Document Type

Paper

Start Date

5-10-2023 4:30 PM

End Date

5-10-2023 5:15 PM

Abstract

The European Union’s Regulation 2023/1115, which requires firms to undertake significant due diligence efforts to minimize the import of deforestation-linked commodities into the Union, relies heavily on cutting-edge geospatial technologies. While several studies of zero-deforestation supply chain efforts have pointed to the political and logistical challenges such initiatives face, there has been less consideration of the role that the measurement errors that are unavoidable in geospatial technologies might affect emerging zero-deforestation governance systems. Using information on misclassification error rates for forest areas from 20 recent validation studies of global land-cover datasets, I simulate the misclassification risks we might expect from a proposed global system designed to facilitate implementing the regulation. I argue that misclassification errors could pose problems for small operators, particularly in mosaic forest landscapes, as well as producers engaging in agroforestry. While some strategies could mitigate these risks, they are unlikely to be eliminated entirely, and their inevitability should be considered in implementing the Regulation.

DOI

10.5703/1288284317779

Share

COinS
 
Oct 5th, 4:30 PM Oct 5th, 5:15 PM

The Challenge of Misclassification Error in the European Union’s Deforestation Regulation

The European Union’s Regulation 2023/1115, which requires firms to undertake significant due diligence efforts to minimize the import of deforestation-linked commodities into the Union, relies heavily on cutting-edge geospatial technologies. While several studies of zero-deforestation supply chain efforts have pointed to the political and logistical challenges such initiatives face, there has been less consideration of the role that the measurement errors that are unavoidable in geospatial technologies might affect emerging zero-deforestation governance systems. Using information on misclassification error rates for forest areas from 20 recent validation studies of global land-cover datasets, I simulate the misclassification risks we might expect from a proposed global system designed to facilitate implementing the regulation. I argue that misclassification errors could pose problems for small operators, particularly in mosaic forest landscapes, as well as producers engaging in agroforestry. While some strategies could mitigate these risks, they are unlikely to be eliminated entirely, and their inevitability should be considered in implementing the Regulation.