Incorrect short-wave infrared to red reflectance ratio affects the application of the dark dense vegetation method for operational atmospheric correction in the hemiboreal region

Dark dense vegetation (DDV) is widely used in Earth observation to determine the aerosol optical depth in the atmosphere. The approach assumes that aerosol has a negligible effect on DDV reflectance in the shortwave infrared (SWIR, [Formula: see text]) and that its reflectance in red ([Formula: see text]) forms a constant fraction of [Formula: see text]. We used hyperspectral airborne measurements and Landsat-8/9 OLI and Sentinel-2 MSI time series of Estonian hemiboreal forests to study the seasonal variations in the optical properties of DDV-covered pixels. Forest inventory data allowed us to identify the effect of tree species composition on the [Formula: see text] ratio of DDV pixels selected based on either SWIR or red reflectance. The results show that [Formula: see text] is systematically less than the commonly assumed constant of 0.5, with a strong seasonal dependence driven by deciduous and evergreen habits of trees. The selection of pixels with the lowest reflectance in SWIR does not always lead to pixels with the lowest reflectance in red. Assuming [Formula: see text] yields underestimated aerosol optical depth and overestimated surface reflectance in the visible part of the reflectance spectrum for the studied forests.