Consumption-based Carbon Emissions for Deforestation (2001–2020)

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Researchers can use this dataset to analyze spatial heterogeneity in consumption-driven deforestation emissions and link trade to land-use change.

🏢実務担当者:Corporate sustainability teams can use this data to identify deforestation risks in their supply chains and inform Scope 3 reporting.

🏛政策担当者:Policymakers can leverage this dataset to design targeted policies for reducing imported deforestation and meeting climate commitments.

<b>Dataset description</b>This dataset provides a global, gridded product of consumption-based deforestation carbon emissions for seven major countries (CHN, DEU, FRA, GBR, JPN, KOR, USA) over the period 2001–2020 at a spatial resolution of 30 arc-seconds (approximately 1 km at the equator). It first quantifies consumption-based deforestation emissions at the national level and then spatially allocates the emissions associated with major consuming countries to individual grid cells. Each grid cell value represents the annual amount of deforestation-related carbon emissions at that location driven by the final consumption of a specific country. The dataset supports spatially explicit analyses of consumption-driven deforestation, and enables users to link international trade, land-use change, and carbon emissions at sub-national to global scales. It can also serve as input to climate and carbon cycle models to assess the contribution of consumption-driven deforestation to global carbon dynamics and climate change. By providing high-resolution spatial information on consumption-based deforestation emissions, this dataset addresses a critical gap in existing global datasets and facilitates improved analysis of spatial heterogeneity and policy-relevant assessments. <b>Data generation and processing overview</b>The dataset was generated by integrating multiple data sources, including satellite-derived annual deforestation maps, forest carbon flux estimates, road network information, and country-level multi-regional input–output (MRIO) trade statistics. National consumption-based deforestation emissions were first estimated using an MRIO framework and subsequently allocated to grid cells using geospatial indicators. To improve the spatial plausibility of the allocation, we propose an export likelihood indicator constructed from deforestation patch size and road density. Optimal parameter values were identified through a spatial plausibility validation procedure (see below), after which national consumption-based emissions were distributed to grid cells according to their relative export likelihood.The resulting dataset consists of annual global raster layers, providing a consistent spatial representation of consumption-driven deforestation carbon emissions of seven major countries worldwide over the period 2001–2020. It addresses the lack of fine-resolution global data on consumption-based deforestation emissions and enables spatially explicit analyses across regions and time. <b>Spatial plausibility validation</b>The spatial plausibility of the gridded deforestation emissions footprints was assessed using independent global accessibility datasets. First, accessibility to cities was assessed using the global travel time to cities dataset (Weiss et al., 2018). Cities were classified into four population-based categories (&lt;10,000; 10,000–20,000; 20,000–50,000; &gt;50,000 inhabitants), and travel time to the nearest city was grouped into four intervals (&lt;1 h, 1–2 h, 2–5 h, and &gt;5 h). Following Hochard and Barbier (2017), deforestation emissions footprints located more than five hours from the nearest city were considered less consistent with export-oriented production, as market access and trade participation decline substantially beyond this threshold.Second, accessibility to ports was evaluated using the Global Accessibility to Ports dataset (Nelson et al., 2019). Ports were classified into five categories (Large, Medium, Small, Very small, and Any), and travel times to ports were grouped into four intervals (&lt;24 h, 24–48 h, 48–72 h, and &gt;72 h). Footprints located at very long travel times to ports were considered less plausible for export-driven deforestation, consistent with findings in previous studies (e.g., Gries et al., 2009).These accessibility indicators were used to compare multiple candidate parameter combinations, including different deforestation patch size thresholds and road density levels. The optimal parameter set was selected by maximizing the share of deforestation emissions footprints located in accessible areas (i.e., closer to cities and ports) and minimizing allocations in remote regions.Finally, the spatial distribution derived from the optimized parameters was compared with that generated by a conventional proportional allocation approach that does not consider accessibility or patch size. The comparison focuses on the relative shares of footprints in accessible versus remote areas and demonstrates the effectiveness of the proposed approach in enhancing spatial plausibility. <b>Usage notes</b>The dataset can be directly used in standard GIS and data analysis platforms, including QGIS, ArcGIS, R, Python, MATLAB, and Google Earth Engine. Typical applications include:mapping and analyzing the spatial distribution of consumption-driven deforestation carbon emissions;assessing the environmental impacts of international trade and consumption patterns;supporting supply-chain sustainability assessments and footprint analyses;combining with land-use, biodiversity, or climate datasets for integrated spatial analyses;serving as input to climate and carbon cycle models to assess the contribution of consumption-driven deforestation to carbon dynamics and climate change. <b>Users should be aware of the following limitations: </b>Due to limited traceability in global supply-chain data, the spatial allocation assumes uniform export probabilities within producing countries, consistent with previous consumption-based deforestation studies; The dataset includes only economically driven deforestation and excludes wildfire-related forest loss, which cannot be reliably linked to final consumption; Temporal coverage is constrained by the availability of global input–output tables. The dataset can be extended as updated MRIO data become available. <b>Data records</b>The dataset is provided as a collection of annual global raster files in GeoTIFF format, covering the period 2001–2020. <b>Raster format and spatial properties</b>File format: GeoTIFF (LZW compressed)Spatial resolution: 30 arc-seconds (≈1 km at the equator)Spatial extent: Global land areas (90°S–90°N, 180°W–180°E)Coordinate reference system: WGS 84 (EPSG:4326)Temporal resolution: AnnualTemporal coverage: 2001–2020 <b>Data type and no-data value</b>Unit: Gg CO₂e yr⁻¹Data type: 32-bit floating pointNo-data value: −9999.

• openalex https://doi.org/10.6084/m9.figshare.28091879.v3first seen 2026-08-02 18:19:47