Using the latest GIS and AI technology, AHA’s independent assessment of deforestation risk provides a county-level evaluation of the risk that forest in a given U.S. county has been converted to agricultural land. It is designed to support U.S. hardwood exporters and their customers in demonstrating due diligence under the EU Deforestation Regulation (EUDR), and to give importers confidence in the provenance of the material they receive.
Data sources and methodology
The assessment combines two primary datasets. Forest cover is established from the Joint Research Centre (JRC) Global Map of Forest Cover 2020, which provides a consistent, internationally recognised baseline of forest extent. This is refined using the USDA National Agricultural Statistics Service (NASS) Cropland Data Layer (CDL) — an annual, geo-referenced, crop-specific land cover dataset produced from satellite imagery and extensive agricultural ground reference data — to isolate the deciduous forest component by removing softwood forests.
The CDL has been updated annually since 2008, with new data typically released in January or February each year, reflecting the crop and forest types occupying land during the previous growing season. It draws on satellite data from the NASA/USGS Landsat Program and the European Space Agency’s Copernicus Sentinel-2 mission, at spatial resolutions of 30m and 10m. This allows the rate and immediate drivers of deforestation — in terms of the crops or other land uses that replace forest — to be readily identified, without the need for AHA to undertake time-consuming and costly work to compile satellite data and classify land uses independently.
The refined base map is compared against the most recent 10m-resolution CDL to identify potential areas of deforestation. These results are then further refined using historic CDL data and other sources to filter out areas that are unlikely to have been forest in the recent past, and which therefore cannot be deemed to have contributed to deforestation under the EUDR.
Expert review
Even the best algorithms can struggle to distinguish accurately between sites subject to sustainable timber harvesting operations; sites damaged by pests, fires or windthrow; and sites undergoing genuine conversion. To address this, AHA operates the Expert Eye tool, which brings expert human review to bear on sites flagged by the automated analysis as potentially at risk of deforestation. Reviewers with knowledge of local forestry practices assess these sites using higher-resolution satellite data, and where necessary supplement this with ground-based verification. Over time, the results of this “truthing” exercise are used to refine the analytical procedures.
Risk classification
Drawing on the analysis and in consultation with stakeholders, AHA classifies each U.S. county according to the annual rate of forest area with a deciduous component potentially converted to agricultural land:
Negligible risk (green) – deforestation confirmed as less than 0.1% per annum. No further risk mitigation measures are required by U.S. hardwood operators.
Unspecified risk (amber) – deforestation possibly more than 0.1% per annum. Further assessment by AHA is needed before negligible risk can be confirmed.
Specified risk (red) – deforestation confirmed as more than 0.1% per annum following further assessment by AHA.
To facilitate broad acceptance in the global market, the assessment is based, as far as possible, on internationally recognised definitions of deforestation from the United Nations Food and Agriculture Organisation (UN FAO). It also takes account of the definitions used in the EUDR. As the focus is on the risk of commodity-driven deforestation of hardwood stands, the analysis quantifies the extent of conversion of deciduous and mixed deciduous/coniferous forest to agricultural land.
Versions of the assessment
The assessment is updated periodically, and each version is named after the year of the CDL data it is derived from. All versions are published and hosted here, so the basis for any given assessment remains permanently available. Each version is accompanied by a description of the data sources and methodology specific to that release, as the underlying datasets and analysis evolve over time. From 2027, the assessment will be updated twice a year, in April and October, to reflect the latest available data.
How this links to AHA statements
Every AHA statement is linked to the version of the assessment that was current at the time of shipment. When you receive a statement as an importer, you can click on any of the counties identified as a source for that shipment and view its risk classification on the online map, exactly as it stood for the relevant version. This ensures that the risk information attached to a statement always reflects the assessment in force when the shipment was made, rather than any later revision.
Integration with U.S. Forest Service data
AHA’s analysis is complemented by an analysis of deforestation risks and drivers prepared by the U.S. Forest Service, using the extensive sample-based inventory data collected across the U.S. as part of the Forest Inventory and Analysis (FIA) Program. The U.S. Forest Service is implementing a strategic plan for enhanced integration of remote sensing information into the FIA Program, with particular emphasis on so-called “small area estimation” (SAE) techniques. This effort engages a broad range of stakeholders to assist the U.S. Forest Service in meeting a Congressional directive to “implement procedures to improve the statistical precision of estimates at the sub-State level.” As in AHA’s own analysis, the focus is on improving the level of precision and access to data at county level, to support regulatory compliance and environmental claims in markets for forest products and agricultural commodities.
Learn more about the innovative work going on behind the scenes.
