Mapping data gaps to estimate biomass across Brazilian Amazon forests

Forest Ecosystems

Table 2 Main characteristics of the Amazon forest AGB density maps

Map	Scale	Spatial resolution	Temporal scale (year)	Field forest plots/ source	Study area plots/ sampled area (ha)	Remote Sensing products/ other inputs	Model
Saatchi et al. (2011)	Pantropical	1 km	2000	4079^b (493 for calibration)/ Many sources	~ 707/ ~1770^d	MODIS (NDVI, LAI, % tree cover), LiDAR from GLAS/Forest height map	MaxEnt
Baccini et al. (2012)	Pantropical	500 m	2007–2008	283^b/ Measured	No Data	MODIS, LiDAR from GLAS, SRTM	Random Forest
Nogueira et al. (2015)	Brazilian Amazon	1 km (landscape level)	1970^a	2317^c/ RadamBrasil and literature	2373/ 2317	No/ Vegetation map (SIVAM 2002)	None
MCT (2016)	Brazilian Amazon	1 km (landscape level)	1973–1983^a	1682 plots/ RadamBrasil	1682/ 1682	No/ Vegetation (MCT 2010), Soils (Bernoux et al. 2002)	Inverse distance weighting
Avitabile et al. (2016)	Pantropical	1 km	2000–2013^a	648/ RAINFOR, TEAM and Sustainable Landscapes	~ 500/ No data	No/ High-resolution AGB maps	Fusion model

Source: Based on Tejada et al. (2019)
RAINFOR Amazon Forest Inventory Network, TEAM Tropical Ecology, Assessment and Monitoring, MODIS moderate resolution imaging spectroradiometer, NDVI normalized difference vegetation index, LAI leaf area index, GLAS Geoscience Laser Altimeter System, LiDAR light detection and ranging, SRTM Shuttle Radar Topography Mission, and JERS-1 Japanese Earth Resources Satellite 1
^aAGB field measurements
^bwe did not have access to the locations of the plots
^cin the case of the RadamBrasil plots, we had the location of only 1682 plots
^dthe total area of the plots was estimated because the plots had different sizes