DATASET

  Title

Onshore Coastline Habitats of the Recherche Archipelago

  Custodian

 Department of Conservation and Land Management (CALM)

  Jurisdiction

 Western Australia

DESCRIPTION

  Abstract

 This dataset consists of linework detailing the onshore coastline habitats of Southern Western Australia's Recherche Archipelago and covers approximately 983km of coastline extending from stokes inlet (121oE) eastward to Isrealite Bay (124o15'E).
The coastline is delineated into areas of beach, rocky shore, N/A or beach + rocky shore and are further attributed with the direction the habitat is facing, habitat length (metres), beach width (maximum and average in metres). The base coastline linework was provided by DOLA.  A source attribute has been added to each line based on the coastline source and/or method used to derive the coastline in any area.
Habitats were identified from digital orthophoto mosaics, unrectified hardcopy and digital aerial photography and local knowledge of the physical characteristics of the islands in the area. Beach widths were measured from the digital orthophoto mosaic and aerial photography using ArcViews measure tool and using a scaled adjustable magnifying glass for the hardcopy photography.  Habitat facing directions were determined using an avenue script to compute a bearing from North.  The dataset was compiled by Ben Lamb in March-June 2002.

  Search Word(s)

 

Geographic Extent Name(s) 

 WA South Coast (WSC) and Eucla (EUC) IMCRA regions

DATA CURRENCY

  Begin Date

 22 February 1991

  End Date

 30 March 2000

DATASET STATUS

  Progress

 In Progress

  Maintenance & Update Frequency

As required

ACCESS

  Stored Data Format

 DIGITAL ArcView shapefile, projected to Map Grid of Australia Zone 50 from the Geocentric Datum of Australia 1994 (GDA94). 

  Available Format Type

 DIGITAL ArcView 3.2 shapefile

  Access Constraint

Data available for external use subject to transfer fee and license conditions.
Data is not to be distributed without authorisation from CALM. 
Contact CALM's database administrator for further details.

DATA QUALITY

  •   Lineage
  • The classification was undertaken in three stages.  Stage 1 was the classification of the coastline using three digital orthophoto mosaics supplied by DOLA and the selection of existing photography to be supplied by DOLA in a digital form.  Stage 2 involved the basic georeferencing of the unrectified digital aerial photography supplied by DOLA and using this as the base for habitat classifications.  Stage 3 was the classification of the remainder of the coastline from older B&W hardcopy aerial photography and the classification of offshore islands using Kevin Bancrofts local knowledge of the physical characteristics of the islands.  Details for all stages are supplied below.
    A standard broad scale marine habitat classification system was developed by CALM Marine Conservation Branch (see T:\144-Marine Conservation Branch\Shared Data\Current_papers\Habitat Classification\SMECCS_291101.doc). 
    NOTE: The classification system is being refined on an on-going basis.

    STAGE 1 (Carried out by Ben Lamb in March 2002)
    1. The dataset 'coastline_wa_20010228_ll_gda94'  was clipped to the cover the extent of the proposed study area for the proposed Recherche Archipelago marine park.   As there was no High Water Mark only the line representing Mean High Water Mark was required for the classification so the Low Water Mark lines were removed.  This coastline was used as the base linework for the coastline habitats to be 'split' and attributed as necessary. 
    2.  This shoreline was then projected to MGA Zone 51 coordinates using the projection functionality of the extension 'calm_add_funk_2001.avx' in Arcview3.2 to facilitate overlaying the orthophotos with the shoreline. 
    2.   The digital orthophoto mosaics (esperanceE1.ecw, esperanceE2.ecw and esperanceW.ecw) were then used as a spatially rectified reference to delineate habitat boundaries on the base coastline. 
    Habitat identification was determined visually from the digital orthophoto. Habitat boundaries were delineated (according to CALM's standardised habitat classification system) into areas of beach, rocky shore, N/A or beach + rocky shore and attributed accordingly.  Delineation was done using the 'split line' functionality of Arcview 3.2 and was done where there was a change in habitat or a significant change in direction of the coastline.
    Lines containing beach habitats were then attributed with beach width (average and maximum). This was measured on the digital orthophoto using Arcviews  measure tool functionality.  The lines were all split whilst moving from East to West along the coast or in the case of an island in an anti-clockwise direction.  This systematic splitting of the coast was done to enable the used of scripts to automate the facing direction calculations.

    The digital orthophoto mosaics were the attribute source for 272.7 km of the total 983.1 km of shoreline.

    STAGE 2 (Carried out by Ben Lamb in March 2002)
    1.  DOLA supplied 215 colour aerial photos scanned at 453 dpi.  These photos were taken from February 1996 to March 2000. These digital photos then underwent basic georeferencing by Ben Lamb.  The georeferencing was done by cropping the image to the actual picture area and recording the size in pixels of the cropped image.  The true scale of each photo was calculated using the focal length of the camera and the flying height, this was then used to calculate the ground dimensions of a pixel separately for each frame.  This equared to approximately 1.43m.  The coordinate for the upper left pixel was calculated by using the photo centre supplied by DOLA in AGD84 Zone 51 coordinates, half the number of pixels in the x and y directions and the ground size of the pixels.  A world file in the required Arcview format was then generated using this information to allow the display of the photos in Arcview and overlaying the shoreline. 
    NOTE: The images have been georeferenced only and have not undergone any rectification.  The accuracy of the georeferencing is low and the georeferenced photos should only be used for applications where the high spatial accuracy is not required
    2.   Habitat identification was determined visually from the digital aerial photography and inaccuracies in the georeferenceing were accounted for in the classification. Habitat boundaries were delineated (according to CALM's standardised habitat classification system) into areas of beach, rocky shore, N/A or beach + rocky shore and attributed accordingly.  Delineation was done using the 'split line' functionality of Arcview 3.2 and was done where there was a change in habitat or a significant change in direction of the coastline.
    Lines containing beach habitats were then attributed with beach width (average and maximum). This was measured on the digital orthophoto using Arcviews  measure tool functionality. The lines were all split whilst moving from East to West along the coast or in the case of an island in an anti-clockwise direction.  This systematic splitting of the coast was done to enable the used of scripts to automate the facing direction calculations.

    The digital aerial photography were the attribute source for 229.4 km of the total 983.1 km of shoreline.

    STAGE 3 (Carried out by Ben Lamb in May/June 2002)
    1.  Hardcopy B&W aerial photography (Job no. 900401  runs 49 and 50 flown Feb1991)  supplied by the Department for Planning and Infrastructures Marine Division and used as the base for the shoreline habitat classification for the Stokes Inlet area.
    2.  Habitat boundaries were delineated (according to CALM's standardised habitat classification system) into areas of beach, rocky shore, N/A or beach + rocky shore and attributed accordingly.  The classifications were made by correlating the shoreline onscreen to the hardcopy photos.  Delineation was done using the 'split line' functionality of Arcview 3.2 and was done where there was a change in habitat or a significant change in direction of the coastline.
    Lines containing beach habitats were then attributed with beach width (average and maximum).   This was measured on the aerial photos using an adjustable scaled magnifying glass.Due to difficulty in accurately identifying beach widths for individual segments of beach the beach widths generated from this photography were calculated over the total length of the beach and applied to the individual segments.  All other attribution has been performed to the same standards as was used in previous stages.  The lines were all split whilst moving from East to West along the coast or in the case of an island in an anti-clockwise direction.  This systematic splitting of the coast was done to enable the used of scripts to automate the facing direction calculations.
    3.  The shoreline which comprised of the offshore islands that was not covered by any of the aerial photography was identified by Kevin Bancroft as being all rocky shore with the exception of beaches on Sandy Hook Island and Middle Island.  The habitat boundaries were delineated (according to CALM's standardised habitat classification system) into areas of beach or rocky shore and attributed accordingly.  The lines were all split whilst moving in an anti-clockwise direction.  This systematic splitting of the coast was done to enable the used of scripts to automate the facing direction calculations.

    The hardcopy photography was the attribute source for 51.2 km  and evin Bancroft was the attribute source for 359.7 km of the total 983.1 km of shoreline

    4. Habitats were further attributed with direction the habitat is facing. This was determined using an avenue script to calculate the direction of the line, which was then converted to a facing direction eg. North, South, East, West, North East, North West, South East or South West.  See 'Attribute Accuracy' below for further details.
    5.  The dataset was then rationalised to merge adjacent polylines with identical attributes and reduce the size of the dataset.  This was done by creating a grouping field by concatenating all the other fields.  A script (autounion.ave) was then used to union all the polylines based upon their group attribute being identical.  This polylines were then converted back to a single part shapes using the Xtools 'convert multipart to single part' functionality.  This resulted in in a reduction in the number of polylines from 9112 to 6869.  Manual verification was undertaken on numerous samples of the dataset to ensure that the integrity of the dataset had been maintained.
    6.  Habitat length was calculated using the CALM added functionality extension (Calc length).  The dataset was then converted to GDA94 Lat/Long coordinates using the Change Projection funcionality of the CALM Added Functionality Extension 2001 extension.

     Positional Accuracy

    Habitat extents were delineated as accurately as is possible to determine from the available digital and hardcopy sources. Further Ground-truthing would need to be undertaken to determine them with greater accuracy.

    Accuracy of DOLA's coastline linework is not stated. DOLA has compiled this dataset from a number of sources of best available information including 1:2 000, 1:50 000 and 1:100 000 scale datasets. These source datasets were collected at various levels of positional accuracy, and as such each individual line is as accurate as its source data. The source attribute in the shapefile identifies these sources.

      Attribute Accuracy

    - Habitat classifications were determined visually from digital orthophoto mosaics, digital georeferenced aerial photography and hardcopy aerial photography, and while every effort has been made to assign these correctly, there may be some errors. The habitats are accurate as best determined at the time of aerial photography (Feb 1991 to Mar 2000). Coastlines by their very nature are dynamic and subject to change due to natural and man-made forces. This is particularly the case with the movement of sand up and down the coast, and as such beaches may form and dissappear with the passing of time.Ground-truthing needs to be undertaken to determine habitats with greater accuracy.

    - Beach widths estimated from the digital and hardcopy aerial photography are estimated to be accurate to within 10 metres. Beach widths estimated from the digital orthophotomosaics are estimated to be accurate to within 5 metres.  Ground-truthing needs to be undertaken to determine them with greater accuracy.

    - Habitat directions were determined using an avene script to calculate the bearing between the first and last points of the individual polylines.  All polylines were split in a uniform direction so as the polyline direction placed the sea to the right of the line and the land to the left.   Directions were assigned based on the seaward direction the habitat was facing. The following domain was used based on the seaward perpendicular of the bearing of the habitat. 
    east = 337.5o- 22.5o
    south-east = 22.5o - 67.5o
    south = 67.5o - 112.5o
    south-west = 112.5o - 157.5o
    west = 157.5o - 202.5o
    north-west = 202.5o - 247.5o
    north = 247.5o- 292.5o
    north-east = 292.5o - 337.5o

    - Habitat lengths were determined based on the arc length of the DOLA coastline (after splitting) and calculated automatically using the CALM extension for Arcview3.2. Lengths were rounded to the nearest metre.

      Logical Consistency

     Attribute names have been checked and validated for consistency across all shoreline habitat datasets.

    Attribute values have been checked and validated for consistency, and checked for logic in relation to attribute names. All attributes that require values have had values assigned.

    The shapefile has been compiled carefully to avoid overlaps or duplication of points.

    Completeness

     The dataset will be upgraded as priorities, time and resources permit. 
    Further work needs to be undertaken on ground-truthing.

    CONTACT INFORMATION

      Contact Organisation

     Department of Conservation and Land Management, Marine Conservation Branch

      Contact Position

     Marine GIS Co-ordinator

      Mail Address 1

     47 Henry Street

      Mail Address 2

      Suburb or Place or Locality

     Fremantle

      State or Locality 2

     WA

      Country

     Australia

      Postcode

     6160

      Telephone

     08 9336 0109

      Facsimile

     08 9430 5408

      Electronic Mail Address

     rayl@calm.wa.gov.au

    METADATA DATE

      Metadata Date

     5 June 2002

    ADDITIONAL METADATA

      Additional Metadata

    - See accompanying habitat classification documentation further describing the classification system(see T:\144-Marine Conservation Branch\Shared Data\Current_papers\Habitat Classification\SMECCS_291101.doc).
    NOTE: This system is still in development. Any feedback would be appreciated.

    - See DOLA coastline Metadata report for further details on the coastline.