• Energy Research
• Open Access close

This file collection contains the estimated spatial distribution of the above-ground biomass density (AGB) by the end of the 21st century across the Brazilian Atlantic Forest domain and the respective uncertanty. To develop the models, we used the maximum entropy method with projected climate data to 2100, based on the Intergovernmental Panel on Climate Change (IPCC) Representative Concentration Pathway (RCP) 4.5 from the fifth Assessment Report (AR5). The dataset is composed of four files in GeoTIFF format: calibrated-AGB-distribution.tif: raster file representing the present spatial distribution of the above-ground biomass density in the Atlantic Forest from the calibrated model. Unit: Mg/ha estimated-uncertanty-for-calibrated-agb-distribution.tif: raster file representing the estimated spatial uncertanty distribution of the calibrated above-ground biomass density. Unit: percentage. projected-AGB-distribution-under-rcp45.tif: raster file representing the projected spatial distribution of the above-ground biomass density in the Atlantic Forest by the end of 2100 under RCP 4.5 scenario. Unit: Mg/ha estimated-uncertanty-for-projected-agb-distribution.tif: raster file representing the estimated spatial uncertanty distribution of the projected above-ground biomass density. Unit: percentage. Spatial resolution: 0.0083 degree (ca. 1 km) Coordinate reference system: Geographic Coordinate System - Datum WGS84

Summary Tropical forests are critical for protecting global biodiversity and carbon stores. While forest degradation and fragmentation cause negative impacts on trees, many woody lianas benefit, with associated negative effects on carbon storage. Here, we focus on the key question of how abiotic environmental changes resulting from tropical forest fragmentation mediate the allocation of carbon into trees and lianas. We focus on the globally threatened Brazilian Atlantic Forest, in forest fragments spanning 13–23 442 ha in area and at fragment edges and interiors. Within each fragment, we established two transects: one at the edge and one in the interior. Each transect consisted of ten 10 × 10 m plots spaced at 20 m intervals. Within each plot, we sampled living trees with diameter ≥4·8 cm at 1·3 m above ground, living lianas with diameter ≥1·6 cm at 10 cm above ground, and several microclimatic and soil variables. Fragmentation changed a broad suite of abiotic environmental conditions recognized as being associated with forest carbon stocks: edges and smaller fragments were hotter, windier, and less humid, with more fertile and less acid soils at edges. Tree carbon stocks were thus higher in forest interiors than at edges, and were positively related to fragment size in interiors, but were not impacted by fragment size at edges. Trees and lianas showed different responses to fragmentation: in interiors of small fragments, tree carbon stocks declined whereas liana carbon stocks increased; and at edges, tree carbon stocks were not affected by fragment size, whereas liana carbon stocks were highest in smaller fragments. These patterns were strongly related to changes in abiotic environmental conditions. We conclude that the abiotic changes across the fragmentation gradient, rather than liana proliferation, were more likely to reduce tree carbon stocks. Cutting of lianas is frequently promoted for restoring forest carbon in human‐modified tropical forests. However, this approach may not be effective for restoring forest carbon stocks in fragmented forests.