Riparian vegetation map of the Pilbara region, WA

Author Identifiers

MASOOMEH ALAIBAKHSH

https://orcid.org/0000-0002-4266-1267

Publication Date

2017

Document Type

Dataset

Publisher

CSIRO

School or Research Centre

School of Engineering

Funders

The research is funded by Edith Cowan University (ECU) and supported by Commonwealth Science and Industrial Organisation (CSIRO) Land and Water, WA.

Description

Landsat 5 TM Standard Terrain Corrected (Level 1T) imagery was used to delineate riparian vegetation. The study area of the research (Pilbara region except the eastern desert) is covered by 16 intersecting Landsat scenes. For each LS, three cloud-free Landsat images acquired at the end dates of the three hydrological years 1992, 1999 and 2009 (one image per year), were downloaded from the USGS website (www.usgs.gov) and calibrated for top of atmosphere reflectance.

Max latitude: 20°0′0″ S Min latitude: 25°0′0″ S Max longitude: 121°0′0″ E Min longitude: 113°25′0″ E Coordinate reference system: WGS84

DOI

10.4225/08/5976b156bfd56

Research Activity Description

The project proposed methods for mapping riparian vegetation and Groundwater Dependent Vegetation of an arid and semi-arid region within the Western Australia, the Pilbara region. Further information and details on the methods, data and study area are available in the related publications.

Project leader

Olga Barron

Organisations

CSIRO (Australia), Edith Cowan University (Australia)

The team contributed to the project: Dr Masoomeh Alaibakhsh (ECU; CSIRO L&W); Dr Irina Emelyanova (CSIRO,Energy); Dr Olga Barron (CSIRO,L&W); Dr Mehdi Khiadani (ECU)

Methodology

Lineage

NDVI was calculated for each Landsat scene and used to composite a 3-dry-date NDVI datasets. PCA was applied to the NDVI datasets generated for each Landsat scene. It was anticipated that PC1 would reveal the range of values representing those areas without vegetation cover (lowest pixel values) to areas of consistent high reflectance vegetation (highest values). Hence, it was expected that riparian vegetation, with its frequent access to water, would be denoted by a high magnitude PC1 value. Then, Otsu thresholding method was applied to PC1 grids to separate riparian vegetation from other land cover types.

Start of data collection time period

1992

End of data collection time period

2009

Language

eng

File Size

12 files and 2 images

Share

 
COinS