Abstract: Spatial heterodyne spectroscopy tomographic spectrometer can characterize atmospheric profiles, including temperature, density and moisture profiles, and provide data support for aerospace engineering and exploration activities. Limited availability of payload data has constrained both the research of the instrument performance and inversion algorithms. Initially, this study establishes a numerical simulation model, followed by a comparison between the simulated interference images and those acquired from the prototype. Subsequently, we adopt the Sobol method, the EFAST and the random forest to analyze weighting coefficients of image quality degradation. A strong consistency is identified between the simulated and experimentally measured images, with a relative error of 13.82%. Three complementary methods verify that Littrow error and zero optical path difference error represent the principal sources of image quality degradation, and the EFAST is determined to be the most well-suited methodology for investigating the underlying degradation process. This investigation will underpin payload design, performance prediction, and the advancement of data processing algorithms.