Area of Applicability Summary, Landsat Seasonal Fractional Cover Version 3.0, Australia Coverage 

The Area of Applicability (AOA) describes the area to which a predictive model can reliably be applied, based on the predictor space covered by the underlying training data. It was evaluated following the approach proposed by Meyer and Pebesma (2021).

The JRSRP seasonal surface reflectance composites between winter 2014 and winter 2024 were used as a proxy for the range of representative surface reflectance values likely to be encountered across the continent under varying environmental conditions from which fractional cover predictions are made. The AOA of the FCv3 model was computed for each seasonal surface reflectance composite, then summarised as a frequency map representing the proportion of seasons that a location was outside the AOA.

For each state, five files are provided: an annual product summarising the AOA across all seasons, and four showing seasonal AOA frequencies for summer, autumn, winter, and spring. 

We at TERN acknowledge the Traditional Owners and Custodians throughout Australia, New Zealand and all nations. We honour their profound connections to land, water, biodiversity and culture and pay our respects to their Elders past, present and emerging. 

This dataset is a summary of the Area of Applicability (AOA) of the JRSRP Fractional Cover version 3.0 model, for the period winter 2014 to winter 2024. 

The training data for the FC3 model (SLATS Star Transects – Australian field sites), consisting of approximately 4000 field observations of fractional cover collected across Australia between 1997 to 2018, were matched to the nearest date, clear Landsat 5 TM,7 ETM+ or 8 OLI overpass, and the surface reflectance values calculated using the procedures described in Flood et. al. (2013), for a 3x3 pixel window.

Using the approach described in Meyer and Pebesma (2021), minimum distances between weighted predictors in the training data were calculated. An AOA threshold defined as the outlier removed maximum dissimilarity index (DI) of the training data was derived via cross-validation. The value of this threshold was found to be 0.128.

A DI image was then calculated to quantify the dissimilarity between each seasonal surface reflectance composite and the reflectance values of the FC3 model calibration data. The final AOA summary data was then calculated by stacking all seasonal DI images and recording the number of times the AOA threshold was exceeded. The count and percentage of AOA threshold exceedance was stored as band 1 and band 2 respectively in the output raster.

Finally, ocean pixels were masked and set to the no data value of 255.