Work

Forest 2200 nm/660 nm reflectance seasonality for improved satellite-image atmospheric correction

Year: 2025

Type: article

Abstract: A widely used algorithm, Sen2Cor, for the atmospheric correction of Sentinel-2 MSI images assumes a constant ratio of red (0.6 μm) and near-infrared (2.2 μm) spectral reflectance, taking ρ 2.2 / ρ 0.6 = 2. In fact, however, this ratio exhibits substantial dependence on vegetation phenology. The discrepancy produces errors in predicted surface reflectance. We analyse the reflectance of 92,230 forest stands, using full time series of Landsat-8/9 OLI and Sentinel-2 MSI images taken over Estonia. We find the ρ 2.2 / ρ 0.6 value for evergreen needleleaf forests to be rather stable over time, with a mean value around 2.25. In broadleaf deciduous forests, on the other hand, we find ρ 2.2 / ρ 0.6 to increase rapidly in the spring, to remain on average close to 3 or greater, and then to decrease in the autumn. We offer parametric and tabulated models of ρ 2.2 / ρ 0.6 for eventual integration with Sen2Cor.

Source: Forestry Studies / Metsanduslikud Uurimused

DOI: 10.2478/fsmu-2025-0001

Authors: Mait Lang, Lea Hallik, Allan Sims

Institutions: Estonian University of Life Sciences, Tartu Observatory, University of Tartu

Language: English

Topic: Remote Sensing in Agriculture

Subfield: Ecology

Field: Environmental Science

Domain: Physical Sciences

SDGs: Life on land, Climate action

Open Access status: diamond

Funders: Eesti Teadusagentuur, U.S. Geological Survey