Predictions of coastal evolution driven by episodic and persistent processes associated with storms and relative sea-level rise (SLR) are required to test our understanding, evaluate our predictive capability, and to provide guidance for coastal management decisions. Previous work demonstrated that the spatial variability of long-term shoreline change can be predicted using observed SLR rates, tide range, wave height, coastal slope, and a characterization of the geomorphic setting. The shoreline is not suf- ficient to indicate which processes are important in causing shoreline change, such as overwash that depends on coastal dune elevations. Predicting dune height is intrinsically important to assess future storm vulnerability. Here, we enhance shoreline-change predictions by including dune height as a vari- able in a statistical modeling approach. Dune height can also be used as an input variable, but it does not improve the shoreline-change prediction skill. Dune-height input does help to reduce prediction uncer- tainty. That is, by including dune height, the prediction is more precise but not more accurate. Comparing hindcast evaluations, better predictive skill was found when predicting dune height (0.8) compared with shoreline change (0.6). The skill depends on the level of detail of the model and we identify an optimized model that has high skill and minimal overfitting. The predictive model can be implemented with a range of forecast scenarios, and we illustrate the impacts of a higher future sea-level. This scenario shows that the shoreline change becomes increasingly erosional and more uncertain. Predicted dune heights are lower and the dune height uncertainty decreases.
Additional publication details
|Publication Subtype||Journal Article|
|Title||Coupling centennial-scale shoreline change to sea-level rise and coastal morphology in the Gulf of Mexico using a Bayesian network|
|Series title||Earth's Future|
|Contributing office(s)||St. Petersburg Coastal and Marine Science Center|
|Online Only (Y/N)||N|
|Additional Online Files (Y/N)||N|