1. Australian Antarctic Data Centre
2. Discover and Manage Data
3. Records
4. HI_VEG_ORTHO_VEGMAP

Description

Vegetation mapping from orthophotos derived from non-metric photography for the north eastern, south eastern, south, west and north west coastal areas of Heard Island.

More information about the vegetation mapping process is documented in a pdf report entitled 'Notes for Heard Island Vegetation Mapping project 2002-2006' available for download at the provided URL. The report is an updated version of the December 2004 pdf report entitled 'Heard Island Vegetation Mapping Report'.

The vegetation mapping project was undertaken between 1986 and 1988 (field mapping), 2002 and 2005 (digitising) and 2003-2004 (limited field checking). The data for the mapping done on orthophotos are contained in two shapefiles as follows:

- Vegetation mapping polygons.
- Reliability (comments on quality of orthophoto coverage, areas of overlapping cover, etc).

Associated shapefiles are as follows:
- Aerial photo coverage, showing polygons for the areas where each orthophoto was used for mapping (see metadata record with ID HI_VEG_ORTHOPHOTOS_VEGMAP).
- Photo locations of 35mm photos used during mapping (see metadata record with ID HI_VEG_PHOTOS_VEGMAP).

Quality

Digitising:
The original field mapping overlays prepared by Dr Jenny Scott between 1987 and 1988 were rectified with the aerial photos which they overlaid. When the digitising phase of the project commenced, it was thought that these overlays would simply be digitised and then checked and adjusted by Scott. However, it was found to be more effective for Scott to re-map and screen-digitise the vegetation boundaries straight onto the rectified images, with the overlays being used merely as a guide and a form of 'ground-truthing'. Another very useful form of 'ground-truthing' was the use of 35mm slides of the areas (HI_VEG_PHOTOS_VEGMAP). Most of the slides were taken by Scott in January 1987 when the Linhof airphotos were flown, with the specific purpose of using them for interpretation when finalising the vegetation mapping. Wherever possible, a panoramic set of oblique aerial photos was taken of each area from a helicopter. These proved very useful for reference during digitising. Additional slides were taken at various viewpoints on the ground for each area, and these were useful for checking vegetation detail at specific locations. The slides have been scanned and linked with the digitised boundaries as part of the data set (HI_VEG_PHOTOS_VEGMAP).

Resolution of mapping and level of accuracy:
a) Resolution - when the field mapping was done in 1987-88 the minimum polygon size was defined as 50m in any one direction, ie. irrespective of the polygon shape. This definition worked well in the field, and was retained for the current Arcview mapping phase.

b) Accuracy - boundaries were digitised and corrected to less than 5m accuracy on flat to undulating ground and probably 5-15m accuracy on steeply sloping ground and breaks of slope (related to accuracy of the existing DEM). This level of accuracy was possible on the 1987 aerial photos but not on the 1980 aerial photos. The 1980 aerial photos were useful as additional checks when digitising (eg. sometimes the differing contrast and shadow conditions aided interpretation), but were not digitised unless 1987 coverage was unavailable. The 1980 aerial photos were rectified and used for digitising for part of the Spit (central section of Dovers Flats), the northern part of Paddick Valley, Cape Gazert, Laurens Peninsula and Atlas Cove area.

Consistency:
The minimum size of polygons digitised was checked regularly for consistency by using the measuring tool. As mapping progressed, the digitised colour images and their vegetation boundaries were printed out and referred to when decisions were being made on combinations of categories, to help maintain consistency when assigning the combined categories. Differences in vegetation texture/colour etc were often quite subtle, and also varied between areas. Due to these subtleties, maintenance of consistency through one person mapping all the vegetated areas, at least initially, is seen as very important for this project. See attached pdf report for further detail.

Mapping:
The minimum size of polygons mapped (50m in any one direction) is relevant when deciding on the correct vegetation mapping category. For example, if a patch of vegetation with over 75% Azorella cover is greater than 50m in length, it will be mapped as 'closed cushionfield'. If the patch is one of several Azorella patches each less than 50m in length, with bare ground in between, then it will be mapped as 'open cushionfield', with the polygon boundary enclosing an area where Azorella has an overall cover of 50-75%.

Vegetation categories:
The vegetation categories are a selection of those used initially in the 1987 field mapping. The basic categories are defined in Scott (1989), and relate to the current provisional vegetation classification for the island outlined in Scott and Bergstrom (2006). See attached pdf report. Names of the provisional vegetation communities were refined after the 2003-04 field season. Subantarctic vegetation, especially on Heard Island, is characterised by very few vascular plant species (12 on Heard) mostly with wide ecological amplitude. This means that most species grow in a wide range of environments and exist in a number of different combinations, often leading to intergrading vegetation communities. This sometimes makes mapping and assigning of categories difficult, and it meant that a large number of provisional combinations of the basic categories were used during the initial field mapping phase in 1986-87-88 (Scott 1989). The number of combinations was reduced during the digitising phase, and the final combined categories were able to be mapped fairly consistently. Variations in the vegetation occur between the major ice-free areas, due to environmental differences such as topography and aspect, and the extent and type of wildlife influence. The mapping categories were designed to be general enough to encompass these differences.

Additional information is included in the pdf report 'Notes for Heard Island Vegetation Mapping Project 2002-2006'.

GIS Notes
The original data were digitised by Dr Jenny Scott into individual shapefiles for each region using ArcMap. The individual files were converted to Arc/Info coverages and cleaned. There was some overlap in the coverages so ambiguous areas demonstrating least consistency with adjoining areas were removed. This has been ratified by Dr Jenny Scott. The polygons were cleaned using a maximum fuzzy tolerance of half a metre. The data were also adjusted to match the topographic data digitised from the same photography by a photogrammetrist.

Access

Vegetation mapping and reliability shapefiles are available for download from the provided URLs.
A pdf document further detailing these data is available for download from the provided URL. The document has been updated from an earlier version dated December 2004.

Temporal Coverages

• Start date: 1980-01-01 - Stop date: 1987-12-31

Spatial Coverages

Data Resolution

Science Keywords

• EARTH SCIENCE > BIOSPHERE > VEGETATION > DOMINANT SPECIES
• EARTH SCIENCE > BIOSPHERE > VEGETATION > FOREST COMPOSITION/VEGETATION STRUCTURE
• EARTH SCIENCE > BIOSPHERE > VEGETATION > INDIGENOUS VEGETATION
• EARTH SCIENCE > BIOSPHERE > VEGETATION > VEGETATION COVER
• EARTH SCIENCE > BIOSPHERE > VEGETATION > VEGETATION INDEX
• EARTH SCIENCE > BIOSPHERE > VEGETATION > VEGETATION SPECIES

Use Constraints

This data set conforms to the CCBY Attribution License (http://creativecommons.org/licenses/by/4.0/).

Please follow instructions listed in the citation reference provided at http://data.aad.gov.au/aadc/metadata/citation.cfm?entry_id=HI_VEG_ORTHO_VEGMAP when using these data.