Eradiate: Supporting metrology for Earth observation with highly accurate radiative transfer simulation
by Dr. Vincent Leroy, Mr. Yvan Nollet, Mr. Sebastian Schunke, Mr. Nicolas Misk, Dr. Yves Govaerts

Abstract

Many metrological activities for Earth observation applications (e.g. vicarious calibration, uncertainty quantification) require the use of numerical models. However, radiative transfer models (RTM) are usually not designed for the purpose of being used in a metrological processing chain. Matching the performance of modern or upcoming hyperspectral instruments dedicated to calibration such as TRUTHS or CLARREO requires an accuracy of simulated results better than 1%. This can currently not be achieved with the radiative transfer models commonly used for cal/val. Getting to a better accuracy level requires to relax strong modelling assumptions (e.g. plane-parallel geometry, smooth surface), as well as advanced molecular absorption modelling and high-quality input data. Eradiate [www.eradiate.eu], a new open-source radiative transfer model, was designed to support metrological activities. It prioritises accuracy and ships traceable data. It integrates surface and atmosphere modelling consistently, solving a single, fully coupled 3D problem. The simulated scene consists of a 1D atmosphere (plane-parallel or spherical-shell geometry) and a detailed 3D surface geometry. Care is taken in atmospheric absorption modelling, simulated using a correlated-k distribution method with narrow spectral bins (1 nm or 10 nm resolution). Eradiate is extensively compared against other models and actual satellite data. Eradiate’s modern Python interface make it easy to integrate in a processing chain. Eradiate is consistently tested, maintained and improved, augmented with new data, algorithms and basic scene elements in full transparency thanks to its open-source license (LGPLv3) and a rigorous documentation policy. Eradiate is currently used for cal/val studies, sensitivity analysis and general radiative transfer physics studies and the development roadmap includes the addition of features and data specifically dedicated to supporting metrological activities (e.g. guaranteeing the SI-traceability of input data). Its high accuracy makes it a valuable tool for satellite calibration (inter-instrument or vicarious). Examples of satellite observations acquired over bright desert pseudo-invariant calibration sites (PICS) will be presented. The development of Eradiate was funded by the MetEOC-3 project (EMPIR grant 16ENV03) and the Copernicus programme managed by ESA (contract no 4000127201/19). Eradiate is used to contribute to the MetEOC-4 project (EMPIR grant 19ENV07 MetEOC-4).