‘Uncertainty audit’ for ecosystem accounting: Satellite-based ecosystem extent is biased without design-based area estimation and accuracy assessment

The SEEA EA lacks guidelines for quantifying uncertainty, leading to potentially biased area estimates in accounting tables. This study uses Oslo municipality in Norway to show best practices for obtaining unbiased area estimates with design-based statistical methods. Comparing custom Sentinel-2 and globally available Dynamic World maps from 2015, 2018, and 2021, the study found that current pixel counting methods cause significant biases, especially for ecosystem conversions. Locally calibrated maps were more accurate than global ones. The study highlights the need for rigorous accuracy assessments and standards for auditing uncertainty to ensure credible ecosystem accounts.