Quantifying spatial and temporal trends in beach-dune volumetric changes using spatial statistics

Jordan B R Eamer, Ian Walker

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Spatial statistics are generally underutilized in coastal geomorphology, despite offering great potential for identifying and quantifying spatial-temporal trends in landscape morphodynamics. In particular, local Moran's Ii provides a statistical framework for detecting clusters of significant change in an attribute (e.g., surface erosion or deposition) and quantifying how this changes over space and time. This study analyzes and interprets spatial-temporal patterns in sediment volume changes in a beach-foredune-transgressive dune complex following removal of invasive marram grass (Ammophila spp.). Results are derived by detecting significant changes in post-removal repeat DEMs derived from topographic surveys and airborne LiDAR. The study site was separated into discrete, linked geomorphic units (beach, foredune, transgressive dune complex) to facilitate sub-landscape scale analysis of volumetric change and sediment budget responses. Difference surfaces derived from a pixel-subtraction algorithm between interval DEMs and the LiDAR baseline DEM were filtered using the local Moran's Ii method and two different spatial weights (1.5 and 5. m) to detect statistically significant change. Moran's Ii results were compared with those derived from a more spatially uniform statistical method that uses a simpler student's t distribution threshold for change detection. Morphodynamic patterns and volumetric estimates were similar between the uniform geostatistical method and Moran's Ii at a spatial weight of 5. m while the smaller spatial weight (1.5. m) consistently indicated volumetric changes of less magnitude. The larger 5. m spatial weight was most representative of broader site morphodynamics and spatial patterns while the smaller spatial weight provided volumetric changes consistent with field observations. All methods showed foredune deflation immediately following removal with increased sediment volumes into the spring via deposition at the crest and on lobes in the lee, despite erosion on the stoss slope and dune toe. Generally, the foredune became wider by landward extension and the seaward slope recovered from erosion to a similar height and form to that of pre-restoration despite remaining essentially free of vegetation.

Original languageEnglish (US)
Pages (from-to)94-108
Number of pages15
JournalGeomorphology
Volume191
DOIs
StatePublished - Jun 1 2013
Externally publishedYes

Fingerprint

dune
beach
morphodynamics
digital elevation model
erosion
deflation
sediment budget
volume change
sediment
geomorphology
pixel
student
trend
statistics
grass
method
vegetation
removal

Keywords

  • Ammophila
  • Beach
  • Coastal geomorphology
  • DEM
  • Foredune
  • Spatial statistics

ASJC Scopus subject areas

  • Earth-Surface Processes

Cite this

Quantifying spatial and temporal trends in beach-dune volumetric changes using spatial statistics. / Eamer, Jordan B R; Walker, Ian.

In: Geomorphology, Vol. 191, 01.06.2013, p. 94-108.

Research output: Contribution to journalArticle

@article{acd0400acbb546b8a861bb3d67a5bfa7,
title = "Quantifying spatial and temporal trends in beach-dune volumetric changes using spatial statistics",
abstract = "Spatial statistics are generally underutilized in coastal geomorphology, despite offering great potential for identifying and quantifying spatial-temporal trends in landscape morphodynamics. In particular, local Moran's Ii provides a statistical framework for detecting clusters of significant change in an attribute (e.g., surface erosion or deposition) and quantifying how this changes over space and time. This study analyzes and interprets spatial-temporal patterns in sediment volume changes in a beach-foredune-transgressive dune complex following removal of invasive marram grass (Ammophila spp.). Results are derived by detecting significant changes in post-removal repeat DEMs derived from topographic surveys and airborne LiDAR. The study site was separated into discrete, linked geomorphic units (beach, foredune, transgressive dune complex) to facilitate sub-landscape scale analysis of volumetric change and sediment budget responses. Difference surfaces derived from a pixel-subtraction algorithm between interval DEMs and the LiDAR baseline DEM were filtered using the local Moran's Ii method and two different spatial weights (1.5 and 5. m) to detect statistically significant change. Moran's Ii results were compared with those derived from a more spatially uniform statistical method that uses a simpler student's t distribution threshold for change detection. Morphodynamic patterns and volumetric estimates were similar between the uniform geostatistical method and Moran's Ii at a spatial weight of 5. m while the smaller spatial weight (1.5. m) consistently indicated volumetric changes of less magnitude. The larger 5. m spatial weight was most representative of broader site morphodynamics and spatial patterns while the smaller spatial weight provided volumetric changes consistent with field observations. All methods showed foredune deflation immediately following removal with increased sediment volumes into the spring via deposition at the crest and on lobes in the lee, despite erosion on the stoss slope and dune toe. Generally, the foredune became wider by landward extension and the seaward slope recovered from erosion to a similar height and form to that of pre-restoration despite remaining essentially free of vegetation.",
keywords = "Ammophila, Beach, Coastal geomorphology, DEM, Foredune, Spatial statistics",
author = "Eamer, {Jordan B R} and Ian Walker",
year = "2013",
month = "6",
day = "1",
doi = "10.1016/j.geomorph.2013.03.005",
language = "English (US)",
volume = "191",
pages = "94--108",
journal = "Geomorphology",
issn = "0169-555X",
publisher = "Elsevier",

}

TY - JOUR

T1 - Quantifying spatial and temporal trends in beach-dune volumetric changes using spatial statistics

AU - Eamer, Jordan B R

AU - Walker, Ian

PY - 2013/6/1

Y1 - 2013/6/1

N2 - Spatial statistics are generally underutilized in coastal geomorphology, despite offering great potential for identifying and quantifying spatial-temporal trends in landscape morphodynamics. In particular, local Moran's Ii provides a statistical framework for detecting clusters of significant change in an attribute (e.g., surface erosion or deposition) and quantifying how this changes over space and time. This study analyzes and interprets spatial-temporal patterns in sediment volume changes in a beach-foredune-transgressive dune complex following removal of invasive marram grass (Ammophila spp.). Results are derived by detecting significant changes in post-removal repeat DEMs derived from topographic surveys and airborne LiDAR. The study site was separated into discrete, linked geomorphic units (beach, foredune, transgressive dune complex) to facilitate sub-landscape scale analysis of volumetric change and sediment budget responses. Difference surfaces derived from a pixel-subtraction algorithm between interval DEMs and the LiDAR baseline DEM were filtered using the local Moran's Ii method and two different spatial weights (1.5 and 5. m) to detect statistically significant change. Moran's Ii results were compared with those derived from a more spatially uniform statistical method that uses a simpler student's t distribution threshold for change detection. Morphodynamic patterns and volumetric estimates were similar between the uniform geostatistical method and Moran's Ii at a spatial weight of 5. m while the smaller spatial weight (1.5. m) consistently indicated volumetric changes of less magnitude. The larger 5. m spatial weight was most representative of broader site morphodynamics and spatial patterns while the smaller spatial weight provided volumetric changes consistent with field observations. All methods showed foredune deflation immediately following removal with increased sediment volumes into the spring via deposition at the crest and on lobes in the lee, despite erosion on the stoss slope and dune toe. Generally, the foredune became wider by landward extension and the seaward slope recovered from erosion to a similar height and form to that of pre-restoration despite remaining essentially free of vegetation.

AB - Spatial statistics are generally underutilized in coastal geomorphology, despite offering great potential for identifying and quantifying spatial-temporal trends in landscape morphodynamics. In particular, local Moran's Ii provides a statistical framework for detecting clusters of significant change in an attribute (e.g., surface erosion or deposition) and quantifying how this changes over space and time. This study analyzes and interprets spatial-temporal patterns in sediment volume changes in a beach-foredune-transgressive dune complex following removal of invasive marram grass (Ammophila spp.). Results are derived by detecting significant changes in post-removal repeat DEMs derived from topographic surveys and airborne LiDAR. The study site was separated into discrete, linked geomorphic units (beach, foredune, transgressive dune complex) to facilitate sub-landscape scale analysis of volumetric change and sediment budget responses. Difference surfaces derived from a pixel-subtraction algorithm between interval DEMs and the LiDAR baseline DEM were filtered using the local Moran's Ii method and two different spatial weights (1.5 and 5. m) to detect statistically significant change. Moran's Ii results were compared with those derived from a more spatially uniform statistical method that uses a simpler student's t distribution threshold for change detection. Morphodynamic patterns and volumetric estimates were similar between the uniform geostatistical method and Moran's Ii at a spatial weight of 5. m while the smaller spatial weight (1.5. m) consistently indicated volumetric changes of less magnitude. The larger 5. m spatial weight was most representative of broader site morphodynamics and spatial patterns while the smaller spatial weight provided volumetric changes consistent with field observations. All methods showed foredune deflation immediately following removal with increased sediment volumes into the spring via deposition at the crest and on lobes in the lee, despite erosion on the stoss slope and dune toe. Generally, the foredune became wider by landward extension and the seaward slope recovered from erosion to a similar height and form to that of pre-restoration despite remaining essentially free of vegetation.

KW - Ammophila

KW - Beach

KW - Coastal geomorphology

KW - DEM

KW - Foredune

KW - Spatial statistics

UR - http://www.scopus.com/inward/record.url?scp=84876962720&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876962720&partnerID=8YFLogxK

U2 - 10.1016/j.geomorph.2013.03.005

DO - 10.1016/j.geomorph.2013.03.005

M3 - Article

VL - 191

SP - 94

EP - 108

JO - Geomorphology

JF - Geomorphology

SN - 0169-555X

ER -