Variations in Self-Gravity Wake Structures Across Saturn's Rings

Optical depths measured in stellar occultations by Saturn's rings depend on viewing geometry due to the presence of aligned, trailing, elongated ephemeral clumps of particles known as self-gravity wakes. Combining observations from multiple viewing geometries makes it possible to untangle the properties of the self-gravity wakes, such as their orientation, aspect ratio, mutual spacing, and inter-wake optical depth. Simple geometric models (Colwell et al. 2006, Geophys. Res. Lett. 33, L07201; Hedman et al. 2007, Astron. J. 133, 2624-2629) have explained most of the variation in optical depths as a function of viewing geometry. Many more occultations have been observed since those initial models were published: more than 100 have been observed by Cassini UVIS, while the initial model results were based on only ~10 measurements. In particular, some measurements made by UVIS at high elevation angle did not agree with predictions from the initial self-gravity wake "granola bar" model of Colwell et al. (2006). Here we present results of a systematic re-analysis of the self-gravity wakes in Saturn's rings taking advantage of more than 80 UVIS occultations with a strong signal and including different geometric models. We find no evidence for self-gravity wakes in the C ring or in the Cassini Division. While we cannot rule out the presence of some preferential orientation of particle structures from these data alone, the theoretically expected wavelength for self-gravity wakes in these regions is comparable to a particle size (~ 1 m), consistent with our non-detection of aligned clumps. We use three different geometric models of self-gravity wakes: an elliptical cross-section (Hedman et al. 2007), a rectangular cross-section (Colwell et al. 2006), and a rectangular cross-section with Gaussian "wings" in optical depth on the self-gravity wakes. The model with wings on the wakes, despite having an extra free parameter, does not provide a better overall fit to the data outside the central A ring. We find significant differences in the results between ingress and egress occultations (except in the central A ring) that cannot be due to calibration or background signal issues. Some differences are due to the different geometries of these occultation sets. In addition, we find differences in the derived wake structure between the different geometric models. These results suggest that the structure of the wakes is more complicated than what is captured by these simple models in most ring regions.