Abstract by Brandon Cook
A Nonstationary, Anisotropic Model for Sand Dune Data
Saturn’s largest moon Titan has a dense atmosphere with equatorial winds that create particle dunes. The Cassini-Huygens fly-by in 2004 collected data and sent back high-resolution images of these surface features. Measurements of dune characteristics at approximately 90,000 locations were gathered from the images. The goal of the research is estimation of dune widths and spacing across Titan’s sand seas. These variables (and covariates) are highly spatially correlated, but in varying directions and magnitudes. Using dune direction as a reference, a nonstationary, anisotropic model is fit to portions of the data using a simplification of kernel convolution proposed by Paciorek and Schervish (2006) and implemented in the convoSPAT package in R.