NSW Woody Vegetation Extent 2011 Derived from SPOT-5 Imagery

2011 statewide binary classification of woody vegetation derived from multitemporal 5-meter SPOT-5 satellite imagery. The dataset was produced by the NSW Department of Climate Change, Energy, the Environment and Water using a logistic regression model trained on statistics from imagery over 2008-2012 and over 26000 training points. It underwent manual corrections throughout 2013-2014.

Use Cases

• Identify isolated tree crowns and contiguous forest based on the 5-meter binary classification.
• Map woody foliage projection cover (FPC) density using this dataset as a spatial constraint.
• Assess vegetation extent for ecological studies based on the 2011 statewide snapshot.
• Validate remote sensing models using the reported 90.1% accuracy compared to Lidar datasets.

Strengths

• High-resolution 5-meter classification capable of identifying small features like paddock trees.
• Overall state accuracy recorded at 90.1% when compared to Lidar datasets.
• Derived from a model utilizing statistics from SPOT-5 imagery over a 5-year period (2008-2012) and referenced over 26000 training points.

Limitations

• Accuracy for specific regions may vary considerably despite extensive manual corrections.
• Intended for use in non-urban environments; accuracy for urban environments has not been assessed.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

NSW Department of Climate Change, Energy, the Environment and Water

Collection Method

Derived from user-driven thresholds on an index of woody probability calculated from a binomial logistic regression model on multi-temporal SPOT-5 data.

Time Range

2011

Geography

New South Wales, Australia