Spectral Mixture Modeling with Laboratory Near-Infrared Data II: Effects of Grain Size and Implications for Europa

Abstract

Spectral analysis using linear mixture (LM) and radiative transfer-based (RT) intimate mixture modeling based on Hapke theory at near-infrared wavelengths are applied to estimate the abundance of surface materials on Europa. Previously, Emran (2026) compared these approaches against the laboratory spectra of H$_2$O ice and H$_2$SO$_4$$\cdot$8H$_2$O mixtures with $\sim$100 $μ$m grains. Here, the effect of particle size on spectral modeling accuracy was assessed using laboratory spectra of H$_2$O ice mixtures with small ($\sim$70 $μ$m spherical) and coarse ($\sim$1 mm irregular) grains, measured over the $\sim$1.2-2.5 $μ$m wavelength range at 100 K and 120 K (Stephan et al., 2021). Modeled abundance estimates at both temperatures show consistent trends across all mixing ratios, with only minor temperature-dependent variations. The discrepancy in abundance estimates from both LM and RT models remains within $\pm$10% across all mixtures, with the error reduced to $\pm$5% when fine grains dominate. Across all mixtures, the average difference between RT- and LM-derived abundance estimates remains within $\pm$2% for mixtures containing both small and large grains. In contrast, mixtures composed solely of smaller grains render larger deviations between the models, with RT producing more accurate estimates (Emran, 2026) -- indicating that the presence of coarse H$_2$O ice grains minimizes abundance differences between LM and RT modeling. Thus, I posit that Hapke-based RT modeling is the preferred spectral modeling approach -- regardless of grain size or compositional mixture -- for constraining Europa's surface composition. Nonetheless, LM modeling remains a reliable approach for compositional analysis of terrains containing H$_2$O ice with $\sim$mm-sized grains.