Development of an automatic delineation of cliff top and toe on very irregular planform coastlines (CliffMetrics v1.0)

Autor: A. Payo, B. Jigena Antelo, M. Hurst, M. Palaseanu-Lovejoy, C. Williams, G. Jenkins, K. Lee, D. Favis-Mortlock, A. Barkwith, M. A. Ellis
Jazyk: angličtina
Rok vydání: 2018
Předmět:
Zdroj: Geoscientific Model Development, Vol 11, Pp 4317-4337 (2018)
Druh dokumentu: article
ISSN: 1991-959X
1991-9603
DOI: 10.5194/gmd-11-4317-2018
Popis: We describe a new algorithm that automatically delineates the cliff top and toe of a cliffed coastline from a digital elevation model (DEM). The algorithm builds upon existing methods but is specifically designed to resolve very irregular planform coastlines with many bays and capes, such as parts of the coastline of Great Britain. The algorithm automatically and sequentially delineates and smooths shoreline vectors, generates orthogonal transects and elevation profiles with a minimum spacing equal to the DEM resolution, and extracts the position and elevation of the cliff top and toe. Outputs include the non-smoothed raster and smoothed vector coastlines, normals to the coastline (as vector shape files), xyz profiles (as comma-separated-value, CSV, files), and the cliff top and toe (as point shape files). The algorithm also automatically assesses the quality of the profile and omits low-quality profiles (i.e. extraction of cliff top and toe is not possible). The performance of the proposed algorithm is compared with an existing method, which was not specifically designed for very irregular coastlines, and to manually digitized boundaries by numerous professionals. Also, we assess the reproducibility of the results using different DEM resolutions (5, 10 and 50 m), different user-defined parameter sets related to the degree of coastline smoothing, and the threshold used to identify the cliff top and toe. The model output sensitivity is found to be smaller than the manually digitized uncertainty. The code and a manual are publicly available on a GitHub repository.
Databáze: Directory of Open Access Journals
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje