Advection and diffusion in shoreline change prediction

We added longshore advection and diffusion to the simple cross-shore rate calculation method, as used widely by the USGS and others, to model historic shorelines and to predict future shoreline positions; and applied this to Hawaiian Island beach data. Aerial photographs, sporadically taken throughout the past century, yield usable, albeit limited, historic shoreline data. These photographs provide excellent spatial coverage, but poor temporal resolution, of the shoreline. Due to the sparse historic shoreline data, and the many natural and anthropogenic events influencing coastlines, we constructed a simplistic shoreline change model that can identify long-term behavior of a beach. Our new, two-dimensional model combines the simple rate method to accommodate for cross-shore sediment transport with the classic Pelnard-Considère model for diffusion, as well as a longshore advection speed term. Inverse methods identify cross-shore rate, longshore advection speed, and longshore diffusivity down a sandy coastline. A spatial averaging technique then identifies shoreline segments where one parameter can reasonably account for the cross-shore and longshore transport rates in that area. This produces model results with spatial resolution more appropriate to the temporal spacing of the data. Because changes in historic data can be accounted for by varying degrees of cross-shore and longshore sediment transport - for example, beach erosion can equally be explained by sand moving either off-shore or laterally - we tested several different model scenarios on the data: allowing only cross-shore sediment movement, only longshore movement, and a combination of the two. We used statistical information criteria to determine both the optimal spatial resolution and best-fitting scenario. Finally, we employed a voting method predicting the relaxed shoreline position over time.