Comments on: Process modeling for slope and aspect with application to elevation data maps

Abstract

The authors should be commended on their methodological development of stochastic processes for slope and aspect. Their development of basic distribution theory needed to study these two processes, and the sufficient conditions that ensure independence and non-informative induced priors, provide a thorough contribution to the collection of work on gradient processes. The fully model-based approach for inference for slope and aspect enables the propagation of uncertainty to environmental process models of interest that would use these variables as explanatory variables in a regression. I greatly appreciate the opportunity to comment on this exciting work and offer some additional model considerations and applications. Specifically, I reaffirm the importance of scalability of the methodology to large datasets, offering a few considerations with regard to model specification. Next, I discuss the unique challenges of using the predictive distributions of the slope and aspect processes as input variables in spatial regression models. Finally, I offer possible applications and extensions of this work that might provide innovative insights into environmental processes.