Coastline Complexity: A Parameter for Functional Classification of Coastal Environments

To understand the role of the world's coastal zone (CZ) in global biogeochemical fluxes (particularly those of carbon, nitrogen, phosphorus, and sediments) we must generalize from a limited number of observations associated with a few well-studied coastal systems to the global scale. Global generalization must be based on globally available data and on robust techniques for classification and upscaling. These requirements impose severe constraints on the set of variables that can be used to extract information about local CZ functions such as advective and metabolic fluxes, and differences resulting from changes in biotic communities. Coastal complexity (plan-view tortuosity of the coastline) is a potentially useful parameter, since it interacts strongly with both marine and terrestrial forcing functions to determine coastal energy regimes and water residence times, and since 'open' vs. 'sheltered' categories are important components of most coastal habitat classification schemes. This study employs the World Vector Shoreline (WVS) dataset, originally developed at a scale of 1:250,000. Coastline complexity measures are generated using a modification of the Angle Measurement Technique (AMT), in which the basic measurement is the angle between two lines of specified length drawn from a selected point to the closest points of intersection with the coastline. Repetition of these measurements for different lengths at the same point yields a distribution of angles descriptive of the extent and scale of complexity in the vicinity of that point; repetition of the process at different points on the coast provides a basis for comparing both the extent and the characteristic scale of coastline variation along different reaches of the coast. The coast of northwestern Mexico (Baja California and the Gulf of California) was used as a case study for initial development and testing of the method. The characteristic angle distribution plots generated by the AMT analysis were clustered using LOICZVIEW, a high dimensionality clustering routine developed for large-scale coastal classification studies. The results show distinctive differences in coastal environments that have the potential for interpretation in terms of both biotic and hydrogeochemical environments, and that can be related to the resolution limits and uncertainties of the shoreline data used.