
Note to Operating Authorities - Climate Change Impacts” (DEFRA, 2006) has continued to
be used in many cases in Scotland due to the difficulties of directly relating changes in river
flow and sub-daily duration rainfall to the UKCP09 results. Recent SEPA and UKWIR
projects provide updated guidance on these areas, making use of the DEFRA (2006) study
no longer appropriate to inform strategic decision making. There have also been recent
improvements in understanding how climate change may affect mean sea level since the
publication of UKCP09.

The Met Office is currently developing a new set of UK climate projections, UKCP18, which
will update UKCP09. The projections are expected to provide improved information on how
climate change may affect short duration rainfall and sea level rise. Publication is due in
March 2018, and this guidance will be updated to reflect this and any other improvements in
scientific understanding.

10.3 Fluvial

Key Points for Responsible Authorities

• New information on how climate change may affect river flows is available. This
information is probabilistic, and varies between river basin regions.
• SEPA’s fluvial hazard maps used the 2080 high emissions scenario 67th
percentile (i.e. uplifts in peak flow that are “unlikely to be exceeded”

SEPA commissioned CEH to assess the vulnerability of Scottish river catchments to climate
change (Kay, Crooks, Davies, & Reynard, 2011). The study comprised a sensitivity analysis
to determine how catchments would respond to changes in temperature and the amount and
seasonality of rainfall. Projections for rainfall and temperature from the UKCP09 projections
were combined with the sensitivity analysis to produce a set of probabilistic estimates for
change in river flow for river basin regions across Scotland. These cover high, medium and
low emissions scenarios for the 2020s, 2050s and 2080s time horizons. The UKCP09 river
basins used are shown in Figure 10-1, together with the corresponding hydrometric areas.
Results for the medium emissions scenario for the 2050s are shown in Table 10-1, and
results for the low, medium and high emissions scenarios for 2080s are shown in Table
10-2. It should be noted that uplifts for the medium emissions scenario 50th percentile
in 2080s in the west of Scotland are considerably higher than the 20% uplift
recommended by the DEFRA (2006) guidance. A full copy of the report is available from
SEPA’s website https://www.sepa.org.uk/media/219493/ceh_report_final_sepa.pdf together
with a non-technical summary https://www.sepa.org.uk/media/219494/ceh-cc-report-wp1-
overview-final.pdf.

SEPA’s fluvial hazard maps used the 2080 high emissions scenario 67th percentile;
this is a relatively conservative approach which is considered appropriate for strategic level
mapping. The choice of scenario and probability should be appropriate to the purpose of the
study for instance, a modelling study to inform the design of a flood defence around a site of
critical national infrastructure may wish to use a more conservative climate change
allowance. The scenario or scenarios used should be justified in the modelling report and,

Flood Modelling Guidance for Responsible Authorities v1.1 96

ideally a sensitivity analysis to different allowances should be carried out. If a different
climate change allowance is used to that in SEPA’s national hazard maps, the model may
need to be rerun with a 2080 high emissions scenario 67th percentile uplift in order to gain a
consistent picture of the potential impact of climate change across Scotland. This
consistency is required for incorporation into SEPA’s National Flood Hazard Maps and the
Flood Risk Management Strategies.

For studies at the coast, climate change projections for sea level rise should be considered
as in section 10.5.1.

River
basin Hydrometric Areas
region
North
1,2,3,4,5,6,7,8
Highland
9,10,11,12,13
North east
(northern)
13 (southern),
Tay
14,15,16
17,18,19,20,21
Forth
(coastal)
Tweed 21
Orkney
and 107,108
Shetland
West
93,64,95,105,106
highland
87,88,89,90,91,92,104
Argyll
(Kintyre), 105
82,83,84,85,86,104
Clyde
(Arran)
Solway 77,78,79,80,81

Figure 10-1: UKCP09 river basin regions covering Scotland, for which probabilistic estimates
are available. The hydrometric areas falling within each river basin region are given in the
table.

Flood Modelling Guidance for Responsible Authorities v1.1 97

 % change in flood peak (thresholded based on exceedance likelihood)
Probability Exceedance Orkney/ N W NE
Scenario (%) Likelihood Shetland Highland Highland Scotland Argyll Tay Clyde Forth Solway Tweed
very likely to
10 13 7 11 2 11 6 7 6 7 4
be exceeded
likely to be
33 19 13 19 9 20 11 14 12 13 10
exceeded
is as likely as
50 not to be 25 16 25 12 26 14 18 16 17 13
exceeded
unlikely to be
MEDIUM

67 29 20 31 15 31 18 23 20 21 17
exceeded
very unlikely to
90 34 29 42 21 42 27 32 29 30 24
be exceeded
Table 10-1: Percentage uplifts for the medium emissions scenario 2050s, results from Kay, Crooks, Davies, & Reynard (2011).

Flood Modelling Guidance for Responsible Authorities v1.1 98

 % change in flood peak (thresholded based on exceedence likelihood)

W Highland
N Highland
Shetland

Scotland
Orkney/

Solway

Tweed
Argyll

Clyde

Forth
Tay
NE
Scenario Probability (%) Exceedence Likelihood
10 very likely to be exceeded 15 7 12 2 12 4 8 5 6 5
33 likely to be exceeded 20 14 23 10 23 12 16 13 13 11
50 is as likely as not to be exceeded 27 18 30 13 30 16 20 17 18 14
67 unlikely to be exceeded 30 24 36 16 36 20 26 22 23 19
LOW

90 very unlikely to be exceeded 38 33 50 24 50 31 35 32 35 28

10 very likely to be exceeded 16 10 15 3 15 7 11 7 8 6
33 likely to be exceeded 27 18 29 11 29 15 20 16 16 13
MEDIUM

50 is as likely as not to be exceeded 30 23 36 14 37 20 27 21 22 17

67 unlikely to be exceeded 34 29 44 18 45 25 32 27 28 22
90 very unlikely to be exceeded 45 40 60 28 60 37 45 40 45 32
10 very likely to be exceeded 18 12 20 4 20 11 15 11 13 9
33 likely to be exceeded 29 23 36 12 36 20 27 22 25 18
50 is as likely as not to be exceeded 33 29 45 17 45 26 34 28 32 23
67 41 37 56 24 56 35 44 40 44 33
HIGH

unlikely to be exceeded
90 very unlikely to be exceeded 53 50 >60 33 >60 50 60 54 60 45
Table 10-2: Percentage uplifts of the high, medium and low emissions scenarios for the 2080s, results from Kay, Crooks, Davies, & Reynard,
(2011). Values shown in bold are those used in SEPA’s national fluvial hazard maps.

Flood Modelling Guidance for Responsible Authorities v1.1 99

10.4 Pluvial

Key Points for Responsible Authorities

• New information on how climate change may short duration rainfall events is
available.
• SEPA’s pluvial hazard maps used a 20% uplift for extreme rainfall for the 2080s.
The new information suggests a larger uplift may be appropriate for future,
strategic studies.

The models used to develop the UKCP09 climate projections did not have sufficient
resolution to analyse the type of rainfall events typically responsible for surface water
flooding. SEPA’s pluvial hazard maps therefore used a 20% uplift based on DEFRA
guidance (DEFRA, 2006), which represented the best understanding at that time.

A recent study by UKWIR, Rainfall Intensity for Sewer Design, provides new
recommendations for percentage uplifts in sub-daily duration rainfall depths for climate
change scenarios (Bennett, Blenkinsop, Dale, Fowler, & Gill, 2015). This study used two
approaches to estimate predicted changes in rainfall depths;

• A comparison of the present day rainfall with that at a “climate analogue”, another
location which has a current climate similar to the projected climate. This was
undertaken for selected locations only, and results for Glasgow and Newcastle are
advised for use in the west and east of Scotland respectively;
• Analysis of a high resolution climate model simulation which is of sufficient resolution
to resolve the type of rainfall responsible for pluvial flooding. The model used does
not cover Scotland or Northern England, but has similar results to the “climate
analogue” approach elsewhere.

The study provides low, central, and high projections. The central projection is an
average of the climate analogue and high resolution model projections, while the low and
high projections give the spread in projections from the different approaches (note that the
high and low projections do not constitute a full probabilistic assessment and may not
capture the full spread of possibly changes). Uplift values from the UKWIR study for use in
Scotland are given in Table 10-3.

Location Water & Duration Epoch

sewerage (hours) 2030s 2050s 2080s
company
L C H L C H L C H
applicability
1 14 19 25 16 27 37 23 50 88
NW Scottish
3 0 7 10 0 6 11 11 22 29
(Glasgow) Water (west)
6 8 12 18 0 7 20 4 20 36
1 20 28 40 24 44 75 45 50 60
NE Scottish
3 12 15 16 18 29 41 35 53 76
(Newcastle) Water (east)
6 5 7 10 8 17 30 33 51 75
Table 10-3: Percentage change in rainfall depth for different locations and different epochs
recommended by the UKWIR study, Rainfall Intensity for Sewer Design (Bennett, Blenkinsop,
Dale, Fowler, & Gill, 2015). The percentage uplifts are based on a recent climate baseline (e.g.
1981-2010) and for use with the 2015 release of the FEH DDF model and are given for Low (L),
Central (C) and High (H) projections.

Flood Modelling Guidance for Responsible Authorities v1.1 100

The central projections for 2080 from the UKWIR study are significantly higher than those
recommended in the DEFRA guidance. Responsible Authorities are recommended to
use the uplifts from the UKWIR study for any future surface water studies as these are
based on the latest science. A risk-based approach should be adopted to make use of the
high and low estimates. It should be noted that the high and low estimates of change are
not absolute, with other sources of uncertainty being present and should be considered.

Changes in mean sea level or river flows may have an impact on the duration and frequency
of tide locking of surface water drainage systems. If the conceptual model identifies tide
locking as important, it may be necessary to consider climate change projections for sea
level rise and river flows, section 10.5.1 and section 10.5.1.

10.5 Coastal

Key Points for Responsible Authorities

• Climate change can affect coastal flood risk through changes in mean sea level or
changes in storminess which affects storm surges and waves.
• Recent projections of global sea level rise are greater than those used for the
UKCP09 climate projections.
• SEPA used the 2080 high emissions scenario 95th percentile, relative sea level rise
for the national coastal hazard maps.
• There are large uncertainties in the projected change in the UK wave climate due
to climate change. It is not possible to recommend climate projections for waves
however, sensitivity analysis should be undertaken where appropriate.
• The projected impact of climate change on surge is small compared to projected
changes in mean sea level and can usually be ignored.

Climate change may impact coastal flooding through changes in mean sea level or through
changes in storminess, which affect surge and waves.

10.5.1 Extreme Still Water Level Rise

Climate change can affect extreme still water levels through changes in mean sea level or
changes in storminess which may affect the frequency and magnitude of surges.

UKCP09 provides projections of absolute and relative sea level rise. The relative sea level
rise predictions are of most use in flood risk management and account for movement in the
land surface. Results are available for the 5th, 50th and 95th percentile, high, medium and
low emissions scenarios on a 12 km grid around the coast. Projected sea level rises are
provided for every year up to 2100 from a base year of 1990.

UKCP09 projections of trends in storm surge are less detailed than for sea level rise. Long
term linear trends in mm/yr are provided for the period 1951-2099 for the medium emissions
scenario 5th, 50th and 95th percentile only. In most locations the projected change in surge
in the UKCP09 results is small compared to the projected changes in mean sea level, and
may not be distinguishable from natural variability so that consideration of changes in mean
sea level only is sufficient.

Since the UKCP09 climate projections were published, improvements in scientific

understanding, particularly in the likely contribution of land ice melt to sea level rise, have led

Flood Modelling Guidance for Responsible Authorities v1.1 101

to higher projections of global sea level rise (IPCC, 2013). The UKCP09 scenarios may
therefore underestimate the potential range in sea level rise. A risk-based approach should
be adopted in using the UKCP09 sea level rise projections, but the medium emission
scenario 50th percentile should not be considered a central estimate. Acknowledging the
changes in scientific understanding since publication of the UKCP09 sea level projections,
SEPA used the 2080 high emissions scenario 95th percentile relative sea level rise for
national coastal hazard maps. UKCP09 also provides a H++ projection of combined sea
level rise and surge which is beyond the likely range but within physical plausibility. The
H++ scenario may be appropriate as a sensitivity test, particularly for critical infrastructure in
coastal locations.

The CFB boundaries suggested for use in section 6.5.1 have a reference year of 2008, so
the change between 1990 and 2008 in the UKCP09 results should not be included if the
CFB boundaries are used as a model input.

The UKCP09 sea level rise grids do not cover the upstream extent of some estuaries and
sea lochs. If hydrodynamic modelling of the loch or estuary is not undertaken to establish
extreme sea levels inland, the adjacent downstream UKCP09 sea level rise grid predictions
should be ‘borrowed’ and used directly at the estuary/loch site of interest. If hydrodynamic
modelling is undertaken, the sea level rise estimates should be applied to the offshore
boundary of the hydrodynamic model.

10.5.2 Waves
There are large uncertainties in the projected change in the UK wave climate due to climate
change; the UKCP09 projections have changes in the annual maxima of between –1.5 m
and +1 m (Lowe, et al., 2009). The Marine Climate Change Impacts Partnership report
“Impacts of climate change on storms and waves”. Woolf & Wolf (2013) reviews current
understanding and identifies knowledge gaps, including:

• How changes in the mid latitude storm tracks due to climate change.
• How results from global climate models can be best used to investigate local
changes in wave climate.
• How changes in offshore waves have an impact at the coast.

Due to the uncertainty it is not possible to recommend climate projections for waves,
however appropriate sensitivity analysis should be undertaken.

Flood Modelling Guidance for Responsible Authorities v1.1 102

11 Defences

Further guidance on modelling defended and undefended scenarios is currently being

developed.

Flood Modelling Guidance for Responsible Authorities v1.1 103

12 Natural Flood Management

12.1 Introduction
Natural Flood Management (NFM) is an important component of the FRM Act, which
requires it to be considered in the development of flood protection measures. SEPA’s
Natural Flood Management Handbook (Scottish Environment Protection Agency, 2016)
provides a practical guide to the delivery of NFM measures. This chapter is intended to
compliment chapters 5 and 6 of the NFM handbook and the rest of this guidance document
by providing further information on how NFM may be included as part of a modelling study.
The relationship between the modelling process in this guidance and the NFM
implementation process outlined in the NFM handbook is shown in Figure 12-1. The
following discussion assumes that opportunity areas for NFM have already been identified
(step 3 in the NFM implementation process).

Despite recent research there are still gaps in scientific understanding of how the potential
effects of NFM measures may be assessed, particularly for measures in the wider
catchment. Dealing with these knowledge gaps may require innovative approaches to be
adopted for NFM studies other than those considered here, and this chapter will be
developed as scientific understanding improves.

12.2 Modelling Approaches for NFM Studies

Key Points for Responsible Authorities

• Any modelling should be proportionate to the study objectives, the likely scale of
impact of the NFM measures and the achievable confidence in the approach and
outcomes.

Modelling may be required at 3 stages of the NFM implementation process, short listing of
measures (part of the NFM scoping stage), options appraisal/detailed assessment and
detailed design. At all stages the investment in any modelling should be proportionate to the
study objectives, the likely scale of impact of the NFM 3 measures and the achievable
confidence in outputs.

12.2.1 Short Listing of Measures

Where resources permit, modelling can be undertaken at this stage to determine how much
change in flood flows or levels NFM measures would have to deliver in order to achieve a
given reduction in flood risk. A possible approach to this is scenario testing with catchment
scale hydraulic models to assess the changes in inflow hydrographs which would be
necessary to deliver a given reduction in flood risk. Further information on this is given in the
NFM handbook. Hydraulic models used for this approach should be unsteady and should
be calibrated for flows, levels and travel times where data permits. If models are not
calibrated, an indication of impact is still possible but there will be a reduced level of
confidence in the results. Either 1D or 1D-2D hydraulic models could be used for this type of
assessment, although the longer runtimes of 1D-2D models may prevent sufficient numbers
of scenarios being tested. A combination of hydraulic and routing models may also be used.

Flood Modelling Guidance for Responsible Authorities v1.1 104

 Stages in Implementing an NFM Project Relevant Flood Modelling Guidance

Step 1 – Need/aspiration
Identification of NFM need or aspiration by Local
Authority in FRM Strategies and Local FRM Plans or
by land manager, NGO or local stakeholders.

Step 2 – Engagement
Land manager engagement to assess level of interest
and obtain buy in, plus wider stakeholder engagement
and awareness raising – will continue throughout
process.

Step 3 - Identification of opportunity

areas
High level assessment of opportunity areas for NFM,
including a desk based study of GIS maps (e.g.
SEPA’s NFM maps).

Step 4 – Scoping study

Identification and prioritisation of NFM measures within
a catchment or coastline, informed by
catchment characterisation, a high level appraisal of
the effects of the measures identified and
feasibility/land manager considerations

Catchment/Coastal Characterisation Scoping a Modelling Study

(Chapter 3)
Long Listing of Measures
Strategic or catchment scale
Short Listing of Measures
modelling study

Step 5 –Options Appraisal

Assessment of the various options to implement the Catchment or local scale
prioritised measures and the relative advantages
and disadvantages of each option, informed by surveys modelling study
and modelling as required

Local scale modelling study. FRA

Step 6 –Design
Production and agreement of design (including in accordance with (SEPA, 2015)
permissions) including the production of engineering
drawings where required, informed by surveys and (Nb. Catchment scale modelling may also be
modelling as required required in some cases to determine wider
impacts of measures)

Step 7 – Implementation of works

Implementation of measures on the ground (ground
works).

Step 8 - Management and monitoring

Ongoing management and maintenance of measures,
including monitoring of effect to inform adaptive
management

Figure 12-1: Relationship between the modelling process as set out in this guidance document
and the NFM implementation process set out in the NFM handbook.

Flood Modelling Guidance for Responsible Authorities v1.1 105

Expert judgement can be used to compare the required changes in inflow hydrographs with
that likely from long listed NFM measures. This could also be informed by sensitivity testing
using hydrological models to determine a plausible range of changes to peak flows and
hydrograph shape resulting from the implementation of NFM measures. Hydrological
models may also be used at this stage to give an indication of the areas in the catchment
where NFM measures may be most effective.

In some cases the same scenario testing process can also be used to inform short listing of
hard engineering measures.

12.2.2 Options Appraisal

Modelling should generally be undertaken to inform options appraisal. The NFM handbook
provides guidance on the type of modelling tool which may be appropriate for assessing
different types of NFM measure.

For measures which are in or adjacent to a river channel, hydraulic modelling can be used to
provide a quantitative estimate of the effects of the NFM measure on water levels,
hydrograph timing, flood extents and damages relative to the baseline scenario.

For measures in the wider catchment, hydrological modelling can be used to provide a
qualitative indication of the effect of the NFM measure on peak flood flows and hydrograph
timing. Current limitations in scientific understanding and assessment tools mean that
a quantitative assessment of NFM measures in the wider catchment is not possible.

12.2.3 Detailed Design

At this stage a flood risk assessment complying with SEPA’s Technical Flood Risk Guidance
for Stakeholders (SEPA, 2015) should be produced. Further detailed modelling may also be
required to inform the design. Modelling and flood mapping showing the effect of the
measure may be used to inform visualisations with and without the measure in place.

12.3 Consideration of NFM within Flood Studies

Key Points for Responsible Authorities

• Consider NFM at the scoping stage for any modelling study.

• Undertake the catchment/coastal characterisation and long listing of potential
measures described in the NFM handbook prior to commissioning any modelling,
and use this to inform the scope.

NFM should be considered at the scoping stage of flood modelling studies, as the location
and type of possible NFM measures may affect the modelling approach taken and the study
area. Developing a conceptual model of a catchment is particularly important for NFM
studies as measures designed with a poor understanding of catchment flooding mechanisms
could inadvertently increase flood risk, for example, through increasing the synchronicity of
flood peaks or increasing the risk of structure blockages. The study area should be sufficient
to cover all upstream and downstream effects of proposed NFM measures.

It is recommended that the catchment/coastal characterisation and long listing of potential

measures described in sections 6.4.1 and sections 6.4.2 of the NFM handbook is
undertaken and delivered prior to commissioning any modelling. This characterisation can
be used to inform the scope of any modelling and to determine whether investment in any
modelling is required. Contractors used for the catchment/coastal characterisation and long
listing of measures may also be asked to scope any modelling study.

Flood Modelling Guidance for Responsible Authorities v1.1 106

 Case Study – Floodplain reconnection upstream on the River Nith

The Whitesands area of Dumfries experienced significant flooding from the River Nith in
1962, 1977, 1982, 2009 and 2013. There are extensive agricultural flood embankments
on the Nith upstream of Dumfries and breaching these to reconnect the flood plain and
provide additional storage was identified as a potential NFM option to reduce flood risk
within the town.

To assess this option a model of Dumfries was constructed covering the proposed areas
of flood plain reconnection and the town of Dumfries. The model showed that although
breaching the embankment reduced water levels in the town for 10% and 4% AEP events
they were increased for 1% and 0.5% AEP events. The breach in the embankments
allowed the water to flow into the storage area, behind the embankment before the peak of
the event so as the event peaked, the storage area, which was already full, was unable to
store more water, causing the water in the storage area to flow back into the watercourse.
Without the breach the area behind the embankments only flooded during the peak of the
event reducing water levels downstream during the peak (Mouchel, 2011).

Considering potential NFM options during scoping allowed the study area to be extended
to cover the area identified for the NFM measures. The unexpected detrimental effect of
this NFM option during larger flood events highlights the importance of developing a
conceptual model of the catchment flooding mechanisms in order to identify all possible
effects of a measure. In other situations the conceptual model may identify positive
impacts of NFM which otherwise may not have been identified during scoping.

12.4 Hydraulic Modelling as part of an NFM Study

Key Points for Responsible Authorities

• NFM is expected to be most effective for more frequent flood events so these
should be considered in the flood modelling study.
• Representation of some NFM measures in hydraulic models is uncertain so
additional sensitivity testing may be required.

Flood Modelling Guidance for Responsible Authorities v1.1 107

 Key Points for Contractors

Where hydraulic modelling is being used to assess NFM measures;

• The hydrological analysis should be extended to cover flow events which occur
more than once a year on average.
• The scenarios run should be sufficient to determine if there is a change to the
frequency of out of bank flows or the depth or duration of flooding during frequent
events.
• The scenarios run need not include less frequent flood events if the measure has
no measurable effect for more frequent events (e.g. if a measure has no effect in
a 3.3% AEP event and a 2% AEP event it would not be necessary to model a 1%
AEP event. However, a 0.5% AEP event will usually be required for a flood risk
assessment at detailed design stage in line with (SEPA, 2015)).
• Results for the 50% AEP event and more frequent events should be mapped for
the baseline model. These maps should be compared with landowner or land
manager knowledge of frequent flooding.
• The schematisation used to represent NFM measures should be described in the
modelling report, and justified with reference to available research literature.
• Sensitivity tests should cover;
o The schematisation of each NFM measure. This should consider the full
range of plausible parameter values.
o Seasonality and maturity of the NFM measure, where relevant.
o Blockages at key structures downstream of the measure if the NFM
measure may increase debris supply.

The guidance in chapters 6, 7 and 8 on model schematisation, boundary conditions and

calibration uncertainty analysis also applies to developing models used to assess the impact
of NFM measures. In general, to assess the effects of measures within the channel, a 1D
or 1D-2D model is required while, for measures adjacent to the channel, a 1D-2D or 2D
model would be required. All models used for NFM studies should be unsteady.

The benefits of many NFM measures are likely to be greatest for more frequent events while
some will have no measurable impact during major flood events. Some NFM measures may
affect the frequency, depth or duration of flooding of agricultural land which is already
subject to frequent flooding; providing an indication of the scale of these effects may also
assist in consultations with landowners. Where data is available, calibration of the baseline
model without any NFM measures in place should be as described in chapter 8. Landowner
and land manager knowledge should be used as an additional source of historic flood event
data for model verification.

Flood Modelling Guidance for Responsible Authorities v1.1 108

 Schematisation Uncertainty in Hydraulic Models

For many NFM measures there is a limited body of scientific research available to
support model schematisation, so modellers have to use expert judgment in choosing
how to represent NFM measures. For instance debris dams could potentially be
represented in a 1d hydraulic model by;
• a constriction in individual a cross section,
• increased roughness along a river reach,
• as a weirs, spill or gate

For each of these possible schematisations there are a range of parameter values which
may be appropriate (percentage constriction, roughness value, weir coefficients etc).

The level of confidence in the modelling of any particular NFM measure will depend on the
type of NFM measure being assessed and the body of scientific research available to
support model schematisations and the choice of parameter values. The effectiveness of
some NFM measures may depend on the season and the maturity of the measure. For
instance vegetation will increase roughness more in summer than in winter, and roughness
for established riparian woodland will be greater than for newly planted woodland. There
may be concerns that, in some cases, NFM may increase debris supply, which may in turn
lead to structure blockages.

12.5 Hydrological Modelling as part of an NFM Study

Key Points for Responsible Authorities

• The additional cost associated with the use of more complicated approaches such
as development of bespoke tools or the use of distributed hydrological models
should be justified by an expected reduction in uncertainty.
• Contractors should be able to demonstrate sufficient experience in the use of any
models or techniques proposed and their application to NFM. This is particularly
important for distributed hydrological models which historically have been used
primarily for academic research and where engineering and modelling consultants
may have limited experience.
• Contractors should be able to demonstrate that any bespoke or new tools have
been checked and reviewed and are appropriate to the study.

Flood Modelling Guidance for Responsible Authorities v1.1 109

 Key Points for Contractors

• Uncertainty in the modelling should be tracked at an appropriate level for the

study so it is clear if predicted effects of the measure are greater than uncertainty
in the modelling.
• Sensitivity testing should consider the full range of plausible parameter values for
the present day and NFM conditions.
• The approach taken for model calibration should take account of the possibility of
multiple sets of parameter values giving the same fit to data and models not
being well calibrated for change conditions.
• Any bespoke tools should be checked and reviewed. This should include a check
of the code and the scientific assumptions used to develop the tool. A record of
the checks carried out should be provided in the modelling report.

With all types of hydrological models there is uncertainty regarding the application to NFM
and how model parameters should be modified to represent proposed catchment changes.
At present, the use of hydrological models is restricted to providing an indication of the
sensitivity to any proposed change and where in the catchment changes are likely to have
most effect. There is insufficient confidence in the application of hydrological models to
assessing NFM measures to provide predictions in the change in flow due to NFM for a
particular rainfall event. The NFM handbook gives examples of three approaches which
may be useful in NFM studies;

• FEH and ReFH2 rainfall runoff models (see section 6.3.2.2),

• Use of more complex commercially or freely available distributed hydrological
models, or
• Development of bespoke tools.

The approach adopted will depend on the purpose of the study, the available data, the size
of the project and the potential impacts. It should be noted that more complex approaches
may not necessarily lead to a significant reduction in uncertainty.

Flood Modelling Guidance for Responsible Authorities v1.1 110

 Uncertainty in Hydrological Models

Uncertainty analysis is particularly important where hydrological models are used in NFM
studies. Uncertainties arise because:
• The relevant physical processes and inputs may vary over a much smaller scale
than the available data. There may also be gaps or inaccuracies in the required
input datasets.
• The model resolution may not be sufficient to capture local processes, for instance
,in a distributed model there may be several small incised channels within a model
grid square.
• Not all relevant processes may be included in a chosen model
• A large number of model parameters may be adjusted through calibration though,
typically, only a small amount of calibration data is available. Several choices of
parameters may give similar fits to data but they may respond differently to
change scenarios. Where models are well calibrated to current conditions, it is not
certain that they will remain well calibrated for the future condition with the
inclusion of the NFM measure.
• It may not be clear how input datasets should be altered to account for change
scenarios.
• Even physically based models which include a detailed representation of
hydrological processes involve some form of parameterisation, for example
vegetation may be divided into types and certain properties such as canopy
storage would be associated with a particular vegetation type. These
parameterisations may not be relevant to all catchments.

A further discussion of the issues involved is provided in (O’Donnell, O’Connell, & Quinn,
2004).

Flood Modelling Guidance for Responsible Authorities v1.1 111

13 Flood Mapping

13.1 Introduction
Flood maps can be produced from hydraulic model results to spatially represent data such
as flood extent, depth, velocity and hazard for sources such as pluvial, fluvial, coastal and
sewer flooding. This chapter;

• describes how flood maps are produced from different types of model;
• describes some of the issues which can occur with different types of maps;
• suggests what should be considered in a review of flood maps;
• describes the post processing required for consistency with SEPA’s national hazard
maps.

13.2 Fluvial

13.2.1 1D flood Mapping

The method of fluvial flood mapping is dependent on the hydraulic modelling package. Many
hydraulic modelling packages include functionality to produce flood maps from 1D results
(e.g. Infoworks RS, FloodModeller, MIKE 11) while others such as HEC-GeoRas contain
add-ons to GIS packages. Flood maps can also be produced solely in GIS.

The basic data requirements to create a flood extent map include maximum water levels, a
DTM and cross section locations. If the model includes reservoir or storage units then these
will need to be represented separately in order to represent the water level within the
reservoir unit as opposed to the cross section at this location and, in this case, a plan of the
reservoir locations is also required.

A triangulated irregular network (TIN) is created from the cross sections and reservoir areas
and the water level at each cross section is assigned to the relevant nodes (vertices) of the
TIN. This water level is then interpolated between the TIN nodes to create a water level
surface. The DTM is subtracted from the water level surface to produce a depth grid. Areas
with negative depth are dry and removed from the outputs. Flood extents are then produced
by contouring the processed depth grid.

The resolution of the DTM used determines the resolution of the flood maps. As such it
should be appropriate for the level of detail in the model and should not lead to excessively
large file sizes for the depth grids. It is recommended that 1D flood maps have a maximum
resolution of 5 m.

Due to the interpolation of level results between model cross sections several issues may
occur in 1D flood maps and a careful check against the 1D model results is required. The
maps should be examined for the following features and if, necessary, the model should be
amended accordingly.

• Isolated patches of flooding which are not well connected to the river;
• Flood extents which appear constrained by cross section extents or reservoir extents;
• Flood extents which are greater than the area covered by the cross section extents
or reservoir extents.

1D flood models do not have the functionality to produce hazard ratings or floodplain
velocity. Where these are required a 2D model should be used.

Flood Modelling Guidance for Responsible Authorities v1.1 112

13.2.2 2D Flood Mapping
The production of flood maps from 2D models is simpler than from 1D models as water
levels, depth, velocity and hazard can be output directly from 2D models. Flood extents are
then produced by contouring the processed depth grid.

For 1D-2D models, flood extents and depth and level grids for the 1D component only
should be produced as for 1D models and added to the 2D grid. For consistency with
SEPA’s hazard mapping the 1D component should be assigned a value of ‘200’, velocity
for the 1D component should be assigned a value of ‘200’ for the magnitude and ‘-9999’ for
the direction.

13.3 Coastal Flood Mapping

13.3.1 Horizontal Projection Method

SEPA’s national coastal hazard maps used a basic horizontal projection methodology from
still water levels, Figure 13-1. The Coastal Flood Boundary (CFB) dataset was used to
provide estimates of design sea levels every 2 km around the coast with some points
included within firths and estuaries where available. From each CFB sea level point, a cross
section was drawn to link to high ground with land above the relevant return period on the
DTM. Everything below this line is within the flood extent. From the still water level data
attached to the cross sections, a water surface was interpolated between the sections to
give a continuous water surface reflecting the values of the CFB data. Flood depth maps
were created by subtracting the DTM containing the ground elevation data from the water
surface layer. It should be noted that this method can overestimate flood risk in some areas
as it assumes an infinite momentum and volume of water. Flood waters will keep moving
inland until they hit the required contour on the DTM. In other areas the method may
underestimate if wave action is considered important. This method is suitable for national
scale flood mapping projects, but it is not expected that Local Authority flood studies will use
the same approach.

Figure 13-1: Flood map generation from horizontal projection method.

13.3.2 2D Coastal Flood Mapping

As with 2D fluvial flood mapping, water levels, depth, velocity and hazard can be output
directly from 2D models and flood extents are produced by contouring the processed depth
grid.

Flood Modelling Guidance for Responsible Authorities v1.1 113

13.4 Pluvial Mapping
In general 2D models will be used for pluvial flood mapping and water levels, depth, velocity
and hazard can be output directly. Pluvial models involve the application of rainfall to every
model grid cell resulting in a shallow depth of water at all grid cells. To produce flood maps
which do not show the entire model grid area as flooded it is necessary to threshold the
results to remove shallow depths. The velocity, hazard and water level results should also
be processed to remove results below the depth threshold. For consistency with SEPA’s
pluvial hazard mapping a 0.1 m depth threshold should be used.

13.5 Post-Processing
Post processing of flood maps is required to ensure an appropriate representation of flood
risk. The requirements of post-processing will vary dependent on the purpose and scale of
the flood map; however typically post processing is carried out to:

Remove dry islands below an area threshold as confidence in flood extents at a small
spatial scale is likely to be low and these areas would be isolated during a flood event.
Numerous small holes also increase the complexity of storing the data in GIS.

Remove puddles below an area threshold as confidence in flood extents at a small spatial
scale is likely to be low. However, the reason for the puddles should be understood before
any post processing as this can indicate incorrect initial conditions, frequent wetting and
drying of the model or general instability. Numerous small puddles also increase the
complexity of storing the data in GIS.

Show bridges as wet or dry depending on whether or not there is flow over the bridge
deck. This is used to assess flood risk to transport routes. Depending on the DTM, the
bridge representation in the model and the method of flood mapping this may require manual
post processing.

Depth threshold pluvial model results so that the entire model domain is not included in
the flood extent.

SEPA’s national hazard maps have been post processed to:

• Remove dry islands and isolated wet areas less than 200 m2 which are not
connected to the floodplain. Dry islands have been assigned a depth of 0.01 m, a
velocity of 0.01 ms-1 and a hazard of 0.1.
• Remove results below a depth threshold of 0.1 m for pluvial flooding

13.6 Quality Checking

Flood maps should be sense checked. Key considerations include (list not exhaustive):

• Are the depths/velocities/water levels reasonable?

• Does the inundation extent reflect the topography?
• Are there any areas with particularly large depths/velocities?
• Is the river included in the floodplain extent?
• Are there any false blockages e.g. areas where flow has been prevented through a
structure which would not occur in reality?
• Does the flood extent/depths increase with return period as you would expect?
• Do isolated wet areas connect to the main floodplain through for example a drainage
channel?
• Do flood extents extend across structures which would obstruct flow?

Flood Modelling Guidance for Responsible Authorities v1.1 114

14 Quality Assurance and Quality Control

14.1 Introduction
Errors in data, model schematisation and analysis can have a major effect on study results.
At worst, if these errors are not identified, decisions can be made based on incorrect
modelling, for instance development could be permitted in areas at risk or flood defences
could be built to the wrong level. Where errors are identified during a project this can lead to
significant rework and result in time delays while, if errors are identified after a project is
finished and “accepted” by a Responsible Authority it can be difficult to get contractors to
revisit the work.

To ensure good quality output, quality control and quality assurance should be built in to all
stages of a modelling project, by both the Responsible Authorities and contractors Figure
14-1. This will require the Responsible Authority to review outputs and provide input at key
stages of modelling project.

This chapter recommends quality control and quality assurance activities which may be
carried out by contractors and Responsible Authorities at different stages of a modelling
project. However, it is does not seek to replace Responsible Authorities’ or contractors’
quality assurance and quality control procedures.

14.2 Scoping and Commissioning a Study

Key Points for Responsible Authorities

• Consider quality criteria at scoping stage.

• Ensure the appointed contractor is proposing to use qualified and experienced
staff.
• Ensure risks to quality are included in a contractor’s risk register.

The Responsible Authority should consider the required quality criteria for the modelling in
terms of calibration tolerances, and this should be stated in the SoR.

The contractor should set out the quality control and quality assurance processes which will
apply to the project in their tender. Risks to quality should be included in the risk register.

The Responsible Authority should ensure that the appointed contractor is proposing to use
qualified and experienced staff for modelling, hydrology and project management.
Identification of less experienced staff as part of a project team is acceptable provided that
sufficient time is allocated for more experienced staff to provide technical input. In addition
to the core project team a contractor should identify an internal reviewer who is not directly
involved in the project.

Where the study involves development or use of novel tools of methodologies to meet
Responsible Authority requirements in developing areas of flood risk science such as NFM
or Climate Change additional levels of review are likely be required. The contractor should
set out how any novel tools will be or have been reviewed in their tender, including review of
the concept, coding and usage of the tool. If there is any concern regarding novel
approaches, please contact SEPA.

Flood Modelling Guidance for Responsible Authorities v1.1 115

 Responsible Contractor Activities
Authority Activities

Quality criteria Scoping

specified in SoR. (Chapter 3)

Suitably qualified and

Commissioning experienced project team,
a study (Chapter including internal reviewer.
4)
Risk register includes risks
to quality
Data
Collection Data register
(Chapter 5)

Agree contractor’s
conceptual model
and methodology Model Conceptual model and
Schematisation proposed methodology
Review hydrology (Chapters 6 and reviewed by experienced
technical note 7) staff.

Model log files kept.

Model well commented.
Calibration Key assumptions and
Review calibration and sensitivity decisions recorded.
outputs testing
(Chapter 8)

Design runs
(Chapter 9)

Full model audit and

Review flood maps Flood Mapping technical review certificate
and levels (Chapter 13)

Review final report Reporting Modelling report comprises

Review final (Chapter 15) audit trail for the modelling.
delivered models Purpose and limitations of
the modelling clearly stated
to avoid inappropriate
future use.

Figure 14-1: Contractor and Responsible Authority QA and QC activities at different stages of
a modelling project.

Flood Modelling Guidance for Responsible Authorities v1.1 116

14.2.1 Conceptual Model
Key Points for Contractors

• The conceptual model and proposed methodology should be reviewed and

signed off by an experienced modeller and agreed with the Responsible Authority.

The contractor should develop its own conceptual model for the study area and decide on an
appropriate methodology also considering the study purpose. The conceptual model and
proposed methodology should be reviewed and signed off by an experienced modeller and
agreed with the Responsible Authority. A review by suitably experienced staff at this stage
should ensure that an appropriate approach is adopted from the start of the study, that
potential problems are identified and appropriate mitigation is put in place. This should
include identifying key points in the study where an internal technical review of the modelling
and analysis by the contractor would be beneficial.

Available data should also be reviewed at this stage to determine if it is suitable to meet the
objectives of the study or if there are any issues with data quality and availability. Where the
available data is not sufficient to meet the study objectives and quality criteria the contractor
should make the Responsible Authority aware of the quality implications.

14.3 Data Collection

A data register should be kept as this can help in understanding any implications to the
project output if quality issues are identified with any of the input datasets at a later date. To
ensure that any data used for the project is of an appropriate quality the contractor should
review all data before use and a record should kept of any checks carried out and any
problems identified.

14.3.1 Model Build

Contractors should ensure that their work procedures for model build follow normal good
practice. This includes keeping log files, ensuring file naming structures are logical and
including comments in models where appropriate.

14.3.2 Model Audit

Key Points for Contractors
• Carry out an internal review of any models and calculations, using an independent
reviewer.

Key Points for Responsible Authorities

• Ask contractors to provide evidence that an internal audit of any models and
calculations has been carried out.

Errors in models and calculations can significantly affect study results. The contractor
should carry out an audit of the modelling and any calculations. This should be carried out by
an experienced modeller who has not been involved with constructing the original model or
models. The exact checks carried out will depend on the level of study and the methodology
used.

Flood Modelling Guidance for Responsible Authorities v1.1 117

For catchment or local scale studies a detailed audit should be carried out for all calculations
and for all models constructed. In some case an external audit by an independent
consultant may also be appropriate for some local scale and flood defence design studies.

For strategic scale studies such involving multiple models, such as the nearly 3500 models
created for SEPA’s national fluvial hazard mapping project, it may not be feasible to carry
out a detailed review of all models. In this case it may be appropriate to use automated tools
to screen for numerically unstable models and any physically unrealistic or inconsistent
results. However, the method used and implementation of the method as well as the results
from any screening should be reviewed by an experienced modeller. If automatic checks
are used, detailed manual checks should also be carried out on a subset of models to
ensure that any systematic errors not identified by the automatic screening are detected.

The Responsible Authority should ask for evidence that this audit has been carried out; this
can be in the form of a signed technical review certificate, or a model audit report, with a
record of actions taken. An example review certificate from SEPA’s regional pluvial hazard
mapping is included in appendix C.1. A pro forma for a fluvial flood estimate review is given
in Appendix C.2.

As a minimum the audit should cover the following areas;

Model schematisation - including roughness, structure representation, boundary conditions

and flood plain representation where relevant.

Numerical solution – including mass balance, sensitivity to parameters which affect the
numerical solution, model convergence.

Documentation – log file documenting all model version and key assumptions, data
register, comments in model etc.

Model Calibration – fit to data, parameters adjusted within physical range.

Results and sensitivity tests - behaviour of model as expected, results consistent between
different AEPs.

Due to potential problems with instability of 1D-2D models, Responsible Authorities are
advised not to accept models unless the mass balance for all model runs is reported in the
model audit and is within acceptable limits.

14.3.3 Responsible Authority Review of Results

Key Points for Responsible Authorities
• Review outputs at key points in a project.
• Ensure you receive all model run and results files.

Review of interim outputs at key points by the Responsible Authority can help identify
potential problems. As a minimum it is recommended that Responsible Authorities review
the following outputs;

• the contractor’s conceptual model and proposed methodology,

• a technical note on the hydrology or tidal/coastal boundary conditions.
• technical note on model calibration,
• flood maps and levels for design runs and sensitivity tests,

Flood Modelling Guidance for Responsible Authorities v1.1 118

 • final report and deliverables,
• final delivered models.

In reviewing outputs Responsible Authorities should use their local knowledge to check that
results are physically realistic however, they should be aware flood models can show the
correct behaviour for smaller events within Responsible Authority experience but may not
exhibit the correct sensitivity for larger events. Responsible Authorities should be satisfied in
their review of outputs that there is no evidence of the common problems for different types
of model described in sections 7.2.3, 7.3.4 and 7.4.1. SEPA may be able to assist with the
review of outputs if required.

It is important that complete model run and results files are provided by a contractor as set
out in Section 15.4 as it is not possible for a Responsible Authority, SEPA or any external
reviewer to review a model only from the modelling report. Any survey data or photographs
should also be provided to enable model schematisation to be checked against the survey.

14.4 Reporting
Key Points for Contractors
• The modelling report should clearly state the purpose of the modelling and any
limitation.
• There should be sufficient detail in the modelling report and appendices for any
experienced modeller to reproduce the analysis.

To avoid inappropriate future use of the study outputs, the modelling report should clearly
state the purpose of the modelling. Any limitations of the study, the data and the method
used which may affect use of the results should be highlighted, and recommendations for
future improvements to the modelling should be made.

The modelling report and appendices should comprise an audit trail for the modelling,
providing sufficient detail of the methods and datasets used for any experienced modeller to
reproduce the analysis.

Flood Modelling Guidance for Responsible Authorities v1.1 119

15 Deliverables

15.1 Introduction
It is important that the correct deliverables are specified in the statement of requirement
(SoR). To enable future reuse of the data this should include all model outputs in a GIS
format, full model results, reports and run files. The details of what is required should be
explicitly stated in the SoR. Where historic flood events have been used for calibration,
results from the calibration models should also be supplied, together with the data used for
calibration. For projects involving 2D modelling it is likely to be necessary to supply a hard
drive for transfer of the data.

15.2 Reporting Requirements

Model documentation is used by decision makers and other users of models to understand
the way in which a model was developed and how a model can be used.

The model documentation should normally consist of four reports:

• Technical report
• Non-technical report or summary
• Model hand over report
• Model audit report

In addition it is recommended that technical notes on hydrology/coastal boundary conditions

and model calibration are requested at appropriate points of the project.

15.2.1 Technical Report

The technical report is generally intended for an ‘expert’ audience. The technical report
should address the modelling objectives set out in the SoR. It should provide a record of
the hydrological and hydraulic analysis including all key modelling decisions, input data,
calibration or sensitivity analysis, a commentary on model confidence and key limitations
and recommendations for future development and use of the model.

The report should include appropriate maps of the study area, cross-section locations and
plans of the model results. If a separate non-technical summary is not requested, this may
be included as a chapter within the main report.

15.2.2 Non-Technical Summary

The non-technical report or summary is generally intended for a ‘lay’ audience. It will provide
an overview of the study, the outcomes and recommendations for the use of the information
generated.

15.2.3 Model hand over report

A model hand over report should be produced providing sufficient information for an
experienced modeller to rerun the model and interrogate the results.

15.2.4 Audit report

A model audit report or technical review certificate should be supplied to provide
confirmation that the model has been assessed and that any issues identified have been
addressed. This applies for internal model audits carried out by a contractor and external
model audits carried out by either SEPA or a contractor appointed for peer review. See
section 14.3.2 for further details.

Flood Modelling Guidance for Responsible Authorities v1.1 120

15.2.5 Format
Model reports should be provided in an appropriate word processing package for use with
partner organisation IT systems. A copy of the report and all figures should also be provided
in .pdf format for sharing with other organisations. Hard copies of reports may also be
requested.

15.3 Results Files

Study outputs and results can be provided in a range of formats. Requesting output in GIS
format allows the data to be queried and manipulated after the end of the project.

The specific outputs which are to be received should be stated in the SoR. SEPA
recommends that the extent and depth, and hazard, velocity and flow direction (as
appropriate) is supplied as a minimum. This should be requested in the relevant proprietary
format the Responsible Authority uses and, if being sent to SEPA, ideally within the ESRI
shapefile/raster format. It is advised that ESRI geodatabases are not used as compatibility
issues with older versions of software may prevent data being shared easily between
Responsible Authorities.

15.3.1 Gridded output

The following gridded outputs should be requested as deliverables;

1D models: water level and depth

2D models: water level, depth, velocity (including flow direction) and hazard

These are standard outputs from hydraulic modelling packages commonly used in the UK,
and are absolute minimum required for interpreting model results.

To facilitate data sharing between Responsible Authorities, all data should be in a suitable
format for import into GIS either ESRI ascii grid format, GeoTIFF or .bil format, and -9999
should be used as no data value.

For direct rainfall (surface water) models where rainfall is applied to every point of the model
grid, large areas of the model will be covered by a shallow depth of water. In this case the
results should be requested with a 0.1 m depth minimum threshold in addition to the un-
thresholded depth results in order to remove large areas of very shallow flooding.

15.3.1.1 Depth and Elevation

Grids of both elevation and depth should be supplied.

Flood depth and elevation grids should not contain no data values at building centroids in
order for depth damage calculations to be carried out. This is discussed in section 7.3.2.1.

15.3.1.2 Hazard
For consistency with SEPA’s national flood hazard maps, hazard should be calculated using
the flood hazard formula in Defra report FD2321/TR1 Flood Risks to People (HR
Wallingford; Flood Hazard Research Centre, Middlesex University; Risk & Policy Analsysts
Ltd., 2006).

HR = d(v+1.5)+DF

Where: HR = (flood) hazard rating;

d = depth of flooding (m);
v = velocity of floodwaters (m/sec); and

Flood Modelling Guidance for Responsible Authorities v1.1 121

DF = debris factor (= 0, 1, 2 depending on probability that debris will lead to a significantly
greater hazard)

For consistency with SEPA’s flood hazard mapping and modelling the debris factor DF
should be 0.

15.3.1.3 Velocity
Grids of both speed and direction should be requested. The maximum velocity should be
‘maximum velocity’, not ‘velocity at maximum depth’. Where the models wetting and drying
is leading to high velocities for shallow depths of water, a depth threshold for tracking
maximum velocity may be applied.

15.3.1.4 Other Gridded Outputs

Several other gridded outputs may be selected for 2D modelling software. The exact range
of outputs which can be selected depends on the software. The following may be useful:

• Flow for assessing the flow split between out of bank flow paths
• Froude number for assessing if flow is sub or supercritical. Some calculation
methods are less accurate as flows become supercritical so this may affect model
confidence, or help in identifying model issues.
• Duration of flooding for emergency planning or detailed damage calculations
• Time of onset for emergency planning
• Bed shear stress for assessing the potential for erosion.

If additional outputs are required this should be discussed with the contractor prior to starting
final model runs as if additional output is not selected at this stage the model may need to be
rerun in order to generate the output.

15.3.2 Flood Extents and Area of Benefit

The Area of Benefit should be calculated as the difference between the defended and
undefended outlines at the SoP of the defence. An appropriate allowance should be made
for freeboard in calculating the SoP.

Flood extents and the Area of Benefit from a flood defence should be requested in ESRI
shapefile format, as well as any proprietary format required by the partner organisation’s GIS
format. ESRI shapefiles can be imported into most other GIS packages and this facilitates
sharing of data between Responsible Authorities.

Flood extents and Area of Benefit should not be simplified or smoothed and should match
the supplied depth grids.

Flood extents should not show bridge decks as flooded unless there is flow across the
bridge deck in the model.

15.3.3 Miscellaneous Outputs

Other outputs which may be requested include time before evacuation routes are cut, or
travel times between gauges and receptors. These outputs can be useful for emergency
planning.

Flood Modelling Guidance for Responsible Authorities v1.1 122

15.3.4 Animations
Where unsteady modelling has been undertaken, animations of the model output can be
requested. These can provide a useful tool in understanding flooding mechanisms and in
engaging with the public. Animations may be particularly effective for breach modelling.

15.3.5 Other formats

Contractors may be able to supply data in a number of other formats which may aid
interpretation of the results.

Interactive PDFs can provide an improved visual representation of flood risk and
understanding into the mechanisms of flooding. They can allow users to click on different
scenarios and pre-defined storm durations to allow the user to visualise flood risk. Interactive
PDFs can be used for various applications such as flood protection scheme designs,
displaying model run information and even combined events i.e. a fluvial and coastal flood
events.

3D Visualisations can be used to help convey complex technical information to local

communities (e.g. flood prevention scheme designs) and can be used to present different
design options and the benefits of each approach or even the complexities involved.
However, they can be time consuming and costly depending on the scale and complexity of
the visual representation and careful consideration should be made into the level of detail
required.

15.3.6 Tables of results and long sections

For 1D models, tables of water elevation and flow at each 1D node should be requested.
Long sections and cross sections of modelled water level may also be useful, to allow for the
visualisation of the water level across the reach

For 2D models it is possible to select time series output of level, depth and velocity at point
locations, and flow through cross section lines. As a minimum, the flow time series should
be extracted at all gauges. Locations where point output and flow through cross section
lines are required should be discussed with the contractor prior to final model runs. This is
because it can be difficult and less accurate to calculate these from the model output files.

15.4 Model Files

Complete sets of raw model results and run files, including check/diagnostic and mass
balance files, should be provided for each scenario run including calibration and sensitivity
runs. For 1D-2D models the mass balance output should be supplied for both the 1D and 2D
components of the model and include the flow across the link.

A description of the files which would be expected to be received for modelling software
commonly used in Scotland is given in Appendix E. SEPA can provide assistance in
checking that all the expected files have been supplied if required.

A model log file should also be provided stating which run files were used to produce each
output.

Comments, including names, of all significant structures within the model should also be
provided together with comments on any modification to structure coefficients.

Derivation calculations (e.g. spreadsheets) used for any model inputs (e.g. for boundary
conditions) must supplied. The merged DTM used in construction of the model and for
production of flood maps must be supplied. This will allow the model to be re-run and
mapped for other scenarios at a later date if required.

Flood Modelling Guidance for Responsible Authorities v1.1 123

15.5 Survey
If a survey is undertaken as part of the contract, the survey should also be included as a
deliverable. The survey formats should be as per the EA Technical Specifications.

15.6 Photographs
Any photographs taken in the course of the study (e.g. on site visits) should be supplied
along with an appropriate caption / commentary to establish what the photographs depict.
These should ideally be georeferenced to the OSGB 1936.

15.7 Other deliverables

If the Responsible Authority wishes to adopt the model and reuse in house, it may be
beneficial to specify a model hand over session. Additionally training in the use of the model
or software may be included as part of the scope.

Once the deliverables have been approved by the reviewer and approver (i.e. project
executive) signoff of the products can be undertaken.

Flood Modelling Guidance for Responsible Authorities v1.1 124

Appendices

Flood Modelling Guidance for Responsible Authorities v1.1 125

A Template Flood Study SoR

Please contact strategic.floodrisk@sepa.org.uk for a copy of the template flood study SoR in
an editable format. It is intended to include the SoR here in a later version of the guidance,
however SEPA wish to receive feedback on the usefulness and appropriateness of the
present version of the SoR prior to making it publically available.

Flood Modelling Guidance for Responsible Authorities v1.1 126

B Example Survey SoR

Please contact strategic.floodrisk@sepa.org.uk for a copy of the template survey SoR in an

editable format. Text highlighted in yellow should be amended by the Responsible Authority.
This SoR is based on Environment Agency’s Standard Technical Specifications version 3.2
(Environment Agency, 2013), and Responsible Authorities are advised to refer to the
Environment Agency Specification while developing any SoR using this template. SEPA
cannot guarantee that survey according to this SoR will be suitable for all flood studies, and
Responsible Authorities should consider their own requirements when developing a survey
SoR.

Flood Modelling Guidance for Responsible Authorities v1.1 127

 Schedule 1

Statement of Requirements

1 Background

Explain the general project background.

2 Introduction

Responsible Authority Name wishes to carry out a topographic survey of a reach of the River
XXX, specifically between XXXX and XXXX, along with associated tributaries including the
XXXX and XXXX. Cross section spacing and length should be informed by best practice and
knowledge. Survey work should be carried out by the use of GPS surveying instruments and
methods. All survey work should be carried out in line with best practice and in line with the
Royal Institution of Chartered Surveyors (RICS) guidelines for surveying. The information
gathered will be used for constructing a computer model of the watercourses and will be
used in conjunction with existing LiDAR information to produce a 1D-2D model for the area.

3 Aims and Objectives

The aim of this project is to undertake a topographic survey of a reach of the upper XXXX
catchment from XXXX to XXXX including the XXXX and the XXXX. The information gathered
will be used for constructing a computer model of the watercourses and will be used in
conjunction with existing LiDAR information to produce flood extents for the area.

The study area is outlined in the XXXX Study Location Map – see Appendix A. Costings for
the work should be produced.

4 Method of Undertaking Research – Scope of Work

The successful tenderer appointed in due course as the surveyor (the “Surveyor”) shall
provide all services required to satisfy the objectives of this study. The services will include,
but not necessarily limited to, the main task of undertaking a topographic survey.

5 Land Ownership

Prior to work commencing Responsible Authority Name will obtain the permission of each
landowner or tenant to undertake the survey. Responsible Authority Name will also provide
in writing, proof that the surveyor is working on their behalf.

6 Quality Assurance

The Surveyor shall apply quality management procedures to ensure that the information and
materials provided under this contract adhere to the Specifications and are fit for purpose in
terms of quality, completeness, standard of presentation and timely delivery.

The Surveyor shall be responsible for adopting full quality control and assurance procedures
at each stage of the work to ensure that mistakes, errors and omissions are identified and
corrected prior to the delivery of the results. The Survey shall not be considered delivered
until received in a form that complies with the specification.

Flood Modelling Guidance for Responsible Authorities v1.1 128

7 Ecological Considerations

Ecological sensitivities should be considered for the catchment and stated if required –
example below.

Consideration will be given to ecological sensitivities in the catchment such as spawning

redds and freshwater pearl mussels. Prior to work commencing, Responsible Authority
Name will make initial contact with the appropriate associations. The Surveyor shall also be
expected to make contact prior to undertaking any survey work and will be required to adopt
best working practices that will prevent disturbance.

8 Survey Specification

The following specification should be adhered to, (any deviations not agreed with
Responsible Authority Name will possibly involve additional survey work by the surveyor at
their own time and cost). Reference should be made to the Environment Agency Survey
Specifications (Environment Agency, 2013, National Standard Contract and Specification for
Surveying Services Standard Technical Specifications Version 3.2).

8.1 Reference System

• All coordinates shall be related to Grid (OSGB1936)

• All levels shall be Newlyn datum (mAOD)
• The specified measurement tolerance is +/-5cm

8.2 Survey Controls

• Permanent ground marks shall be established on firm ground as required.

• Paint must not be used for marking survey control stations and wooden pegs should
not be left protruding from the ground unless they are to be removed on the same
day as this creates a hazard. Permission of landowners should be gained before
establishing a survey control station.
• Control shall be related to the Ordnance Survey OS Net.
• The planimetric co-ordinates of directly surveyed points shall be correct to ± 0.05m
RMSE on carriageways and hard surfaces, and ± 0.10m RMSE on all other surfaces.

8.3 Definitions and Control of Works

The following definitions shall apply:

• WATERCOURSE CENTRELINE is determined from the lowest point in the bed level;
• A CROSS SECTION is normal to the watercourse centreline;
• LEFT and RIGHT are determined either side of the watercourse centreline
when viewed toward the downstream direction of the watercourse;
• SKEW ANGLES are estimated clockwise from the direction of stream;
• CHANNEL WIDTH is determined between natural river bank edges;
• HARD BED LEVEL is that to which a staff, pole or rod with a base area of 0.0005 to
0.0025 square metres can be driven to refusal;
• SOFT BED LEVEL is that to which a staff, pole or rod first meets resistance
underwater;

Flood Modelling Guidance for Responsible Authorities v1.1 129

SIGNIFICANT CHANGE IN SLOPE is deemed to be noticeable when walking the slope.

8.4 Channel Cross Sections

• Levelled cross sections are to be taken across the channel. Cross sections should be
perpendicular to the channel/flow direction and viewed downstream. As a general
guide, cross sections should be undertaken at XXX m spacing reducing to XX-XX m
and X m and to capture physical changes to the river channel respectively (Insert
location map reference). All levels shall be accurate to +/- 10 mm. Cross channel
chainage shall be accurate to +/- 100 mm and longitudinal chainage between cross
sections to be accurate to +/ -1000 mm.
• Where it is not practical to survey a section at the prescribed position or interval the
position of the section may be moved. However, the interval between two adjacent
sections shall not exceed the prescribed interval.
• Cross section levels shall be taken at straight line normal to the watercourse
centreline with all changes in slope recorded. Section survey points should be taken
at each significant change in slope and at chainages not exceeding 2 m across the
channel
• Cross-sections are to be surveyed viewed downstream. The origin (zero chainage)
must be established on the left side of the section.
• Cross sections should extend both sides of the water courses to the true land level,
extending 5 m beyond bank top where possible. Where possible, it is essential that
all sections are measured into open spaces clear of trees and dense vegetation
cover to a maximum distance of 50 m to allow tie in with LIDAR data. In those
instances where a bank top is raised above the surrounding ground (flood plain),
sections should be measured to 5 m beyond the landward toe of the crest; the crest,
defined as the line along the bank top over which water will spill form the river onto
the surrounding ground.
• Water level should be recorded at each section on the day of the survey with the date
and time recorded each day. Channel bed levels and bank levels either side are to
be recorded.
• Bed levels will be measured directly whenever and wherever possible. Where direct
measurement is impossible, where, for instance, the water depth is too great or other
causes make it impractical, then other methods to be considered include
measurement by boat or reading the depth of water against a staff and relating these
readings to a measured water level.
• Where silt occurs both the hard bed and the silt top will be measured at the same
points. The hard bed will be shown as a pecked line and labelled "H" in the digital
data. The silt top will be shown as a solid line.
• Each individual cross-section, including structure sections should be given a unique
identifier.
• The sections will be plotted to a vertical scale of 1:100 and horizontal scale of 1:200.

8.5 Flood Plain Sections

• Any flood plain sections required are denoted in insert cross section location
reference. Flood plain sections will be taken normal to the centre line of the valley

Flood Modelling Guidance for Responsible Authorities v1.1 130

 and not necessarily at right angles to the centre line of the channel. Because of this,
flood plain sections may appear ‘dog-legged’ on the key plan. These sections may be
defined on the contract mapping.

8.6 Hydraulic Structures

• Cross sections will be taken immediately upstream and downstream of each

structure and in the case of bridges ensure that dimensions of bridges openings and
flood arches are included. Structures include all those shown on the attached map
and drawings as well as any significant bridges, weirs, culverts, mill lades, major pipe
crossings and impounding structures identified in the field.
• Where structures extend beyond the top of bank, then the complete upstream
elevation will be surveyed with its cross section.

Bridges

• For bridges, the springing level, soffit level, abutments, parapets, deck level and
internal arch or flow area dimensions should be recorded and marked on the cross
section plan. The Surveyor should survey the bed level where the structure enters
the bed. Details of any bridge piers must also be included and the length of the
bridge or tunnel is to be measured parallel to the watercourse.
• The downstream elevation will be presented as viewed looking downstream and is
required to be surveyed when specifically requested or where it is different from the
upstream side. Even when a downstream elevation is not required, the downstream
soffit, top of parapet, invert, bed level and bank crests are to be measured and added
to the longitudinal section
• Where structures are skewed across the channel, the skew span will be measured
together with the appropriate skew angle and marked on the associated topographic
drawing. The length of the bridge tunnel will therefore be the channel length through
the bridge, not the distance at right angles to the roadway.
• Where a structure extends 10 m beyond the top of bank then the complete elevation
will be surveyed with its cross section. Where a bridge spans the flood plain, then all
relevant flood arches (and other openings that could take flood water) must be
included in the cross section.
• In situations where the bridge is not going to be overtopped and/or reduce
conveyance with increasing water depth, a full bridge survey is not required; bridge
parapet, soffit and springing levels can be omitted. Bridges identified for survey will
be discussed at the inception meeting.

Culverts

• Complete dimensions of the inlet and outlet elevations of culverts are to be taken
alongside the channel section as done with the bridge structure. For pipe culverts,
internal pipe diameter, invert, soffit and crown of pipe levels should be recorded
upstream and downstream. The length of the culvert should also be measured if safe
to do so. Details of any trash screen and flaps, including dimensions, number of bars,
bar width and bar spacing should be recorded and noted on the cross section plan.

Flood Modelling Guidance for Responsible Authorities v1.1 131

Weirs

• A cross section will be taken along the crest of the weir and structure details will be
taken and annotated on the associated topographic drawing. For weirs that do not
cross the watercourse in a straight line perpendicular to the watercourse, the actual
length of the weir shall b stated clearly on the cross section drawing. A long section
of the weir will be produced extending both upstream and downstream to the natural
river bed. The weir long section will have the following information:

o upstream and downstream water level;

o upstream and downstream bed level;
o weir crest;
o upstream and downstream extent of any apron;
o water and bed levels at the tail of any weir pool

8.7 Longitudinal (top of bank) survey – flood bank level

• A longitudinal survey is required along the top of both banks of all watercourses.
Levels will be taken at a minimum of 25 m (or as agreed) or where there are sudden
or pronounced changes in ground level e.g. collapsed embankment.
• Where flood defences or embankments are present, this should be taken as the top-
of-bank levels and general details on the condition of the flood defences or
embankment should be noted i.e. if there is a gap where water could escape.
• Where there is no embankment / wall the ground level should be given:

o Where the river is fenced, at the fence line;

o Where a road or path runs along the river, at the centre line;
o Elsewhere, at 5 m away from the river bank

• Where applicable (e.g. at agricultural embankments) additional longitudinal surveys

should be taken; this should include the base of the embankment/wall on both sides
(access permitting).
• Photographs of all structures and embankments surveyed will be taken and
georeferenced with a time and date.

8.8 Additional Cross Sections

• Additional cross sections should be undertaken at the following SEPA gauging

stations:
o Insert gauging station and grid reference;
• At each gauging station there is a minimum requirement for four cross sections
across the floodplain; three downstream of the ramp gauge or post and one under
the cableway/winch where available. A cross section will be taken at the downstream
control of each gauging station. A cross section will also be taken upstream of the
cableway/winch.
• Where standard cross section spacing does not cover the above then additional
cross sections should be undertaken.

Flood Modelling Guidance for Responsible Authorities v1.1 132

8.9 Data Format and Key Deliverables

• An electronic copy of the survey on CD-ROM in PDF, CSV, 2D and 3D DWG (2013
or earlier) with each surveyed point to have an X, Y and Z value.
• All data to be presented graphically on key plan/section location maps, cross
sections, structure sections and long sections should be made available as a CSV
xyz file, in DWG format and in a format compatible with Flood Modeller/HEC-RAS
software (.txt/.dat and EACSD) – see Appendix B. This cross section data will contain
the following:

o Cross section with unique identifier incorporating chainage from most

downstream end of reach;
o Level (mAOD);
o Chainage across the sections/structures working from left to right bank
(viewed downstream);
o National Grid northing;
o National Grid easting;
o Applicable survey code;
o Reference to any photographic or anecdotal evidence

• The longitudinal data will contain the following:

o The deepest bed level at each section, both hard bed (solid) and silt line
(pecked).
o The water level at each section.
o The bank crest levels derived from crest point levels shown on the cross-
sections, the left bank as a pecked line and the right bank as a bold line.
o The extent and level of any concrete sill or apron together with appropriate
label. The section number and chainage of each section and the altitudes of
each of the plotted points. The chainage shall be quoted to the nearest metre
except when the scale of the survey makes it appropriate to quote the
chainage to decimetres.
o All structure with their critical levels (soffit, invert, deck, crest etc.)
o Tributary channels should be included where surveyed
o Where changes in the levels of bank, bed or water level occur between cross-
sections, these changes are to be measured and added to the longitudinal
section. The longitudinal section should represent an accurate and complete
profile of the channel to ensure that low spot and level changes are identified.

• A GIS shapefile clearly showing the survey route, the uniquely identified cross
sections and any survey gaps.
• Digital copies of georeferenced photographs of cross section locations,
embankments and structures.
• A digital key plan based on suitable Ordnance Survey grid is to be produced showing
clearly the extent of survey. In addition, scale, a north point and sheet coordinates
are to be indicated.

Flood Modelling Guidance for Responsible Authorities v1.1 133

 • A quality statement will be provided detailing confidence in the survey and any
associated uncertainties.

9 Data

• The following data will be supplied to the successful tenderer by Responsible

Authority:
o Cross section location drawings;
o Shapefiles of the study area and the cross section locations;
o Bridge and other structures data including drawings annotating the structures
to be included in the survey.

10 Security

• The highest classification of data for this contract will be OFFICIAL: COMMERCIAL.
• From the onset of the contract all Consultant staff (or any contractor or sub-
contractor appointed by it) who have access to Responsible Authority data must as a
minimum be fully compliant with the requirements of the Baseline Personal Security
Standard (BPSS).
• Responsible Authority require confirmation of the office location(s) from where the
work will be undertaken both by the Consultant and Sub Contractors for this contract.
• At tender stage details of how project data will be accessed, stored, transmitted and
handled within your organisation is required. This should include both electronic and
hard copy data and meet the Cabinet Office Security Policy Framework requirements
as a minimum.

11 Meetings

An allowance should be made for an inception meeting. Thereafter, contact will be made
primarily over email or telephone to discuss progress or any issues that have arisen which
may lead to a delay in the delivery date.

12 Delivery Timescales

The tender return should include a programme of work that takes the following milestones
and key dates into consideration:

Task
Main tasks Date
No.

Flood Modelling Guidance for Responsible Authorities v1.1 134

13 Budget

The projected cost of the contract is expected to be within the range of £XXXX to £XXXX
including VAT.
• Fixed Price
The surveyor shall submit a fixed price. Each price shall be deemed to include, inter
alia, the following:-
o All travel and subsistence costs.
o All media and consumable costs.
o Field work and data processing.
o Traffic management and maintenance.
o Liaison with the landowners / tenants for access.
o Tender preparation costs.
o Weather downtime.
o All post, telephone, fax and e-mail costs.
o Controls established to OS GPS Network.
o Where reflector less total station is adopted the surveyor shall include for a
detailed visual inspection of the site to ensure all features (e.g. manholes,
gullies etc) are included. Any visual inspection and additional survey work
shall be deemed to adopt a safe system of work as noted above.
• The tender price should be made up based on the form of data being collected as
follows:
o Channel survey
o Topographical survey

14 Project Management

The appointed contractor will be responsible for taking and distribution of minutes and
agenda for all meetings and telecom’s.

Any compensation claim due to change in scope needs to be provided and agreed in writing
before commencing work.

Requests for changes to key project staff must be provided in writing for approval.

15 Intellectual Property Rights

All copyright and Intellectual Property Rights (IPR) will be transferred to Responsible
Authority Name in accordance with the Terms and Conditions.

16 Sustainability

Responsible Authority Name is committed to working in a sustainable manner. For example,

public transport should be used whenever possible for meetings, waste should be kept at a
minimum, and a recycling policy should be in use. Paper used as part of the project should
be from a variety of sources, including recycled.

17 Tender Submission

Tender submissions should include the following:

• Methodology statement

Flood Modelling Guidance for Responsible Authorities v1.1 135

 • Project outputs
• Key staff.
• Costing
• Project programme including timetable taking into consideration the key dates
detailed in Section 11.

All tender submissions should cover the above requirements in a maximum of 10 pages.

Schedule 2

Tender Evaluation

Tenders will be evaluated using the following criteria and weightings.

Technical Criteria (overall weighting 80%)

Financial Criteria (overall weighting 20%)

Schedule 3

Price Summary - Template

Tenderers are required to submit a firm price for the service detailed in Schedule 1 excluding
VAT. All costs appropriate to the proposal must be included or summarised here. Costs
which appear elsewhere in the proposal but which are not summarised here will be
presumed to have been waived.

Price £
Activity Person Hours (excl. VAT)

Total Price (excl. VAT) £

Tenderers must also provide a breakdown of the staff involved in this contract.

Hours
Personnel Activity Hourly Rate
Input

£
£
£
£
£
£

Flood Modelling Guidance for Responsible Authorities v1.1 136

 £
£
£
£

Flood Modelling Guidance for Responsible Authorities v1.1 137

C Quality Control

C.1 Example technical review certificate

Flood Modelling Guidance for Responsible Authorities v1.1 138

Flood Modelling Guidance for Responsible Authorities v1.1 139
Flood Modelling Guidance for Responsible Authorities v1.1 140
Flood Modelling Guidance for Responsible Authorities v1.1 141
Flood Modelling Guidance for Responsible Authorities v1.1 142
Flood Modelling Guidance for Responsible Authorities v1.1 143
Flood Modelling Guidance for Responsible Authorities v1.1 144
Flood Modelling Guidance for Responsible Authorities v1.1 145
C.2 Fluvial Flood Estimate Review

Key

Action required Potential to have a significant impact on study

outcomes.
Action maybe required Potential to affect study outcomes. Amendments
may be required. Otherwise further explanation is
required in the report.
No Action required Negligible impact on the study outcomes.

1. Project Details

Project Name
Report(s) being reviewed
Date and version of report
Author and Company
Reviewer
Date of review
Review Status

2. Concept Review

Item Check Comments

Is their evidence that a concept review was undertaken at the

start of the project?
Did the preliminary methodology get sign off from SEPA and
an internal principle hydrologist? Reviewers? Date?
Does the report adequately define the hydrology
methodology, and provide details of updates required
throughout the life cycle of the project?
Does the report capture the key catchment processes and
flood mechanisms?
What is the chosen methodology, FEH RR, FEH Statistical,
Hybrid, REFH2, Other.

3. Review of Data

Item Check Comments

Has a report of the available data been undertaken and is

outlined in sufficient detail in the report?
Does the review include both flow and level stations? Current
and closed stations?
Have stations outside the Hi-Flows-UK dataset been
considered?

Flood Modelling Guidance for Responsible Authorities v1.1 146

Has the flood peak data in Hi-Flows been reviewed against
the SEPA raw data, and extended to present day where
necessary?
Has the data been assigned a quality rating or has an opinion
of the robustness of the data been provided?
Where necessary, have rating reviews been undertaken?
What was the outcome? Where rating reviews deviate from
SEPA rating curves, have the differences been explained,
and has the new rating curve been approved by SEPA?
Does the review include rain gauge data?
Has the rain gauge data review looked at data from sources
outside SEPA? i.e Met Office?
Has RADAR data been considered for calibration events, and
compared against SEPA/Met Office observed data.
Is the review supported by plots, i.e
hydrographs/hyetographs or time series?

4. Calibration Events

Item Check Comments

Have calibration events been selected? How many? Dates?

Has the quality of the data from all sources over the period of
the calibration events been reviewed?
Is flow/level and rain gauge data available for the chosen
events, and has a quality rating been applied to the data?
Has RADAR data been considered for calibration events? If
so has this been compared against SEPA/Met Office
observed data.
Has MORECS data been obtained, reviewed and used to
calibrate antecedent conditions?

5. Boundaries & Reconciliation Points.

Item Check Comments

Is there evidence that a catchment schematisation exercise

has been undertaken?

Have all major and minor inflow locations been represented

and detailed in the report, and is the schematisation
appropriate?
Have flow estimation points for the purpose of flow
reconciliation been outlined in the report?
Do the locations of the flow estimation points provide a
robust hydrological assessment, i.e are located at the
locations of interest (e.g. confluences, areas of known flood
risk, gauging stations, model boundaries)

Where tributaries have not been represented as a discrete

boundary inflow, has this been detailed and adequately
explained?
Have lateral inflows been included, if so, have they been

Flood Modelling Guidance for Responsible Authorities v1.1 147

adequately explained and justified?

6. Catchment Descriptors

Item Check Comments

Has a review of the catchment been undertaken, and the

presence of significant land use or catchment factors been
described?
Have catchment descriptors been extracted for flood
estimation, and, if so are they clearly presented?
Is yes, have the parameters been sense checked, and details
provided of any changes? – Area, Urbext etc
Where changes have been made, have these been properly
documented and justified?

7. Estimation of QMED

Item Check Comments

How has QMED been derived? i.e using observed data –

AMAX/POT or by Catchment Descriptors. Has the method
been justified?
Has a climatic variation adjustment been applied?

Have confidence intervals been presented?

Has the revised QMED equation (CEH, 2008) been used

where QMED is estimated from catchment descriptors?
Does the calculation record state whether URBEXT1990 or
URBEXT2000 has been used?
Have donor sites been used?
Is there an adequate audit trail, documenting how donor sites
have been chosen? i.e. have similarity of catchment
descriptors been documented, quality of gauge, as well as
distance from catchment centroids.
Has the revised method of data transfer (CEH, 2008) been
used in the selection of donor sites and calculation of
adjustment factors for QMED?
Is the assessment considered robust?

8. FEH Statistical Method – Estimation of Growth Curves

Item Check Comments

What methodology has been applied? i.e Single Site, Pooling

Group, Enhanced Single Site for pooling groups at rural
gauges?
Is the chosen methodology adequately justified in the context
of catchment dynamics, data available?
If pooling group has been used, have multiple pooling groups
been used for the different catchments, or a single for the
entire catchment?

Flood Modelling Guidance for Responsible Authorities v1.1 148

Has the composition of the pooling group(s) been recorded?
Has an audit trail of decision making for pooling groups been
presented? Reasons for exclusion explained?
What distribution has been applied? GL, GEV, Pearsons
Type 3?

9. Rainfall-runoff approaches

Item Check Comments

What technique has been used? Hybrid, i.e FEH Statistical

with FEH RR, or REFH2. Has the chosen approach been
justified with reference to key catchment characteristics?
If the FEH RR method has been used, have the parameters
Tp(0) and SPR been adjusted using donor catchments if
available?
If a flood event or lag analysis has been carried out for
estimating Tp(0) or SPR, is a list of the events and
description of the results provided?
Has the derived FEH RR critical duration been reviewed?
And compared against observed data if available?
If ReFH2 has been applied at or near a flow gauging station,
have model parameters been estimated from flow and rainfall
data?
If ReFH2 parameters have been estimated from observed
data, is a list of the events and description of the results
provided?
If REFH2 has been used is it documented whether the alpha
factor has been applied? SEPA guidance recommends that
the factor should not be invoked for Scottish catchments.
Does the calculation record state what storm durations were
used for the rainfall-runoff calculations?
Are different critical storm durations appropriate for different
parts of the catchment, and have they been applied?
Does the report consider the hydrograph shape and
volumes? i.e Archer Approach.

10. Small Catchment methods – IOH 124, ADAS

Item Check Comments

Has a non FEH method for small catchment hydrology been

applied?
If, yes, which method has been used? And has the
application of the method been justified with reference to
catchment characteristics?

11. Results

Item Check Comments

Flood Modelling Guidance for Responsible Authorities v1.1 149

Have derived QMED values been compared to gauged
records upstream/downstream to ensure consistency with
observed data?
Have design flows been considered at a catchment scale,
and do flows increase downstream, if not are the reasons
adequately explained, i.e flood plain storage?
Have design growth curves for the range of return periods
assessed been presented and plotted for multiple
methodologies in order to validate the chosen approach?
Have design flows been checked against observed flood
peaks and the theoretical return period sense checked?
If there are any other studies in the catchment, has the report
presented these and discussed similarities/differences?
Are confidence limits presented for the chosen design flows?
Have limitations of the approach been outlined and
recommendations presented for future improvements?
Is the final approach clearly defined, and the final design
flows clearly outlined in tabular and graphical format against
where possible observed AMAX data?
Does the report clearly outline how the principle flows are
input into the hydrodynamic model and how intervening
catchments are to be included?

12. Conclusion

Item Check Comments

Is the hydrology methodology clearly defined, such that it can

be reproduced in the future?
Is the methodology appropriate and proportionate to the
study in question?
Has all the available data been used in order to inform the
approach?
Has the approach been developed in conjunction with SEPA
staff and got company sign off?
In your professional opinion is the chosen approach
acceptable for the study in question?
Any additional comments/changes required, over and above
comments already included in the review?

Flood Modelling Guidance for Responsible Authorities v1.1 150

D Report Template

A report template will be provided here at a later date.

Flood Modelling Guidance for Responsible Authorities v1.1 151

E Model Deliverables

E.1 General
Model file paths should be kept below 50 characters where possible, whilst ensuring
meaning or logical structure is not lost.

A logical and descriptive naming structure for models and scenarios should be adopted. File
and scenario names should include the following information where appropriate.

• River reach identifier e.g. Tay. For long rivers abbreviate the name.
• Version
• Return period
• Storm duration
• Scenario identifiers – D for defended, ND for undefended, S_N for sensitivity to
roughness, S_Q for sensitivity to flow. Climate change scenarios to be labelled with
the scenario run e.g. 2080H.

Eg. Tay_V1_10yrs_10hrs_D.ied

E.2 InfoWorks(CS and ICM)

A compact transportable database (.iwc) and migration file (.cs2icm) should be for Infoworks
CS models. For ICM models a transportable database (.icmt) should be supplied. These
transportable databases should include all the information which was used to run the model.
This will include, but may not be limited to, the following information:
• The model network (s)
• Inflow files
• Rainfall files
• Trade/waste flows
• Initial conditions
• Level boundaries
• Run files
• Ground model
• Results files (unless prohibitively large)

A description should be provided to accompany any scenarios used in ICM modelling.

E.3 MIKE Flood

Please contact SEPA for details.

E.4 HECRAS

File Type File Description

extensions
to contain
Model .prj Project file. Contains title of project, unit system, list of
files associated with project and list of default
variables.
.p01 Plan files. Each plan will represent a different specific

Flood Modelling Guidance for Responsible Authorities v1.1 152

 set of geometric and flow data.
.g01 Geometric data. Geo-referenced cross sections and
information on structures.
.f01 Steady flow data (if used). This will contain inflow
locations, values and reach boundary conditions.
.u01 Unsteady flow data (if used). This will contain inflow
locations, hydrographs and reach boundary conditions.
.q01 Quasi-steady flow data (if used). This will contain
inflow locations, hydrographs and reach boundary
conditions.
Results .log Log file for the project.
.b01 Boundary condition file used in unsteady flow
simulations.
.bco Unsteady flow log output file.
.ic.001 Initial condition file used for each unsteady flow plan.
Runs .r01 Model run file for steady state simulations.
.x01 Model run file for unsteady state simulations.
.o01 Output file for each plan.

E.5 FloodModeller 14 1D

File Type File Description

extensions
to contain
Boundaries .ied Hydrological boundary conditions or operating rules for
structures..ied files should be used running multiple
design events through the same model rather than
importing boundary conditions to the .dat file.
Model .dat Model data file. There should be a single .dat file for
each different model geometry used.
.gxy Georeferenced model schematic.
.ixy Diagrammatic model schematic. Not compulsory,
largely superseded by .gxy file.
.iic or .zzs Initial condition files. These may not exists as it is
good practice to include initial conditions in the.dat file
unless multiple initial conditions are being run with the
same geometry. The .ief files state if a separate initial
conditions file has been used. These must be supplied
if used.
Results .zzs Model results file for steady state simulations
.zzn Model results files for unsteady simulations. Note that
.zzl both the .zzn and .zzl are required in order to open the
files.
.zzd Diagnostics file. Containing error messages and

14
Formerly ISIS.

Flood Modelling Guidance for Responsible Authorities v1.1 153

 warnings.
.zzx Supplementary results file. Must be included for
FloodModeller 1D-2D or FloodModeller-Tuflow
simulations.
Runs .ief Model run parameter file, one per scenario for design
runs.
GIS GIS format. For models containing non georeferenced cross
sections, spills extracted from a DTM, or reservoir
units shapefiles showing the extent and location of
these should be supplied. These are not necessary for
running the model but are necessary for flood mapping
and auditing purposes. Any GIS files used in model
construction must be supplied.

E.6 FloodModeller 2D

Folder/File File Description

extensions

Runs .xml Model run files. There should be one of these per
scenario run.
GIS .shp .asc There should be a GIS folder containing all model GIS
etc. inputs. The exact files required will be specified in the
.xml file. The same GIS folder and files should be
referenced by multiple scenarios
Boundaries .ied Hydrological boundary conditions if these are not
contained in the .xml file.
Results .asc, .dat, The results folder name matches the .xml file name
and check .sup, .2Dm and is created in the Run directory. Check files have
files the extension .chk*.asc and should be provided for all
runs to enable checking. Model results are in .dat, .sup
and .2Dm format. There is one .dat file for each output
and one .sup and .2Dm file for each model run. All 3
components of the results files are required.

Flood Modelling Guidance for Responsible Authorities v1.1 154

E.7 TUFLOW

Folder File Description

extensions
bc_dbase .csv, .xls Boundary condition data.
check Model check files.
model .tbc Boundary conditions control file
.tgc Geometry control file.
.tmf Sets model roughness.
mi mi or shp. Model GIS files. All files must be
included.
results Model results and log files.
runs .tcf, .ecf Simulation run files

E.8 FloodModeller-TUFLOW
Files are to be included as in the ISIS1D and TUFLOW descriptions above.

E.9 FloodModeller 1D-2D

Files are to be included as in the ISIS1D and ISIS2D descriptions above.

Flood Modelling Guidance for Responsible Authorities v1.1 155

For all coupled ISIS-TUFLOW models the volume output options should be selected in the
additional output tabs. The 1D volume output and save interval should be the same as the
volume output interval in the 2D model.

Flood Modelling Guidance for Responsible Authorities v1.1 156

F Model Node Naming Structure

For models which require names to be entered for model cross sections the industry
standard [XX][CHAINAGE][CHAR] node naming strategy should be adopted where:

XX is the river identifier. This is usually some abbreviation of the river name e.g. Tay or F for
the River Tay or River Forth. Use a separate identifier for each tributary in the model.

CHAINAGE is the chainage. The chainage should be 0 at the d/s end of a river, or at a
confluence. Chainage should be measured along the centerline of the river from the tidal
limit.

CHAR is an optional additional descriptor that can contain letter or numbers. This is usually
used for structures such as bridges e.g. BrUp – upstream bridge node, Wr1Dn –
downstream node of weir one. Try to be consistent within a model, but there are no hard
and fast rules as some software restricts node name length, typical names are given in
Table 15-1, and an example of node naming around a bridge in an ISIS 1D model is given in
Figure 15-1.

Tributary inflows should be named after the tributary e.g. Pow for the Pow Burn or Devon for
the River Devon.

Table 15-1: Typical abbreviations added to Isis node labels.

Typical Abbreviation Structure/Item
Wr,W Weir
Sp,S Spill
Br,B Bridge
Cu,C Culvert
Up,U,u Upstream
Dn,D,d Downstream

Flood Modelling Guidance for Responsible Authorities v1.1 157

Figure 15-1: Example ISIS node labels around a bridge with a spill at chainage 50000 m.

Flood Modelling Guidance for Responsible Authorities v1.1 158

G Glossary
Term Description
AMAX Series A series of the largest event in any given year
The probability that an event of the same or greater
Annual exceedance
magnitude will occur in any one year. This is the reciprocal
probability
of the return period.
Antecedent Conditions
The wetness of a catchment prior to a flood event.

An area which has benefited from a flood defence and is

Area of benefit now at a reduced risk of flooding relative to the scheme’s
standard of protection.
Average Recurrence The average period between events of a same or greater
Interval magnitude.
The effect on water level upstream of a structure or
constriction in flow where the depth is raised above the
Backwater Effect normal depth for the flow. The backwater length is the
distance upstream of the constriction or structure before
normal depth is re-established.

Catchment All the land drained by a river and its tributaries.

The process of adjusting model parameters to make a

model fit with measured conditions (e.g. measured flows).
Calibration
This process should be followed by validation using a
different set of data to that used in the calibration.
Conceptual models are simple qualitative descriptions of a
Conceptual model system as a chain of concepts or processes, which are
used to help understand how the system works.
An estimated range of values which is likely to include the
value of an unknown parameter (e.g. the 0.5% AEP design
Confidence interval
flow). The width of the confidence interval indicates how
certain the value of the unknown parameter is.
A flood event of a given annual exceedance probability
Design event against which the suitability of any proposed development
and mitigation measures are assessed.
An artificial raising of the natural bank height of a water
Embankment
body.
Flood Risk Assessments are detailed studies of an area
where flood risk may be present. These are often used to
Flood Risk
inform planning decisions, develop flood schemes and they
Assessment (FRA)
also contributed to the National Flood Risk Assessment.
They detail site specific flood risk.
Legislation which transposes the EC Floods Directive into
Flood Risk
Scots Law and aims to reduce the adverse consequences
Management
of flooding on communities, the environment, cultural
(Scotland) Act 2009
heritage and economic activity.
A term used in the FRM Act. Flood Risk
Flood Risk Management Plans set out the actions that will
Management be taken to reduce flood risk in a Local Plan
Plan District. They comprise Flood Risk Management
Strategies, developed by SEPA, and Local

Flood Modelling Guidance for Responsible Authorities v1.1 159

 Flood Risk Management Plans produced by
lead local authorities.
Sets out a long-term vision for the overall reduction of flood
Flood Risk risk. They will contain a summary of flood risk in each Local
Management Plan District, together with information on catchment
Strategy characteristics and a summary of objectives and measures
for Potentially Vulnerable Areas.
Area of land that borders a watercourse, an estuary or the
sea, over which water flows in time of flood, or would flow
Floodplain
but for the presence of flood defences and other structures
where they exist.
Fluvial flooding Flooding from a river or other watercourse.
A dimensionless parameter which represents the ratio
Froude Number
between inertial and gravity forces in a fluid.
Locations within a hydraulic model where the peak flow is
constrained to match a hydrological estimate. These are
Reconciliation Points
typically located at gauging stations or at key receptors
such as flood defences.
A hazard is a source of potential damage or harm. In terms
Hazard of the FRM Act, hazard refers to the characteristics (extent,
depth, velocity) of a flood.
A function of depth, velocity and a debris factor used to
Hazard rating
assess the risk to people from flooding.
Generic software program , which can be used for different
Software (model code)
study areas without modifying the source code
Local Flood Risk Management Plans, produced by lead
local authorities, will take forward the objectives and actions
Local Flood Risk set out in Flood Risk Management Strategies. They will
Management Plans provide detail on the funding, timeline of delivery,
arrangements and co-ordination of actions at the local level
during each 6 year planning cycle.
Person who applies of a software to a particular study area,
including input data and parameter values
Modeller
1. the developer of a model
2. someone working with a model
Modelling Making a model or working with a model
Site application of a software to a particular study area,
including input data and parameter values;
Note: ‘model’ is often referred to as a computer program (a
Model
model program) with corresponding input. However, the
word ‘model’ may also refer to some notes on paper, a
mathematical model, a diagram or a figure
Natural flood A set of flood management techniques that aim to work with
management natural processes (or nature) to manage flood risk.
Flooding that results from rainfall runoff flowing or ponding
over the ground before it enters a natural (e.g. watercourse)
or artificial (e.g. sewer) drainage system or when it cannot
Pluvial flooding
enter a drainage system (e.g. because the system is
already full to capacity or the drainage inlets have a limited
capacity).
Peak Over Threshold All events over a given threshold
Series (POT)
Potentially Vulnerable Areas based on interconfluence catchments that contain

Flood Modelling Guidance for Responsible Authorities v1.1 160

Areas significant flood risks, sufficient to justify further assessment
and appraisal of flood management actions. The NFRA has
identified 243 of these for Scotland.
Designated in the FRM Act as Local Authorities, Scottish
Water and from 21 December 2013 the National Park
Authorities and Forestry Commission Scotland.
Responsible Authority
Responsible authorities, along with SEPA and Scottish
Ministers, have specific duties in relation to their flood risk
related functions.

Flood Modelling Guidance for Responsible Authorities v1.1 161

 The average interval between years containing an event of
Return Period
the same or greater magnitude.
A measure of the combination of the likelihood of flooding
occurring and the associated impacts on people, the
Risk
economy and the environment. For a hazard to become a
risk there have to be receptors.
Sensitivity testing involves varying an element of the
modelling and assessing how this alters the model results.
Sensitivity Analysis
This helps develop an understanding of the confidence in
the model and its outputs.
The lowest probability flood event that a defence will
withstand to a high degree of confidence throughout its
Standard of Protection design life, This allows for uncertainty in the assessment
and physical processes such as settlement. This is not the
same as the threshold of flooding.
A hydraulic model in which the flow at any point in the
model is constant with time. This type of model cannot
Steady–state model
estimate the effects of storage on flood levels or
downstream flows
In coastal studies, the water level due to a combination of
astronomical tide and surge. Still water levels and waves
Still Water Level are often treated separately, however waves may increase
still water levels at the coast due to a process called wave
setup.
Sub critical flow Flow for which the Froude number is less than 1.
Supercritical flow Flow for which the Froude number is greater than 1.
The most probable flood event at which a defence will be
Threshold of
overtopped. This is not the same as the standard of
Flooding/Overtopping
protection.
Upstream/Downstream The limits of the model assessment upstream and
boundary downstream of the site of interest
Velocity The speed and direction that the water travels.
The process of checking a numerical solution generated by
the software against one or more analytical solutions or
other numerical solutions to determine its accuracy.
Verification
Verification ensures that the computer programme
accurately solves the equations that constitute the
mathematical model. The software can be verified.
The process of demonstrating that a given site-specific
model is capable of making accurate predictions for periods
outside a calibration period. A model is said to be validated
Validation
if it accuracy and predictive capability in the validation
period have been proven to be within acceptable limits or
errors.
Hydrological analysis in the UK is typically based on water
years. UK water years are defined as 1st October to 30th
Water Year
September. The 2015 water year starts on 1st October
2015.

H Bibliography

Apel, H., Aronica, G., Kreibich, H., & Thieken, A. (2009). Flood risk analyses – how detailed
do we need to be? . Natural Hazards, 79-98.

Flood Modelling Guidance for Responsible Authorities v1.1 162

Apel, H., Thieken, A. H., Merz, B., & Blöschl, G. (2004). Flood risk assessment and
associated uncertainty. Natural Hazards and Earth System Sciences(4), 295-308.
Archer, D., Foster, M., Faulkner, D., & Mawsdley, J. (2000). The synthesis of design flood
hydrographs. Flooding Risks and Reactions. Proceedings of the Water Environment
2000 Conference (pp. 45-57). London: Institution of Civil Engineers.
Bales, J. D., & Wagner, C. R. (2009). Sources of uncertainty in flood inundation maps.
Journal of Flood Risk Management, 2, 139-147.
Balkham, M., Fosbeary, C., Kitchen, A., & Rickard, C. (2010). CIRIA C689, Culvert design
and operation guide. London: CIRIA.
Bayliss, A. C., & Reed, D. W. (2001). The use of historical data in flood frequency
estimation. Wallingford: Centre for Ecology and Hydrology.
Bennett, J., Blenkinsop, S., Dale, M., Fowler, H., & Gill, E. (2015). Rainfall Intensity for
Sewer Design - Guidance for water companies. London: UK Water Industry
Research Limited 2015.
Beven, K. (2011). FRMRC: Framework for Addressing Uncertainty in Fluvial Flood Risk
Mapping. FRMRC Research Report SWP1.7. Flood Risk Management Research
Consortium. Retrieved from
http://frmrc.hw.ac.uk/downloads/Framework_for_Assessing_Uncertainty_Fluvial_Floo
d_Risk_Mapping_WP1_7.pdf
Brunner, G. (2010). HEC-RAS River Analysis System User's Manual Version 4.1. Davis, CA:
US Army Corps of Engineers.
Castellarin, A., Di Baldassarre, G., Bates, P. D., & Brath, A. (2009). Optimal Cross-Sectional
Spacing in Preissmann Scheme 1D Hydrodynamic Models. Journal of Hydraulic
Engineering, 135(2), 96-105. doi: http://dx.doi.org/10.1061/(ASCE)0733-
9429(2009)135:2(96)
Chow, V. (1959). Open CHannel Hydraulics. New Jersey: Blackburn Press.
Climate Change (Scotland) Act. (2009). Retrieved from
http://www.legislation.gov.uk/asp/2009/12/contents
Crowder, R., Pepper, A., Whitlow, C., Sleigh, A., Wright, N., & Tomlin, C. (2004).
Benchmarking of Hydraulic River Modelling Software Packages R&D Technical
Report W5-105/TR0. Bristol: Environment Agency.
DEFRA. (2006). Flood and Coastal Defence Appraisal Guidance FCDPAG3 Economic
Appraisal Supplementary Note to Operating Authorities - Climate Change Impacts.
Environment Agency. (2013). National Standard Contract and Specification For Surveying
Services Standard Technical Specifications version: 3.2. Bristol: Environment
Agency. Retrieved from http://www.eacsd.co.uk/Docs/Nat_Specs_V3.2.pdf
Faulkner, D. (1999). Flood Estimation Handbook, Volume 2: Rainfall frequency estimation.
Wallingford: Institue of Hydrology.
Fisher, K., & Dawson, H. (2003). Reducing Uncertainty in River Flood Conveyance
Roughness Review Project W5A-057. Lincoln: Defra.
Gouldby, B., Méndez, F. J., Guanche, Y., Rueda, A., & Mínguez, R. (2014). A methodology
for deriving extreme nearshore sea conditions for structural design and flood risk
analysis. Coastal Engineering, 88, 15-26.
Halcrow. (2011). Selkirk Flood Protection Scheme - Preferred Scheme Report. Scottish
Borders Council.
Hawkes, P. (2005). Use of Joint Probability Methods in Flood Management A Guide to Best
Practice R&D Technical Report FD2308/TR2. London: Defra.
Hawkes, P., & Svensonn, C. (2005). Joint Probability: Dependence Mapping and Best
Practice: Techncal report on dependence mapping. R&D Technical Report
FD2308/TR1. London: Defra.
Hedges, T., & Reis, M. (1998). Random wave overtopping of simple seawalls: a new
regression model. Water, Maritime and Energy Journal, Proc. ICE, 130(1), 1-10.
HR Wallingford. (2004). Best Practice in Coastal Flood Forecasting. London: Defra.

Flood Modelling Guidance for Responsible Authorities v1.1 163

HR Wallingford; Flood Hazard Research Centre, Middlesex University; Risk & Policy
Analsysts Ltd. (2006). Flood Risks to People Phase 2: FD2321/TR1 The Flood Risks
to People Methodology. London: Defra.
Institute of Hydrology. (1975). Flood Studies Report (5 volumes). Wallingford: Institute of
Hydrology.
Institute of Hydrology. (1999). Flood Estimation Handbook. 5 Volumes. Wallingford: Institute
of Hydrology.
IPCC. (2013). Summary for Policymakers. In T. Stocker, D. Qin, G.-K. Plattner, M. Tigno, S.
Allen, J. Boschung, . . . P. Midgley (Eds.), Climate Change 2013: The Physical
Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New
York, NY, USA: Cambridge University Press.
JBA Consulting. (2014). Derivation of a Regional Pluvial Flood Hazard Dataset, Scotland -
Methodology Report. SEPA.
Kay, A., Crooks, S., Davies, H., & Reynard, N. (2011). An assessment of the vulnerability of
Scotland's river catchments and coasts to the impacts of climate change.
Wallingford: Centre for Ecology and Hydrology.
Kjeldsen, T., Jones, D., Bayliss, A., Spencer, P., Surendran, S., Stefan, L., . . . McDonald, D.
(2003). Improving the FEH statistical method. Flood & Coastal Management
Conference 2008. Manchester: Environment Agency/Defra.
Lowe, J. A., Howard, T. P., Pardaens, A., Tinker, J., Holt, J., & Wakelin, S. (2009). UK
Climate Projections science report: Marine and coastal projections. Exeter, UK: Met
Office Hadley Centre.
Maier, H. R., & Ascough, J. C. (2006). Uncertainty in environmental decision-making: issues,
challenges and future directions. Proceedings of the iEMSs Third Biennial Meeting:
Summit on Environmental Modelling and Software. Burlington, USA.
Mason, D. C., Schumann, G., & Bates, P. (2001). Data utilization in flood inundation
modelling. In G. Pender, & H. Faulkner (Eds.), Flood risk science and management
(pp. 211-233). Oxford: Wiley-Blackwell. doi:10.1002/9781444324846.ch11
McMillan, A., Batstone, C., Worth, D., Tawn, J., Horsburgh, K., & Lawless, M. (2011).
Coastal flood boundary conditions for UK mainland and islands Project:
SC060064/TR2: Design sea levels. Bristol: Environment Agency.
Mouchel. (2011). Whitesands Flood Protection Scheme - Hydraulic Modelling Report.
Dumfries: Dumfries and Galloway Council.
Neelz, S., & Pender, G. (2013). Benchmarking the latest generation of 2D hydraulic
modelling packages. Bristol: Environment Agency.
O’Donnell, G. M., O’Connell, P. E., & Quinn, P. F. (2004). R&D Technical Report FD2114/TR
Review of impacts of rural land use and management on flood generation, Appendix
B: data analysis and modelling at the catchment scale. London: Defra.
Penning-Rowsell, E., Priest, S., Parker, D., Morris, J., Tunstall, S., Viavattene, C., . . . Owen,
D. (2013). Flood and Coastal Erosion Risk Management - A Manual for Economic
Appraisal. Abingdon: Routeledge.
Pullen, T., Allsop, N., Bruce, T., Kortenhuase, A., Schüttrumpf, H., & van der Meer, J.
(2007). EurOtop Wave Overtopping of Sea Defences and Related Structures:
Assessment Manual.
Royal Haskoning DHV and JBA. (2013). Derivation of a National Coastal Flood Hazard
Dataset. SEPA.
Royal HaskoningDHV. (2013). Eymouth Overtopping and Flood Study. Scottish Borders
Council.
Samuels, P. G. (2004). Afflux at bridges and culverts, Review of current knowledge and
practice, Annex 2: Hydraulic Model Implementation of Bridge and Culvert Afflux and
Blockage, R&D Project Record W5A-061/PR2. Bristol: Environment Agency.
Retrieved from http://www.river-conveyance.net/aes/documents/phase1/7_W5A-
061%20-%20PR2%20-%20Annex%202%20-%20Samuels.pdf

Flood Modelling Guidance for Responsible Authorities v1.1 164

Sayers, P. B., Gouldby, B. P., Simm, J., Meadowcroft, I., & Hall, J. (2003). Risk,
Performance and Uncertainty in Flood and Coastal Defence – A Review. R&D
Technical Report FD2302/TR1. London: DEFRA.
Scottish Environment Protection Agency. (2016). Natural Flood Management Handbook.
Stirling: Scottish Environment Protection Agency.
Scottish Government. (2011). Flood Risk Management (Scotland) Act 2009: Delivering
Sustainable Flood Risk management. Edinburgh: Scottish Government.
SEPA. (2015). Technical Flood Risk Guidance for Stakeholders (Refernece:SS-NFR-P-002).
Retrieved from http://www.sepa.org.uk/media/162602/ss-nfr-p-002-technical-flood-
risk-guidance-for-stakeholders.pdf
Stewart, E. J., Jones, D. A., Svensson, C., Morris, D., Dempsey, P., Dent, J., . . . Anderson,
C. (2010). Reservoir Safety – Long Return Period Rainfall. London: DEFRA.
Stokes, C., Masselink, G., & Conley, D. (2016). Investigating Opportunities for Improving
SEPA’s National Coastal Flood Hazard Dataset. Final report CRW2016_09. CREW.
Retrieved from crew.ac.uk/publications
Syme, W. J. (2001). Modelling of Bends and Hydraulic Structures in a Two-Dimensional
Scheme. Conference on Hydraulics in Civil Engineering. Hobart: The Institution of
Engineers, Australia. Retrieved from
http://tuflow.com/Download/Publications/Modelling%20of%20Bends%20and%20Hydr
aulic%20Structures%20in%20a%202D%20Scheme,%20Syme,%202001.pdf
Syme, W. J. (2008). Flooding in Urban Areas - 2D Modelling Approaches for Buildings and
Fences. 9th National Conference on Hydraulics in Water Engineering. Darwin:
Engineers Australia. Retrieved from
http://www.tuflow.com/Downloads/Publications/2008.09%20-
%202D%20Modelling%20Approaches%20for%20Buildings%20and%20Fences.Sym
e.pdf
Walker, W., Harremoës, P., Romans, J., van der Sluus, J., van Asselt, M., Janssen, P., &
Krauss, M. K. (2003). Defining uncertainty. A conceptual basis for uncertainty
management in model-based decision support. Integrated Assessment, 4, 5-17.
WaPUG. (2009). INTEGRATED URBAN DRAINAGE MODELLING GUIDE. London:
CIWEM. Retrieved from http://www.ciwem.org/knowledge-networks/groups/urban-
drainage/publications/modelling-guides.aspx
Woolf, D., & Wolf, J. (2013). Impacts of climate change on storms and waves. MCCIP
Science Review, 20-26. doi:doi:10.14465/2013.arc03.020-026

Flood Modelling Guidance for Responsible Authorities v1.1 165