Figures captions

Figure 1. The location and climatic context of Yanga National Park and Barmah National Park (image courtesy of Google Earth on Aug/2018).

Figure 2. Scatterplots showing LiDAR FTCC against predicted FTCC (a) Yanga and (b) Barmah using individual site-specific models (RFYanga and RFBarmah). The results of predicted FTCC using the model trained using both Yanga and Barmah data (RFall) are shown in (c) for Yanga and (d) for Barmah. The blue line indicates the regression line.

Figure 3. Spatial outputs for the Yanga region of interest, mapping (a) predicted fractional tree canopy cover (FTCC); (b); LiDAR FTCC and (c) the difference between the two images. Similarly, spatial outputs for the Barmah region of interest, mapping (d) predicted fractional tree canopy cover (FTCC); (e); LiDAR FTCC and the (f) difference between the two images.

Figure 4. Bands of importance for Sentienl-1 and Sentinel-2 for FTCC estimation for the RFYanga, RFBarmah and RFAll Random forest models. The importance score was scaled between 0.1 and 1.