TY - JOUR
T1 - Analysis and modeling of field data of coastal morphological evolution over yearly and decadal time scales
T2 - Part 1. Background and linear techniques
AU - Larson, Magnus
AU - Capobianco, Michele
AU - Jansen, Henk
AU - Rózyński, Grzegorz
AU - Southgate, Howard N.
AU - Stive, Marcel
AU - Wijnberg, Kathelijne Mariken
AU - Hulscher, Suzanne
PY - 2003
Y1 - 2003
N2 - A number of statistical techniques to analyze and model coastal morphological evolution over yearly and decadal (i.e., long-term) time scales based on field data are presented. After a general introduction to long-term morphological modeling, mainly linear methods are discussed, whereas nonlinear methods are treated in a companion paper (South-gate et al, 2001). The theoretical background to the methods introduced is summarized and examples of field applications are given to illustrate each method. High-quality field data sets from different sites in the world, including Germany, The Netherlands, and United States, were employed in these examples. The analysis and modeling techniques used encompassed bulk statistics (mean, standard deviation, correlation etc), random sine functions, empirical orthogonal functions, canonical correlation analysis, and principal oscillation pattern analysis. Besides an evaluation of how suitable respective technique is for analyzing and modeling long-term morphological evolution, some general observations are presented regarding scales of morphological response as derived from the field applications. Data describing the evolution of both natural and anthropogenically affected coastal systems were studied. All methods investigated proved their usefulness for extracting characteristics of long-term morphological evolution, as well as for modeling this evolution, when applied under the right circumstances. However, more sophisticated techniques rely on more data in time and space, which is typically the limiting factor in the application of statistical methods as those presented here.
AB - A number of statistical techniques to analyze and model coastal morphological evolution over yearly and decadal (i.e., long-term) time scales based on field data are presented. After a general introduction to long-term morphological modeling, mainly linear methods are discussed, whereas nonlinear methods are treated in a companion paper (South-gate et al, 2001). The theoretical background to the methods introduced is summarized and examples of field applications are given to illustrate each method. High-quality field data sets from different sites in the world, including Germany, The Netherlands, and United States, were employed in these examples. The analysis and modeling techniques used encompassed bulk statistics (mean, standard deviation, correlation etc), random sine functions, empirical orthogonal functions, canonical correlation analysis, and principal oscillation pattern analysis. Besides an evaluation of how suitable respective technique is for analyzing and modeling long-term morphological evolution, some general observations are presented regarding scales of morphological response as derived from the field applications. Data describing the evolution of both natural and anthropogenically affected coastal systems were studied. All methods investigated proved their usefulness for extracting characteristics of long-term morphological evolution, as well as for modeling this evolution, when applied under the right circumstances. However, more sophisticated techniques rely on more data in time and space, which is typically the limiting factor in the application of statistical methods as those presented here.
KW - Data analysis
KW - Modeling
KW - Bulk statistics
KW - Principal component analysis
KW - Long
KW - Principal oscillation patterns
KW - Canonical correlation analysis
KW - Thogonal functions
KW - 2023 OA procedure
M3 - Article
SN - 0749-0208
VL - 19
SP - 760
EP - 775
JO - Journal of coastal research
JF - Journal of coastal research
IS - 4
ER -