A SENSOR INVARIANT ATMOSPHERIC CORRECTION: SENTINEL-2/MSI AND LANDSAT 8/OLI 1 Feng Yin 1* , Philip E Lewis 1,2 , Jose L G ´ omez-Dans 1,2 Qingling Wu 1 1 Department of Geography, University College London, Gower Street, London WC1E 6BT, United Kingdom 2 National Centre for Earth Observation (NCEO), NERC, United Kingdom * Corresponding author: feng.yin.15@ucl.ac.uk Abstract 1 Mitigating the impact of atmospheric effects on optical data is 2 a critical for monitoringland processes. Consistent approaches to 3 different sensors, which also quantify uncertainty, are required to 4 combine surface reflectance observations from heterogeneous 5 sensors. This paper provides a sensor agnostic approach to 6 atmospheric correction, called SIAC. It exploits operational global 7 datasets on (i) coarse resolution spectral surface bi-directional 8 reflectance distribution function (BRDF) and (ii) coarse resolution 9 atmospheric composition. The method infers aerosol optical 10 thickness (AOT) and total columnar water vapour (TCWV) from 11 top of atmosphere (TOA) reflectance observations, using a 12 Bayesian framework that exploits the MODIS MCD43 BRDF 13 descriptor product and the Copernicus Atmosphere Monitoring 14 Service (CAMS) operational forecasts of AOT and TCWV to 15 provide an a priori estimate. Spatial smoothness constraints are 16 assumed for AOT and TCWV, and efficient statistical 17 approximations (so-called emulators) to atmospheric radiative 18 transfer (RT) codes are used to efficiently invert the parameters. 19 BRDF descriptors are used to provide an estimation of surface 20 directional reflectance (SDR) at a coarse resolution, and linear 21 spectral mappings to convert to the target sensor spectral 22 configuration. The method is demonstrated on Sentinel 2 and 23 Landsat 8 data. AOT retrieval for both S2 and L8 shows a very 24 high correlation to AERONET estimates (r 2 > 0.9, RMSE < 0.025 25 for both sensors), although with a small underestimate of AOT. 26 1 Preprint submitted to Remote Sensing of Environment 1/42