Hello! I thought that I’d create a post/thread for summarizing the advances in polarimetric wave slope sensing that were shared during WISE 2025. This thread can also aid our collaborative efforts as we work to move the technique forward.
Polarimetric slope sensing presents a special (perhaps unique) opportunity to observe the spatiotemporal characteristics of short-scale waves in the real ocean. But there are a number of challenges associated with the technique-- some well-known, others newly emerging. Here is a list of some of those challenges. I’ll update it as the discussion evolves.
- Light rays encounter the focal plane array at an angle oblique to the polarizing filter. Pistellato & Bergamasco (2024) provide a geometric model to mitigate the issues associated with this challenge.
- The relationship between Degree of Linear Polarization (DoLP) and the angle of incidence (AOI) is strongly nonlinear. DoLP becomes doubly-defined with respect to AOI near Brewster’s angle, which might cause a DoLP->AOI inversion to “wrap”. Perhaps more insidiously, the sensitivity of the technique to AOI will diminish as AOI approaches 0 or Brewster’s angle; i.e., a very small amount of variation in DoLP contains a large variation in AOI near those extrema. This latter concern was mentioned as a concern by Zappa et al. 2012. This likely has a negative effect on signal-to-noise ratio.
- Absent external measurements of incoming (sky-leaving) and upwelling (surface-leaving) polarization state, we have an incomplete set of information for inferring the surface normal vectors. Barsic & Chinn (2012) provided a framework for incorporating measurements of sky-leaving polarization state into the PSS technique. Mara discussed a technique for using stereo processing to obtain a first-order quantification of surface slope, iterating and refining to estimate the incoming and upwelling Stokes vector.
- Sea surface slope variability on scales of the order of the image pixel has a negative impact on the recovered wave spectrum and associated statistics (Laxague et al., 2025). This impact is seen in analysis of Division of Amplitude (DoAm) polarimetric data, high-resolution polarimetric ray-tracing modeling, and the newest Division of Focal Plane (DoFP) polarimetric data. The effect is strongest for the DoFP imager, given the sparse nature of the polarization measurements made with that detector. These results may depend on the choice of interpolator; however, our application of larger kernels (e.g., the 12-pixel kernel of Ratliff et al., (2009)) has not seemed to mitigate the negative effect.
Code repositories
Collaborative effort between Laxague group at UNH (CASSLL) and Zappa group at LDEO (OASIS). Code that gets you from raw image intensity to surface slope fields (albeit without directly addressing the challenges identified in the text above).
… more to come