 |
Metadata : Provisional Vegetation Types over Litchfield National Park |
Index:
Metadata Details:
Name:AS/NZS ISO 19115 Geographic Information - Metadata, ANZLIC Metadata Profile
Version:1.0
Date Metadata Extracted:2024-04-25
Date Metadata Last Updated:2014-05-02
Current URL (HTML format) : http://www.ntlis.nt.gov.au/metadata/export_data?type=html&metadata_id=F8606A2E4DC98FA8E040CD9B21440F2C
Current URL (XML format) : http://www.ntlis.nt.gov.au/metadata/export_data?type=xml&metadata_id=F8606A2E4DC98FA8E040CD9B21440F2C
Citation
ANZLIC Identifier:F8606A2E4DC98FA8E040CD9B21440F2C
Title: Provisional Vegetation Types over Litchfield National Park
Citation Date:2014-05-01
Date Type:publication
Custodian:Department of Environment, Parks and Water Security
Description
This spatial dataset describes vegetation types over Litchfield National Park (plus a 2km buffer). The survey area of 1,905km2 is located west of the towns Batchelor and Adelaide River and is approximately 70km south of Darwin.
This project was completed in response to a need for vegetation mapping over the whole of Litchfield National Park, specifically for integrated strategic planning and particularly in regards to fire management. The project was completed using existing data and has not been verified in the field. This dataset is provided as the result of a rapid, best-effort mapping approach. Users are strongly advised to note the accuracy limitations of the dataset, particularly the attribute accuracy. The final dataset has not been validated.
- biota
- environment
- imageryBaseMapsEarthCover
Bounding Coordinates:
- North Bounding Coordinate:-12.998
- South Bounding Coordinate:-13.5597
- East Bounding Coordinate:131.1166
- West Bounding Coordinate:130.6369
Dataset Status
Dataset ID:
Language:English
Character Set: Latin 1
Progress:completed
Maintenance and Update Frequency:notPlanned
Data Currency Start Date:2014-02-28
Data Currency End Date:2014-04-07
Access Constraint:
Download the spatial data package (contains spatial data, maps and reports) via data.nt.gov.au/
You are licensed to use the DENR geospatial products on the terms and conditions set out in: Creative Commons Attribution 4.0 International Public License (CC BY 4.0) at: creativecommons.org/licenses/by/4.0/legalcode
If you do not agree to the terms and conditions you must not use the geospatial products.
You are free to copy, communicate and adapt the licensed material, provided that you abide by the licence terms (including Attribution) and attribute the licensed material using the statement:
Supplied by the Department of Environment and Natural Resources. Copyright Northern Territory Government.
Data Quality
Lineage:These boundaries and attributes of vegetation types were derived from two datasets of Landsat 8 imagery taken in 2013 (19th April and 9th August) and other data.
Supplementary datasets included the 1-second DEM (http://www.ga.gov.au/meta/ANZCW0703014016.html) and a Landsat-derived model of the persistent green vegetation fraction (http://www.auscover.org.au/xwiki/bin/view/Product+pages/Persistent+Green-Vegetation+Fraction).
Further datasets used were Land Units (Litchfield National Park [Survey Code: TTRAN] and Warrai Catchment [Survey Code: WARRA], both at 1:50,000), Geology (1:250,000), Wetlands (1:100,000), Melaleuca (1:100,000) and Rainforests (1:80,000).
Field site data used were extracted from the Northern Territory Vegetation Site Database - Survey Codes: TTRAN, WARRA, LITCH, COLGA and LFP. Most of the sites had been surveyed between 1981 and 1992, and a few between 1995 and 2006. A classification of the LFP sites was provided by the Parks and Wildlife Commission.
Classification of the Landsat data was performed using a step-wise process including a nearest-neighbour classification guided by sample points, as described in the metadata field Supplementary Information.
Positional Accuracy:This spatial dataset should not be used at scales greater than 1 : 100,000. The final dataset has not been validated.
Sample points were used for the nearest-neighbour classification and were derived from the field sites and additional points. Initially field sites were classified according to site descriptions and plant species cover data. The spatial positioning of most sites was only approximate due to the lack of location-recording tools available at the time of data collection. Thus the positioning of each point was manually checked and adjusted where necessary, according to visual interpretation of the Landsat imagery and other modelled spatial datasets. Additional sample points were created using a visual interpretation of the Landsat imagery and other modelled datasets.
Attribute Accuracy:Classes of vegetation type (Attribute VEG_TYPE) were defined according to the planning needs of the Parks and Wildlife Commission, field data, and visible features in the Landsat data.
The assignation of classes of vegetation type had varying degrees of accuracy due to the availability of field data and the ability to visually identify each class in the datasets. In particular, there were few field data points representing the classes of Wet heath and Dry heath, both of which could not easily be identified in the Landsat imagery. This is reflected in the value of Low for the attribute Accuracy in areas classified as those two vegetation types. Wet heath and Dry heath occupy biophysically similar parts of the landscape and have similar vegetation cover to the vegetation type Sandstone woodland, so areas classified as these three vegetation types may have been confused with one another. The classes of Alluvial grassland, Melaleuca woodland and Drainage woodland had few sample points, though were more easily visually identified in the imagery, reflected in the value of Medium Accuracy for most areas mapped as those classes. For these three classes, and Riparian, segments were assigned High Accuracy if they spatially coincided with relevant map units in the Melaleuca, Riparian or Wetland datasets. The remainder of the dataset was assigned Medium Accuracy.
Logical Consistency:GIS editing tools were used to rectify all overshoots and undershoots in line work. Polygon topology rules applied are; must not overlap, must not have gaps and each polygon has only one label.
Completeness:Complete
Contacts
Name Organisation Position Role Phone Fax Data Requests Officer Department of Environment, Parks and Water Security Geospatial Services Branch (on behalf of department) distributor datarequests.depws@nt.gov.au
Data Dictionary
No data dictionary defined for this dataset
Supplementary Information
Department Survey Code (Vegetation Land Cover): LITCHV_100. There is no accompanying technical report with this spatial dataset.
Bibliographic Reference: Trueman, M and Cuff, N (2014) Provisional Vegetation Types over Litchfield National Park (digital data only); Rangeland Division, Department of Land Resource Management, Palmerston, Northern Territory.
Data Classification process:
Like-pixel segments were grouped (software: eCognition Developer 8, scale factor 20, shape 0.1, compactness 0.5) and then classified into the vegetation types as follows:
. Manual separation of sandstone and non-sandstone substrates guided by the geology and land units datasets and manual interpretation of the April Landsat dataset (Attribute SUBSTRATE).
. Manual delineation of fire scars and smoke in each Landsat dataset using visual analysis (Attribute FIRE_MASK) for the purpose of excluding these data from the nearest-neighbour classification.
. Nearest-neighbour supervised classification based on sample points, all Landsat bands, DEM and persistent green. This classification was restricted by Elevation, Substrate and FireMask as follows:
.. Elevation: Veg_type classes of Tabletop open forest, Tabletop woodland, Upland sedgeland and Upland swamp community were restricted to areas with an elevation of over 200m (tabletop).
.. Substrate: Veg_type classes of Dry heath, Sandstone woodland, Tabletop open forest, Tabletop woodland, Upland sedgeland, Upland swamp community and Wet heath were restricted to sandstone areas. Veg_type classes of Alluvial grassland, Drainage woodland, Lowland woodland and Melaleuca woodland were restricted to non-sandstone areas. The remaining Veg_type classes of Riparian, Rock and Water were not restricted by Substrate.
.. Where no FireMask was present, both Landsat datasets were used. Where a FireMask was present, the nearest-neighbour classification excluded the Landsat dataset that was affected by fire.
. Riparian areas were expanded to include persistent green percentage cover values between 50-100%, but not including any segments that had already been mapped as Tabletop woodland or open forest, Upland swamp community, Drainage woodland or Melaleuca woodland.
. Segments remaining unclassified and segments that were obviously incorrectly classified as Upland swamp community or Upland sedgeland were manually assigned a class based on visual assessment of the data.
This record maintained using the NTLIS Metadata Tool