Assessment of Methane Retrieval Algorithms for EnMAP and PRISMA Shortwave Infrared Observations

Kurzfassung

Quantitative Comparison of Methane Retrieval Algorithms from EnMAP Shortwave Infrared Observations The Environmental Mapping and Analysis Program (EnMAP) satellite provides hyperspectral measurements with high spatial resolution. Despite its rather low spectral resolution for atmospheric remote sensing in the shortwave infrared (SWIR) range, it is possible to detect local methane (CH4) enhancements from strong point-like sources such as coal mines or landfills. In this study, data-driven models (using fast linear least squares) and physics-based retrievals (using iterative estimation schemes) are applied to EnMAP measurements of identified CH4 point sources. The methods include Generalized Least Squares, Separable Least Squares, Nonlinear Least Squares, Matched Filters, and Singular Value Decomposition. The results of the different retrieval schemes are assessed with regards to their uncertainties and statistical significance. Particular attention is given to addressing degeneracies between surface reflectivity and molecular absorption features caused by the sensor's low spectral resolution. Retrieval algorithms are tested under diverse environmental backgrounds with both homogeneous and heterogeneous spectral reflectivity. Retrieval errors are quantified and evaluated using various statistical methods. Well established approaches , successfully applied to other platforms such as airborne HySpex measurements, are adopted for this analysis.