Technical Report 2005/24 July 2005 I L Turner by COLLAROY

COLLAROY-NARRABEEN COASTAL IMAGING SYSTEM REPORT 1

SYSTEM DESCRIPTION, ANALYSIS OF SHORELINE VARIABILITY AND EROSION/ACCRETION TRENDS: JULY 2004 - JUNE 2005 by

I L Turner

Technical Report 2005/24 July 2005

THE UNIVERSITY OF NEW SOUTH WALES SCHOOL OF CIVIL AND ENVIRONMENTAL ENGINEERING WATER RESEARCH LABORATORY

COLLAROY-NARRABEEN COASTAL IMAGING SYSTEM REPORT 1

SYSTEM DESCRIPTION, ANALYSIS OF SHORELINE VARIABILITY AND EROSION/ACCRETION TRENDS: JULY 2004 – JUNE 2005

WRL Technical Report 2005/24 July 2005

I L Turner

Water Research Laboratory School of Civil and Environmental Engineering Technical Report No 2005/24 University of New South Wales ABN 57 195 873 179 Report Status Final King Street Date of Issue July 2005 Manly Vale NSW 2093 Australia

Telephone: +61 (2) 9949 4488 WRL Project No. 02092.01 Facsimile: +61 (2) 9949 4188 Project Manager Ian L Turner

Title Collaroy-Narrabeen Coast Coastal Imaging System - Report 1: System Description, Analysis of Shoreline Variability and Erosion/Accretion Trends: July 2004 – June 2005

Author(s) Ian L Turner

Client Name Warringah Council

Client Address Civic Centre, 725 Pittwater Road, Dee Why NSW 2099

Client Contact Daylan Cameron – Catchment Management Team

Client Reference

The work reported herein was carried out by the Water Research Laboratory, School of Civil and Environmental Engineering, University of New South Wales, acting on behalf of the client.

Information published in this report is available for general release only by permission of the Director, Water Research Laboratory, and the client.

WRL TECHNICAL REPORT 2005/24

CONTENTS

1. INTRODUCTION 1-1 1.1 General 1-1 1.2 Report Outline 1-2 2. BACKGROUND 2-1 2.1 Environmental Setting 2-1 2.2 Study Area 2-1 3. DESCRIPTION OF THE COLLAROY-NARRABEEN COASTAL IMAGING SYSTEM 3-1 3.1 What is Coastal Imaging? 3-1 3.2 The Difference between Coastal Imaging and a 'surfcam' 3-1 3.3 The ARGUS Coastal Imaging System 3-2 3.4 Design of Optimum Camera Layout for Collaroy-Narrabeen 3-2 3.5 Installation 3-3 3.6 Calibration 3-3 3.6.1 Lens/Camera Calibration 3-4 3.6.2 Ground Control Point (GCP) Survey 3-5 4. IMAGE TYPES AND IMAGE PROCESSING TECHNIQUES 4-1 4.1 Image Types 4-1 4.1.1 Snap-Shot 'snap' Images 4-1 4.1.2 Time-Exposure 'timex' Images 4-1 4.1.3 Variance 'var' Images 4-2 4.1.4 Day Time-Exposure 'daytimex' Images 4-2 4.2 Basic Image Processing – Merge and Rectification 4-2 4.3 Shoreline Detection and Analysis 4-3 4.3.1 Overview of the ‘PIC’ shoreline identification technique 4-3 4.3.2 Standardised Procedure for Shoreline Mapping 4-4 5. COASTAL IMAGING WEB SITE 5-1 5.1 Coastal Imaging Home Page 5-1 5.2 Image Archive 5-1 5.3 On-Line ‘Beach Analysis System’ 5-2 6. MORPHODYNAMIC DESCRIPTION OF COLLAROY-NARRABEEN: JULY 2004 – JUNE 2005 6-1 6.1 A Morphodynamic Classification of Beaches 6-1 6.2 Morphodynamic Interpretation of Daily Images 6-2 6.2.1 July 2004 6-3 6.2.2 August 2004 6-4 6.2.3 September 2004 6-4 6.2.4 October 2004 6-4 6.2.5 November 2004 6-5 6.2.6 December 2004 6-6 6.2.7 January 2005 6-6 6.2.8 February 2005 6-7 6.2.9 March 2005 6-7 6.2.10 April 2005 6-8

- i - WRL TECHNICAL REPORT 2005/24

6.2.11 May 2005 6-8 6.2.12 June 2005 6-8 6.3 Visual Assessment of Net Beach Width Changes (July 2004 – June 2005) 6-9 7. QUANTITATIVE ANALYSIS OF SHORELINE CHANGES: JULY 2004 – JUNE 2005 7-1 7.1 Weekly Shorelines 7-1 7.2 Shoreline Variability – Mean, Maximum, Minimum, Standard Deviation 7-2 7.3 Time-Series of Beach Widths at Control Transects 7-3 7.4 Future reporting 7-4 8. ANALYSIS OF EROSION-ACCRETION TRENDS 8-1 8.1 Methodology 8-1 8.2 Monthly Beachface Bathymetric Mapping 8-1 8.3 Monthly Erosion-Accretion Trends 8-2 8.4 Net Erosion-Accretion Trends: June 2004 - November 2004 8-4 9. SUMMARY AND CONCLUSIONS 9-1 9.1 Qualitative Visual Assessment 9-1 9.2 Shoreline Variability and Beach Width Analysis 9-2 9.3 Erosion/Accretion Trends 9-3 10. ACKNOWLEDGEMENTS 10-1 11. REFERENCES 11-1

APPENDIX A – WEEK-TO-A-PAGE: JULY 2004-JUNE 2005

APPENDIX B – MONTHLY WAVE CLIMATE SUMMARIES: JULY 2004–JUNE 2005

APPENDIX C – Aarninkhof et al., (2003). Coastal Engineering, Vol.49(4), p.275-289.

- ii - WRL TECHNICAL REPORT 2005/24

LIST OF FIGURES 2.1 Locality 2.2 Collaroy - Narrabeen Embayment, NSW 2.3 Precinct 3 2.4 Coastal Erosion – 1950’s , 1970’s and 2002 3.1 Schematic of an ARGUS Coastal Imaging System 3.2 Design of Optimum Five-Camera Layout 3.3 Location of Coastal Imaging System – Flight Deck Building 3.4 Cameras Mounted at an Elevation of Approximately 50 m 3.5 Calibration of Intrinsic (CCD-Lens) Camera Model Parameters 3.6 GCP Survey (Composite Images - Camera 1 and Camera 5) 4.1 Snap-Shot, Time-Exposure and Variance Image Types (29/06/05) 4.2 Pan and Plan View Images: Five-Camera Merge/Rectification (29/06/05) 4.3 Identification of ‘Shoreline’ Feature from Colour Images 4.4 Standardised Shoreline Mapping Procedure 5.1 Coastal Imaging Web Site – Home Page 5.2 Coastal Imaging System Web Site – Image Archive 5.3 On-Line Beach Analysis System – ‘Week-to-a-Page’ 5.4 On-Line Beach Analysis System – ‘Beach Width Analysis’ 6.1 Morphodynamic Beach State Model (after Wright and Short, 1983) 6.2 Snap Images from Camera 1 (South): 15/07/04 and 29/06/05 6.3 Snap Images from Camera 5 (North): 15/07/04 and 29/06/05 7.1 Weekly Shorelines: July 2004 – June 2005 7.2 Weekly Beach Width: July 2004 – June 2005 7.3 Statistical Summary of Beach Width Changes: July 2004 – June 2005 7.4 Time-Series of Beach Width: July 2004 – June 2005 7.5 On-Line Beach Width Analysis: June 2005 8.1 Definition Sketch - Intertidal Bathymetry from Hourly Waterlines 8.2 Beachface Mapping – July, August 2004 8.3 Beachface Mapping – September, October 2004 8.4 Beachface Mapping – November, December 2004 8.5 Beachface Mapping – January, February 2005 8.6 Beachface Mapping – March, April 2005 8.7 Beachface Mapping – May, June 2005 8.8 Monthly Erosion/Accretion: July - October 2004 8.9 Monthly Erosion/Accretion: October 2004 – January 2005 8.10 Monthly Erosion/Accretion: January – April 2005 8.11 Monthly Erosion/Accretion: April – June 2005 8.12 Net Erosion/Accretion: July 2004 – June 2005

- iii - WRL TECHNICAL REPORT 2005/24 1-1.

1. INTRODUCTION

This report was prepared by the Water Research Laboratory (WRL) of the University of New South Wales for Warringah Council. It is the first in a series of annual data reports, to describe, quantify and analyse the regional-scale coastline variability and erosion/accretion trends that occur at Collaroy-Narrabeen Beaches. This annual summary supplements the on-line and ‘real-time’ monitoring program information that is available to Council and the community via the world-wide web (refer Section 5). It is intended that the growing database describe herein, of qualitative and quantitative coastal monitoring information, will inform and enhance the current and future management of the Collaroy-Narrabeen embayment.

1.1 General

In July of 2004, an ARGUS coastal imaging system was installed at the Collaroy- Narrabeen site for an initial period of three years. This leading-edge technology was selected by Warringah Council to provide regional-scale, continuous and long-term monitoring of this ‘high value’ coastal embayment. It is the ability to provide quantitative as well as qualitative information that distinguishes the ARGUS coastal imaging system from conventional 'webcam' or 'surfcam' technology.

Collaroy-Narrabeen is the first of a growing number of coastal management sites in New South Wales that now utilise coastal imaging technology and associated digital image techniques to monitor regional-scale coastal response to natural and engineered coastal impacts. Coastal imaging stations have been operating at coastal management sites in Australia since the first site was installed at the northern Gold Coast in 1999, and at the present time there are a total of 9 stations operating throughout Queensland and New South Wales (Turner et al., 2005).

Electronic copies of this and all future monitoring reports (as they are produced) are made available for public viewing and download in PDF format at:

Æ www.wrl.unsw.edu.au/coastalimaging/narrabn (link: monitoring reports)

The purpose of this first report is to describe the ARGUS system now installed at the Collaroy-Narrabeen embayment, and to present a summary of the results of shoreline

WRL TECHNICAL REPORT 2005/24 1-2. change analysis and erosion/accretion analysis for the initial 12-month monitoring period July 2004 to June 2005.

1.2 Report Outline

Following this introduction, Section 2 of this report provides a brief description of the Collaroy-Narrabeen embayment, and in the particular the ‘Precinct 3’ study area.

Section 3 contains a description of the ARGUS coastal imaging system, including the design, installation and calibration of the system at the Collaroy-Narrabeen site.

The image types that are collected on a routine basis are illustrated in Section 4, along with an overview of the digital image processing techniques used to analyse the images. More detailed description is provided of the key image analysis method that is used on a routine (weekly) basis to map the shoreline and beach width at the Precinct 3 study area.

The web site that was established to promote and distribute the images collected by this monitoring program is introduced in Section 5. Description includes the web-based image archive that provides unrestricted public access to all images, weekly-updated quantitative analysis of current coastline conditions, and 'time-lapse' animation files that are updated on a monthly basis.

Section 6 introduces the beach morphodynamic classification model of Wright and Short (1983), which is then used to describe in a qualitative manner the beach changes observed using the time-series of daily images for the period covered by this report, July 2004 – June 2005.

The quantitative analysis of shoreline change for the 12 month period July 2004 to June 2005 is detailed in Section 7.

The application of an image analysis technique that enables patterns of beach erosion and accretion to be identified and quantified within the Precinct 3 study area on a regular (monthly) basis is presented in Section 8. The principal findings of this first annual monitoring period are summarised in Section 9.

WRL TECHNICAL REPORT 2005/24 2-1.

2. BACKGROUND

2.1 Environmental Setting

Collaroy/Narrabeen Beach is located 16 km north of Sydney’s Central Business District, within Warringah Council Local Government Area. The beach is approximately 3.6 km in length from the Collaroy ocean pool in the south to the entrance of Narrabeen Lagoon in the north (Figure 2.1).

The Collaroy/Narrabeen embayment is characterised by having the most intense and highly capitalised shoreline development in Warringah (Figure 2.2). Development along the beach is further characterised as being at risk of impact by coastal processes and coastal erosion.

Several processes cause movement of sand within the Collaroy/Narrabeen Beach system. These include natural processes such as longshore movement of sediment, offshore movement of sediment into deeper water by wave action, and lagoon infilling by wave and tidal action, as well as human activities such as Narrabeen Lagoon Entrance Clearance Works. The only likely natural sources of sediment supply to the beach are the near-shore sand body and biogenic shell production associated with seabed reefs.

2.2 Study Area

The specific study area for the coastline monitoring extends along the embayment with particular focus on Precinct 3 as described in the Collaroy/Narrabeen Coastline Management Plan (1997) and shown in Figures 2.2 & 2.3. It is within this area that existing development encroaches to the greatest degree into the active beach environment. Periodic storm damage to beachfront property has occurred within this area of the Collaroy- Narrabeen embayment since it was first developed (Figure 2.4). A boulder wall has been proposed to protect beachfront property, however this option was subsequently rejected by the Warringah community.

NT New A L I South A QLD WA T R Wales S Narrabeen U Collaroy A SA Sydney NSW ACT Canberra VIC

TAS

ELANORA HEIGHTS NORTH De ep NARRABEEN C reek

S O U T H NARRABEEN Middle Creek Narrabeen Lakes Narrabeen Beach

COLLAROY PLATEAU WHEELER Collaroy HEIGHTS Beach CROMER HEIGHTS k COLLAROY e re C Wheeler P A C I F I C

C r CROMER e e k

h t

Dee Why Lagoon

BEACON HILL DEE WHY NARRAWEENA

O C E A N NORTH CURL CURL Gr ALLAMBIE BROOKVALE eendal C e re HEIGHTS ek Curl Curl Lagoon CURL CURL ALLAMBIE N

Manly NORTH ly n reek MANLY Dam a C M HARBORD

QUEENSCLIFF 0 0.5 1km

NORTH Scale BALGOWLAH Manly Lagoon MANLY VALE

Figure WRL LOCALITY 2.1

Report No. 2005/24 02092-2-1.cdr

PRECINCT 3

Figure WRL COLLAROY - NARRABEEN EMBAYMENT, NSW 2.2

Report No. 2005/24 02092-2-2.cdr FLIGHTFLIGHT DECK

WRL Figure PRECINCT 3 2.3

Report No. 2005/24 02092-2-3.cdr Figure WRL COASTAL EROSION 1950’s, 1970’s AND 2002 2.4 Report No. 2005/24 02092-2-4.cdr WRL TECHNICAL REPORT 2005/24 3-1

3. DESCRIPTION OF THE COLLAROY-NARRABEEN COASTAL IMAGING SYSTEM

3.1 What is Coastal Imaging?

'Coastal imaging' simply means the automated collection, analysis and storage of pictures, that are then processed and analysed to observe and quantify coastline variability and change.

Aerial photography has been the tool most commonly used by coastal managers to monitor regional-scale coastal behaviour. This is expensive and as a result, coverage is often ‘patchy’ and incomplete. Pictures are only obtained when visibility from the airplane is satisfactory, often resulting in a limited number of suitable pictures per year (at most), with no information about the behaviour of the beach between flights.

In contrast, with the recent development of digital imaging and analysis techniques, one or more automated cameras can be installed at a remote site and, via a telephone and internet connection, be programmed to collect and transfer to the laboratory a time-series of images. These images, taken at regular intervals every hour of the day for periods of months and years, can cover several kilometres of a coastline. Not every image need be subjected to detailed analysis, but by this method the coastal manager can be confident that all 'events' will be recorded and available for more detailed analysis as required.

3.2 The Difference between Coastal Imaging and a 'Surfcam'

At the core of the coastal imaging technique is the ability to extract quantitative data from a time-series of high quality digital images. In contrast, conventional 'surfcams' are very useful to applications where a series of pictures of the coastline is sufficient, and these types of images can be used to develop a qualitative description of coastal evolution.

The extraction of quantitative information from the coastal imaging system is achieved by careful calibration of the cameras and the derivation of a set of mathematical equations that are used to convert between two-dimensional image coordinates and three-dimensional ground (or 'real world') coordinates. Sophisticated digital image processing techniques are then applied to extract and quantify information contained within the images.

WRL TECHNICAL REPORT 2005/24 3-2

3.3 The ARGUS Coastal Imaging System

The ARGUS coastal imaging system has developed out of fifteen years of ongoing research effort based at Oregon State University, Oregon USA (Holman et al., 1993). A schematic of a typical ARGUS station is shown in Figure 3.1. The key component of an Argus station is one or more cameras pointed obliquely along the coastline. The camera(s) are connected to a small image processing computer (Silicon Graphics SGI workstation), which controls the capture of images, undertakes pre-processing of images, and automatically transfers the images via the internet from the remote site to the laboratory. The cameras installed at the Collaroy-Narrabeen site are fitted with high quality lenses. A switching interface between the cameras and computer maintains synchronisation of the captured images. The SGI workstation incorporates an internal analogue I/O card that enables all images to be captured, stored and distributed in standard JPEG digital image file format.

At WRL, a dedicated host computer (dual-processor Linux workstation) stores all images as they are received from the remote site within a structured archive. This workstation is also integrated to a world-wide-web server, with the images made available to all visitors to the web site to view and download within minutes of their capture and transfer from Collaroy- Narrabeen Beaches to WRL. Post-processing of the images is completed using a variety of Linux and PC computer hardware and custom image processing software within the MATLAB programming environment.

3.4 Design of Optimum Camera Layout for Collaroy-Narrabeen

Prior to the installation of the coastal imaging system at Collaroy-Narrabeen, modelling was undertaken to determine the optimum camera/lens specifications. The following criteria for the system were identified:

• Coverage to include as much as possible of the entire embayment, with particular emphasis on Precinct 3.

• Image resolution sufficient to enable 1 m or better resolution of the cross-shore position of the shoreline within Precinct 3.

• Cameras to be mounted at a single location only.

It was determined that a minimum of five cameras were required to obtain the necessary coverage. The results of this modelling are shown in Figure 3.2. In this figure the upper panel depicts the cross-shore pixel resolution of the optimized system, and the lower panel

WRL TECHNICAL REPORT 2005/24 3-3 depicts the longshore pixel resolution (the better the pixel resolution, the higher the accuracy of mapped shorelines along the Collaroy-Narrabeen beachfront). For reference, the modelled location of cameras is at the longshore coordinate y = 0 m, with the shoreline immediately in front of the cameras located at the cross-shore coordinate of approximately x = 0 m. The region being modelled therefore extends 3000 m alongshore from Collaroy ocean pool to North Narrabeen, for a distance of approximately 500 m offshore.

The conclusion from this modelling was that cross-shore pixel resolution (and hence position of the mapped shoreline) of better than 1 m could be achieved within the Precinct 3 study area. Longshore resolution deteriorates away from the cameras, however for the task of shoreline mapping and calculation of beach width, this is less critical. The lenses required to obtain this optimum layout require focal lengths (from north to south) of 12.5 mm, 9 mm, 6 mm, 9 mm and 12.5 mm respectively.

3.5 Installation

The ARGUS coastal imaging system was installed at Collaroy-Narrabeen in early July 2004. The system is located at an elevation of approximately 50 m above mean sea level, within the roof services area of the Flight Deck building (Figure 3.3). Flight Deck is located approximately 10 m - 20 m landward of the frontal dune, approximately 700 m to the north of the Collaroy ocean pool.

The cameras are mounted externally on a single frame that stands on the roof of the building, and are protected within weatherproof housings (Figure 3.4). The SGI workstation is housed within a pump services room, where 240 V power and a dedicated phone line connection to the internet are provided. The system is designed to run autonomously, and is self-recovering should an interruption to the mains power supply occur. Routine maintenance of the system is achieved by connection to the remote system via the internet from WRL. Occasional cleaning of the camera lenses is required.

3.6 Calibration

Fundamental to the use of image data to quantify coastal behaviour is the ability to convert between image coordinates (i.e., individual pixels) and real-world ground coordinates (Holman et al, 1993; Holland et al., 1995). For any particular object located by its three- dimensional (3-D) ground coordinates, the associated two-dimensional (2-D) image

WRL TECHNICAL REPORT 2005/24 3-4 location can be found uniquely using one of several transformation algorithms. The opposite process, the determination of the 3-D ground location of a 2-D image feature, is undetermined in a mathematical sense, and further information is needed.

A common photogrammetric solution is to use stereo techniques, requiring two or more independent cameras focussed on the point of interest from two or more different locations. Alternatively, several situations occur in which image features are naturally constrained. For example, at the coastline, waves can be assumed to propagate on a horizontal plane, the elevation of this plane determined by a local tide gauge.

A requirement for these calculations is knowledge of lens distortion and the camera location, field of view, azimuth, tilt, focal length and roll. While these latter parameters can be measured directly, a least-squares inverse algorithm – based upon control points at known ground and measurable image locations – is the preferred method to solve for geometric image parameters.

The full details of the camera model calibration technique are documented in Todd et al. (1995). Summarised below is the methodology applied at the Palm Beach site to derive the required geometric image parameters.

3.6.1 Lens/Camera Calibration

Video image acquisition requires the digitisation of analogue video signals as individual frames. This digitisation process involves the determination of image scale factors, which essentially determine the ‘squareness’ of pixels. The process of digital image capture is not perfect, and the error introduced is referred to as the ‘horizontal scale factor’. A pixel being slightly out of square is not important if the object of interest is close to the camera, but as objects of distances up to several 1000s of metres are important for coastal imaging, a small error will become very much magnified as the area covered by each individual pixel increases with increasing distance from the camera.

Lens distortion can also introduce significant errors. Radial symmetric lens distortion (distortion along radial lines from the centre of the image) has been shown to be the largest source of distortion error. In addition, the image centre coordinates are often chosen as the centre of the image frame, however, the true image centre for CCD cameras can deviate significantly from this position.

WRL TECHNICAL REPORT 2005/24 3-5

To calculate the above intrinsic camera model parameters (i.e., horizontal scale factor, lens distortion and image centre error), the cameras and lenses were individually calibrated in the laboratory to determine the appropriate calibration factors. To achieve this, each camera / lens pair was mounted in an optical rail and focussed on a control-point test field, comprising regularly-spaced white circles on a black background (Figure 3.5). Calibration parameters were then derived using proprietary software to solve for the deviations between predicted and observed image coordinates.

3.6.2 Ground Control Point (GCP) Survey

Having installed the cameras at the Flight Deck building, the remaining extrinsic camera model parameters were obtained by the completion of a detailed survey. The objective of such a survey is to identify ground control points (GCPs) in the camera fields of view, as well as the precise locations of cameras. This survey was undertake using high-accuracy RTK-GPS (‘real time kinematic’ global positioning system) equipment.

By constraining the camera coordinates in 3-D space to their surveyed position, the system of equations used to solve for the problem of converting between image and world coordinates can be uniquely determined given two GCPs only. Having more control points leads to an over-determined system in the mathematical sense, with redundant information that can be solved in a least-squares sense.

GCP surveys of control points was completed by the WRL survey team moving across the beach within the field of view of each of the five cameras to create temporary GCPs, which were surveyed at the same time that images of the beach were captured. To illustrate, Figure 3.6 shows a composite image of all mobile GCPs surveyed within the field of view of camera C5 (north) and camera C1 (south). By this method, GCPs and resulting camera geometries were obtained for all five cameras.

WEB WORLD WIDE WORLD Y orkstation OR T TER RESEARCH LABORA A W - image archive - image post-processing - web server (image distribution) Linux Dual-Processor W Internet Modem Camera 5 Camera 4 Camera 1 orkstation REMOTE SITE REMOTE - image capture - image pre-processing SGI W ideo (Flight Deck Building) (Flight Camera 2 Camera 3 Interface A/D V

WRL Figure SCHEMATIC OF AN ARGUS COASTAL IMAGING SYSTEM 3.1

Report No. 2005/24 02092-3-1.cdr WRL Figure DESIGN OF OPTIMUM FIVE-CAMERA LAYOUT 3.2

Report No. 2005/24 02092-3-2.cdr Figure WRL LOCATION OF COASTAL IMAGING SYSTEM 3.3 FLIGHT DECK BUILDING Report No. 2005/24 02092-3-3.cdr Figure WRL CAMERAS MOUNTED AT AN ELEVATION 3.4 OF APPROXIMATELY 50m Report No. 2005/24 02092-3-4.cdr Figure WRL CALIBRATION OF INTRINSIC (CCD-LENS) CAMERA MODEL PARAMETERS 3.5 Report No. 2005/24 02092-3-5.cdr Figure WRL GCP SURVEY (COMPOSITE IMAGES - CAMERA 1 AND CAMERA 5) 3.6 Report No. 2005/24 02092-3-6.cdr WRL TECHNICAL REPORT 2005/24 4-1

4. IMAGE TYPES AND IMAGE PROCESSING TECHNIQUES

4.1 Image Types

The ARGUS coastal imaging system installed at Collaroy-Narrabeen is presently configured to collect three different types of images on a routine hourly basis. A fourth image type is created by automated post-processing at the completion of each day of image collection.

Images are collected every daylight hour. The image collection procedure is fully automated and controlled by the SGI workstation at the remote site. Prior to commencing the hourly image collection routines, a test is undertaken to determine if there is sufficient daylight to proceed with image collection. If the ambient light threshold is exceeded, image collection commences. The reason for first checking for daylight conditions is to avoid unnecessary image collection at night, without excluding image collection earlier in the morning and later in the evening during extended summer daylight hours.

4.1.1 Snap-Shot 'snap' Images

The simplest image type is the snap-shot image. This is the same image obtained if a picture of the beach were taken using a conventional digital camera. Snap-shot images provide simple documentation of the general characteristics of the beach, but they are not so useful for obtaining quantitative information. An example of a snap image obtained in late June 2005 is shown in Figure 4.1 (upper panel).

4.1.2 Time-Exposure 'timex' Images

A much more useful image type is the time-exposure or 'timex' image. Time-exposure images are created by the 'averaging' of 600 individual snap-shot images collected at the rate of one picture every second, for a period of 10 minutes.

A lot of quantitative information can be obtained from these images. Time exposures of the shore break and nearshore wave field have the effect of averaging out the natural variations of breaking waves, to reveal smooth areas of white, which has been shown to provide an excellent indicator of the shoreline and nearshore bars. In this manner, a quantitative 'map' of the underlying beach morphology can be obtained. An example of a timex image is shown in Figure 4.1 (middle panel).

WRL TECHNICAL REPORT 2005/24 4-2

4.1.3 Variance 'var' Images

At the same time that the timex images are being collected, an image type called a variance or 'var' image is also created. Whereas the time-exposure is an 'average' of many individual snap-shot images, the corresponding variance image displays the variance of light intensity during the same 10 minute time period.

Variance images can assist to identify regions which are changing in time, from those which may be bright, but unchanging. For example, a white sandy beach will appear bright on both snap-shot and time-exposure images, but dark in variance images. Because of this, other researchers have found that variance images are useful at some specific coastal sites for analysis techniques such as the identification of the shoreline, as the (bright) changing water surface is readily identifiable against the (dark) beach. An example of a var image is shown in Figure 4.1 (lower panel).

4.1.4 Day Time-Exposure 'daytimex' Images

The fourth image type routinely created from the coastal imaging system installed at Collaroy-Narrabeen is referred to as a daytimex image. It is created at the end of each day of image collection, by the averaging of all hourly timex images collected that day. This has the effect of 'smoothing' the influence of tides, and for some conditions may enhance the visibility of the shore break and bar features in the nearshore.

4.2 Basic Image Processing – Merge and Rectification

As noted earlier in Section 3.2, the key feature of coastal imaging technology that distinguishes it from conventional webcam systems is the ability to extract quantitative information from the images. As described previously, this is achieved through the solution of the camera model parameters to extract 3-D real-world position from 2-D image coordinates, and the application of image processing techniques to identify, enhance and manipulate the image features of interest.

Image merging is achieved by the solution of camera model parameters for individual cameras, then the boundaries of each image are matched to produce a single composite image. Image rectification is then undertaken, whereby the dimensions of the merged image are corrected so that each pixel represents the same area on the ground, irrespective of how close to or how far from the camera position it may be. (In contrast, for an

WRL TECHNICAL REPORT 2005/24 4-3 unrectified image the area represented by each pixel increases with increasing distance from the camera).

Image rectification is achieved by using the calculated camera model parameters to fit an image to a regular grid that defines longshore and cross-shore distance. The rectification of merged images produces a 'plan view' of the area covered by all five cameras. This is illustrated in Figure 4.2. Also shown in this figure is a pan image, which provides an alternative wide-angle (but distorted) image of the coastline. The merged and rectified plan image created from five oblique images is analogous to a montage of distortion-corrected photographs taken from an airplane flying directly overhead the Collaroy-Narrabeen embayment. For convenience, the longshore and cross-shore dimensions of this image are referenced (in metres) to the location of the cameras. The pixel resolution of the merged/rectified images created at Collaroy-Narrabeen is 5 m; that is, a single pixel represents an area 5 m × 5 m on the ground.

4.3 Shoreline Detection and Analysis

To map the position of the shoreline and its changing location through time, a rigorous image analysis methodology is required to enable the extraction of this information from the database of hourly ARGUS images.

4.3.1 Overview of the ‘PIC’ shoreline identification technique

Comprehensive description of the PIC shoreline identification technique is provided in Aarninkhof (2003), Aarninkhof and Roelvink (1999) and Aarninkhof et al (2003). Briefly, the technique aims to delineate a shoreline feature from 10 minute time exposure images, on the basis of distinctive image intensity characteristics in pixels, sampled across the sub- aqueous and sub-aerial beach. Raw image intensities in Red-Green-Blue (RGB) colour- space, sampled from a region of interest across both the dry and wet beach, are converted to Hue-Saturation-Value (HSV) colour space, to separate colour (Hue, Saturation) and grey scale (Value) information. The HSV intensities are filtered to remove outliers and scaled between 0 and 1, to improve the contrast between two clusters of dry and wet pixels. Iterative low-passing filtering of the spiky histogram of scaled intensity data yields a smooth histogram with two well-pronounced peaks Pdry and Pwet, which mark the locations of the two distinct clusters of dry and wet pixels (Figure 4.3).

WRL TECHNICAL REPORT 2005/24 4-4

The filtered histogram is used to define a line to distinguish between Hue Saturation information used for colour discrimination (Figure 4.3a), or Value information in the case of luminance-based discrimination (Figure 4.3b). For both discriminators, the line defined in this manner crosses the saddle point of the filtered histogram, and thus provides the means to separate objectively the two clusters of dry and wet pixels within the region of interest. With the help of this line, a discriminator function Ψ is defined such that Ψ = 0 along this line (see Figure 4.3). The areas of dry and wet pixels are then mapped, and the boundary between the two regions defines the resulting shoreline feature of interest.

4.3.2 Standardised Procedure for Shoreline Mapping

To obtain the highest possible accuracy of measurements of the shoreline position using digital images obtained at Collaroy-Narrabeen, the oblique images (rather than lower- resolution merged-rectified images) from each of the five cameras are analysed separately. By this approach, the sub-metre cross-shore pixel resolution information contained within these images (refer Section 3.4) is fully exploited. Subsequent rectification and merging of the individual shoreline segments obtained from each camera is then used to produce a continuous shoreline alongshore.

The procedure used to map the shoreline within Precinct 3 at Collaroy-Narrabeen is summarised in Figure 4.4. At weekly (nominal seven day) intervals, observed tide information is used to determine the hourly timex images that correspond to mid-tide (0 m AHD). The database of wave information is also searched to determine the rms ('root mean square') wave height (Hrms) and spectral peak wave period (Tp) that correspond to these daily mid-tide images.

Based on a seven day cycle, the corresponding mid-tide images are checked to confirm that the wave height satisfies the low-pass criteria Hrms ≤ 1.0 m (~Hsig ≤ 1.4 m). This wave height criteria was used for shoreline mapping as, above this wave height, wave runup at the beach face increases and the width of the swash zone widens, introducing a corresponding uncertainty in the cross-shore position of the waterline. If the rms wave height is less than 1.0 m, then the shoreline is mapped. If the Hrms wave height exceeds the 1.0 m threshold, then the mid-tide images for the preceding day is checked. If these images still do not satisfy the wave height criteria, then the following day's mid-tide images are checked. This process is repeated for up to ± 3 days, to locate the mid-tide images for which the rms wave height did not exceed 1.0 m. If no mid-tide images are available in any one seven day cycle that satisfy this criteria, then no shoreline is mapped for that week.

WRL TECHNICAL REPORT 2005/24 4-5

Once the mid-tide images to be processed have been identified, the PIC method is applied and the shoreline feature is mapped. Beach width is then calculated relative to the alignment of the previously proposed seawall (refer Section 2.2), which is approximately equivalent to the dune line. By repeating this procedure every seven days, a growing database is developed that contains the time-series of weekly shoreline positions within the Precinct 3 study area. These data are then subjected to a range of quantitative analyses as described in the following Sections 7 & 8.

snap

timex

var

Figure WRL SNAP-SHOT, TIME-EXPOSURE AND VARIANCE IMAGE TYPES (29/06/2005) 4.1 Report No. 2005/24 02092-4-1.cdr Figure WRL PAN AND PLAN VIEW IMAGES: FIVE-CAMERA MERGE/RECTIFICATION (29/06/2005) 4.2 Report No. 2005/24 02092-4-2.cdr Source: Aarninkhof (2003) Figure WRL IDENTIFICATION OF ‘SHORELINE’ FEATURE 4.3 FROM COLOUR IMAGES Report No. 2005/24 02092-4-3.cdr Narrabeen - Collaroy Coastal Imaging System

Gold Coast create daily merged/rectified tide data image at mid tide

determine corresponding Gold Coast wave conditions wave data

does image select image No satisfy wave for proceeding/ height threshold? preceeding day (Hs <1m)

Yes

MAP SHORELINE

WRL Figure STANDARDISED SHORELINE MAPPING PROCEDURE 4.4

Report No. 2005/24 02092-4-4.cdr WRL TECHNICAL REPORT 2005/24 5-1

5. COASTAL IMAGING WEB SITE

5.1 Coastal Imaging Home Page

To promote the dissemination of information about the Collaroy-Narrabeen coastal monitoring project, to provide a convenient means to distribute images as they are collected, and to facilitate ‘real-time’ access to the regularly-updated results of shoreline monitoring and beach width analysis, a coastal imaging project site was established on the world-wide web at the following address:

Æ www.wrl.unsw.edu.au/coastalimaging/narrabn

The Collaroy-Narrabeen coastal imaging home page is shown in Figure 5.1. The most recent snap images are displayed here and updated every hour, enabling visitors to the site to observe the current beach conditions. This page also includes a number of links to a variety of background information including a description of the coastal imaging system, image types and image processing techniques. Links are also provided to the Warringah Council web site, wave monitoring, local weather conditions, and tide predictions.

For general interest, a record is maintained of the number of visitors to the WRL coastal imaging web site and the countries they are from. At the time of writing, approximately 156,000 hits to the main WRL coastal imaging web pages have been recorded, with around 5,500 hits to the Collaroy-Narrabeen site since it first came on-line in July 2004. Visitors from Australia account for approximately half the total visitors, with the remaining visitors coming from approximately 80 countries world-wide.

5.2 Image Archive

The current snap, timex images and var images are updated and available at the project web site every hour.

All present and past images can be accessed via the on-line image archive. This provides a convenient and readily navigable structure to quickly locate the image(s) of interest. Figure 5.2 shows an example of a daily page contained within the image archive. These images are provided freely to encourage their use by students, researchers, managers and other non-commercial organisations.

WRL TECHNICAL REPORT 2005/24 5-2

5.3 On-Line ‘Beach Analysis System’

On-line access to ‘real time’ beach monitoring and analysis is made available at the Collaroy-Narrabeen coastal imaging web site. This capability results from the on-going research and development effort underway by the coastal imaging team at WRL. The purpose of this system is to provide regularly-updated results of the beach monitoring program to Warringah Council and the general public on a routine basis, via the world wide web.

A detailed description of the capabilities of this system was detailed in Anderson et al (2003). To summarise, the features available at the project web site include the ability to view the latest mid-tide plan images; access to a zoom tool feature that enables zooming-in and panning through the current oblique and rectified images; full on-line access to all past and present monitoring reports; and two products specifically designed to assist both the qualitative and quantitative interpretation of images, shoreline data and the results of beach width analysis.

An example of the first of these products called ‘week-to-a-page’ is illustrated in Figure 5.3. Every Monday morning, this figure is generated and made available for viewing (and download, if required) via the project web site. The figure is pre-formatted to fit on a standard A4 page, to assist reporting. This figure compiles daily mean sea level plan view images of the Collaroy-Narrabeen embayment for that week, into a compact one-page summary. This product provides coastal managers a means of quickly and efficiently interpreting the daily changes in beach morphology and shoreline position, without continual recourse to the hourly images. An archive of these weekly figures is also maintained and available on-line.

The second product that is also updated each Monday morning and made available via the project web site is ‘Beach-Width-Analysis’ (Figure 5.4). This figure in graphical format summarises quantitative information of the mean shoreline position for that week; shoreline variability by comparing the current shoreline position with previous weeks and months; beach width along pre-defined monitoring transects; and beach width trends throughout the history of the monitoring project.

WRL Figure COASTAL IMAGING WEB SITE - HOME PAGE 5.1

Report No. 2005/24 02092-5-1.cdr WRL Figure COASTAL IMAGING SYSTEM WEB SITE - IMAGE ARCHIVE 5.2

Report No. 2005/24 02092-5-2.cdr Figure WRL ON-LINE BEACH ANALYSIS SYSTEM ‘WEEK-TO-A-PAGE’ 5.3 Report No. 2005/24 02092-5-3.cdr Figure WRL ON-LINE BEACH ANALYSIS SYSTEM ‘BEACH WIDTH ANALYSIS’ 5.4 Report No. 2005/24 02092-5-4.cdr WRL TECHNICAL REPORT 2005/24 6-1

6. MORPHODYNAMIC DESCRIPTION OF COLLAROY-NARRABEEN: JULY 2004 – JUNE 2005

From the daily images obtained by the ARGUS coastal imaging station atop of the Flight Deck building, it is self-evident that Collaroy-Narrabeen beaches are dynamic and continually changing. Bars move onshore and offshore and vary in shape from straight to crescentic, rips emerge and disappear, and the shoreline changes shape and translates landward and seaward in response to varying wave conditions. This section is included to provide a qualitative description of the observed beach changes during the past 12-month monitoring period July 2004 to June 2005. The ‘week-to-a-page’ summary figures that are updated every week and made publicly available for inspection and download via the project web site, are used in this section to illustrate the observed beach changes. The objective is not to describe every characteristic of Collaroy-Narrabeen beaches during this period, but rather provide an overview of general trends and predominant features that were observed during this time.

To summarise beach changes in some structured manner, it is useful to first outline a systematic beach classification scheme with which to undertake this qualitative analysis.

6.1 A Morphodynamic Classification of Beaches

Despite the seemingly endless range of changes observed at any sandy coastline, in fact it has been shown that beaches tend to exhibit certain characteristics that vary in a systematic and predictable way. One such scheme for describing these changes is the 'Morphodynamic Beach State Model' first outlined by Wright and Short (1983). This beach classification scheme was developed in Australia, and is now the most widely-used descriptive beach model internationally. The term 'morphodynamics' derives from the combination of the words 'morphology' and 'hydrodynamics', emphasising the strong linkage between the shape of a beach and the associated wave and current conditions.

Beaches can be classified as being in one of six beach 'states' at any given point in time. The generalised cross-section and plan form characteristics of these six beach states are summarised in Figure 6.1. A brief description of each of these states is provided below.

At one extreme is the dissipative beach state (Figure 6.1a), which is characterised by a very low profile slope and wide surf zone. Dissipative beaches are generally composed of fine sand and occur along coastlines exposed to high wave energy. Nearshore bathymetry

WRL TECHNICAL REPORT 2005/24 6-2 is usually characterised by one or more straight and shore-parallel bars. The term 'dissipative' is used to describe beaches that exhibit these characteristics because wave energy is essentially dissipated by extensive wave breaking across the surf zone, before it can reach the shoreline.

At the other end of the beach state spectrum, reflective beaches (Figure 6.1f) are invariably steep, with no nearshore bars. Waves tend to break close to or right at the shoreline, and hence very little wave energy is dissipated; instead it is reflected by the beach face and propagates offshore. These beaches tend to be composed of coarse sediments and/or are generally located in protected or low wave energy coastal regions.

Between the dissipative and reflective extremes, four intermediate beach states can be identified. These incorporate elements of both the reflective and dissipative domains. The four intermediate beach types are referred to as longshore bar-trough LBT (Figure 6.1b), rhythmic bar and beach RBB (Figure 6.1c), transverse bar and rip TBR (Figure 6.1d) and low tide terrace LTT (Figure 6.1e). Together, these intermediate beach types form a sequence of characteristic beach states related to the movement of sand onshore (decreasing wave steepness) and offshore (increasing wave steepness). The onshore-offshore movement of sand is most easily recognised by the movement and changing shape of bars within the nearshore zone.

Following the characteristic offshore movement (ie., erosion) of sediment during a major storm, typical post-storm beach recovery includes the gradual onshore migration of nearshore bars and the development of weak and then stronger rips (LBT Æ RBB Æ TBR). If low wave conditions persist, bars ultimately disappear as the bar becomes welded to the beach to form a terrace (LTT). Beaches of the moderately high energy east Australian open coast are typically observed to transfer between these four intermediate morphodynamic beach states, in response to lower wave conditions interspersed by episodic storm events.

6.2 Morphodynamic Interpretation of Daily Images

All week-to-a-page figures for the period July 2004 to June 2005 are presented in Appendix A. Each of these figures shows a week (seven days) of sequential mid-tide plan images, with the date of each indicated. All images are obtained at the same stage of the tide (mean sea level), to enable the direct comparison between different days and weeks. The region shown in these figures extends 3000 m alongshore, from Collaroy ocean pool north along North Narrabeen beach.

WRL TECHNICAL REPORT 2005/24 6-3

To assist the interpretation of these images, Appendix B contains monthly summaries of wave height and period, obtained from the Sydney Waverider buoy and supplied to WRL by Manly Hydraulics Laboratory.

6.2.1 July 2004

The month of July was dominated by a large storm event commencing the 18th July, when offshore significant wave heights rose rapidly from around 1 m up to 7 m, with the corresponding maximum offshore wave height during this time recorded at 10.3 m. The Collaroy-Narrabeen ARGUS station became operational on the 14th of July immediately prior to the July storm. At that time the central embayment exhibited a complex TBR-RBB intermediate beach state, with the beach to the south characterised by the single nearshore bar observed closer to the shoreline, and the beach grading to increasingly dissipative morphology northward.

The storm peaked on the 19th and then declined rapidly through to the 21st, when offshore significant wave heights returned to around 2 m. By the 24th the offshore significant wave height had reduced further to around 1 m, with these milder conditions generally persisting through to the end of the month.

During the peak of the storm a wide and highly energetic surf zone was observed, but no clear evidence of beach erosion was discernable. Further, by the 21st of July when the single nearshore bar was visible again due to the declining wave energy, the beach and nearshore morphology throughout the Collaroy-Narrabeen embayment remained similar to the conditions that prevailed prior to the high energy storm. Along the northern half of the embayment the single bar had moved a short distance further seaward, but a complex TBR- RBB beach state was still dominant within the central embayment. It is perhaps initially surprising that this intermediate beach state persisted through the high energy storm event. However, the explanation for this observation can be found by examining the characteristics of this storm more closely. During the large wave event the peak wave period increased to 15 seconds, indicating that the storm was associated with very long- period swell waves. So despite the high wave energy, the associated long wave periods caused the storm waves to break further seaward than is usual, and effectively dissipate their energy prior to impacting upon the beachface and inner surf zone.

WRL TECHNICAL REPORT 2005/24 6-4

6.2.2 August 2004

The wave climate through the month of August exhibited two distinct phases. Through the first two weeks of the month offshore significant wave heights were generally in the range of 1.5 – 2.5 m, declining briefly on the 13th – 14th before rising up to 5 m on the 15th. Wave heights then rapidly declined, so that the second half of the month was characterised by offshore significant wave heights generally in the range of 0.5 – 1.5 m. As per the previous month, August was characterised by generally swell wave conditions, with peak wave periods typically around 12 seconds.

Complex bar and rip features continued to characterised the central region of the Collaroy- Narrabeen embayment, as the sandbar moved landward to form TBR morphology. To the south, more reflect beach conditions prevailed, with closely-spaced beach cusps persisting in this region. During the second half of the month the relatively mild wave energy conditions caused the RBB-TBR morphology to spread progressively northward, as rips formed and the nearshore bar developed an increasingly crescentic form. The beach characteristics at this time were indicative of generally accretionary beach conditions.

6.2.3 September 2004

Wave data for a two week period in mid September was not available due to operational issues with the Sydney wave-rider buoy at that time. However, the wave conditions that prevailed prior to and following this period, along with the daily observed beach images available throughout the month, are all indicative of generally mild and accretionary beach conditions.

The offshore significant wave height rose briefly to around 3 m on the 2nd, then for the reminder of the month (where data were available) remained in the range of 1 – 2 m. TBR beach state morphology prevailed through the first half of the month, with multiple and complex rip systems along the beachfront. As sand moved onshore through the surf zone the single bar system welded to the beachface, and by the second half of the month the central and southern beach had accreted toward a LTT beach state. This accretion was accompanied by the growth of closely-spaced cusps along the entire southern half of the Collaroy-Narrabeen embayment.

6.2.4 October 2004

The wave climate conditions favouring beach accretion that had prevailed in the proceeding months were interrupted in October by the occurrence of four storm events, all occurring

WRL TECHNICAL REPORT 2005/24 6-5 within the one month period. During three of the four storms the offshore significant wave heights peaked in the range of 5 – 6 m, with the fourth and smaller storm peaking at around 3 m. the maximum wave heights recorded during these storms were up to 11.3 m. Peak wave periods during these storm events were generally around 10 – 11 seconds, resulting in high energy and erosive conditions along the Collaroy-Narrabeen beachfront.

From the lower-energy intermediate beach states that were present at the beginning of the month, by the end of October the beach had changed dramatically. The complex and three- dimensional terrace, rip channel and transverse-bar features had been completely removed, as sand moved offshore from the beachface and seaward through the surf zone. A linear LBT offshore bar system developed, separated from the beachface by a wide and linear trough. The offshore bar was formed by the removal of sand from the upper beach and inshore, resulting in the accumulation of this eroded sediment across the outer surf zone. The observed rapid transition in beach state through October from the prevailing lower intermediated LTT-TBR beach states to a high energy LBT beach state, is indicative of significant erosion of the beachface and nearshore. This erosion was caused by the succession of closely-spaced storms during the four week period, with little opportunity for recovery between these events.

6.2.5 November 2004

In contrast to the preceding month, November was characterised by mild to moderate wave energy conditions, with offshore significant wave heights generally in the range of 1 – 1.5 m. Around the 4th – 6th and again on the 20th and 23rd-25th offshore significant wave height rose above 2 m, but declined again within the following 24 hour periods. Peak wave periods decreased from proceeding months, varying through November in the range of 5 – 10 seconds.

The decline in wave energy caused the initially linear LBT morphology to transition to a more complex and lower energy RBB state by around the 10th, as the sand bar moved landward and began to develop crescent features. The bar continued to more shorewards, and by the 20th had welded to the beachface in places within the central region of the beach, while remaining further seaward in the northern (and more exposed) region of the embayment. The TBR morphology that had emerged within the central region of the embayment then persisted through to the end of the month, characterised by a series rips and channels interspersed by sand bars.

WRL TECHNICAL REPORT 2005/24 6-6

6.2.6 December 2004

Wave energy increased again in December, with a series of moderate storms occurring at approximately weekly intervals through the month. During each of these storms offshore significant wave heights peaked in the range of 3 – 4 m, with maximum wave heights up to 7 m. Early in the month shorter (and more erosive) peak wave periods were in the range of 5 – 10 seconds, but by around the 12th this had generally stabilised at around 9 - 11 seconds, which generally persisted to the end of December.

In response to the first storm commencing around the 7th (note incomplete wave information at this time, due to a problem with the waverider buoy), erosion of the beachface and nearshore was observed. By the 10th the previous TBR beach state had transitioned to the higher-energy LBT morphology, indicative of the seaward movement of sediment to form the linear, offshore bar.

By the 20th the bar had begun to exhibit accretionary crescentic features again, which increased during the following two week period as RBB morphology emerged. By the end of December the central region of the beach was again characterised by a series of rip channel features, while along the northern embayment the bar remained more linear and separated from the beachface by a near-continuous trough.

6.2.7 January 2005

Wave information is unfortunately available only for the first half of January, due to a fault with the Sydney waverider buoy. During the two week period for which wave data is available, mild to moderate wave energy conditions prevailed, with offshore significant wave heights in the range 1 – 2 m.

From visual inspection of the daily mid-tide plan images of the Collaroy-Narrabeen embayment, increased wave energy conditions occurred around the 15th and again on the 20th, causing the bar to briefly detach fully from the beachface to form LBT morphology, characterised by a linear trough separating the bar from the beach. Following these two events, by around the 24th a RBB state had developed, that transitioned to TBR two days later as sand moved onshore again, and the bar began to weld to the beachface. These lower energy and accretionary conditions prevailed to the end of the month, with multiple rip systems occurring alongshore.

WRL TECHNICAL REPORT 2005/24 6-7

6.2.8 February 2005

The Sydney waverider buoy remained out of operation until the 16th February, from which time the wave climate remained mild, characterised by offshore significant wave heights generally in the range of 1 – 1.5 m, and peak wave periods rising to around 12 seconds, before decreasing to 5 – 10 seconds in the last few days of the month.

The mild wave energy conditions resulted in little change to the prevailing beach conditions through February, with the beach remaining in a lower energy and accretionary TBR state. The bar generally progressed onshore, so that by the last week of February the bar had largely welded to the beachface, with the emergence of LTT morphology characterised by small rip channel systems cutting seaward across the low tide terrace feature.

6.2.9 March 2005

The wave climate at the Collaroy-Narrabeen embayment in March was characterised by an initial 2.5 week period of mild and accretionary conditions, followed by a higher energy period of beach erosion. Up to the 17th the incident wave energy was mild, with offshore significant wave heights generally around 1 m and rising to 2 m intermittently, with the peak wave period increasing to 10 – 12 seconds during this time. After the 17th wave heights rose, peaking with the occurrence of a storm on the 23rd – 24th, when offshore significant wave heights peaked at over 6 m, and maximum offshore wave heights up to 9.5 m were recorded. During the peak of the storm the peak wave period had decreased to around 9 – 10 seconds, indicating erosive conditions during this time.

The lower energy and accretionary LTT beach state persisted through the first half of the month, with small and decreasing numbers of rips cutting across the narrow surf zone. The increase in wave energy around the 17th followed by the onset of the storm around the 22nd cased the bar to detach and move seaward through the surf zone, as the beach transitioned to a higher-energy RBB-TBT state, indicative of the movement of a significant volume of sand offshore. The detached bar was separated from the beachface by an initially linear trough, which as the storm passed and incident wave energy decreased again, along with the bar developed increasingly rhythmic and crescentic features. By the end of the month, the bar had moved landward and partially welded to the beachface again, resulting in the emergence of an intermediate RBB-TBR beach state.

WRL TECHNICAL REPORT 2005/24 6-8

6.2.10 April 2005

Mild wave conditions prevailed throughout all but the first few days of April. On the 4th offshore significant waves rose for a period of a few hours to 4 m, but then rapidly declined again. For the remainder of the month offshore significant wave heights were typically in the range of 1 – 1.5 m, with the peak wave period in the general range of 8 – 12 seconds. Accretionary conditions prevailed along the Collaroy-Narrabeen embayment, with the bar continuing to move onshore as the beach state progressed from RBB-TBR to lower energy LTT by the end of the month. Rips became closer together and smaller in scale, then largely disappeared from the embayment, replaced by extensive beachface cusp development along the southern and central regions of the embayment.

6.2.11 May 2005

Mild wave conditions continued through the first two weeks of May, but then rose again for the remainder of the month, with the beach reverting from accretionary to erosive conditions. Offshore significant wave heights peaked at around 4 m on at least three occasions (data gaps due to missing data from the Sydney waverider), with maximum offshore wave heights up to 6 m at these times. Peak wave periods were variable during the latter higher-energy conditions, ranging from 8 seconds up to 17 seconds.

The onset of higher wave energy cased the bar to again detach from the beachface, resulting in the formation of a trough as sand moved offshore. In the central region RBB morphology emerged, with a more dissipative LBT state towards the more exposed northern end of the embayment. This morphology persisted to the end of the month, with large rip channel systems characterising the central region of Collaroy-Narrabeen, grading to a more linear bar-trough system toward the north.

6.2.12 June 2005

Through the first three weeks of June moderate to mild wave energy conditions prevailed, resulting in the general landward movement of sediment in the surf zone. Rips decreased in size and the sand bar moved toward the shore, with the beach state again reverting to a lower intermediate TBR state. During this time the offshore significant wave heights were typically in the range of 1 – 2 m, dipping to 0.5 m from the 20th to the 23rd. Peak wave period was generally around 11 – 13 seconds, characteristic of accretionary conditions during this time.

WRL TECHNICAL REPORT 2005/24 6-9

On the 24th June offshore significant wave heights rose rapidly to exceed 5 m, with offshore maximum wave height up to 8 m. Offshore significant wave heights then decreased to 1 m by the 29th, before rising steeply again to exceed 4 m by the end of the following day. During these high wave events, peak wave periods dropped to around 10 seconds, which combined with the prevailing storm wave heights, was indicative of the onset of erosive conditions. Sand moved rapidly offshore as the bar detached fully from the shoreline to exhibit a linear and uniform-alongshore LBT state at the end of the month. The linear sandbar was separated from the beachface by a wide a deep trough, with all rips and associated channels completely removed from the nearshore.

6.3 Visual Assessment of Net Beach Width Changes (July 2004 – June 2005)

The first stormy period commenced with the occurrence of a large storm event in mid July 2004, when offshore significant wave heights rose to 7 m. Moderate wave conditions through the first half of August peaked with the occurrence of a second storm in the middle of the month (offshore significant wave heights up to 5 m). Wave energy then declined, with generally mild conditions through the second half of August, and throughout September. The second stormy period commenced in October, with the occurrence of four storm events in the single month, when offshore significant wave heights up to 6 m were recorded. In contrast, November again saw a return to mild conditions, followed by a series of moderate storms in December, when offshore significant wave heights of 3 – 4 m were observed. Moderate wave conditions prevailed through January 2005, followed by generally milder conditions in February. The third stormy period commenced in the second half of March, when offshore significant wave heights again exceeded 6 m. Mild to moderate wave conditions return in April and continued through to late June, interrupted in mid May by a period of elevated wave energy. At the conclusion of this present 12 month monitoring period wave heights again increased, suggesting the commencement of a fourth period of higher wave energy and corresponding beach erosion.

WRL TECHNICAL REPORT 2005/24 6-10

Figure 6.2 shows the snap images obtained at mid-tide from Camera 1 (south) on 15/07/04 and 29/06/05 respectively. The corresponding snap images of the northern embayment obtained from Camera 5 are shown in Figure 6.3. Along the southern beach (Figure 6.2) the width of the subaerial (or ‘dry’) beach in June 2005 appears very similar to the conditions that prevailed 12 months earlier in July 2004. No retreat of the dune is apparent, and a modest widening of the beach in the vicinity of the Collaroy Surf Club is discernable. The general alignment of the beach in this region appears unchanged. Looking north along the Collaroy-Narrabeen embayment (Figure 6.3), again the width of the subaerial beach appears comparable from July 2004 to June 2005, however a general trend of beach flattening is observed. The scarped berm that was present in July 2004 is absent, with the beach exhibiting a flatter and more uniform profile in June 2005. The sand coverage over the existing sections of boulder wall within Precinct 3 in places has increased, with section of the rock no longer visible. However, elsewhere along this section of the beachfront the exposure of rocks at the toe of structures is more apparent.

A quantitative and more detailed assessment of the response of the beach within Precinct 3 for the 12-month period July 2004 to June 2005, is detailed in the following Sections 7 & 8.

Figure WRL MORPHODYNAMIC BEACH STATE MODEL (after WRIGHT and SHORT, 1983) 6.1 Report No. 2005/24 02092-6-1.cdr Figure WRL SNAP IMAGES FROM CAMERA 1 (SOUTH): 15/07/2004 AND 29/06/2005 6.2 Report No. 2005/24 02092-6-2.cdr Figure WRL SNAP IMAGES FROM CAMERA 5 (NORTH): 15/07/2004 AND 29/06/2005 6.3 Report No. 2005/24 02092-6-3.cdr WRL TECHNICAL REPORT 2005/24 7-1

7. QUANTITATIVE ANALYSIS OF SHORELINE CHANGES: JULY 2004 – JUNE 2005

The primary function of the coastal imaging system installed at the Flight Deck building is to quantify shoreline changes and beach variability within Precinct 3, in order to better understand the behaviour of this most vulnerable section of the Collaroy-Narrabeen embayment. Quantitative analysis of weekly shoreline position and beach width provide an objective measure to assess the beach amenity and storm buffer seawards of the existing property boundaries.

7.1 Weekly Shorelines

All weekly shorelines for the period 01/07/04 to 30/06/05 are shown in Figure 7.1. A total of 50 mid-tide shorelines were mapped during this period (for shoreline mapping method and procedure refer Section 4.3). For reference, these measured shorelines are overlaid onto a representative merged/rectified timex image (image date: 29th June, 2005). The image represents a 3500 m length of the Collaroy-Narrabeen embayment, extending from the Collaroy ocean pool in the south to North Narrabeen in the north. The ARGUS station is located at coordinate [0,0], with the black ‘shadow’ region in this vicinity caused by the region of beach immediately in front of and adjacent to the building being outside the cameras’ field of view. For reference, the crest alignment of the back-beach boulder wall (subsequently rejected by the community in 2003 - refer Section 2.2) is here used as the landward reference line for the calculation of beach width, and on Figure 7.1 is also indicated by the landward-most red line.

To see more clearly the range of shoreline positions mapped during this twelve month period, Figure 7.2 shows a plot of the position of the weekly shorelines within Precinct 3, relative to the proposed boulder-wall alignment. The distance of these shorelines from the wall alignment is plotted in the upper panel, and for convenience the alongshore position in this figure is relative to the location of the ARGUS station (0 m). In the lower panel of this figure the same mid-tide timex image used in the previous figure is shown for reference. Note that, due to the Flight Deck building being located so close to the beachfront, on all but a limited number of occasions, the shoreline between approximately ±100 m alongshore could not be mapped, due to it being obscured from the cameras by the edge of the building.

WRL TECHNICAL REPORT 2005/24 7-2

During the present monitoring period 01/07/04 – 30/06/05 it can be seen from Figure 7.2 that the beach along the Precinct 3 oceanfront varied in width (relative to the alignment of the proposed boulder wall) from approximately 60 m at the northern end, to 0 m at the location 300 m north of the Flight Deck building. The envelope of beach width changes at any particular point along the entire embayment was in the range of 30 - 40 m during this period.

7.2 Shoreline Variability – Mean, Maximum, Minimum, Standard Deviation

The alongshore variability of the measured shoreline positions within Precinct 3 during the monitoring period 01/07/04 – 30/06/05 is further quantified in Figure 7.3. The upper panel of this figure shows a plot of the mean, maximum and minimum shoreline position at 5 m increments alongshore. For reference, in the lower panel the mean shoreline position during this period is overlaid on to a merged/rectified timex image of the Precinct 3 region (image date = 29th June 2005). Again, data for the region immediately in front of and adjacent to the Flight Deck building is not plotted, as limited shoreline data only could be obtained in this location during the present monitoring period.

Referring to Figure 7.3, the median beach width at mid-tide (relative to the wall alignment) was of the order of 30 m toward the northern end of Precinct 3. This reduced to 10 - 20 m along a 20 m length of the beach centred at 300 m north of Flight Deck, then south of this region increased again to around 30 m mean beach width. Referring to Figure 7.2, it can be seen that the back-beach alignment from which beach widths are measured tends to project further seaward in the region identified above as corresponding to the region of narrowest beach, due to existing boulder wall structure(s) in this region of Precinct 3.

The analysis of maximum and minimum beach width for Precinct 3 (upper panel, Figure 7.3) reveals a trend of uniform beach width variability alongshore. Throughout Precinct 3 the beach width varied by up to ±15 m from the mean shoreline position.

The middle panel of Figure 7.3 shows the standard deviation (s.d.) of weekly shorelines from the mean shoreline position during the period 01/07/04 – 30/06/05. The standard deviation of weekly shorelines also shows a uniform trend alongshore, with the calculated s.d. in the narrow range of 8 – 10 m at all positions alongshore. The uniform envelope of max. – min. variability and constant s.d. indicates that the beachfront along Precinct 3 during the present 12 month monitoring period exhibited no distinctive alongshore

WRL TECHNICAL REPORT 2005/24 7-3 variability, and no area in this region was more or less prone to erosion-accretion in response to the passage of storms.

7.3 Time-Series of Beach Widths at Control Transects

The variations in shoreline position measured at four representative transects within Precinct 3 for the twelve month monitoring period July 2004 to June 2005 are shown in Figure 7.4. For convenience, the locations of transects were chosen to coincide with the eastern end of beachfront streets. Figure 7.4 plots the weekly shoreline position at Devitt Street, Mactier Street, Stuart Street and Jenkins Street, with the alongshore position of each of these beach transects shown in the accompanying merged/rectified image (image date: 29/06/2005). To aid interpretation, a 3-point running mean has been fitted to the weekly data, to assist interpretation of the predominant trends.

For the sake of completeness Figure 7.5 is also included here, that show the same data as presented in Figure 7.4, but in the on-line graphical format (‘Beach Width Analysis’) that are updated each week, and are available for public viewing (and download) via the monitoring project web site (refer Section 5.3). The top and bottom panels in these figures are equivalent to the two panels in Figure 7.4, with the additional inclusion of selected shorelines to show the most recent shoreline movements.

The beach response throughout the period July 2004 to June 2005 shows similar trends at all four transects (further confirming the conclusion from the previous section that during the present 12 month monitoring period the beachfront along Precinct 3 exhibited no distinctive alongshore variability). From July to September 2004 the beachfront increased in width by 10 m to 20 m, then in response to the succession of storms in October a rapid decline in beach width was recorded, as the beach receded by some 30 m.

An early indication of beach recovery in November was reversed in December, by the occurrence of a further series of moderate storms. Six months after the commencement of the monitoring program in July 2004, by the end of December 2004 the beach along Precinct 3 had eroded back of the order of 20 m. The moderate to mild wave climate through January – February 2005 saw a period of partial beach recovery, with the beach accreting of the order of 10 m during this time. The storm in March again generally cut back the beach by 5 – 10 m, followed by an accretionary period through to mid May, with the beach continuing to accrete up to the conditions that existed 8 months previously at the end of September 2004. Increased wave energy in mid May caused the beach to again be

WRL TECHNICAL REPORT 2005/24 7-4 eroded by some 15 – 20 m at the majority of locations alongshore, with the exception being at Stuart Street, where the beach eroded by around 30 m. In the final month of June a general trend of beach recovery was recorded. Despite the relatively large magnitude fluctuation of beach width due to erosion-accretion events over the preceding 12 month period, by the end of June 2005 the beach width along Precinct 3 had returned to within 5 m of the conditions that prevailed 12 months previously in July 2004.

7.4 Future reporting

This is the first in a series of annual data summary reports. In future reports, it is intended that the above analysis will be extended to include an additional quantitative comparison of shoreline position and variability during each current monitoring period, relative to observe changes measured during the proceeding monitoring period. In addition, as the available database expands to cover multiple years, the analysis of net beach changes since the commencement of this monitoring program in mid 2004 will be presented, as well as the critical assessment of seasonal versus longer-term beach trends.

eotN.2005/24 No. Report WRL EKYSOEIE:JL 04-JN 2005 JUNE - 2004 JULY SHORELINES: WEEKLY

WEEKLY SHORELINES: Jul2004 − Jun2005 narrabn

600

400

200

distance cross−shore (m) 0

2500 2000 1500 1000 500 0 −500 −1000 distance along−shore (m) Figure 02092-7-1.cdr 7.1 BEACH WIDTH: Jul04 − Jun05 narrabn 100

distance from dune/ref line (m) 20

0 800 700 600 500 400 300 200 100 0 −100 −200 distance along−shore (m)

600

400

200

distance cross−shore (m) 800 700 600 500 400 300 200 100 0 −100 −200 distance along−shore (m)

WRL Figure WEEKLY BEACH WIDTH: JULY 2004 - JUNE 2005 7.2

Report No. 2005/24 02092-7-2.cdr BEACH WIDTH STATISTICS: Jul04 − Jun05 narrabn

mean max−min (m) 0 800 700 600 500 400 300 200 100 0 −100 −200 alongshore (m)

5 standard dev. (m) 0 800 700 600 500 400 300 200 100 0 −100 −200 alongshore (m)

600

400

200

distance cross−shore (m) 800 700 600 500 400 300 200 100 0 −100 −200 distance along−shore (m)

Figure WRL STATISTICAL SUMMARY OF BEACH WIDTH CHANGES: 7.3 JULY 2004 - JUNE 2005 Report No. 2005/24 02092-7-3.cdr Devitt St 60

2500 50

20 beach width (m) 2000 10

Mactier St 60

1500 50

20 1000 beach width (m) 10 Devitt St 0 y (m) Stuart St Mactier St 60 500 50

40 Stuart St 30

0 beach width (m) 10 Jenkins St 0

Jenkins St 60 −500 50

−1000 beach width (m) 0 200 400 600 10 x (m) 0 Jul04 Aug04 Oct04 Dec04 Mar05 May05 Jul05 date

Figure WRL TIME-SERIES OF BEACH WIDTH: 7.4 JULY 2004 - JUNE 2005 Report No. 2005/24 02092-7-4.cdr Figure WRL ON-LINE BEACH WIDTH ANALYSIS: 7.5 JUNE 2005 Report No. 2005/24 02092-7-5.cdr WRL TECHNICAL REPORT 2005/24 8-1

8. ANALYSIS OF EROSION-ACCRETION TRENDS

A new routine image analysis technique has been implemented at Collaroy-Narrabeen that now enables patterns of beach erosion and accretion to be identified and quantified. On a monthly basis, hourly images throughout a single spring tide are analysed and a 3-D bathymetry of the beach face extending from the low tide waterline to the high tide waterline is derived. These data are further analysed to assess regions of beachface erosion and deposition within Precinct 3.

8.1 Methodology

A detailed description of the analysis techniques used to derive three-dimensional beachface bathymetry from two-dimensional image analysis is provided in Appendix C. In summary, throughout a single spring tide cycle, the shoreline mapping technique is applied to locate the waterline in successive hourly images. The elevation corresponding to the detected waterlines is calculated on the basis of concurrent tide and wave information, which is incorporated in a model that combines the effects of wave setup and swash, at both incident and infragravity frequencies. As illustrated in Figure 8.1, if this process is repeated at all points alongshore throughout a complete tide cycle, a three-dimensional bathymetry of the beachface - between the high tide and low tide waterlines - can be derived. The beachface is the most dynamic region of sediment movement within the coastal system, and sand changes observed in this area are indicative of the total profile.

8.2 Monthly Beachface Bathymetric Mapping

The main application to coastal management of monthly intertidal bathymetries derived from image analysis, is the ability to compare successive months in order to identify and quantify areas of erosion and accretion. However, the monthly beachface bathymetries used to achieve this, do provide a degree of interest in their own right. The bathymetries derived at monthly intervals along Precinct 3 are shown in Figure 8.2 (July-August 2004), Figure 8.3 (September-October), and Figure 8.4 (November-December), Figure 8.5 (January-February 2005), Figure 8.6 (March-April) and Figure 8.7 (May-June 2005). These data reveal two distinct ‘modes’ in the bathymetry of the intertidal region along the Precinct 3 beachfront. Following periods of high wave energy (e.g. October 2004 – Figure 8.3 and April 2005 – Figure 8.6), the beach exhibits a uniform and lower gradient profile alongshore, indicative of the movement of sediment offshore and dissipative beach

WRL TECHNICAL REPORT 2005/24 8-2 conditions. In contrast, follow periods of relatively mild incident wave energy (e.g. February 2005 – Figure 8.5) a more complex and variable beachface morphology prevails, indicative of sand accretion and more intermediate to reflective beach conditions.

8.3 Monthly Erosion-Accretion Trends

As noted above, by further processing of the monthly bathymetries shown in Figures 8.2 - 8.7, a quantitative measure of the net change in sand volumes across the beachface throughout Precinct 3 can be obtained. Figures 8.8 to 8.11 show the results of these calculations to determine the net change in bed elevation between July 2004 and June 2005.

The top and middle panels of Figure 8.8 confirm that between July and September Precinct 3 experienced a period of net beachface accretion throughout much of the embayment, with the greatest areas of beachface development (1+ m vertical accretion) toward the northern end of this region of the beachfront.

In contrast, during the period September – October (lower panel, Figure 8.8), significant erosion in excess of 1 m vertical elevation occurred along the entire beachfront, in response to the succession of storm events commencing in early October.

From late October 2004 to the end of January 2005 (Figure 8.9) only minor net changes in sand volume (less than 0.5 m change in elevation) were recorded. January – February (top panel, Figure 8.10) marked the commencement of a period of more marked beachface accretion, with the northern half of Precinct 3 increasing in elevation in excess of 1 m. Accretion continued through February – March (middle panel, Figure 8.10), with a strong accretionary trend (great than 0.5 m increase in beachface elevation) continuing in the southern half of Precinct 3 during March-April. This contrasted to a lesser degree of erosion (less than 0.5 m decrease in beachface elevation) along the northern half of Precinct 3 recorded during this same period (lower panel, Figure 8.10).

April – May saw a return to erosional conditions (Figure 8.11, upper panel), with the beachface throughout Precinct 3 decreasing in elevation by up to 1 m. During the final period May – June 2005 mildly accretionary conditions again prevailed, accretion (up to 0.8 m increase in elevation) occurring within a localised region approximately 200 m north of Flight Deck.

WRL TECHNICAL REPORT 2005/24 8-3

The data presented in Figures 8.8 – 8.11 enables the calculation of the monthly totals of intertidal (+0.7 to -0.5 mAHD) sand volume lost or gained within Precinct 3, along with the equivalent loss or gain of sand volume per metre length of the beach. The results of these analyses are summarised in Table 8.1 below.

NET CHANGE INTERTIDAL (+0.7 to -0.5 mAHD) SAND VOLUME – PRECINCT 3 period cubic m (total) cubic m per m shoreline July - August 2004 +2,240 +2.1 August – September +7,330 +6.9 September - October -25,200 -23.7 October – November -1,320 -1.3 November – December +3,770 +3.6 December – January 2005 +74 0 January - February +10,960 +10.3 February – March +5,240 +4.9 March – April +4,660 +4.4 April – May -14,600 -13.8 May – June +3,400 +3.2

Table 8.1: Total and Net Intertidal Sand Volumes Changes: July 2004 – June 2005

Referring to Table 8.1 above, the larger erosion events in October 2004 and May 2005 can been seen to have resulted in the loss of the order of 25,200 m3 and 14,600 m3 of sand respectively from the beachface in Precinct 3. These total volumes equate to a respective loss of around 24 m3 per m and 14 m3 per m of the shoreline (between +0.5 & -0.5 mAHD). In contrast, the largest monthly accretion values were in August-September 2004 (around 7,000 m3 total or 7 m3 per m) and January-February 2005 (around 11,000 m3 total or 10 m3 per m), confirming that the loss of sediment from the beachface during storms occurs at a considerably greater rate than the subsequent recovery of the beach.

In future reports, the net change in intertidal sand volume within Precinct 3 will be plotted and tracked through time, providing an objectives means to measure the net loss or gain of sediment from this region of the Collaroy-Narrabeen embayment.

WRL TECHNICAL REPORT 2005/24 8-4

8.4 Net Erosion-Accretion Trends: June 2004 - November 2004

The net change in beachface bathymetry that was measured through the entire twelve month monitoring period July 2004 - June 2005 is summarised in Figure 8.12. This analysis confirms the general conclusion suggested by the previous shoreline and beach width analysis (Sections 7.2 and 7.3). Over the total 12 month monitoring period, despite the occurrence of a number of large storm events and resulting fluctuations of the shoreline, by June 2005 beach conditions in Precinct 3 were similar to the conditions that prevailed 12 months earlier in July 2004. From the more detailed analysis presented in Figure 8.12, it is apparent that the majority of Precinct 3 experienced a slight net erosional trend during this period, with the beachface elevation lowering in the vertical range of 0 m to 0.4 m during this time. The exception to this trend was at a very localised region located approximately 650 m north of Flight Deck, where up to 0.6 m vertical beach accretion was observed.

Dune Waterline high tide

Waterline low tide

Y Wet beach

Figure WRL DEFINITION SKETCH INTERTIDAL BATHYMETRY FROM HOURLY WATERLINES 8.1 Report No. 2005/24 02092-8-1.cdr JULY 2004

intertidal contour map: narrabn.28.7.2004.mat 150 > 0.70 m

0.50 m

100 −0.4

−0.20

0.2 0.4 0.6

0.25 m 50

−0.4

−0.20 0.00 m AHD 0 0.40.2 0.6 Cross−shore distance (m) −0.4 −0.20 −0.4 0.40.2 0.6 0.40.2 −0.20 0.6 −50 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

AUGUST 2004

intertidal contour map: narrabn.26.8.2004.mat 150 > 0.70 m

0.50 m 100 0 0.2 0.40.6 −0.2

−0.4 0.25 m 50

0 0.2 0.40.6 −0.2 −0.4 0.00 m AHD 0 Cross−shore distance (m) 0 −0.4 −0.2 0.2 −0.2 −0.4 0.40.6 0.20 0.40.6 −50 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - JULY, AUGUST 2004 8.2

Report No. 2005/24 02092-8-2.cdr SEPTEMBER 2004

intertidal contour map: narrabn.29.9.2004.mat 150 > 0.70 m

0.50 m 100 0.20 −0.20.4

−0.4 0.6

0.25 m 50

0 −0.20.40.2 −0.4 0.6 0.00 m AHD 0 −0.4 Cross−shore distance (m) −0.2 −0.4 −0.20.20 0 0.4 0.6 0.40.2 0.6

−50 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

OCTOBER 2004

intertidal contour map: narrabn.26.10.2004.mat 150 > 0.70 m

0.50 m 100

−0.4

−0.2 00.2 −0.4

0.4 0.25 m 50 0.6

−0.4

−0.2 0 0.00 m AHD 0 0.2 0.4 0.6 Cross−shore distance (m)

−0.4 −0.2 00.2 −0.2 0 −0.4 0.40.6 −0.25 m −50 0.40.6 0.2

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - SEPTEMBER, OCTOBER 2004 8.3

Report No. 2005/24 02092-8-3.cdr NOVEMBER 2004

intertidal contour map: narrabn.28.11.2004.mat 150 > 0.70 m

0.50 m 100

−0.4 −0.2

0.4 0.6 00.2 0.25 m 50

−0.4 −0.2 0.00 m AHD 0 0.40.6 00.2 Cross−shore distance (m)

−0.2−0.4 −0.4 0 00.2 −0.2 0.2 0.4 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

DECEMBER 2004

intertidal contour map: narrabn.27.12.2004.mat 150 > 0.70 m

0.50 m 100

−0.4

−0.20 0.2 0.40.6 0.25 m 50

−0.4

−0.2 0.00 m AHD 0 0.20 0.40.6 Cross−shore distance (m)

−0.2 −0.4 −0.4 −0.20 −0.2 0.20.4 0.20 0.6 −50 0.40.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - NOVEMBER, DECEMBER 2004 8.4

Report No. 2005/24 02092-8-4.cdr JANUARY 2005

intertidal contour map: narrabn.29.1.2005.mat 150 > 0.70 m

0.50 m 100

−0.4 −0.204 0.20..6 0

0.4 0.6 00.2 0.25 m 50 −0.4 −0.2

−0.4 0−0.2 0.20.4 0.6 0.00 m AHD 0 0.40.6 00.2

Cross−shore distance (m) −0.4 −0.2

−0.2−0.4 −0.4−0.2−0.4 0 0.20 000.2 −0.2 0.2 0.40.6 0.6 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

FEBRUARY 2005

intertidal contour map: narrabn.25.2.2005.mat 150 > 0.70 m

0.50 m 100

0 −0.4−0.40.20.40 −0.20.6 0.20.4 0.6

0.4 0.6 00.2 0.25 m 50 −0.4 −0.2 −0.2

0 −0.4 0.4−0.4 0.200.6−0.2 0.20.4 0.6 0.00 m AHD 0 0.40.6 00.2

Cross−shore distance (m) −0.2 −0.4 −0.4 −0.2 −0.4 0 0.20.4 0 −0.2 0.6 −0.2−0.4 0.40.6 −0.4−0.2−0.4 0 0.20 000.2 −0.2 0.2 0.40.6 0.6 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - JANUARY, FEBRUARY 2005 8.5

Report No. 2005/24 02092-8-5.cdr MARCH 2005

intertidal contour map: narrabn.12.3.2005.mat 150 > 0.70 m

0.50 m

100 0

0 4 −0.4−0.40.20.0 6 −0.20.4 0.20..6 0

0.4 0.6 −0.4 −0.2 00.2 0.4 0.25 m 0.6 50 0.2 −0.4 −0.2 −0.2

0 0 −0.4 0.4−0.4 0.200.6−0.2 0.20.4 0.6 0.00 m AHD 0 −0.2 −0.4 0.40.6 0.4 00.2 0.2 0.6 Cross−shore distance (m) −0.2 −0.4 −0.4 0 −0.2 −0.4 −0.4 0 0 −0.2 0.20.4 0 −0.2 0.2 0.6 0.4 −0.2−0.4 0.40.6 0.6 −0.4−0.2−0.4 0 0.20 000.2 −0.2 0.2 0.40.6 0.6 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

APRIL 2005

intertidal contour map: narrabn.23.4.2005.mat 150 > 0.70 m

0.50 m −0.40 100 0.20.4 0.60

0 −0.4−0.40.20.40 −0.20.6 0.20.4 0.6

0.4 0.6 −0.4 −0.2 00.2 0.4 0.25 m 0.6 50 −0.2 0.2 −0.4 −0.2 −0.4 0 −0.2 0.20.4 0.60 0 −0.4 0.4−0.4 0.200.6−0.2 0.20.4 0.6 0.00 m AHD 0 −0.2 −0.4 0.40.6 0.4 00.2 −0.2 0.2 0.6 Cross−shore distance (m) −0.4 −0.2 0 −0.4 −0.4 0 −0.2 −0.4 −0.4 00.2 0−0.4 −0.2 0.20.4 0 −0.2−0.2 0.2 0.60.4 0.4 0.6−0.2−0.4 0.40.40.6 0.6 −0.4−0.2−0.4 0 0.20 000.20.6 −0.2 0.2 0.40.6 0.6 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - MARCH, APRIL 2005 8.6

Report No. 2005/24 02092-8-6.cdr MAY 2005

intertidal contour map: narrabn.23.5.2005.mat 150 > 0.70 m

−0.4 0.50 m 0 4 0.2 100 0..6 −0.4−0.20 0 0 0.20.4 0.60 4 −0.4−0.40.20.0 6 −0.20.4 0.20..6 0

0.4 0.6 −0.4 −0.2 00.2 0.4 0.25 m 0.6 50 −0.2 0.2 −0.4 −0.2 −0.4 0 −0.2 0.20.4 0.60 0 −0.4 −0.4 0.4−0.4−0.20 0.200.6−0.20.20.4 0.6 0.20.4 0.6 0.00 m AHD 0 −0.2 −0.4 0.40.6 0.4 00.2 −0.2 0.2 0.6 Cross−shore distance (m) −0.4 −0.2−0.4 0 −0.4 −0.4 0 −0.2 −0.4 −0.4 00.2 0−0.4 −0.2 0.20.4 0−0.2 −0.2−0.2 0.2 0.6−0.20.20.40 0.20 0.4 0.40.6−0.2−0.4 0.40.40.6 −0.4 0.6 0.6 0.4 −0.4−0.2−0.4 0 0.20 000.20.6 −0.2 0.2 0.40.6 0.6 0.4 0.6 −50 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

JUNE 2005

intertidal contour map: narrabn.16.6.2005.mat 150 > 0.70 m −0.4 −0.4

−0.4

−0.4 0.50 m

−0.4 100 −0.40 −0.2 0.2 00.20.4 0 0.6 −0.4 0.6 −0.4 −0.2 0.4 −0.20.2 −0.4 −0.2 −0.2

−0.2 0.25 m 50

−0.4 −0.2 −0.4 −0.2 −0.4 −0.2 0 −0.2 −0.4 00.2 0.6 00.20.4 −0.20.20.40.6 −0.4 0.00 m AHD 0 Cross−shore distance (m)

−0.4 −0.4−0.4 −0.4 −0.20 −0.2−0.200.20.4 0.6 00.20.2 0 −0.2−0.2 0 0.4 0.20.20 0.2 0.6 0.4 0.4 −50 0.6 0.6 −0.25 m

−100 < −0.50 m 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

WRL Figure BEACHFACE MAPPING - MAY, JUNE 2005 8.7

Report No. 2005/24 02092-8-7.cdr intertidal erosion/accretion: narrabn.28.7.2004.mat − narrabn.26.8.2004.mat 150 1.0

0.8

2 −0. 100 0.6

0.4

−0.2 50 0.2

0.2 0.2 0.0 m 0.4

0.8 0.6 0.4

0 0.2 −0.2 Cross−shore distance (m)

0.2 −0.4 0.2 0.4 0.4 −0.2 0.8 0.60.2 −50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.26.8.2004.mat − narrabn.29.9.2004.mat 150 1.0

0.8

100 0.6

0.4 0.2

50 0.2

0.0 m 0.2

0.4 0.2 0 −0.2 −0.2

Cross−shore distance (m) 0.6 0.2 0.2 0.4 0.4 −0.4 1 0.6 1 0.8 0.6 0.8 0.6 0.20.4 0.20.4

−50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.29.9.2004.mat − narrabn.26.10.2004.mat 150 1.0

0.8

100 0.6 −1 −0.2 −0.6 −0.4 −1 −0.8 0.4 −0.6−0.2

−0.8−0.4 50 0.2

−0.2 0.0 m −0.6 −1 −0.4−1 −0.8 −0.2 0 −0.6 −0.2 −0.8−0.4

Cross−shore distance (m) −0.4 −0.8 −0.6−0.2 −0.2 −0.4−0.6 −0.4 −1 −0.8−1 −1 −0.6 −1 −0.8 −0.2 −0.8 −1 −0.4 −50 −0.4−0.6 −0.6 −0.2

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

Figure WRL MONTHLY EROSION/ACCRETION: JULY - OCTOBER 2004 8.8 Report No. 2005/24 02092-8-8.cdr intertidal erosion/accretion: narrabn.26.10.2004.mat − narrabn.28.11.2004.mat 150 1.0

0.8

100 0.6

0.4

50 0.2

0.0 m

0 −0.2 −0.2 −0.4 Cross−shore distance (m) −0.2 −0.4

0.2 0.2 −50 0.6 0.4 −0.6 0.20.4

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.28.11.2004.mat − narrabn.27.12.2004.mat 150 1.0

0.8

100 0.6

0.4

50 0.2 0.2

0.0 m 0.2

0 −0.2

Cross−shore distance (m) 0.2 0.40.2 0.6 −0.4

0.2 0.2 0.4 −0.2 −50 0.2 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.27.12.2004.mat − narrabn.29.1.2005.mat 150 1.0

0.8

100 0.6

0.4

50 0.2

0.0 m 0.2 0.2

0 0.6 −0.2 0.4

Cross−shore distance (m) −0.2

0.4 −0.4 0.2 −0.2 −0.2 0.2 0.2 −50 −0.2 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

Figure WRL MONTHLY EROSION/ACCRETION: OCTOBER 2004 - JANUARY 2005 8.9 Report No. 2005/24 02092-8-9.cdr intertidal erosion/accretion: narrabn.29.1.2005.mat − narrabn.25.2.2005.mat 150 1.0

0.8

100 0.6 4 0.

0.2 0.4

50 −0.2 0.2

0.2 0.0 m 0.4

0.6 0 0.2 −0.2 0.8 0.6 Cross−shore distance (m) 1 0.4

0.2 −0.4 0.20.4 0.4 1 0.8 0.6 1 0.6 0.8 1 0.8 1 0.2 0.6 0.4 0.2 −50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.25.2.2005.mat − narrabn.12.3.2005.mat 150 1.0

0.8

100 0.6

0.4 0.20.4

0.6

50 0.4 0.2 0.2

0.0 m

0.2 0 0.4 −0.2 0.6 0.2

Cross−shore distance (m) 0.2 0.4 −0.4 0.40.2

0.4 0.8 0.2 0.60.4

−50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.12.3.2005.mat − narrabn.23.4.2005.mat 150 1.0

0.8

100 0.4 0.6 0.8 0.2 0.4 0.8 0.6 0.4 0.6 1 0.2

50 0.2

0.4 0.2 0.6 0.0 m 0.8 0.4

0.2 0 −0.2 Cross−shore distance (m)

−0.2 −0.2 −0.4

−0.4 −0.2 −0.2 −50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

Figure WRL MONTHLY EROSION/ACCRETION: JANUARY - APRIL 2005 8.10 Report No. 2005/24 02092-8-10.cdr intertidal erosion/accretion: narrabn.23.4.2005.mat − narrabn.23.5.2005.mat 150 1.0

0.8

100 0.6

−0.2 −0.4

−0.6 0.4 −0.8 −0.2

−0.4 50 0.2

−1

−0.6 0.0 m −0.2 −0.8

−1 −0.4 −0.6 −0.8 0 −0.2 −0.4 −0.2 Cross−shore distance (m) −0.2 −1 −0.4 −0.2 −0.6 −0.8−0.6 −0.4 −0.2 −0.4 −0.4−0.6 −0.2 −50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

intertidal erosion/accretion: narrabn.23.5.2005.mat − narrabn.16.6.2005.mat 150 1.0

0.8

100 0.6

−0.2 0.4

50 0.2 0.2 0.2

0.0 m

0 −0.2

0.60.8 0.4 Cross−shore distance (m) 0.20.4 0.2 −0.4

−0.4 −0.2

−0.2 0.4 0.2 −50 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

Figure WRL MONTHLY EROSION/ACCRETION: APRIL - JUNE 2005 8.11 Report No. 2005/24 02092-8-11.cdr NET CHANGE: JULY 2004 - JUNE 2005

intertidal erosion/accretion: narrabn.28.7.2004.mat − narrabn.16.6.2005.mat 150 1.0

0.8

100 −0.2 0.6

−0.2

−0.4 0.4

−0.4 50 0.2

−0.2 0.0 m

0 −0.2 Cross−shore distance (m)

−0.2 −0.4 0.2 0.4 0.6 −0.4 −0.4−0.6 −0.2 −50 −0.2 −0.6

−0.8

−100 −1.0 800 700 600 500 400 300 200 100 0 −100 −200 Longshore distance (m)

Figure WRL NET EROSION/ACCRETION: JULY 2004 - JUNE 2005 8.12 Report No. 2005/24 02092-8-12.cdr WRL TECHNICAL REPORT 2005/24 9-1

9. SUMMARY AND CONCLUSIONS

This summary data report for the twelve month monitoring period July 2004 to June 2005 marks the first of a planned series of annual reports to be produced for the Collaroy- Narrabeen embayment, with particular focus on Precinct 3.

9.1 Qualitative Visual Assessment

Beach and nearshore conditions during the 12 month monitoring period July 2004 to June 2005 were characterised by three distinct stormy periods when the beach was observed to erode, separated by moderate to mild wave conditions, during which times beach recovery was observed. The first stormy period commenced with the occurrence of a large storm event in mid July 2004, when offshore significant wave heights rose to 7 m. Moderate wave conditions through the first half of August peaked with the occurrence of a second storm in the middle of the month (offshore significant wave heights up to 5 m). Wave energy then declined, with generally mild conditions through the second half of August, and throughout September. The second stormy period commenced in October, with the occurrence of four storm events in the one month, when offshore significant wave heights up to 6 m were recorded. In contrast, November again saw a return to mild conditions, followed by a series of moderate storms in December, when offshore significant wave heights of 3 – 4 m were observed. Moderate wave conditions prevailed through January 2005, followed by generally milder conditions in February. The third major stormy period commenced in the second half of March, when offshore significant wave heights again exceeded 6 m. Mild to moderate wave conditions return in April and continued through to late June, interrupted in mid May by a period of elevated wave energy. At the conclusion of this present 12 month monitoring period wave heights again increased, suggesting the commencement of a fourth period of higher wave energy and corresponding beach erosion.

A qualitative visual assessment of the net trends in beach adjustment during this period can be seen by contrasting images of the beach obtained at the start and end of the twelve month monitoring period. Along the southern beach (Figure 6.2) the width of the subaerial (or ‘dry’) beach in June 2005 appears very similar to the conditions that prevailed 12 months earlier in July 2004. No retreat of the dune is apparent, and a modest widening of the beach in the vicinity of the Collaroy Surf Life Saving Club is discernable. The general alignment of the beach in this region appears unchanged. Looking north along the Collaroy-Narrabeen embayment (Figure 6.3), again the width of the subaerial beach appears comparable from July 2004 to June 2005, however, a general trend of beach flattening is observed. The scarped berm that was present in July 2004 is absent, with the

WRL TECHNICAL REPORT 2005/24 9-2 beach exhibiting a flatter and more uniform profile in June 2005. The sand coverage over the existing sections of boulder wall within Precinct 3 in places has increased, with a section of the rock no longer visible. However, elsewhere along this section of the beachfront the exposure of rocks at the toe of structures is more apparent.

9.2 Shoreline Variability and Beach Width Analysis

Based upon the qualitative analysis of weekly shoreline positions within Precinct 3 during the present monitoring period 01/07/04 - 31/06/05 (Figure 7.2), the median beach width at mid-tide (relative to the wall alignment) was of the order of 30 m at the northern end of this region. This reduced to 10 - 20 m along a 20 m length of the beach centred at 300 m north of Flight Deck, then south of this region increased again to around 30 m mean beach width. The back-beach alignment from which beach widths are measured tends to project further seaward in the region identified above as corresponding to the region of narrowest beach, due to existing boulder wall structure(s) in this region of Precinct 3.

The analysis of maximum and minimum beach width (upper panel, Figure 7.3) reveals a trend of uniform beach width variability alongshore through the present 12 month monitoring period. Throughout Precinct 3 the beach width varied by up to ±15 m from the mean shoreline position. The standard deviation of weekly shorelines (middle panel, Figure 7.3) also shows a uniform trend alongshore, with the calculated s.d. in the narrow range of 8 – 10 m at all positions alongshore. The uniform envelope of max. – min. variability and constant s.d. indicates that the beachfront along Precinct 3 during the present 12 month monitoring period exhibited no distinctive alongshore variability, and no area in this region was more or less prone to erosion-accretion in response to the passage of storms.

The variations in shoreline position measured at four representative transects within Precinct 3 for the twelve month monitoring period July 2004 to June 2005 (Figure 7.4 and 7.5) show similar trends at all four transects (further confirming the conclusion that during the present 12 month monitoring period the beachfront along Precinct 3 exhibited no distinctive alongshore variability). From July to September 2004 the beachfront increased in width by 10 m to 20 m, then in response to the succession of storms in October a rapid decline in beach width was recorded, as the beach receded by some 30 m. An early indication of beach recovery in November was reversed in December, by the occurrence of a further series of moderate storms. Six months after the commencement of the monitoring program in July 2004, by the end of December 2004 the beach along Precinct 3 had eroded back of the order of 20 m. The moderate to mild wave climate through January – February

WRL TECHNICAL REPORT 2005/24 9-3

2005 saw a period of partial beach recovery, with the beach accreting of the order of 10 m during this time. The storm in March again generally cut back the beach by 5 – 10 m, followed by an accretionary period through to mid May, with the beach continuing to accrete up to the conditions that existed 8 months previously at the end of September 2004. Increased wave energy in mid May caused the beach to again erode by some 15 – 20 m at the majority of locations alongshore, with the exception being at Stuart Street, where the beach eroded by around 30 m. In the final month of June a general trend of beach recovery was recorded. Despite the relatively large magnitude fluctuation of beach width due to erosion-accretion events over the proceeding 12 month period, by the end of June 2005 the beach width along Precinct 3 had returned to within 5 m of the conditions that prevailed 12 months previously in July 2004.

In future reports, it is intended that the above analysis will be extended to include an additional quantitative comparison of shoreline position and variability during each current monitoring period, relative to observe changes measured during the proceeding monitoring period(s). In addition, as the available database expands to cover multiple years, the analysis of longer-term net beach changes since the commencement of this monitoring program in mid 2004 will be presented, as well as the critical assessment of seasonal versus longer-term beach trends.

9.3 Erosion/Accretion Trends

Based upon the analysis of monthly intertidal bathymetries (Figures 8.2 – 8.7) to derive monthly net erosion-accretion trends (Figures 8.8 – 8.11), it was confirmed that between July and September, Precinct 3 experienced a period of net beachface accretion throughout much of the embayment, with the greatest areas of beachface development (greater than 1 m vertical accretion) toward the northern end of this region of the beachfront. In contrast, during the period September – October, significant erosion in excess of 1 m vertical elevation occurred along the entire beachfront, in response to the succession of storm events commencing in early October.

From late October 2004 to the end of January 2005 only minor net changes in sand volume (less than 0.5 m change in elevation) were recorded. January – February marked the commencement of a period of more marked beachface accretion, with the northern half of Precinct 3 increasing in elevation in excess of 1 m. Accretion continued through February – March, with a strong accretionary trend (greater than 0.5 m increase in beachface elevation) continuing in the southern half of Precinct 3 during March-April. This

WRL TECHNICAL REPORT 2005/24 9-4 contrasted to a lesser degree of erosion (less than 0.5 m decrease in beachface elevation) along the northern half of Precinct 3 recorded during this same period. April – May saw a return to erosional conditions, with the beachface throughout Precinct 3 decreasing in elevation by up to 1 m. During the final period May – June 2005 mildly accretionary conditions again prevailed, accretion (up to 0.8 m increase in elevation) occurring within a localised region approximately 200 m north of Flight Deck.

The net change in beachface bathymetry that was measured through the entire twelve month monitoring period July 2004 - June 2005 (Figure 8.12) confirms the general conclusion that, over the entire total 12 month monitoring period, despite the occurrence of a number of large storm events and resulting fluctuations of the shoreline, by June 2005 beach conditions in Precinct 3 were similar to the conditions that prevailed 12 months earlier in July 2004. The majority of Precinct 3 experienced a slight net erosional trend during this period, with the beachface elevation decreasing vertically in the range of 0 m to 0.4 m during this time. The exception to this trend was at a very localised region located approximately 650 m north of Flight Deck, where up to 0.6 m vertical beach face accretion was observed.

Based upon the analysis of monthly intertidal bathymetries, calculations of the monthly net changes in intertidal (+0.7 to -0.5 m AHD) sand volumes within Precinct 3 are tabulated below.

NET CHANGE INTERTIDAL (+0.7 to -0.5 mAHD) SAND VOLUME– PRECINCT 3 period cubic m (total) cubic m per m shoreline July - August 2004 +2,240 +2.1 August – September +7,330 +6.9 September - October -25,200 -23.7 October – November -1,320 -1.3 November – December +3,770 +3.6 December – January 2005 +74 0 January - February +10,960 +10.3 February – March +5,240 +4.9 March – April +4,660 +4.4 April – May -14,600 -13.8 May – June +3,400 +3.2

WRL TECHNICAL REPORT 2005/24 9-5

The larger erosion events in October 2004 and May 2005 can been seen to have resulted in the loss of the order of 25,200 m3 and 14,600 m3 of sand respectively from the beachface within Precinct 3. These total volumes equate to a respective loss of around 24 m3 per m and 14 m3 per m of sand from the beachface (between the elevations of +0.7 and -0.5 mHD). In contrast, the largest monthly accretion values were in August-September 2004 (around 7,000 m3 total or 7 m3 per m) and January February 2005 (around 11,000 m3 total or 10 m3 per m), confirming that the loss of sediment from the beachface during storms occurs at a considerably greater rate than the subsequent recovery of the beach.

In future reports, the net changes in intertidal sand volume within Precinct 3 will be plotted and tracked through time. This will provide an objective means to identify and quantify any underlying trends that may characterise the net loss or gain of sediment from this region of the Collaroy-Narrabeen embayment, where existing development encroaches to the greatest degree into the active beach environment.

WRL TECHNICAL REPORT 2005/24 10-1

10. ACKNOWLEDGEMENTS

This project was commissioned and funded by Warringah Council.

The Body Corporate of Flight Deck is thanked for permitting the ARGUS system to reside on the roof of the building.

Manly Hydraulics Laboratory is acknowledged for the ongoing provision of wave and tide data.

Doug Anderson of WRL continues to assist with wave and tide data processing, computer operations for remote communications, image storage, off-line image archiving and web serving at WRL. Ainslie Frazer and Doug Anderson of WRL were responsible for the weekly analysis and updating of monitoring program information via the project web site. Mitchell Harley completed the monthly waterline analysis for calculation of intertidal bathymetries.

Finally, Professor Rob Holman of Oregon State University and the growing world-wide team of ARGUS users are acknowledged for continuing system development. These research efforts are providing practical tools for coastal monitoring and management.

WRL TECHNICAL REPORT 2005/24 11-1

11. REFERENCES

Aarninkhof, S G J (2003). “Nearshore bathymetry derived from video imagery”. Unpublished PhD thesis, Technical University Delft, 175p.

Aarninkhof, S G J and Roelvink, J A, (1999). “Argus-based monitoring of intertidal beach morphodynamics”. Proceedings, Coastal Sediments Conference, Long Island (NY), USA, 2429-2444.

Aarninkhof, S G J, Turner I L, Dronkers T D T, Caljouw, M and Nipius, L, (2003), “A video-based technique for mapping intertidal bathymetry”, Coastal Engineering, 49, 275- 289.

Anderson, D JA, Turner, I L, Dyson, A, Lawson, S and Victory, S, (2003), “Tweed River Entrance Sandy Bypassing Project: ‘real-time’ beach monitoring and analysis system via the world-wide-web”, Proceedings, 16th Australasian Coastal and Ocean Engineering Conference, IEA,, Auckland, New Zealand, Paper #002 (CD Conference Volume), 9p.

Holman R A, Sallenger Jr A H, Lippmann T C D and Haines J W, (1993), “The Application of Video Image Processing to the Study of Nearshore Processes”, Oceanography, Vol.6, No.3.

Turner, I L, Aarninkhof, S G J and Holman, R, (2005), Coastal imaging research and applications in Australia. Invited submission to Short A D and Thom B G (eds), Australian Coastal Geomorphology 2004. (Special Issue, Journal of Coastal Research).

Wright L D and Short A D, (1983), Morphodynamics of Beaches and Surf zones In Australia, In Komar, P.D., (ed), Handbook of Coastal Processes and Erosion, CRC Press, Boca Raton, 34-64.

WRL TECHNICAL REPORT 2005/24

Appendix A

Week-to-a-Page: July 2004 to June 2005 Week-to-a-Page

2004-07-12

2004-07-13

Figure WRL DAILY MID-TIDE IMAGES 12/07/2004 - 18/07/2004 A1 Report No. 2005/24 02092-A01.cdr Figure WRL DAILY MID-TIDE IMAGES 19/07/2004 - 25/07/2004 A2 Report No. 2005/24 02092-A02.cdr Week-to-a-Page

2004-07-26

2004-07-27

2004-07-30

2004-07-31

Figure WRL DAILY MID-TIDE IMAGES 26/07/2004 - 01/08/2004 A3 Report No. 2005/24 02092-A03.cdr Figure WRL DAILY MID-TIDE IMAGES 02/08/2004 - 08/08/2004 A4 Report No. 2005/24 02092-A04.cdr Figure WRL DAILY MID-TIDE IMAGES 09/08/2004 - 15/08/2004 A5 Report No. 2005/24 02092-A05.cdr Figure WRL DAILY MID-TIDE IMAGES 16/08/2004 - 22/08/2004 A6 Report No. 2005/24 02092-A06.cdr Figure WRL DAILY MID-TIDE IMAGES 23/08/2004 - 29/08/2004 A7 Report No. 2005/24 02092-A07.cdr Figure WRL DAILY MID-TIDE IMAGES 30/08/2004 - 05/09/2004 A8 Report No. 2005/24 02092-A08.cdr Week-to-a-Page

2004-09-06

2004-09-12

Figure WRL DAILY MID-TIDE IMAGES 06/09/2004 - 12/09/2004 A9 Report No. 2005/24 02092-A09.cdr Figure WRL DAILY MID-TIDE IMAGES 13/09/2004 - 19/09/2004 A10 Report No. 2005/24 02092-A10.cdr Figure WRL DAILY MID-TIDE IMAGES 20/09/2004 - 26/09/2004 A11 Report No. 2005/24 02092-A11.cdr 2004-09-30

Figure WRL DAILY MID-TIDE IMAGES 27/09/2004 - 03/10/2004 A12 Report No. 2005/24 02092-A12.cdr 2004-10-09

2004-10-10

Figure WRL DAILY MID-TIDE IMAGES 04/10/2004 - 10/10/2004 A13 Report No. 2005/24 02092-A13.cdr Figure WRL DAILY MID-TIDE IMAGES 11/10/2004 - 17/10/2004 A14 Report No. 2005/24 02092-A14.cdr 2004-10-22

2004-10-23

2004-10-24

Figure WRL DAILY MID-TIDE IMAGES 18/10/2004 - 24/10/2004 A15 Report No. 2005/24 02092-A15.cdr Figure WRL DAILY MID-TIDE IMAGES 25/10/2004 - 31/10/2004 A16 Report No. 2005/24 02092-A16.cdr 2004-11-05

2004-11-06

2004-11-07

Figure WRL DAILY MID-TIDE IMAGES 01/11/2004 - 07/11/2004 A17 Report No. 2005/24 02092-A17.cdr 2004-11-11

Figure WRL DAILY MID-TIDE IMAGES 08/11/2004 - 14/11/2004 A18 Report No. 2005/24 02092-A18.cdr 2004-11-17

2004-11-18

Figure WRL DAILY MID-TIDE IMAGES 15/11/2004 - 21/11/2004 A19 Report No. 2005/24 02092-A19.cdr 2004-11-23

Figure WRL DAILY MID-TIDE IMAGES 22/11/2004 - 28/11/2004 A20 Report No. 2005/24 02092-A20.cdr Figure WRL DAILY MID-TIDE IMAGES 29/11/2004 - 05/12/2004 A21 Report No. 2005/24 02092-A21.cdr Week-to-a-Page

2004-12-06

Figure WRL DAILY MID-TIDE IMAGES 06/12/2004 - 12/12/2004 A22 Report No. 2005/24 02092-A22.cdr Week-to-a-Page

image not available 2004-12-13

image not available 2004-12-14

image not available 2004-12-15

image not available 2004-12-16

image not available 2004-12-17

image not available 2004-12-18

image not available 2004-12-19

Figure WRL DAILY MID-TIDE IMAGES 13/12/2004 - 19/12/2004 A23 Report No. 2005/24 02092-A23.cdr 2004-12-26

Figure WRL DAILY MID-TIDE IMAGES 20/12/2004 - 26/12/2004 A24 Report No. 2005/24 02092-A24.cdr 2005-01-01

2005-01-02

Figure WRL DAILY MID-TIDE IMAGES 27/12/2004 - 02/01/2005 A25 Report No. 2005/24 02092-A25.cdr 2005-01-09

Figure WRL DAILY MID-TIDE IMAGES 03/01/2005 - 09/01/2005 A26 Report No. 2005/24 02092-A26.cdr 2005-01-13

2005-01-16

Figure WRL DAILY MID-TIDE IMAGES 10/01/2005 - 16/01/2005 A27 Report No. 2005/24 02092-A27.cdr 2005-01-23

Figure WRL DAILY MID-TIDE IMAGES 17/01/2005 - 23/01/2005 A28 Report No. 2005/24 02092-A28.cdr 2005-01-28

2005-01-29

2005-01-30

Figure WRL DAILY MID-TIDE IMAGES 24/01/2005 - 30/01/2005 A29 Report No. 2005/24 02092-A29.cdr Figure WRL DAILY MID-TIDE IMAGES 31/01/2005 - 06/02/2005 A30 Report No. 2005/24 02092-A30.cdr 2005-02-13

Figure WRL DAILY MID-TIDE IMAGES 07/02/2005 - 13/02/2005 A31 Report No. 2005/24 02092-A31.cdr 2005-02-20

Figure WRL DAILY MID-TIDE IMAGES 14/02/2005 - 20/02/2005 A32 Report No. 2005/24 02092-A32.cdr Week-to-a-Page

image not available 2005-02-21

image not available 2005-02-22

image not available 2005-02-23

image not available 2005-02-24

image not available 2005-02-25

image not available 2005-02-26

image not available 2005-02-27

Figure WRL DAILY MID-TIDE IMAGES 21/02/2005 - 27/02/2005 A33 Report No. 2005/24 02092-A33.cdr 2005-03-05

2005-03-06

Figure WRL DAILY MID-TIDE IMAGES 28/02/2005 - 06/03/2005 A34 Report No. 2005/24 02092-A34.cdr Figure WRL DAILY MID-TIDE IMAGES 07/03/2005 - 13/03/2005 A35 Report No. 2005/24 02092-A35.cdr 2005-03-18

2005-03-19

2005-03-20

Figure WRL DAILY MID-TIDE IMAGES 14/03/2005 - 20/03/2005 A36 Report No. 2005/24 02092-A36.cdr Week-to-a-Page

2005-03-21

2005-03-24

2005-03-25

2005-03-26

2005-03-27

Figure WRL DAILY MID-TIDE IMAGES 21/03/2005 - 27/03/2005 A37 Report No. 2005/24 02092-A37.cdr Week-to-a-Page

2005-03-28

2005-03-31

Figure WRL DAILY MID-TIDE IMAGES 28/03/2005 - 03/04/2005 A38 Report No. 2005/24 02092-A38.cdr Figure WRL DAILY MID-TIDE IMAGES 04/04/2005 - 10/04/2005 A39 Report No. 2005/24 02092-A39.cdr Figure WRL DAILY MID-TIDE IMAGES 11/04/2005 - 17/04/2005 A40 Report No. 2005/24 02092-A40.cdr Figure WRL DAILY MID-TIDE IMAGES 18/04/2005 - 24/04/2005 A41 Report No. 2005/24 02092-A41.cdr 2005-04-27

2005-04-28

Figure WRL DAILY MID-TIDE IMAGES 25/04/2005 - 01/05/2005 A42 Report No. 2005/24 02092-A42.cdr Figure WRL DAILY MID-TIDE IMAGES 02/05/2005 - 08/05/2005 A43 Report No. 2005/24 02092-A43.cdr 2005-05-12

Figure WRL DAILY MID-TIDE IMAGES 09/05/2005 - 15/05/2005 A44 Report No. 2005/24 02092-A44.cdr Figure WRL DAILY MID-TIDE IMAGES 16/05/2005 - 22/05/2005 A45 Report No. 2005/24 02092-A45.cdr Figure WRL DAILY MID-TIDE IMAGES 23/05/2005 - 29/05/2005 A46 Report No. 2005/24 02092-A46.cdr Figure WRL DAILY MID-TIDE IMAGES 30/05/2005 - 05/06/2005 A47 Report No. 2005/24 02092-A47.cdr Figure WRL DAILY MID-TIDE IMAGES 06/06/2005 - 12/06/2005 A48 Report No. 2005/24 02092-A48.cdr Figure WRL DAILY MID-TIDE IMAGES 13/06/2005 - 19/06/2005 A49 Report No. 2005/24 02092-A49.cdr Figure WRL DAILY MID-TIDE IMAGES 20/06/2005 - 26/06/2005 A50 Report No. 2005/24 02092-A50.cdr Figure WRL DAILY MID-TIDE IMAGES 27/06/2005 - 03/07/2005 A51 Report No. 2005/24 02092-A51.cdr WRL TECHNICAL REPORT 2005/24

Appendix B

Monthly Wave Climate Summaries: July 2004 to June 2005 OFFSHORE WAVE CLIMATE: 01−Jul−2004 to 31−Jul−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY JULY 2004 B1 Report No. 2005/24 02092-B01.cdr OFFSHORE WAVE CLIMATE: 01−Aug−2004 to 31−Aug−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY AUGUST 2004 B2 Report No. 2005/24 02092-B02.cdr OFFSHORE WAVE CLIMATE: 01−Sep−2004 to 30−Sep−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

Figure WRL MONTHLY WAVE SUMMARY SEPTEMBER 2004 B3 Report No. 2005/24 02092-B03.cdr OFFSHORE WAVE CLIMATE: 01−Oct−2004 to 31−Oct−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY OCTOBER 2004 B4 Report No. 2005/24 02092-B04.cdr OFFSHORE WAVE CLIMATE: 01−Nov−2004 to 30−Nov−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

Figure WRL MONTHLY WAVE SUMMARY NOVEMBER 2004 B5 Report No. 2005/24 02092-B05.cdr OFFSHORE WAVE CLIMATE: 01−Dec−2004 to 31−Dec−2004 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY DECEMBER 2004 B6 Report No. 2005/24 02092-B06.cdr OFFSHORE WAVE CLIMATE: 01−Jan−2005 to 31−Jan−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY JANUARY 2005 B7 Report No. 2005/24 02092-B07.cdr OFFSHORE WAVE CLIMATE: 01−Feb−2005 to 01−Mar−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 01 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 01 day

Figure WRL MONTHLY WAVE SUMMARY FEBRUARY 2005 B8 Report No. 2005/24 02092-B08.cdr OFFSHORE WAVE CLIMATE: 01−Mar−2005 to 31−Mar−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY MARCH 2005 B9 Report No. 2005/24 02092-B09.cdr OFFSHORE WAVE CLIMATE: 01−Apr−2005 to 30−Apr−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

Figure WRL MONTHLY WAVE SUMMARY APRIL 2005 B10 Report No. 2005/24 02092-B10.cdr OFFSHORE WAVE CLIMATE: 01−May−2005 to 31−May−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 day

Figure WRL MONTHLY WAVE SUMMARY MAY 2005 B11 Report No. 2005/24 02092-B11.cdr OFFSHORE WAVE CLIMATE: 01−Jun−2005 to 30−Jun−2005 (narrabn) 8

Wave heights Hsig and Hmax (m) 0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

5 Peak Wave Period Tp (s)

0 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 day

Figure WRL MONTHLY WAVE SUMMARY JUNE 2005 B12 Report No. 2005/24 02092-B12.cdr WRL TECHNICAL REPORT 2005/24

Appendix C

Aarninkhof et al., (2003). Coastal Engineering, Vol. 49 (4), p. 275 - 289 Coastal Engineering 49 (2003) 275–289 www.elsevier.com/locate/coastaleng

A video-based technique for mapping intertidal beach bathymetry

Stefan G.J. Aarninkhof a,*, Ian L. Turnerb, Thomas D.T. Dronkersc, Mark Caljouwc, Leann Nipiusc

a WL/Delft Hydraulics, PO Box 177, 2600 MH Delft, The Netherlands b Water Research Laboratory, School of Civil and Environmental Engineering, University of New South Wales, Sydney, Australia c Delft University of Technology, Faculty of Civil Engineering and Geosciences, Delft, The Netherlands Received 28 February 2002; received in revised form 16 May 2003; accepted 13 June 2003

Abstract

Measuring the location of the shoreline and monitoring foreshore changes through time are core tasks carried out by coastal engineers for a wide range of research, monitoring and design applications. With the advent of digital imaging technology, shore-based video systems provide continuous and automated data collection, encompassing a much greater range of time and spatial scales than were previously possible using field survey methods. A new video-based technique is presented that utilises full-colour image information, which overcomes problems associated with previous grey-scale methods, which work well at steep (reflective) sites, but are less successful at flatter (dissipative) sites. Identification of the shoreline feature is achieved by the automated clustering of sub-aqueous and sub-aerial pixels in ‘Hue– Saturation–Value’ (HSV) colour space, and applying an objective discriminator function to define their boundary (i.e., ‘shoreline’) within a time-series of consecutive geo-referenced images. The elevation corresponding to the detected shoreline features is calculated on the basis of concurrent tide and wave information, which is incorporated in a model that combines the effects of wave set-up and swash, at both incident and infragravity frequencies. Validation of the technique is achieved by comparison with DGPS survey results, to assess the accuracy of the detection and elevation methods both separately and together. The uncertainties associated with the two sub-components of the model tend to compensate for each other. The mean difference between image-based and surveyed shoreline elevations was less than 15 cm along 85% of the 2-km study region, which corresponded to an horizontal offset of 6 m. The application of the intertidal bathymetry mapping technique in support of CZM objectives is briefly illustrated at two sites in The Netherlands and Australia. D 2003 Elsevier B.V. All rights reserved.

Keywords: Video-based technique; Intertidal beach bathymetry; Hue–Saturation–Value

1. Introduction

Measuring the location of the shoreline and monitoring foreshore changes through time are core tasks * Corresponding author. Tel.: +31-15-285-88-89; fax: +31-15- 285-87-10. carried out by coastal engineers at many sites around E-mail address: [email protected] the world. The motivation may be research-oriented, (S.G.J. Aarninkhof). but more often is associated with practical applica-

0378-3839/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0378-3839(03)00064-4 276 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 tions that can include: identifying and quantifying The second challenge, of mapping the shoreline shoreline erosion, assessment of the performance of and foreshore from a time-series of digital images, is coastal protection structures, and as a basic input to the focus of the paper. The development and valida- engineering design in the coastal zone. tion of a new colour-based technique to map intertidal Mapping of the shoreline and foreshore has tradi- bathymetry is presented and discussed in some detail. tionally relied upon conventional surveying methods The application of this technique at two sites in The and airphoto interpretation techniques. More recently, Netherlands and Australia to support CZM objectives differential GPS has enabled the more rapid collection is briefly illustrated. of shoreline and foreshore data in the field. With the recent advent of new digital imaging technology, shore-based video systems now provide the additional 2. A new model to map intertidal beach capability of automated data collection, encompassing bathymetry a much greater range of time and spatial scales than were previously possible. Continuous (typically every 2.1. Motivation daylight hour) collection of image data with a resolution of centimetres to meters, extending along regions Models to quantify intertidal beach bathymetry of hundreds of meters to several kilometers, is now from video imagery generally delineate a shoreline routinely undertaken at sites in the USA, Europe, feature from oblique or plan view images and Australia and Asia. estimate the associated elevation from the hydrody- The high temporal and spatial resolution that these namic conditions at the time of image collection new image-based data provide is accompanied by new (e.g. Plant and Holman, 1997). This yields an challenges. Fundamentally, the task is to first identify alongshore elevation contour of the intertidal beach shoreline and foreshore features by the analysis of (Fig. 1). Mapping a time series of contour lines oblique images, and secondly to convert distorted throughout the tidal cycle enables the composition of image (two-dimensional) co-ordinates to their real- the beach surface between the shoreline contours at world (three-dimensional) position. The solution to low and high tide, assuming changes of morphology the second problem is now well established, and is to be small over the period of data sampling (typ- embedded in the Argus coastal imaging system (Hol- ically 6–10 h). man et al., 1993; Aarninkhof and Holman, 1999) that The first generation of video-based shoreline de- underlies the work we present here. tection techniques (e.g. Plant and Holman, 1997;

Fig. 1. Mapping intertidal beach bathymetry from a set of shorelines, derived from time-averaged video observations throughout a tidal cycle. S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 277

Aarninkhof et al., 1997; Davidson et al., 1997) 2.2. Identification of the shoreline feature from video originated from the time that Argus stations regularly collected grey-scale images only. These methods A developmental version of the shoreline identifi- aimed to identify a characteristic pattern in the distri- cation technique was previously described in Aarnink- bution of grey-scale pixel intensities sampled across hof and Roelvink (1999). Briefly, the technique aims the swash zone. The ‘Shore-Line Intensity Maximum’ to delineate a shoreline feature from 10-min time (SLIM) model (Plant and Holman, 1997; Madsen and exposure images, on the basis of distinctive image Plant, 2001) was typical of this approach, using the intensity characteristics in pixels, sampled across the visually observed shoreline break as a proxy for the sub-aqueous and sub-aerial beach, respectively. To location of the shoreline. Application of the SLIM that end, it converts raw image intensities in ‘Red– model was found to be highly robust, easy to auto- Green–Blue’ (RGB) colour space, sampled from a mate and perform well at beaches with a well-pro- region of interest covering both the dry and wet beach, nounced and discrete SLIM feature. to ‘Hue–Saturation–Value’ (HSV) colour space, to These latter criteria occur at steep (reflective) separate colour (Hue, Saturation) and grey scale beaches with mild to rough wave conditions. How- (Value) information (e.g. Russ, 1995). The HSV ever, SLIM features are often diffuse or absent at intensity data are filtered to remove outliers and scaled mildly sloping (dissipative) beaches with emerging between 0 and 1, to improve the contrast between the sandbars. This initiated the development of alternative two clusters of dry and wet pixels. Iterative low-pass shoreline detection techniques, which were still based filtering of the spiky histogram of scaled intensity data upon grey-scale image information, but differed from yields a smooth histogram with two well-pronounced the SLIM approach in that the location of the shore- peaks Pdry and Pwet, which mark the locations of the line was estimated from a characteristic feature in the clusters of dry and wet pixels (Fig. 2a–b). The filtered correlogram of the cross-shore intensity and variance histogram thus obtained is used to define a line l, profile (Aarninkhof et al, 1997) or spatial gradients in formulated as intensity levels in rectified images (Davidson et al., 1997). The absence of a well-pronounced greyscale l: Iy ¼ p1Ix þ p2 ð1Þ contrast between pixel intensities sampled from sub- aerial and sub-aqueous regions of the beach often complicated or prevented unambiguous application of where Ix and Iy represent Hue and Saturation in the the latter techniques. A further difficulty with the case of colour-based discrimination (Fig. 2a), while grey-scale methods was that a site-dependent correc- both represent the Value information in the case of tion to estimate the shoreline elevation was generally luminance-based discrimination (Fig. 2b). For both required (e.g. Plant and Holman, 1997; Davidson et discriminators, the line l crosses the saddle point of al., 1997). the filtered histogram, thus separating the clusters of The problems associated with the application of dry and wet pixels. When applied to an arbitrary first generation shoreline detection models at mildly image, the detection technique performs these oper- sloping beaches like Egmond (The Netherlands) and ations in both the colour (Hue–Saturation) and lumi- the Gold Coast (Queensland, Australia) has initiated nance (Value–Value) domain and assesses what the development of the new model presented here. Our domain provides the highest degree of contrast be- technique identifies the shoreline location using the tween the sub-aqueous and sub-aerial beach. The additional information made available by the use of latter is determined on the basis of a relative measure, full colour images. The horizontal position of the involving the spread of pixel intensities within each shoreline feature is obtained from the visual contrast cluster as regard to the distance between the cluster between the sub-aqueous and sub-aerial beach, while peak P and the discrimination line l. the associated shoreline elevation is estimated from the With the help of l, a discriminator function W(Ix,Iy) offshore tide and wave conditions at the time of image is defined such that W = 0 along l: collection. Both components of this new technique are summarised in Fig. 1, as described below. WðIx; IyÞ¼p1Ix þ p2 Iy ð2Þ 278 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289

Fig. 2. Shoreline detection from contrasting pixel intensity characteristics at the sub-aqueous and sub-aerial beach. Video-based estimates of shoreline locations (c,d) at Noordwijk, Netherlands, determined from the clustering of pixel intensities in the colour (a) and luminance (b) domain, respectively.

Evaluation of Eq. (2) for the scaled intensity data demonstrating the capability of this method to resolve (Ix,Iy) sampled from each individual pixel within the three-dimensional morphology including emerging region of interest yields a map of the function W, intertidal bars. An important attribute of this technique which shows positive (negative) W at the sub-aerial is that no site-specific calibration is required, and the (sub-aqueous) beach, for both colour-based and lumi- method works equally well along reflective (steep) and nance-based cluster distinction. These maps enable the more dissipative (flatter) coastlines. detection of the shoreline feature in image space at the location of the W = 0 elevation contour. Geo-referenc- 2.3. Determination of the shoreline elevation at the ing of this feature is achieved through application of time of image collection sophisticated video-processing techniques (Holland et al., 1997). The result may locally show erroneous During the 10 min of time exposure for image contours at the sub-aerial beach, which are associated collection, the instantaneous location of the waterline with the irregular intensity characteristics of features (i.e. the interface of the sub-aerial and sub-aqueous like water-filled, detached runnel systems or vehicles beach) is affected by the offshore tidal level, storm at the beach. These are removed through application of surge, breaking induced wave set-up and swash oscil- empirical demands on shoreline persistency in both lations (Fig. 3). As described above, the shoreline real world and image space. Example results repre- detection technique analyses time-averaged, colour senting colour-based and luminance-based shoreline video data to identify a beach contour at some location detection at Noordwijk are shown in Fig. 2c–d, xsl within the swash zone of width Dxosc (i.e. the grey- S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 279

Fig. 3. Physical processes affecting the instantaneous waterline location. Artificial timestack of swash run-up a plane beach with slope m, for an energy spectrum dominated by short waves with peak period Tp and long waves with a period 7Tp. The instantaneous waterline elevation zwl is affected by the offshore water level zo outside the surf zone, the breaking induced wave set-up g and an oscillatory component at the time scale of individual waves and wave groups.

shaded area in Fig. 3). In general, xsl represents some measured, and local wind set-up between the tide (as yet unknown) location within the swash zone, gage and the shoreline is limited to a few centimetres. which is associated with a certain level of swash The wave set-up gsl is computed from a standard wave exceedence. The elevation of this shoreline feature decay model (Battjes and Janssen, 1978) that incor- does not necessarily coincide with the 10-min time- porates the roller concept (Svendsen, 1984; Stive and averaged waterline location xavg during this same time De Vriend, 1994) to delay the dissipation of organised period. energy. An inner surf zone bore model (Aarninkhof The formulation for the water level elevation zsl and Roelvink, 1999) is used to extend computations that corresponds to the horizontal position xsl is up to zero water depth. defined by To quantify the contribution of oscillations at incident (frequency f>0.05 Hz) and infragravity gosc ( f < 0.05 Hz) frequencies to the overall swash height zsl ¼ z0 þ gsl þ Kosc ð3Þ 2 gosc (Eq. (3)) at the shoreline, we adopt empirical formulations for the sea swell swash height Rss and where z0 is the tide- and wind-induced offshore water the infragravity swash height Rig. Holman and Sal- level without the contribution of gravity waves, gsl is lenger (1985) found the normalised infragravity swash the wave-breaking induced mean rise of the water height Rig/H0 and the normalised sea swell swash level at the shoreline (hereafter referred to as wave height Rss/H0 both to be linearly related to the Iribar- set-up) and gosc represents the vertical swash height, ren number n0 as related to waterline oscillations at the time scale of individual waves and wave groups. The swash pa- Rig ¼ 0:53n0 þ 0:09 ð4aÞ rameter Kosc is a constant, non-site-dependent empir- H0 ical coefficient that accounts for the level of swash exceedence associated with the beach contour re- Rss turned from the shoreline detection model. ¼ 0:69n0 0:19 ð4bÞ H0 The water level z0 is preferably obtained within pffiffiffiffiffiffiffiffiffiffiffiffiffi approximately 10 km off the coastline, so that varia- where n0 ¼ tanðmÞ= H0=L0, m is the local foreshore tions of the tide and large-scale storm surge are slope and L0 is the deep water wave length, determined 280 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 with the peak period Tp. Eqs. (4a) and (4b) were thus, ignoring the short-wave contribution to swash derived for n0 varying between 0.5 < n0 < 3.5. Stock- oscillations during highly dissipative conditions (n0 < don et al. (2002) confirm the general applicability of 0.275). The latter can be justified from the measure- these empirical relationships over a range of reflective ments by Ruessink et al. (1998), where the average natural beaches. In contrast, at a low-sloping, dissipa- ratio of infragravity to total swash height Rig/R was tive beach at Terschelling (The Netherlands), Ruessink found to be 0.85. As both Rig and Rss are determined as et al. (1998) found a significantly larger dependency a fraction of the offshore significant wave height H0 between Rig/H0 and n0, parameterised as defined as 4r (where r is the standard deviation of the sea surface elevation), the overall gosc can be calculat- Rig ed as ¼ 2:20n0 þ 0:02 ð5Þ H0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 gosc ¼ Rig þ Rss ð8Þ The Ruessink et al. (1998) data set represents n0 n ranging between 0.05 < 0 < 0.35. The increase of the where Rig and Rss are obtained from the empirical constant of proportionality as compared to Eq. (4a) is parameterisations (Eqs. (6)–(7b), respectively). The ascribed to the saturation of the higher infragravity result obtained from Eq. (8) is used to estimate gsl (Eq. frequencies for situations with small n0. To accom- (3)), applying the site-independent empirical constant modate both dissipative and reflective conditions, Kosc to account for the level of swash exceedence Ruessink et al. (1998) suggest a relationship associated with the beach contour obtained from the shoreline detection model. Rig ¼ 0:65tanhð3:38n0Þð6Þ H0 3. Validation of the intertidal beach mapper which reduces to Eq. (5) for highly dissipative conditions. For more reflective conditions, Eq. (6) con- This section discusses the individual and combined verges to a constant value of 0.65, closely matching validation of both sub-models that together comprise the relationship reported by Raubenheimer and Guza the intertidal beach mapper technique (Fig. 1), against (1996), but underestimating the parameterisation a data set of GPS-surveyed shorelines sampled at according to Holman and Sallenger (1985) for n0> Egmond, The Netherlands (Fig. 4). 1.05. To date, the validity of Eq. (6) in the transitional range between dissipative and reflective conditions 3.1. Data set has not been established because of a lack of measured data in the range 0.2 < n0 0.6. Pragmatically, A data set of 52 measured shorelines was collect- being the only parameterisation for Rig that caters for ed at a nourished beach in front of the Egmond both dissipative and reflective conditions, it is adopted boulevard, during the periods November 29–30, here for use with the shoreline elevation model. 1999 and March 14–15, 2000. The shoreline surveys In search of the quantification of Rss, the shoreline were conducted by moving a differential GPS sys- elevation model adopts Eq. (4b) after Holman and tem, mounted on a jeep, over a distance of approx- Sallenger (1985). To prevent the occurrence of nega- imately 2-km alongshore, at the higher part of the tive Rss/H0 as would be obtained for small n0 outside swash zone. These data yield a set of beach contours the measured range of Holman and Sallenger, Rss/H0 is with a vertical accuracy of the order of 2 cm. The set to zero for calculated values of n0 less than 0.275, surveys were carried out on a semi-hourly basis, simultaneously to the recording of time-averaged Rss ¼ 0:69n0 0:19 for n0 > 0:275 ð7aÞ video images by the five camera Argus video station, H0 mounted at 43 m above sea level on top of the Egmond lighthouse ‘Jan van Speyk’ (Fig. 4).Off- Rss shore wave conditions (rms wave height Hrms, peak ¼ 0 for n0 < 0:275 ð7bÞ H0 period Tp and angle of incidence h0) were measured S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 281

Fig. 4. Argus video station ‘Jan van Speyk’, mounted on top of the Jan van Speyk lighthouse in Egmond (Netherlands). with a directional wave buoy at IJmuiden, located video-based and GPS surveyed shorelines are evalu- approximately 15 km to the south of the nourished ated on a grid with 2 m spacing alongshore. site. Offshore tidal levels were interpolated using water level data collected at two tidal stations located 3.2. Error quantification 15 km north and south of Egmond. During the experiment, Hrms ranged between 0.6 and 2.3 m, To separately assess the shoreline position and while the offshore tidal level spanned a range be- shoreline elevation sub-models (Fig. 1), deviations tween 0.7 and 1.0 m NAP (Dutch ordinance between modelled and surveyed shorelines were level). quantified by means of a horizontal offset dd, induced The intertidal beach mapper was used to map by the shoreline detection model, and a vertical offset shorelines from all 260 images collected simulta- de, induced by the elevation model (Fig. 5), neously to the field measurements. On the basis of visual inspection (e.g. people on the beach, poor ddðy; tÞ¼xvðy; tÞxsðy; tÞð9aÞ visibility, etc.), 137 shorelines were accepted for further analysis. Lacking information on the actual deðy; tÞ¼zvðy; tÞzsðy; tÞ; ð9bÞ surf zone bathymetry, gsl (Eq. (3)) was estimated by running the inner surf zone model on an equilibrium where xv( y,t) and xs( y,t) represent the shoreline posi- beach profile (Dean, 1977) calculated for the site. tion x (positive offshore) at alongshore location y and Previous investigations (Janssen, 1997) show that this time t, as identified from the video analysis and field simplification introduces only minor deviations of survey, respectively. Similarly, zv(zs) is the shoreline order 1–2 cm in terms of gsl at the shoreline. The elevation obtained from video (field survey). To foreshore slope m at the intertidal beach was set at a facilitate the inter-comparison, both deviations are fixed value 1:40, which is characteristic for the low interpreted as vertical offsets dzd and dze,which gradient beach at Egmond. Deviations between the demands the mapping of dd on a vertical plane with 282 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289

3.3. Assessment of model performance

The individual performance of the detection (ele- l vation) model is quantified by means of the mean dzd

(ldze) and standard deviation rdzd (rdze)ofdzd (dze) over time, as function of the alongshore location y. The number of shorelines contributing to the statistics at an arbitrary location y* varies with the location Fig. 5. Error quantification against GPS-surveyed shorelines. alongshore. Statistics were only determined if at least Quantification of model deviations by means of detection and 10 shorelines were found in y*, out of a theoretical elevation induced offsets dd and de. Scenario 1 represents perfect maximum of 52. As a result, 4.3% of the grid points, performance of the elevation model, Scenario 2 represents perfect all located in overlap regions between cameras or at performance of the detection model and Scenario 3 the general case the northern end of the area of interest, were excluded with non-zero dd and de. from the statistical analysis. The results show negative l along virtually the the help of the foreshore beach slope m, assuming m dzd to be small such that tan(m) c m: entire area of interest (Fig. 6a), indicating that the horizontal position of video-derived shorelines were located landward of the surveyed shorelines. In abso- dzdðy; tÞ¼m ddðy; tÞð10aÞ l lute sense, the detection induced deviations dzd were generally less than 10 cm (with a relatively constant dzeðy; tÞ¼deðy; tÞð10bÞ rdzd of about 15 to 20 cm) along the major part of the area of interest. The exception to this was in the far- The overall error dz( y,t) is determined as the sum field region to the north of the video station. Appli- of cation of the elevation model (Fig. 6b) typically l r yielded mean deviations dze up to 10 cm, with a dze dzðy; tÞ¼dzdðy; tÞþdzeðy; tÞð11Þ l in the order of 10 to 15 cm. The local increase of dze r and dze near y = 800 m was related to the presence of Note that this approach allows the overall error dz a local seaward morphological extension (clearly to be smaller than the absolute values of the individual visible on plan view images of November 29, components dzd and dze. This can be justified with the 1999), which induced a local increase of zs, hence help of Fig. 5. In the case of a perfect estimate of the negative ldzd. shoreline elevation (Scenario 1, dze = 0), dz is entirely Similar trends were observed when considering the governed by the detection induced error dzd. In the overall model performance in terms of the mean ldz case of perfect estimate of the shoreline location and standard deviation rdz of dz over time (Fig. 7). combined with a poor elevation estimate (Scenario The tendency was for the absolute magnitude of the 2), the opposite occurs. In general, both dzd and dze deviations to decrease, owing to the mutual compen- will be non-zero. If dzd and dze have opposite signs, sation of errors resulting from the individual sub- this yields a decrease of the overall error dz, since models. ldz typically amounted to 10–20 cm along errors resulting from both sub-models compensate for the entire region of interest, except for the far-field each other. This situation occurs for instance if a region north of the video station where ldz locally seaward offset of the video-derived location of the increases up to 30 cm. In an absolute sense, ldz was shoreline (positive dv) is compensated by an underes- less than 15 cm along 85% of the 2 km wide area of timate of the shoreline elevation (Scenario 3). The interest, which corresponded to an horizontal offset of resulting dz is small as compared to the individual 6 m. With rdz in the order of 15 to 20 cm throughout components dzd and dze, which matches the real world most of the region of interest, the scatter of results is situation where shoreline estimate 3 (Fig. 5) is much relatively constant, with larger values again found in closer to the actual bathymetry than the estimates 1 the far-field region to the north and the non-uniform and 2. area near y = 800 m. S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 283

Fig. 6. Error quantification per sub-model. Time-averaged mean (bold dots) of the detection (a) and elevation model (b) induced vertical offsets dzd and dze, as a function of the location alongshore. The fine dots represent the scatter, quantified as the mean F the standard deviation of dzd and dze. The positive y-axis is pointing south, with the video station being located around y = 120 m.

3.4. Sensitivity to the empirical swash parameter Kosc The modelled mean shoreline elevation increased with increasing Kosc, yielding a decrease of the rms value The results presented here were obtained with a of dze for Kosc ranging from 0.2 to 1. For Kosc above 1, parameter setting Kosc = 1.20. To assess model sensi- the rms value of dze increased, indicating that the tivity to variable Kosc, Table 1 summarises model modelled shoreline elevation increasingly exceeded performance for different Kosc, quantified by means the surveyed elevation. However, the combined error, of the rms error of the detection induced dzd, the quantified by means of the rms value of dz, further elevation induced dze and the overall offset dz, in- decreased for Kosc values in exceedence of 1. This volving a total number of 29,551 samples that com- reflects the mechanism of mutual error compensation pose the entire data set of 137 video-based shorelines. noted earlier. Minimum deviations in terms of dz are

Fig. 7. Error quantification overall model. Time-averaged mean (bold dots) of the overall vertical offsets dz, as a function of the location alongshore. The fine dots represent the scatter, quantified as the mean F the standard deviation of dz. The positive y-axis is pointing south, with the video station being located around y = 120 m. 284 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289

Table 1 the histogram of image intensities and the resulting Sensitivity of the intertidal beach mapper to variable Kosc discriminator function (Fig. 2) were biased towards Kosc rms (dzd) rms (dze) rms (dz) the colour intensity characteristics of the near field. 0.20 0.194 0.209 0.318 More difficult to control are atmospheric effects that 0.50 0.194 0.153 0.259 increasingly affect image clarity, hence model perfor- 0.80 0.194 0.122 0.211 mance, in the far-field. The decrease in model perfor- 1.00 0.194 0.125 0.190 1.20 0.194 0.148 0.182 mance in the far-field to the north of the video station 1.40 0.194 0.182 0.188 is further explained by the presence of buildings at the Egmond site, which obscure the visibility of the dry beach in this region, thus further limiting the oppor- obtained with a parameter setting Kosc = 1.2, indicat- tunities for the clustering of dry pixels. ing that the detection technique identified a shoreline Apart from imperfections in the shoreline detection near the higher end of the swash run-up. and elevation models as quantified above, the model accuracy reported here was also affected by the survey 3.5. Discussion data. This observation concerns the survey method rather than the fundamental measurement accuracy of Model validation against a data set of GPS sur- the differential GPS system (1–2 cm), which is an veyed shorelines at Egmond showed that mean verti- order of magnitude better than the image analysis cal deviations, in absolute sense, were less than 15 cm technique described here. The elevation model as- (corresponding to a mean horizontal offset of 6 m) sumes constancy of the time-averaged shoreline ele- along more than 85% of the 2-km-long area of vation during the 10-min of time exposure for image interest. Model deviations increased with increasing collection. So an ideal shoreline measurement would distance from the video station. The bulk statistics of be surveyed along a perfectly horizontal track. This is the overall data set of 29,551 shoreline samples hard to achieve in the field where the instantaneous showed a mean, detection induced, vertical deviation location of the waterline is continually affected by of 8.5 cm (with standard deviation 17.4 cm), which oscillating swash motions over a complex intertidal reflects a landward offset of the video-derived shore- bathymetry. The 137 GPS-surveyed shorelines that line location. This deviation is largely compensated by were used to quantify the offset of the corresponding a mean elevation induced offset of + 7.8 cm (with 137 video-derived shorelines show a mean standard standard deviation 12.6 cm). Considering these stan- deviation of 4.7 cm. In other words, the scatter values dard deviations, it can be seen that the scatter of the rdz reported here due to the image-based shoreline overall model deviations was dominated by uncertain- detection-elevation models represent the upper esti- ties resulting from the shoreline detection technique. mate of these errors, due to the inherent variability in The overall model deviations found here were the the ground truth data. same order of magnitude as the vertical excursion of The relatively large settings Kosc = 1.20 found here the oscillating swash motion, indicating that the indicates that the detection technique identifies a technique was consistently identifying shorelines shoreline feature near the upper end of the swash within the swash region. run-up. However, this may also suggest that the

Theincreaseofldzd and rdzd with increasing empirical swash formulations which form the basis distance from the video station can be explained by of the shoreline elevation model tend to underestimate the discriminator function W becoming less represen- the real-world vertical swash excursion. The latter tative in the far-field of the rather large region of may well be the case, as Eqs. (7a) and (7b) ignore interest used for this comparison. The results pre- the short-wave contribution Rss at dissipative beaches sented here were obtained by sampling pixel intensi- (n0 < 0.275). For reflective conditions characterised by ties directly from oblique images. Owing to a decrease n0>1.05, the Ruessink et al. (1998) parameterisation of the pixel resolution in the far-field, which is as- for infragravity swash (Eq. (6)) underestimates the sociated with less pixels per unit area, far-field infor- Holman and Sallenger (1985) relationship. Both mation is relatively poorly represented. Consequently, aspects contribute to an underestimate of the real- S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 285 world vertical swash excursion over a wide range of the entire monitoring region. In March 2000, this had n0, which is compensated by the large Kosc settings resulted in coastal retreat of 30 m in the north, and more found here. It is concluded that the calibration of Kosc than 40 m in the south. At some locations, virtually no accounts for uncertainties in both the location of the sub-aerial beach was left, despite the nourishment shoreline feature identified from time-averaged video activities which were executed only 9 months previ- imagery as well as the associated elevation estimated ously. Between April and June 2000, a degree of from the empirical parameterisations for the vertical recovery ( f 20 m additional beach width) was ob- swash excursion. served; full restoration of the morphological configuration of June 1999 had however not been achieved. A second beach nourishment was completed by the 4. Model application at Egmond and the gold coast end of June 2000, along the strip between 0 and 800 m south of the video station. Shoreline retreat of 10– 4.1. Video monitoring of a nourished beach at 20 m per month continued along the nourished section Egmond over the period August–September 2000, as the nourished sand transferred from the intertidal beach To mitigate local beach erosion in front of the to restore an erosion hot-spot around 700 m south of township of Egmond aan Zee (Fig. 4), the coastal the video station (Fig. 8). Rough wave conditions authorities undertook combined beach and shoreface (upper panel, Fig. 8) in November and December nourishment in the early summer of 1999. The 200 2000 resulted in further erosion, and a shoreward shift m3/m beach nourishment extended a distance of 1500 of the shoreline by 15–20 m, relative to the initial m alongshore; the 400 m3/m shoreface nourishment, beach conditions measured in June 1999. Early in completed at 5 m water depth at the seaward side of 2001, a ‘slug’ of sand entered the intertidal zone the outer bar, extended 2200 m. To monitor the directly north of the video station (Fig. 8), causing a effectiveness of the combined nourishment and its net local accretion in the order of 50 m, accompanied effect on the evolution of nearshore morphology, an by the flattening of the beach profile and the devel- Argus video station was installed on top of the ‘Jan opment of strong morphological irregularities in the van Speyk’ light house in May 1999. The intertidal intertidal zone. The redistribution of sediments along- beach mapper was used to quantify intertidal beach shore, combined with an overall accretionary trend, bathymetry on a monthly basis, between the NAP 0 resulted in an increasingly smoothed shoreline in and + 1 m elevation contours. April 2001 with this trend continuing through to end The initial intertidal morphology of June 1999, 2 of the summer in August 2001. Interestingly, the plan- months after the completion of beach and shoreface shape morphological configuration of August 2001, nourishment, was characterised by a highly irregular characterised by a seaward extension of the foreshore shoreline, with an erosion hot-spot located approxi- directly north of the video station, and a contrasting mately 500 m south of the video station, and consider- erosion hot-spot approximately 600 m to the south, is able accretion at 200 m north of the station (Fig. 8). The remarkably similar to the initial situation in June width of the beach varied of by more than 60 m within a 1999, despite the nourishment effort in July 2000. distance of 700 m. After a calm summer period during This latter observation suggests that the morphologi- which only minor foreshore changes were observed cal evolution of the intertidal beach is at least partly (Fig. 8), a sequence of storm events in October and governed by larger-scale phenomena, for example the November 1999 (upper panel, Fig. 8) causes significant presence of a depression in the outer bar. erosion of the beach, in particular at the location of In summary, 2 years of video-based monitoring of minimum beach width at about 400 m south of the intertidal coastal changes at a nourished beach on the video station (Fig. 8). The flattening of the beach west coast of The Netherlands have shown significant profile was also observed, identified from the image- variability in the morphodynamic behaviour at the site. derived survey data by the widening of elevation con- Averaged over the area of interest, a strong seasonal tours. During the winter months (December 1999– variability was observed, and considerable spatial February 2000), ongoing erosion was observed along variability occurred through redistribution of sedi- 286 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289

Fig. 8. Video-observed evolution of intertidal bathymetry at Egmond over the period June 1999–August 2001. Monthly, plan view maps of the intertidal beach bathymetry between the NAP + 0 and NAP + 1 m elevation contours. The dry beach is located at the lower side of each panel. The video station is located near the origin of the horizontal coordinate axis, which is positively directed south. Elevations at the sub-aqueous

(sub-aerial) beach are manually set to zero (one). The upper panel shows Hrms for the period of interest. ments within the area of interest. In the past, this chronic erosion problem, and is contributing to the variability had the effect of obscuring a (previously design of more effective mitigation measures. unrecognised) chronic erosion problem in the southern region of the monitored area. This had resulted in the 4.2. Video monitoring of coastal changes at the Gold design and implementation of an inappropriate coastal Coast Reef (Australia) management option (i.e., sand nourishment). The application of the intertidal beach mapper technique In 1997, the ‘Northern Gold Coast Beach Protec- at the site clearly identified and located the underlying tion Strategy’ (Boak et al., 2000) was implemented S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 287 by Gold Coast City Council to maintain and en- shore morphological changes during the construction hance the beaches extending from Surfers Paradise phase of the submerged reef structure. Intertidal beach to Main Beach (Fig. 9). The aim of the Strategy was profiles were mapped on a monthly basis over the 18- to decrease the risk of potential economic loss month period January 2000 to August 2001. The area (infrastructure and tourism) following storms events, of interest focussed on a 1000-m length of the beach, by increasing the volume of sand within the storm centered at the reef construction site. In contrast to the buffer seaward of the existing oceanfront boulder presentation of monthly intertidal beach profiles as wall. The Strategy has the dual objectives of in- shown for the Egmond site (Fig. 8),aquarterly creasing the width of the sub-aerial beach, and summary of relative changes in the intertidal beach improved surfing conditions. The major components profile can also be calculated (Fig. 10) to assess the of the engineering works included an initial 1.2 alongshore distribution of accretion–erosion. Mm3 of beach nourishment along approximately In July 2000, sand nourishment of northern Gold 2000 m of the coastline, and construction at Narrow- Coast beaches was completed, and the video-derived neck (Fig. 9) of a submerged artificial reef structure results shown in Fig. 10 were calculated relative to to provide a coastal ‘control point’ and enhance the beach at that time, encompassing the subsequent surfing opportunities. In August 1999, an Argus 12-month period. An initial phase of accretion was coastal imaging system was installed at the site, observed to have occurred in the region between immediately prior to the start of reef construction, 600 and 1100 m alongshore, coinciding with a and 6 months following the commencement of period of relatively calm waves. After than time, it beach nourishment. was observed that sand generally migrated north- Within the scope of a larger video-based monitor- wards, leaving behind a localised region of accretion ing program (Turner et al., in press), the intertidal in the lee and immediately up-drift of the reef beach mapper technique was applied to assess fore- structure (Fig. 10). The alongshore migration speed

Fig. 9. The Goldcoast Argus video station, mounted on top of the Focus building at Narrowneck (Queensland, Australia). 288 S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289

Fig. 10. Video-observed evolution of intertidal bathymetry at the Gold Coast over the period July 2000–June 2001. Quarterly maps of coastal changes at the intertidal beach, relative to the coastal morphology at the time of completion of the beach nourishment. The dry beach is located at the lower side of each panel. The video station is located near the origin of the horizontal coordinate axis, which is positively directed north. was estimated based upon the monthly accretion– the site on an ongoing basis, to determine when erosion calculations. By this method a net migration additional nourishment will be required. speed of f 1 m/day was determined for calm weather conditions (Hs f 1 m during September 2000–December 2000) and about 8 m/day for rough 5. Conclusions weather conditions (Hs f 2 m during December 2000–February 2001). It is interesting to note that This paper has presented the details and validation these estimates of the migration speed using the of a video-based method for monitoring of morpho- video-derived estimates of foreshore accretion–ero- logical changes at the intertidal beach. sion (Fig. 10) correspond well with the simple The quantification of intertidal beach bathymetry empirical approach proposed by Sonu (1968).By is achieved by mapping multiple shorelines during a this method, the predicted average annual migration tidal cycle. The technique presented here is based speed for this same 12-month period was approxi- upon two independent sub-models. The first identi- mately 6 m/day. fies the location of the shoreline from time-averaged In summary, video-based monitoring of reef con- video images, based on colour differences between struction and nourishment works at a dissipative site the wet and dry beach. The second estimates the on the east coast of Australia, has provided coastal associated vertical elevation from the hydrodynamic managers the ability to assess and quantify the effec- conditions offshore. These sub-models were validated tiveness of the new coastal protection works to meet separately, indicating that the waterline detection sub- project objectives. The application of the intertidal model is largely responsible for deviations between beach mapper technique is continuing to be used at the video-based waterline and the GPS-surveyed S.G.J. Aarninkhof et al. / Coastal Engineering 49 (2003) 275–289 289 waterline. Uncertainties of the overall model are Aarninkhof, S.G.J., Janssen, P.C., Plant, N.G., 1997. Quantitative typically in the order of 15 cm in the vertical sense, estimations of bar dynamics from video images. Proc. Coastal Dynamics Conf., Plymouth, UK, pp. 365–374. which is of the same order of magnitude as the Battjes, J.A., Janssen, J.P.F.M., 1978. Energy loss and set-up due to dimensions of the swash zone. From this validation, breaking in random waves. Proc. of 16th Int. Conf. on Coastal it can be concluded that the model is well suited for Eng. ASCE, New York, pp. 569–587. monitoring intertidal beach changes at the time scale Boak, L., McGrath, J., Jackson, L.A., 2000. IENCE—a case of weeks to months. study—the Northern Gold Coast Beach protection strategy. Proc. of 27th Int. Conf. on Coastal Eng., vol. 4. ASCE, Sydney, Application of the model to two different Argus pp. 3710–3717. stations at Egmond and the Northern Gold Coast have Davidson, M., Huntley, D., Holman, R., George, K., 1997. The shown significant morphological changes in the inter- evaluation of large scale (km) intertidal beach morphology on tidal beach, which can be related to seasonal variabil- a macrotidal beach using video images. Proc. Coastal Dynamics ity, placing of nourishment, storm-based erosion Conf., Plymouth, UK, pp. 385–394. Dean, R.G., 1977. Equilibrium beach profiles: US Atlantic and Gulf events and the construction of a reef respectively. coast. Ocean Engineering Report, vol. 12. University of Dela- The observed evolution of the intertidal beach indi- ware, Newark, DE. 45 pp. cates variability on relatively short temporal and Holland, K.T., Holman, R.A., Lippmann, T.C., Stanley, J., Plant, spatial scales, which are not easily observed with N.G., 1997. Practical use of video imagery in nearshore more traditional field methods. oceanographic field studies. IEEE Journal of Oceanic Engi- neering 22 (1), 81–92. Holman, R.A., Sallenger Jr., A.H., 1985. Setup and swash on a natural beach. Journal of Geophysical Research 90, 945–953. Acknowledgements Holman, R.A., Sallenger Jr., A.H., Lippmann, T.C., Haines, J.W., 1993. The application of video image processing to the study of This study was funded by the Interfaculty Research nearshore processes. Oceanography 6 (3). Janssen, P.C., 1997. Intertidal beach level estimations from video Centre ‘Observation of the Earth and Earth’s crust’ at images. MSc. Thesis. Delft University of Technology, Faculty of Delft University of Technology, the Dutch Ministry of Civil Engineering and Geosciences. Transport and Public Works (Rijkswaterstaat). The Madsen, A.J., Plant, N.G., 2001. Intertidal beach slope predictions Argus station at the Gold Coast is funded by Gold compared to field data. Marine Geology 173, 121–139. Coast City Council. SGJA and TDTD were co- Plant, N.G., Holman, R.A., 1997. Intertidal beach profile estimation using video images. Marine Geology 140, 1–24. sponsored by the Delft Cluster Project Coasts Raubenheimer, B., Guza, R.T., 1996. Observations and predic- (03.01.03) and the EU-funded CoastView project tions of wave run-up. Journal of Geophysical Research 101, (contract number EVK3-CT-2001-0054). ILT collabo- 25575–25587. rated in the preparation of the revised manuscript Ruessink, B.G., Kleinhans, M.G., Van Beukel, P.G.L., 1998. Ob- while working as a guest at WL/Delft Hydraulics. The servations of swash under highly dissipative conditions. Journal of Geophysical Research 103, 3111–3118. Argus video technique has been developed with funds Russ, J.C., 1995. The Image Processing Handbook, 2nd ed. Boca generated by the Coastal Imaging Laboratory, Oregon Raton (Florida), USA. CRC Press (ISBN 0-8493-2516-1). State University. The authors wish to acknowledge Sonu, C.J., 1968. Collective movement of sediment in littoral Prof. Rob Holman, OSU, for actively and generously environment. Journal of Coastal Engineering, 373–400 stimulating the collaboration within the worldwide (Chapter 24). Stive, M.J.F., De Vriend, H.J., 1994. Shear stresses and mean flow Argus research group. in shoaling and breaking waves. Proc. Int. Conf. Coastal Eng. ASCE, New York, pp. 594–608. Stockdon, H.F., Holman, R.A., Sallenger Jr., A.H., 2002. Parame- References terisation of incident and infragravity swash variance. Eos Trans. AGU, vol. 83 (47), p. F746. Fall Meet. Suppl. Svendsen, I.A., 1984. Wave heights and set-up in a surf zone. Aarninkhof, S.G.J., Holman, R.A., 1999. Monitoring the nearshore Coastal Engineering 8, 303–329. with video. Backscatter 10 (2), 8–11. Turner, I.L., Aarninkhof, S.G.J., Dronkers, T.D.T., McGrath, J., in Aarninkhof, S.G.J., Roelvink, J.A., 1999. Argus-based monitoring press. CZM applications of Argus coastal imaging at the Gold of intertidal beach morphodynamics. Proc. of Coastal Sediments Coast, Australia. Journal of Coastal Research. Conf., Long Island (NY), USA, pp. 2429–2444.