NN Overtopping & XGB Over-

topping

User & Technical Manual

Version: 1.0.0.0
SVN Revision: 00

7 February 2023
NN Overtopping & XGB Overtopping, User & Technical Manual

Published and printed by:

Deltares telephone: +31 88 335 82 73
Boussinesqweg 1 fax: +31 88 335 85 82
2629 HV Delft e-mail: info@deltares.nl
P.O. 177 www: https://www.deltares.nl
2600 MH Delft
The Netherlands

For sales contact: For support contact:

telephone: +31 88 335 81 88 telephone: +31 88 335 81 00
fax: +31 88 335 81 11 fax: +31 88 335 81 11
e-mail: software@deltares.nl e-mail: software.support@deltares.nl
www: https://www.deltares.nl/software www: https://www.deltares.nl/software

Copyright © 2023 Deltares

All rights reserved. No part of this document may be reproduced in any form by print, photo
print, photo copy, microfilm or any other means, without written permission from the publisher:
Deltares.
 Contents

Contents
List of Figures v

List of Tables vii

1 Introduction 1

2 NN-Overtopping User Manual 3

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Web Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.2 Calculator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.2.1 Schematic View . . . . . . . . . . . . . . . . . . . . . . 5
2.2.2.2 Input Tab . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.2.3 Output Tab . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.3 Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.4 Help and Download . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.5 Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.6 Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.7 Disclaimer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.8 Messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Installing stand-alone version of NN Overtopping . . . . . . . . . . . . . . . 12

3 XGB-Overtopping User Manual 15

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Web Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.2 Calculator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.2.1 Schematic View . . . . . . . . . . . . . . . . . . . . . . 17
3.2.2.2 Input Fields . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.2.3 Output Fields . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.3 Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.4 Import . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.5 Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.6 Disclaimer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4 Theoretical background 21
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Input variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.3 Schematisation of construction . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3.1 Determination of input parameters . . . . . . . . . . . . . . . . . . 22
4.3.2 Schematisation of complex structures . . . . . . . . . . . . . . . . . 27
4.3.2.1 Example 1: Structure with several slopes . . . . . . . . . . 27
4.3.2.2 Example 2: Structure with more than one berm . . . . . . . 28
4.3.2.3 Example 3: Sloping crest . . . . . . . . . . . . . . . . . . 29
4.4 Validity range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5 Release notes 31
5.1 New Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 Modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2.2 NN Overtopping . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Deltares iii
NN Overtopping & XGB Overtopping, User & Technical Manual

5.2.3 XGB Overtopping . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

6 References 33

iv Deltares
 List of Figures

List of Figures
2.1 Main window (shown after starting the web application) . . . . . . . . . . . . 4
2.2 Calculator menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Input tab, including its components . . . . . . . . . . . . . . . . . . . . . . 6
2.4 Output tab, including its components . . . . . . . . . . . . . . . . . . . . . 7
2.5 Report menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.6 Help and download menu item . . . . . . . . . . . . . . . . . . . . . . . . 8
2.7 Import menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.8 Export menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.9 Disclaimer menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.10 Overview NN Overtopping stand-alone files . . . . . . . . . . . . . . . . . . 12

3.1 Main window (shown after starting the web application) . . . . . . . . . . . . 16

3.2 Calculator menu item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3 Input tab, including its components . . . . . . . . . . . . . . . . . . . . . . 18

4.1 Parameters used for the NN modelling of wave overtopping discharge at coastal
structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Parts of a coastal structure . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.3 Berm schematisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.4 Possible determinations of ht (m) and Bt (m) . . . . . . . . . . . . . . . . 24
4.5 Possible determinations of Rc (m), Ac (m) and Gc (m) . . . . . . . . . . . . 25
4.6 Possible determinations of cot(αd )(−), cot(αu )(−) . . . . . . . . . . . . . 26
4.7 roughness/permeability of common armour layers (Molines and Medina, 2015). 27
4.8 Example 1: Structure with several slopes . . . . . . . . . . . . . . . . . . . 28
4.9 Example 2: Structure with more than one berm . . . . . . . . . . . . . . . . 28
4.10 Example 3: Sloping crest . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Deltares v
NN Overtopping & XGB Overtopping, User & Technical Manual

vi Deltares
 List of Tables

List of Tables
4.1 NN-Overtopping,Input Parameters . . . . . . . . . . . . . . . . . . . . . . 21

Deltares vii
NN Overtopping & XGB Overtopping, User & Technical Manual

viii Deltares
1 Introduction
The document in hand provides the user with a description on how to use the web applications
NN Overtopping and XGB Overtopping in Chapter (2) respectively Chapter (3). Furthermore
a description on the input parameters and how to schematize a structure can be found in
Chapter (4). For information on the latest improvements and/or new features which have
been implemented in the operational version of the web applications the reader is referred to
Chapter (5).

Deltares 1 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

2 of 35 Deltares
 2 NN-Overtopping User Manual

2.1 Introduction
The prediction tool for the estimation of mean overtopping discharges at various types of
coastal structures is developed by WL|Delft Hydraulics. Details of the methodology followed
for the development of the prediction tool are described in van Gent et al. (2007). The output
of this program tool includes the Neural Network prediction of the mean overtopping discharge
at a coastal structure and several other parameters indicating the uncertainty of the prediction.

The study on which this tool is based was co-sponsored by the Commission of the Euro-
pean Communities within the framework of the CLASH project (‘Crest Level Assessment of
Coastal Structures by full scale monitoring, neural network prediction and Hazard analysis on
permissible wave overtopping’).

The predictions based on the Neural Network can be used for the conceptual design of coastal
structures; they may not be used in the final design stage, since the results should be veri-
fied based on dedicated physical model tests for the particular wave conditions and structure
geometry of the structure to be built. The predictions are based on a data-set based on small-
scale physical model tests; the predictions are to some extent affected by model effects,
scale effects (see deliverable D40 from the CLASH project), limited accuracy of measurement
equipment, limited accuracy of wave generation techniques (compared to nowadays state-
of-the-art techniques), inconsistencies in the data-set, and lack of data in certain fields of
application. Although reliability levels are given in addition to the predictions, these reliability
levels do not account for most of these influences. Therefore, the Neural Network predictions
may only be used as first estimates of mean overtopping discharges.

For the definition of input parameters, including the validity ranges, and schematisation of a
structure the reader is referred to Chapter (4).

2.1.1 Background
For the prediction of the mean overtopping discharge a Neural Network model is used. This
model was derived by WL|Delft Hydraulics from 8372 input-output combinations obtained from
measurements performed in hydraulic scale models at several institutes (Aalborg Univer-
sity, Denmark; Danish Hydraulic Institute, Denmark; WL|Delft Hydraulics, The Netherlands;
Hydraulic Research Wallingford, UK; Leichtweiss Institute für Wasserbau, WKS+GWK, Ger-
many; Modimar, Italy; University of Edinburgh, United Kingdom; Universidad Politécnica de
Valencia, Spain; and others in Iceland, Japan, Norway and U.S.A).

2.2 Web Application

The NN Overtopping web application has been embedded as a frame in a Deltares web page.
The web application can be accessed through the URL:https://www.deltares.nl/en/software/
overtopping-neural-network/. The main window of the application is depicted in Figure (2.1).

Deltares 3 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

Figure 2.1: Main window (shown after starting the web application)

In this section the components of the NN Overtopping web application (Figure (2.1)) are pre-
sented. Each component can be accessed through the main menu items which are listed
below. Further a brief description on each menu item is given in a separate sub-section.

⋄ Introduction
⋄ Calculator
⋄ Report
⋄ Help and Download
⋄ Import
⋄ Export
⋄ Disclaimer

2.2.1 Introduction
After a user presses this menu item, the text depicted in Figure (2.1) is shown.

4 of 35 Deltares
 NN-Overtopping User Manual

2.2.2 Calculator
After a user presses this menu item, the window depicted in Figure (2.2) is shown.

Figure 2.2: Calculator menu item

As shown in Figure (2.2) the calculator window contains of the following main components:

⋄ Schematic View
⋄ Input Tab
⋄ Output Tab

2.2.2.1 Schematic View

This component shows a schematic view of a structure, including the meaning of each input
parameter relevant for performing a NN Overtopping calculation. Each parameter is either
visually associated in the schematic view or explained textually.

Deltares 5 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

2.2.2.2 Input Tab

The component marked as INPUT/OUTPUT TABS in Figure (2.2) consists of two separate
tabs, namely one tab associated with the input parameters (Input) and one tab associated
with the corresponding output parameters (Output). This sub-section describes the tab asso-
ciated with the input parameters as depicted in Figure (2.3).

Figure 2.3: Input tab, including its components

The window depicted in Figure (2.3) enables the user to perform a calculation of one or more
overtopping scenarios. A scenario is defined as a set of input parameters which are visual-
ized in Figure (2.2) as row. Each input field is associated with a parameter depicted in the
schematic view in Figure (2.2).

The input tab as shown in Figure (2.3) contains the following components:

⋄ CHECKBOX: A control to select one ore more scenarios, which need to be removed.
⋄ BUTTON: Add Scenario: A button which adds one or more new scenario (row(s)) to
the scenario list.

⋄ BUTTON: Remove selected scenario(s): A button which removes the selected sce-
narios from the scenario list.

⋄ BUTTON: Calculate all scenarios: A button which starts a NN Overtopping calculation

for the user specified scenario list.

⋄ BUTTON: Help: A button to switch to the menu item Help and Download.

Apart from the components listed above, the input tab in Figure (2.3) also shows the informa-
tion marked by the red dashed box. This text indicates if a calculation has finished. If the text
marked by the dashed box is shown the user can swithc to the output tab for the results. Note
that if this text is not shown, the calculation has not finished.

Note:

⋄ The web application can handle a maximum of 500 scenarios but the calculation can
take quite some time. There for it is advised to either reduce the number of scenarios
per run or use the stand-alone version of the NN Overtopping application. The stand-
alone version can be downloaded via the menu item Help and Download.

6 of 35 Deltares
 NN-Overtopping User Manual

2.2.2.3 Output Tab

The component marked as INPUT/OUTPUT TABS in Figure (2.2) consists of two separate
tabs, namely one tab associated with the input parameters (Input) and one tab associated
with the output parameters (Output). This sub-section describes the tab associated with the
output parameters as depicted in Figure (2.4).

Figure 2.4: Output tab, including its components

The window depicted in Figure (2.4) enables the user to inspect the NN Overtopping output
results. These results are presented numerically as well as visually via a chart. Note that
each row on the tab output corresponds with the same row (scenario) on the input tab. The
main part of the output is the mean overtopping discharge (q) predicted by the NN model
in m3 /s/m. Next to that, the output consists of the quantiles q2, 5%, q5%, q25%, q50%,
q75%, q95% and q97, 5%.

As mentioned above, apart from the numerical presentation of the output parameters, the
output results are also presented visually in a chart, which is located at the bottom of the
output tab in Figure (2.4). The contents of the chart depends on the rows selected. To
include/exclude output results from the chart, the user can select/de-select the checkbox con-
trol located at the left of each row. Next to the chart control and the list of rows containing the
output results, the output tab contains the following controls:

⋄ CHECKBOX: A control to select/de-select the output rows to be visualized in the chart.

⋄ BUTTON: Update chart: A button which updates the chart after one or more rows have
been selected/de-selected.

⋄ MESSAGES: A control which shows warnings and remarks generated for a particular
row. For more information on warnings and remarks the reader is referred to section
Messages.

⋄ BUTTON: Help: A button to switch to the menu item "Help and Download".

Deltares 7 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

2.2.3 Report
After a user presses this menu item, the window depicted in Figure (2.5) is shown.

Figure 2.5: Report menu item

The window depicted in Figure (2.5) enables the user to convert the input and output, pre-
sented on the corresponding tabs, into a report as a PDF file. The report window contains the
following components:

⋄ REPORT INPUT: Five report properties to characterize the calculation.

⋄ BUTTON: Full report as PDF: A button which generates a report based on the user
provided report properties, input scenarios and corresponding output.

⋄ BUTTON: Return to calculation: A button which return to the menu item "Calculator".

2.2.4 Help and Download

After a user presses this menu item, the window depicted in Figure (2.6) is shown.

Figure 2.6: Help and download menu item

The window depicted in Figure(2.6) enables to user to download the help file by pressing the
button "Click here to download the help file". Moreover, for the NN Overtopping web applica-
tion the user can also download the stand-alone version of the tool. This is recommended if
the user wants to perform calculations for more than 500 records. To download the set-up the
user presses the button "Click here to download the program (as msi)". For more information
on how to install the program, the reader is referred to Section(2.3).

8 of 35 Deltares
 NN-Overtopping User Manual

2.2.5 Import
After a user presses this menu item, the window depicted in Figure (2.7) is shown.

Figure 2.7: Import menu item

The import dialog depicted in Figure (2.7) enables to user to add one are more new records to
the scenario list located on the input tab by selecting an input file located on a user specified
local drive. The imported records either replace the scenarios already defined on the input
tab or appends them. The latter depends on whether the checkbox marked in Figure (2.7) as
"Append to current scenarios" is selected or not.

The input file for the NN Overtopping application is an ASCII based file and its format is
specified below. Furthermore, each column must be separated by one or more spaces, note
not e.g. tabs! Moreover the file extension of the input file must be ".inp"

!---------------------------------------------------------------------------
! Begin input file
! NN_OVERTOPPING_1.INP
! This is a comment record
! ===============================
! COLUMN#01 Angle of Wave attack
! COLUMN#02 Water depth in front of structure
! COLUMN#03 Significant Wave Height at the toe of structure
! COLUMN#04 Wave period
! COLUMN#05 Water depth at the toe of structure
! COLUMN#06 Width of toe
! COLUMN#07 Roughness coefficient
! COLUMN#08 Angle of down slope
! COLUMN#09 Angle of upper slope
! COLUMN#10 Crest Freeboard in relation to SWL
! COLUMN#11 Berm Width
! COLUMN#12 Water depth at the berm of the structure
! COLUMN#13 Berm slope
! COLUMN#14 Armour Freeboard in relation to SWL
! COLUMN#15 Armour Width
! ===============================
! and this is another comment record
* as well as this one, and the next five ones

Deltares 9 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

!-----------------------------------------------------------------------------
!0 1 02 03 04 05 06 07 08 09 10
!b h Hm0 Tm-1,0 ht Bt gf cotad cotau Rc
!degr m toe(m) toe(s) m m m
!-----------------------------------------------------------------------------

Note: To obtain a default input file for a NN Overtopping calculation, the user can simply go
to the menu item "Export" and press the button "File with the input data (input file)".

2.2.6 Export
After a user presses this menu item, the window depicted in Figure (2.8) is shown.

Figure 2.8: Export menu item

The window depicted in Figure (2.8) enables the user to obtain the input as well as the output
file(s). To obtain the input file the user has to press the button "File with the input data (input
files)". The output file(s) is(are) obtained by pressing the button "File with the results (output
file)".

2.2.7 Disclaimer
After a user presses this menu item, the window depicted in Figure (2.9) is shown.

Figure 2.9: Disclaimer menu item

10 of 35 Deltares
 NN-Overtopping User Manual

2.2.8 Messages
Apart from numerical output (see Section (2.2.2.3)), remarks and warnings can be generated
during a calculation for specific scenarios also. If a message is generated for a particular
input scenario, it is displayed at the end of that particular row on the output tab. In case an
input parameter is specified with an input table (with more than one entry for which remarks
or warnings apply), the remarks or warnings are produced only for the first entry in the table
for which a remark applies. For example, if the wave height is specified in an input table with 3
different values for which the second and the third entry should produce a remark or warning,
the remark or warning for only the second entry is shown.

⋄ ‘WARNING> Parameter <parameter name> out of range of validity. No prediction is

given’
This warning indicates that because the parameter <parameter name> is out of the
range of validity of the NN model, no prediction is given. This warning can appear for
any of the input parameters. For the mean overtopping discharge (q ) the value −1000
is given.

⋄ ‘WARNING> For 10−6 < Q < 10−5 , the NN prediction is less reliable (indicative)’
This remark indicates that the dimensionless value of the mean overtopping discharge
predicted by the NN lies in the region 10−6 < Q < 10−5 , (with Q = dimensionless
q
overtopping discharge, Q = √ 3 ), and therefore, the NN prediction should be con-
gHs
sidered as less reliable, or indicative.

⋄ ‘WARNING> For Q < 10−6 , no prediction is given’

This remark indicates that the dimensionless value of the mean overtopping discharge
predicted by the NN is in the region Q < 10−6 (with Q = dimensionless overtopping,
Q = √ q 3 ), and therefore, the NN prediction does not give any prediction. The default
gHs
value given in this case is −1000.

⋄ ‘REMARK> For prototype, rough-sloping structures, a correction factor (for scale/model/wind

effects) is applied: q = <corrected value> m3 /s/m’
With this remark the user is given the possibility of choosing between the direct output of
the NN model or a corrected value to account for model effects, scale effects and wind
effects in prototype situations (Hm0 > 0.5m), for rough-sloping structures (γf < 0.9
and cot(α) > 1).

⋄ ‘REMARK> For prototype, rough-sloping structures, a correction factor (for scale/model/wind

effects) is applied: q = <corrected value> m3 /s/m, Since 10−6 < Q < 10−5 , the NN
prediction is less reliable (indicative)’
With this remark the user is given the possibility of choosing between the direct output of
the NN model or a corrected value to account for model effects, scale effects and wind
effects in prototype situations (Hm0 > 0.5m), for rough-sloping structures (γf < 0.9
and cot(α) > 1). The user is also warned that the dimensionless value of the mean
overtopping discharge predicted by the NN before the correction is applied lies in the
region 10−6 < Q < 10−5 , and therefore the NN prediction should be considered as less
reliable, or indicative.

⋄ ‘REMARK > For prototype, smooth vertical structures,a correction factor (for scale/model/wind
effects) is applied: q = <corrected value> m3 /s/m’
With this remark the user is given the possibility of choosing between the direct output
of the NN model or a corrected value to account for wind effects in prototype situations
(Hm0 > 0.5m), for smooth (γf ≥ 0.9) and vertical structures (cot(α) ≤ 1).

Deltares 11 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

⋄ ‘REMARK > For prototype, smooth vertical structures, a correction factor (for scale/model/wind
effects) is applied: q = <corrected value> m3 /s/m, Since 10−6 < Q < 10−5 , the NN
prediction is less reliable (indicative)’
With this remark the user is given the possibility of choosing between the direct output
of the NN model or a corrected value to account for wind effects in prototype situations
(Hm0 > 0.5m), for smooth (γf ≥ 0.9) and vertical structures (cot(α) ≤ 1). The user
is also warned that the dimensionless value of the mean overtopping discharge pre-
dicted by the NN before the correction is applied lies in the region 10−6 < Q < 10−5 ,
and therefore the NN prediction should be considered as less reliable, or indicative.

When input is specified with input arrays to investigate the sensitivity of certain parameters
or to find trends, it is possible that (for certain combinations of input parameters) the Neural
Network calculates mean wave overtopping discharges that are physically not entirely sound.
It is therefore recommended to always look at the predicted quantiles. It could well be that the
confidence bands are much wider (indicating a smaller reliability) for certain combinations of
input parameters than for others.

2.3 Installing stand-alone version of NN Overtopping

For the CLASH based NN Overtopping tool, you have the option to download the stand-alone
executable via the menu item "Help and Download". It is advised to use the stand-alone tool,
if you want to run a large number (more than 500 records) of input records.

After a user has pressed the download button a set-up called "nnovertopping_x86.msi" is
available in the download directory. Running this set-up installs the following files on your
local drive (default location "c:
Program Files (x86)
Deltares
NN_Overtopping xxxx
") as depicted in Figure (2.10).

Figure 2.10: Overview NN Overtopping stand-alone files

12 of 35 Deltares
 NN-Overtopping User Manual

To perform an NN Overtopping analysis, you need to specify all your input records (equivalent
to scenarios in the web application) in the file "NN_Overtopping.inp". An example file is lo-
cated in the installation directory. For more information on the format of the input file the reader
is referred to Section (2.2.5). For now, use the same file name for storing your input data as
is used for the example file. Note that you can upload the input file ("NN_Overtopping.inp") to
the web application by selecting the menu item "Import".

For running the NN Overtopping stand-alone version, go to the installation directory and either
open a DOS box (the installed application does not come with a Graphical User Interface) and
run the program within the DOSBOX by typing "NN_Overtopping.exe" on the command-line
or via Explorer double click on the program "NN_Overtopping.exe" to perform the calculation.
After the program has finished, it will produce an output file which extension "*.LIS", which
contains the calculation results, including remarks if applicable.

Deltares 13 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

14 of 35 Deltares
 3 XGB-Overtopping User Manual

3.1 Introduction
The prediction tool for the estimation of mean overtopping discharges at various types of
coastal structures is developed by Deltares. Details of the methodology followed for the de-
velopment of the prediction tool are described by den Bieman et al. (2021). The output of
this web tool includes the XGB-Overtopping prediction of the mean overtopping discharge at
a coastal structure and several other parameters indicating the uncertainty of the prediction.
The predictions based on the XGB-Overtopping model can be used for the conceptual design
of coastal structures; they may not be used in the final design stage, since the results should
be verified based on dedicated physical model tests for the particular wave conditions and
structure geometry of the structure to be built. The predictions are based on a data-set based
on small-scale physical model tests; the predictions are to some extent affected by model
effects, scale effects limited accuracy of measurement equipment, limited accuracy of wave
generation techniques (compared to nowadays state-of-the-art techniques), inconsistencies
in the data-set, and lack of data in certain fields of application. Although reliability levels are
given in addition to the predictions, these reliability levels do not account for most of these
influences. Therefore, the XGB-Overtopping predictions may only be used as first estimates
of mean overtopping discharges.

3.1.1 Background
For the prediction of the mean overtopping discharge a XGBoost (Chen and Guestrin, 2016)
model is used. This model was derived by Deltares from 9511 input-output combinations ob-
tained from measurements performed in hydraulic scale models at several institutes (Deltares
| Delft Hydraulics, The Netherlands; Aalborg University, Denmark; Danish Hydraulic Insti-
tute, Denmark; Hydraulic Research Wallingford, UK; Leichtweiss Institute für Wasserbau,
Germany; Modimar, Italy; University of Edinburgh, United Kingdom; Universidad Politécnica
de Valencia, Spain; Ghent University, Belgium; and other sources from countries including
Iceland, Japan, Norway and U.S.A).

3.2 Web Application

The XGB Overtopping web application has been embedded as a frame in a Deltares web
page. The web application can be accessed through the URL:https://www.deltares.nl/en/
software/overtopping-xgb/. The main window of the application is depicted in Figure (3.1).

Deltares 15 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

Figure 3.1: Main window (shown after starting the web application)

In this section the components of the XGB Overtopping web application as depicted in Figure
(3.1) are presented. Each component can be accessed through the main menu items listed
below. Further a brief description on each menu item is given in a separate sub-section.

⋄ Introduction
⋄ Calculator
⋄ Help
⋄ Import
⋄ Export
⋄ Disclaimer

Note: Nearly all the menu items used in the XGB Overtopping web application are equal to
the items used in the NN Overtopping web application menu bar. However, there are some
differences. The main differences are that there is no menu item "Report" and the menu item
"Help and Download" in the NN Overtopping menu bar does not support downloading the
XGB stand-alone version.

3.2.1 Introduction
After a user presses this menu item, the text depicted in Figure (3.1) is shown.

16 of 35 Deltares
 XGB-Overtopping User Manual

3.2.2 Calculator
After a user presses this menu item, the window depicted in Figure (3.2) is shown. This
window is similar to the one shown in Figure (2.2), except that the roughness coefficient γf
has been split in a roughness coefficient for the upper slope (γf,u ) and for the down slope
(γf,d ). Moreover the slope of the berm - tan(αB ) - is no longer an input parameter, while the
slope of the foreshore - tan(αF ) - has been added as a input parameter.

Figure 3.2: Calculator menu item

As shown in Figure (3.2) the calculator window contains of the following main components:

⋄ Schematic View
⋄ Input Fields
⋄ Output Fields

3.2.2.1 Schematic View

This component shows a user a schematic view of a structure, including the definition of each
input parameter relevant for performing a XGB Overtopping calculation. Each parameter is
either visually associated in the schematic view or explained in the accompanying text.

3.2.2.2 Input Fields

The functionality of the controls available on the XGB Overtopping input tab are equal to the
ones available on the corresponding tab for the NN Overtopping web application. For more
information on how to e.g. add/remove/calculate one or more XGB Overtopping scenarios the
reader is referred to Section (2.2.2.2).

Deltares 17 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

In contrast to the NN Overtopping calculation, for XGB Overtopping the text depicted in Figure
(3.3) is displayed during a calculation. The reason for this is that the calculation can take a bit
longer relative to a NN Overtopping calculation when using the same number of records.

Figure 3.3: Input tab, including its components

3.2.2.3 Output Fields

The functionality of the controls available on the XGB Overtopping outpub tab are equal to the
ones available on the corresponding tab for the NN Overtopping web application. For more
information on how to inspect results of one or more XGB Overtopping scenarios the reader
is referred to Section (2.2.2.3).

3.2.3 Help
After a user presses this menu item, a similar window as depicted in Figure (2.6) is shown,
except the button "Click here to download the program (as msi)" is missing because this
feature is not supported for the XGB Overtopping tool.

3.2.4 Import
After a user presses this menu item, a similar window as depicted in Figure (2.7) is shown.
For more information on how to import the input file the reader is referred to Section (2.2.5).

However, regarding the format of the input file. This differs from the format used for the NN
Overtopping web application. For the XBG Overtopping application supports a CSV formatted
input file. This CSV file starts with a header as specified below. Following the header, the
different calculation scenarios are supplied, with the input parameters in the order specified
by the header. Note that it used the semi-colon (;) as delimiter, so that the CSV files can be
easily read and edited in Excel.

scenario [-];b [degrees];h [m];Hm0 toe [m];Tm-1,0 toe [s];ht [m];

Bt [m];hb [m];B [m];Rc [m];Ac [m];Gc [m];cotad [-];cotau [-];
gf_d [-];gf_u [-];tanaF [-]

Note: To obtain a default input file for a XGB Overtopping calculation, the user can simply go
to the menu item Export and press the button "File with the input data (input file)".

18 of 35 Deltares
 XGB-Overtopping User Manual

3.2.5 Export
After a user presses this menu item, a similar window as depicted in Figure (2.8) is shown. For
more information on how to import the export a file the reader is referred to Section (2.2.6).

3.2.6 Disclaimer
After a user presses this menu item, a similar window as depicted in Figure (2.9) is shown.

Deltares 19 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

20 of 35 Deltares
 4 Theoretical background

4.1 Introduction
In this chapter the input parameters for NN Overtopping as well as XGB Overtopping are
defined in Section (4.2). For information on schematisation of a structure the reader is referred
to Section(4.3). The validity ranges of the input parameters can be found in Section (4.4).

4.2 Input variables

Figure 4.1: Parameters used for the NN modelling of wave overtopping discharge at
coastal structures

The input consists of the following 15 (or 16 for XGB Overtopping) parameters. A more de-
tailed description of these parameters is given hereafter (4.1). Figure 4.1 shows a graphical
illustration of the meaning of the sixteen input parameters.

Table 4.1: NN-Overtopping,Input Parameters

Parameter Description Unit NN XGB

β Direction of wave attack w.r.t. to the normal of the structure (deg)

h Water depth in front of the structure (m)
Hm0,toe Spectral significant wave height at the toe of the structure ob- (m)
1/2
tained from spectral analysis, Hm0,toe = 4m0
Tm−1,0−toe Mean wave period at the toe of the structure obtained from spec- (s)
m
tral analysis, Tm−1,0−toe = m−1
0
ht Water depth on the toe of the structure (m)
Bt Width of the toe of the structure (m)
hb Water depth on the berm (m)
B Berm width (m)
Rc Crest freeboard of the structure (m)
Ac Armour crest freeboard of the structure (m)
Gc Crest width of the structure (m)
cot(αd ) Slope of the structure downward of the berm (−)
cot(αu ) Slope of the structure upward of the berm (−)
continued on next page

Deltares 21 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

continued from previous page

γf Roughness/permeability of the structure (−)

γf,d Roughness/permeability of the structure downward of the berm (−)
γf,u Roughness/permeability of the structure upward of the berm (−)
tan(αB ) Slope of the berm (−)
tan(αF ) Slope of the berm (−)

For each user-supplied set of input parameters

⇒ NN Overtopping: [ Hm0 , Tm−1,0 , β , h, ht , Bt , γf , cot(αd ), cot(αu ), B , hb , tan(αB ),

Rc , Ac , Gc ],
⇒ XGB Overtopping:[ Hm0 , Tm−1,0 , β , h, ht , Bt , γf,d ,γf,u , cot(αd ), cot(αu ), B , hb ,
tan(αF ), Rc , Ac , Gc ]

, the output includes the mean wave overtopping discharge (q ), and 7 other output values
indicating the quantiles of several orders, q2, 5%, q5%, q25%, q50%, q75%, q95% and
q97, 5% The 95% confidence interval is, for instance, given by the quantiles q2.5% and
q97.5%.

4.3 Schematisation of construction

4.3.1 Determination of input parameters

This section briefly describes how to estimate some of the hydraulic parameters required
by the program when these are not available, and how to determine the required structure
parameters for any arbitrary structure and for rather complicated coastal structures. A more
extensive description can be found in Steendam et al. (2004)

Often some of the hydraulic parameters might not be directly available. In these cases, the
following estimations or calculations can be applied:

⋄ If only the deep water wave characteristics are available and not the wave characteris-
tics at the toe of the structure, the calculation of the wave characteristics at the toe of
the structure can be performed with the model ‘SWAN’.

⋄ If only the wave height from time domain analysis, Hs , is available, the wave height
from spectral domain analysis, Hm0 , can be determined with Battjes and Groenendijk
(2000)

⋄ If only the wave period from time domain analysis is available, the estimation of the
wave period from spectral analysis can be made by means of empirically determined
proportions of wave periods available in literature, such as the following (Goda and
Nagai, 1974; Goda, 1985), which can be used for single-peaked wave energy spectra
with a spectral shape similar to Jonswap spectra.

Tp ≈ 1.05T1/3
⋄ ⋄ ⋄

Tp = 1.2Tm
Tm−1,0 ≈ Tp /1.1

22 of 35 Deltares
 Theoretical background

Most coastal structures can be relatively well schematised by means of the 12 structure
parameters. For the correct use of this prediction tool it is important that all parameters are
determined in the same way. In the following, a brief description indicating how to determine
these 12 structure parameters for an arbitrary structure is given.

Three parts can be distinguished in an average coastal structure, the lower part (or toe), the
centre part (eventually with a berm), and the upper part (or crest). The separation of these
three parts of the structure is not always that clear and depends on the hydraulic conditions
and structure shape. In this way, the same structure could have a different schematisation
for a different water level and different wave attack. Figure (4.2) shows the three parts of a
typical coastal structure, where the Centre part corresponds to the area within the vertical
distance 1.5 ∗ Hm0,toe above and below the sea water level, and the Upper and Lower parts
correspond to the areas above and below the "Centre part", respectively.

Figure 4.2: Parts of a coastal structure

The toe of the structure is normally situated in the lower area of the structure. Nevertheless,
in some cases the toe can also be located at the centre part of the structure (e.g. structure
with quite large toe situated in relatively shallow water). In this case, the toe can be taken into
account as a berm, as shown in Figure (4.3).

A berm is always located in the centre area of the structure. If the "berm" is situated lower,
then it is considered as a toe (as shown in Figure (4.3)), while if it is situated higher, then it
is considered as a crest. Some further restrictions exist regarding the definition of a berm.
These are described in TAW (2002).

(a) Berm schematised as a toe (b) Berm schematised as a toe

Figure 4.3: Berm schematisation

A crest of a structure is situated normally in the upper area of a structure. However, there are
also exceptions in this case (e.g. in very low structures the crest can be a part of the centre

Deltares 23 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

area).

In the following, a detailed explanation regarding the determination of each of the 12 structure
parameters is given.

⋄ h(m) : This is the water depth at the toe of the structure, more precisely the water
depth just before the structure.

⋄ ht (m), Bt (m): These are the water depth on the toe and the width of the toe, see
Figure (4.4). The width of the toe is measured on top of the toe. If there is no toe,
the value of the water depth on the toe is the same as the water depth in front of the
structure ht = h. In this case the width of the berm Bt is equal to zero.

Figure 4.4: Possible determinations of ht (m) and Bt (m)

⋄ B(m), hb (m), tan(αB )(−): These parameters describe the berm of the structure. B
is the berm width, measured horizontally. hb is the water depth on the berm, measured
in the middle of the berm. If the berm is situated above swl, hb is negative. T an(αB )
is the tangent of the slope of the berm. If the berm is horizontally, tan(αB ) = 0. If the
structure has no berm, the values of B(m), tan(αB )(−) and hb (m) are all equal to
zero.

⋄ Rc (m), Ac (m), Gc (m): These parameters describe the upper part of the structure.
Rc is the crest freeboard of the structure; that is, the distance, measured vertically from
the still water level (swl) to the highest impermeable point of the structure, see Figure
(4.5). This means that at this point, waves are stopped by the structure. This does
not always have to be the highest point of the structure: Rc can adopt a value larger,
smaller or equal to Ac . Ac is the armour crest freeboard of the structure; that is, the
distance, measured vertically from the swl to the highest point of the armour on the
structure. Gc is the crest width. In the case that a crest is constructed on the structure,
Gc is the width of the armour in front of the crest element.

24 of 35 Deltares
 Theoretical background

Figure 4.5: Possible determinations of Rc (m), Ac (m) and Gc (m)

⋄ cot(αd )(−), cot(αu )(−): These parameters are used to describe the slope(s) of the
structure, see Figure (4.6). The toe and the crest of the structure are already described
in other parameters, therefore they are not included in these two parameters. cot(αd )
and cot(αu ) are the cotangents of the mean slopes in the centre area of the structure
under (cot(αdown )) and above (cot(αup )) the berm respectively. The upper slope au
can be determined by taking the point of the structure at a level of 1.5 ∗ Hm0 toe above
the swl and connecting it with the point of the berm farthest from the sea. If the crest
of the structure is situated in the centre area of the structure (at a distance less than
1.5 ∗ Hm0 toe above swl), then the point of the crest nearest to the sea has to be
used to determine αu . The lower slope αd can be determined by taking the point of the
structure at a level of 1.5 ∗ Hm0 toe under the swl and connecting it with the point of
the berm nearest to the sea. If the toe of the structure is situated in the centre area of
the structure (at a distance less than 1.5 ∗ Hm0 to under swl), then the point of the toe
farthest of the sea has to be used to determine αd .

Deltares 25 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

Figure 4.6: Possible determinations of cot(αd )(−), cot(αu )(−)

26 of 35 Deltares
 Theoretical background

⋄ γf (−),γf,d (−) and γf,u (−) : This parameter gives an indication of the roughness and
the permeability of the structure. The rougher and more permeable a structure is, the
lower the overtopping will be since more energy is dissipated on a rough surface and
more energy will disappear into a permeable structure. The factor γf was originally
introduced as a reduction factor for the roughness and permeability for run-up on a
structure. Values of γf for run-up on dikes with different top-layers are given in TAW
(2002). Overtopping structures are often constructed of rubble mound, artificial armour
units or a combination of both. As no extensive research has been performed before
regarding the roughness of different armour units, some assumptions can be made. Ta-
ble (4.7) gives some suggestions for the roughness-factors of the most common armour
layers.

Figure 4.7: roughness/permeability of common armour layers (Molines and Medina,

2015).

4.3.2 Schematisation of complex structures

In some cases, it is not possible to describe very complicated coastal structures with the
above-mentioned parameters. For these cases, it is then required to make rough schemati-
sations to come to an approximate description of the section with these parameters. Often,
several possibilities exist to make such approximations; therefore it is the task of the user
to choose the best solution for the schematisation of the structure. In the following, some
examples of the schematisation of several structures are given.

4.3.2.1 Example 1: Structure with several slopes

In the case that a structure consists of more than two slopes (e.g. lower, mid and upper
slopes), it should be taken into account that a mid-slope of 1 : 5 is too steep to be a berm.
This slope should be therefore included in the slope of the structure downward of the berm
(parameter cot(αd )). A possible schematisation could be to consider the lower slope as
the toe of the structure with toe width Bt = 0, the mid-slope as the slope downward the
berm (cot(αd )), the upper slope as the slope upward the berm (cot(αu )) and the intersection
between the last slopes as a berm with berm width B = 0.

Deltares 27 of 35
 NN Overtopping & XGB Overtopping, User & Technical Manual

Figure 4.8: Example 1: Structure with several slopes

4.3.2.2 Example 2: Structure with more than one berm

In the case that a structure consists of several sloping and horizontal parts (e.g. two horizontal
parts or horizontal berms with not a too big difference in level between both berms), a rougher
schematisation is required. The schematisation advised in this case is to combine the two
horizontal berms in one mean berm at the average level of both berms. The berm width can
be determined here by lengthen the upper and lower slope up to the level of the mean berm.

If the width of the two berms is very different, a weighted average level of the mean berm is
preferable (so the berm with the largest width has most influence on the level of the mean
berm).

Figure 4.9: Example 2: Structure with more than one berm

28 of 35 Deltares
 Theoretical background

4.3.2.3 Example 3: Sloping crest

The schematisation advised in the case that the crest of the structure is a sloping crest, is
to consider the crest as a horizontal crest at a level corresponding to the middle point of the
crest. The value of the crest width, Gc , in this case results from extending the slope of the
structure up to the level of the horizontal crest.

Figure 4.10: Example 3: Sloping crest

4.4 Validity range

The range of validity of the possible input parameters is given below referring to a significant
wave height of Hm0 = 1m.

1(m) < Hm0 < 1(m)

0.005(−) < sm−1,0 < 0.07(−)
0(deg) < β < 80(deg)
0.9(m) < h < 20(m)
0.5(m) < ht < 20(m)
0(m) < Bt < 10(m)
0.3(−) < γf < 1(−)
0.3(−) < γf,d < 1(−)
0.3(−) < γf,u < 1(−)
0(−) < cot(αd ) < 10(−)
−1(−) < cot(αu ) < 10(−)
0(m) < B < 15(m)
−1(m) < hb < 5(m)
0(−) < tan(αB ) < 0.1(−)
0(−) < tan(αF ) < 0.1(−)
continued on next page

Deltares 29 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

continued from previous page

0.5(m) < Rc < 5(m)

0(m) < Ac < 5(m)
0(m) < Gc < 10(m)

30 of 35 Deltares
 5 Release notes

5.1 New Features

No new features have been implemented for the current version of NN Overtopping and XGB
Overtopping.

5.2 Modifications
For the current version the following modification have been implemented.

5.2.1 General

⋄ To avoid confusion the units used for the discharge output parameters of the NN Over-
topping as well as the XGB Overtopping web application have been set to SI units,
so m3 /s/m instead of l/s/m. Before the NN Overtopping and XGB Overtopping web
application showed the discharge output parameters in l/s/m. Moreover, that the XGB
Overtopping application labelled these values as m3 /s/m in stead l/s/m. However,
the unit of the discharge output parameters stored in the downloadable output files are
presented in m3 /s/m for both applications.

⋄ Updated the manual.

5.2.2 NN Overtopping

⋄ Added a description on the format of the downloadable input file to the input file itself,
⋄ Changed the numerical values in the output tab, including warnings, reporting and chart
such that they are presented in m3 /s/m instead of l/s/m,

⋄ Changed the units specified in the output tab, including warnings, reporting and chart
such that they match the numerical values, meaning m3 /s/m instead of l/s/m.

5.2.3 XGB Overtopping

⋄ Added units to the description of the input parameters specified in the header of the
downloadable input file,

⋄ Added units to the description of the output parameters specified in the header of the
downloadable output file.

⋄ Changed the numerical values in the output tab and chart such that they match the unit
description, meaning m3 /s/m instead of l/s/m.

Deltares 31 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

32 of 35 Deltares
6 References
Battjes, J. A. and H. W. Groenendijk, 2000. “Wave height distributions on shallow foreshores.”
Coastal Engineering 40 (3): 161–182. DOI: https://doi.org/10.1016/S0378-3839(00)00007-
7, ISSN 0378-3839.

Chen, T. and C. Guestrin, 2016. “XGBoost: A Scalable Tree Boosting System.” CoRR
abs/1603.02754. URL http://arxiv.org/abs/1603.02754.

den Bieman, J. P., M. R. A. van Gent and H. F. P. van den Boogaard, 2021. “Wave overtopping
predictions using an advanced machine learning technique.” Coastal Engineering 166:
103830. DOI: https://doi.org/10.1016/j.coastaleng.2020.103830, ISSN 0378-3839.

Goda, Y., 1985. “A review on statistical properties of sea waves.” Doboku Gakkai Ronbunshu
1985 (357): 1–12. DOI: 10.2208/jscej.1985.357_1.

Goda, Y. and K. Nagai, 1974. “Investigation of the statistical properties of sea waves with field
and simulation data.” Rept. Port and Harbour Res. Inst. 13 (1): 3–37.

Molines, J. and J. R. Medina, 2015. “Calibration of overtopping roughness factors for concrete
armor units in non-breaking conditions using the CLASH database.” Coastal Engineering
96: 62–70. DOI: https://doi.org/10.1016/j.coastaleng.2014.11.008, ISSN 0378-3839.

Steendam, G. J., J. W. Van der Meer, H. Verhaeghe, P. Besley, L. Franco and M. R. A. Van
Gent, 2004. “The International Database on Wave Overtopping.” In Coastal Engineering
2004, pages 4301-4313. DOI: 10.1142/9789812701916_0347.

TAW, 2002. Wave run-up and wave overtopping at dikes. Tech. rep., Technische Adviescom-
missie voor de Waterkeringen.

van Gent, M. R. A., H. F. P. van den Boogaard, B. Pozueta and J. R. Medina, 2007. “Neural
network modelling of wave overtopping at coastal structures.” Coastal Engineering 54 (8):
586 - 593. DOI: https://doi.org/10.1016/j.coastaleng.2006.12.001, ISSN 0378-3839.

Deltares 33 of 35
NN Overtopping & XGB Overtopping, User & Technical Manual

34 of 35 Deltares