Markedly divergent estimates of Amazon forest carbon density from ground plots and satellites

Abstract

AimThe accurate mapping of forest carbon stocks is essential for understanding the global carbon cycle, for assessing emissions from deforestation, and for rational land-use planning. Remote sensing (RS) is currently the key tool for this purpose, but RS does not estimate vegetation biomass directly, and thus may miss significant spatial variations in forest structure. We test the stated accuracy of pantropical carbon maps using a large independent field dataset.LocationTropical forests of the Amazon basin. The permanent archive of the field plot data can be accessed at: http://dx.doi.org/10.5521/FORESTPLOTS.NET/2014_1MethodsTwo recent pantropical RS maps of vegetation carbon are compared to a unique ground-plot dataset, involving tree measurements in 413 large inventory plots located in nine countries. The RS maps were compared directly to field plots, and kriging of the field data was used to allow area-based comparisons.ResultsThe two RS carbon maps fail to capture the main gradient in Amazon forest carbon detected using 413 ground plots, from the densely wooded tall forests of the north-east, to the light-wooded, shorter forests of the south-west. The differences between plots and RS maps far exceed the uncertainties given in these studies, with whole regions over- or under-estimated by > 25%, whereas regional uncertainties for the maps were reported to be < 5%.Main conclusionsPantropical biomass maps are widely used by governments and by projects aiming to reduce deforestation using carbon offsets, but may have significant regional biases. Carbon-mapping techniques must be revised to account for the known ecological variation in tree wood density and allometry to create maps suitable for carbon accounting. The use of single relationships between tree canopy height and above-ground biomass inevitably yields large, spatially correlated errors. This presents a significant challenge to both the forest conservation and remote sensing communities, because neither wood density nor species assemblages can be reliably mapped from space.

Multi-year MODIS active fire type classification over the Brazilian Tropical Moist Forest Biome

“…Methodologies currently do not consider the fire type explicitly and typically model emissions on the basis of the dominant land cover or vegetation type where active fires are detected (Freitas et al 2005;van der Werf et al 2008;Ichoku and Ellison 2013;Castellanos, Boersma, and Van Der Werf 2014;Mitchard et al 2014). The spatial distribution of the classified fire types may help provide more reliable parameterization of the biomass loading, combustion completeness, and emission factors that differ among the fire types.…”

Biomass turnover time in terrestrial ecosystems halved by land use

“…1). In the light of the considerable uncertainties between RS products, as well as with site data [79][80][81] , we accepted the uncertainty we introduce through tree-height-based downscaling for the advantage of consistency, because national forest C-stock data are available with global coverage 26 . Other approaches that have been proposed for downscaling national carbon stock information, for example, following the pattern of NPP 82 , would result in one national τbact value for forest.…”

Brazil’s Amazonian forest carbon: the key to Southern Amazonia’s significance for global climate

“…All of these studies represent advances in interpretation of remote sensing data, but remain limited by their datasets for ground truth. Mitchard et al (2014) contrasted the spatial results of the Saatchi et al (2011) and Baccini et al (2012) remote sensing studies, as well as the geographical information system (GIS) analyses derived directly from plot data by Houghton et al (2001), Malhi et al (2006) and their own analysis of RAINFOR (Amazon Forest Inventory Network) plots (e.g., Phillips et al 2009). The results show major differences between all of the resulting maps, including those with largely overlapping ground-based datasets.…”