Premium
Assessing DEM interpolation methods for effective representation of upland stream morphology for rapid appraisal of bed stability
Author(s) -
Schwendel Arved C.,
Fuller Ian C.,
Death Russell G.
Publication year - 2012
Publication title -
river research and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.679
H-Index - 94
eISSN - 1535-1467
pISSN - 1535-1459
DOI - 10.1002/rra.1475
Subject(s) - kriging , interpolation (computer graphics) , channel (broadcasting) , representation (politics) , digital elevation model , multivariate interpolation , stability (learning theory) , geology , triangulation , computer science , algorithm , remote sensing , hydrology (agriculture) , mathematics , geometry , statistics , bilinear interpolation , artificial intelligence , computer vision , geotechnical engineering , machine learning , politics , political science , law , motion (physics) , computer network
Digital elevation models (DEMs) of river channels, built by interpolation between samples of topographic survey points, are widely used to represent surfaces and to derive land‐surface parameters. Differencing between successive DEMs permits quantification of change, which in gravel‐bed rivers is used to construct a morphological budget of lower bound estimates of sediment flux and bed‐stability surrogate. Choice of DEM interpolation method strongly influences DEM quality and realistic representation of channel forms. When comparing morphological budgets between multiple contrasting reaches, e.g. for rapid ecological appraisal, an effective and consistent means of DEM construction is required to avoid digitally generated inconsistencies. An appropriate interpolation method should be suitable for accurate representation of channels contrasting in substrate and hydraulic conditions, surveys of varying data density and distribution, and avoidance of site‐specific parameterization. This paper investigates representation of channel form using a series of DEMs generated within Surfer® by triangulation with linear interpolation, natural neighbours, point kriging, universal kriging, multiquadratic radial basis function, modified Shepard's method and inverse distance to a power on the example of four reaches of mountain streams in New Zealand. These reaches represent a diversity of channel forms, substrate and hydraulic properties. DEMs from triangulation with linear interpolation revealed consistently the best results and this method is recommended for geomorphological and ecological studies of multiple reaches. The main advantage over point kriging and radial basis function is better representation of channel margins and bedforms without introduction of breaklines, while it outperforms natural neighbours in honouring measured points. Copyright © 2010 John Wiley & Sons, Ltd.