Analysis of bathymetrical derived features on the Belgian continental shelf as a support for
marine habitat mapping
Verfaillie Els1, Foster-Smith, Bob2,Van Lancker Vera1
1Ghent
University, Renard Centre of Marine Geology (RCMG), Krijgslaan 281, S8, 9000 Gent, Belgium (Els.Verfaillie@ugent.be, Tel. +32 9 264 45 73, Fax +32 9 264 49 67)
2
Envision, University of Newcastle, Newcastle upon Tyne, United Kingdom
INTRODUCTION
­
study area Westhinder
The study area is a kink area in the middle part of the Westhinder offshore tidal sandbank, where a multibeam dataset is available with a
resolution of 2 m (Figure 1). Deleu et al. (2004) studied the morphology of this area thorougly. Both on a large and a medium scale the
occurrence of features (from sandbank level until medium dune level using the classification of Ashley (1990)) was described and
interpreted. Bedforms were manually drawn based on the bathymetry of the map, indicating their strike and asymmetry.
The aim of this study is to facilitate and automate the geomorphological analysis, comparing different bathymetrical derived
features. Furthermore the geomorphology is analysed in the context of marine habitat mapping, as topographic features are assumed to
be important possible habitats for marine organisms.
5740000
study area Hinderbanken
(
!
Hinderbanken
(
!
Kustbanken
(
!
Vlaamse Banken
(
!
Zeelandbanken
Bathymetry (m)
46
5720000
0
(
!
Noordhinderbank
!
(
Northing (m)
Bligh Bank
(
!
Oosthinderbank
!
(
5700000
Thorntonbank
MATERIALS AND METHODS
(
!
(
!
Westhinderbank
Gootebank
!
(
Fairy Bank
(
!
Akkaertbank
1. Bathymetric position index or BPI:
• A measure of where a location with a defined elevation is relative to the overall landscape (Weiss (2001), Iampietro & Kvitek (2002)
and Lundblad et al. (in press));
• Evaluates bathymetry differences between a focal point and the mean elevation of the surrounding cells within a user-defined
rectangle, annulus or circle;
• Aims to distinguish slopes, valleys, ridges and flat areas; generally calculated on a broad-scale (B-BPI) and a fine-scale (F-BPI);
Figure 1: Position of Westhinder study area and larger
• Applies a scalefactor (= resolution of bathymetry multiplied by outer radius of annulus) of 24 and 250 (because the medium and large Hinderbanken study area
dunes in this area have an average wavelength of 20 to 25 m, the very-large dunes of 250 m);
• Calculated with ArcGIS Benthic Terrain Modeler or BTM (Wright et al., 2005).
2. Topographic feature recognition – Isocluster classification:
B-BPI
• Series of different filters with varying window sizes, reveals different scale
features of the bedform morphology;
-1 stdev < B-BPI < 1stdev
<= -1 stdev
>= 1 stdev
• Subtraction of filtered images from other images, with smaller cell sizes, creates
Depressions
Flats & slopes
Crests
residual filtered images;
F-BPI
F-BPI
Slope
• Filtered images, with varying search radii, reveal bedforms ranging from 10 to
100 m and with heights between 0.4 and 3 m;
< 1 stdev
< 1 stdev
>= 1 stdev
>= 1 stdev
<= 4,01
> 4,01
• Key scales on different scales (ranging from fine-scale until broad-scale
Small scale crests
features) are selected from the filtered images to come to a classified image;
Small scale crests
Depressions
Flats
Crests
Slopes
• Calculated with Idrisi software Isocluster unsupervised classification (based on
maximum likelihood procedure) of the residual images, creating 10 clusters of
Criteria used in BPI classifications
F-BPI
F-BPI
1 stdev = 1 standard deviation,
bedform features.
corresponding to 100 grid value units
(
!
Wenduinebank
Middelkerkebank !
(
(
!
Oostendebank
!
(
!
(
Ravelingen
(
!
5680000
Oostdijck
Kwintebank
!
(
Buiten Ratel
Balandbank
(
!
(
!
0 2.5 5
10
km
UTM31N - WGS84 coordinates
Stroombank
(
!
Nieuwpoortbank
!
(
Smalbank
Den Oever
(
!
5660000
440000
5706000
4
4
4
4
4
4
4
4
Northing (m)
(
4
4
WH0203.025
WH0203.034
4
5704000
WH0203.036
4
WH0203.042
WH0203.041
4
4
WH0203.043
4
WH0203.040
WH0203.049
­
4
4
4
4
4
4
WH0203.050
0
A
km
1
0.25 0.5
0
WH0203.051
WH0203.058
0.25
km
1
0.5
UTM31N - WGS84 coordinates
WH0203.052
WH0203.057
Easting (m)
Easting (m)
Figure 5: Isocluster classification of 4 key scale
filtered images resulting in 10 classes
4
4
4
4
4
4
4
RESULTS
-10
bathymetry (m)
Northing (m)
5704000
4
5706000
4
4
4
4
4
5706000
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
(
4
4
4
4
4
4
4
4
4
4
4
4
4
Northing (m)
WH0203.026
UTM31N - WGS84 coordinates
4
!
(
(
4
4
4
4
4
4
4
(
(
(
(
( (
(
4
!!
(
4
4
4
(
(((! 4
(
(
(
(
(
(
( (
( ((
(
(
(
(
(
(
(
!
((
(
(
(
(
(
(
(
( (
5704000
WH0203.027
WH0203.033
WH0203.048
4
4
4
4
WH0203.022
WH0203.035
4
4
4
­
WH0203.020
WH0203.021
(
(
(
(
(
(
(
(
4
WH0203.019
(
(
(
(
(
(
(
(
(
4
4
4
4
(
WH0203.010
WH0203.028
4
(
4
(
(
(
4
(
(
(
(
(
4
(
(
(
(
(
4
B
WH0203.011
-38.00
4
(
(
(
( (
4
(
( (
4
(
(
(
(
(
!
(
km
1
(
(
(
(
(
(
(
(
( (
(
(
4
4
4
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
W9413
WH0203.007
WH0203.012
WH0203.018
4
( ((
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
( (
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
!!
(
4
(
(
(
!( !
(
(
(
(
(
(
(
(
(
(
(
(
(
(
4
(
(
(
(
(
(
(
(
(
Figure 4: Aspect map of Westhinder area makes it
possible to derive asymmetry of bedforms
(
470000
Easting (m)
WH0203.013
-6.00
!
(
(
(
8
(
(
(
(
(
(
(
0.25 0.5
520000
WH0203.006
grab samples
Bathymetry (m)
(
(
4
wh0203.021
wh0203.010
-20
wh0203.058
wh0203.049
wh0203.042
wh0203.035
wh0203.026
A
-30
W9413
wh0203.007
1. Bathymetric position index or BPI:
• Classification decision tree (Figure 2), based on B-BPI (4 classes: depressions, crests, flats, slopes) and F-BPI classification (1 class:
small-scale crests) results in 5 classes: Figure 3;
• ‘Small-scale crests’ class corresponds to a large extent to the crestlines of the very-large dunes and, partly, to the crestlines of the
medium to large dunes, defined in Deleu et al. (2004) (Figure 3);
• Combined with an aspect map (orientation of the slopes: Figure 4) the asymmetry of bedforms can be derived;
• Figure 6: SW-NE profile through a sandbank, very-large dunes, medium to large dunes and 9 sediment samples shows that most of
morphological features are extracted using the BPI classification;
• Most of the tops/slopes of the very-large dunes (central part of profile between 1000 and 4000 m on Figure 6) are clearly classified
as small-scale crests;
• Depressions/crests are only derived for larger features. For medium to large dunes (at right side of 4000 m on
profile) only the largest crestlines, depressions and steepest slopes are derived based on the classification;
• Application on larger area in Hinderbanken (Figure 7): clear distinction between bedforms on sandbanks and swales.
460000
0
1000
2000
vertical exaggeration of profiles: 40
3000
4000
B
5000
distance (m)
small scale crests
wh0203.021
crests
depressions
wh0203.010
wh0203.058
wh0203.049
flats
wh0203.042
wh0203.035
slopes
wh0203.026
A
W9413
wh0203.007
BPI classification
B
wh0203.021
wh0203.010
wh0203.058
wh0203.049
wh0203.042
wh0203.035
wh0203.026
A
1
6
2
7
3
8
4
9
5
10
W9413
wh0203.007
Isocluster classification
B
Figure 6: SW-NE profile through the study area; comparison
between BPI and Isocluster classification
465000
Easting (m)
470000
475000
­
BPI classes
5705000
small scale crests
crests
depressions
flats
Topographic feature recognition – Isocluster classification:
Figure 5: similar features could be extracted, although sometimes difficult interpretation of classes;
Better extraction of crestlines of medium to large dunes, especially in the NW part of study area;
Class 1, 3 = noise; class 2 ≅ BPI-flats; class 4, 7, 10 ≅ BPI-depressions; class5, 9 ≅ BPI-small-scale crests;
class 6 ≅ BPI-slopes; class 8 ≅ BPI-crests.
FURTHER RESEARCH
• Application of bathymetric derived features on other areas
• Correlation of bathymetric features with other physical data layers such as multibeam backscatter
and sedimentological samples
• Correlation with marine species/communities to test if they are important possible habitats
• Other bathymetric derived features will be tested as input parameters for models (e.g. rugosity = topographic
roughness with a surface area to planar area ratio, as input for sand transport or biological models)
Northing (m)
2.
•
•
•
5700000
slopes
5695000
5706000
!
(
Northing (m)
!
(
(
7
(
5704000
(
(
(
6
10
­
!
(
(
5
9
UTM31N - WGS84 coordinates
0
(
(
(
(
4
(
(
(
((
(
(
(
(
3
(
(
(
(
(
(
(
(
(
(
(
(
(
(
Easting (m)
(
(
Isocluster classes
(
(
(
(
UTM31N - WGS84 coordinates
km
0
0.25 0.5
1
Figure 3: B-BPI and F-BPI classification based on
classification decision tree (Figure 2)
468000
very-large dunes with lee slope to the SW
very-large symmetric dune
(
(
(
(
(
(
(
­
(
(
((
(
(
( (
(
(
(
(
(
(
( (
(
(
(
(
(
(
((
(
(
(
(
(
(
(
(
(
(
(
(
(
(
!
2
(
(
(
(
(
(
(
(
(
(
!
very-large dunes with lee slope to the NE
(
(
(
(
(
4
1
(
(
(
(
(
( ((
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
((
(
(
(
(
(
(((! 4
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
( (
(
( ((
( (
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(( (!
(
(
(
(
(
( (( (
(
(
(
(
(
(
(
(
! ! ! ! !
(( (
(
(
(
(
(
(
(
( (
(
( ((
(
(
(
(
(
(
(
(
(
(
(
(
( (
(
(
(
(
( (
( ( 4
(
( (( (
( !
(
(
(
( (
(
(
(
(
(
(
(
(
(
(
(
(( (
(
( (
(
( (
( ( (
(
(
(
(
(
(
(
( (
(
( (( ! ( !
(
( ((
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
slopes
(
(
(
(
(
(
flats
(
( (
(
(
(
(
depressions
(
!( !
(
(
(
crests
(
(
small scale crests
!
(
(
(
BPI classes
!
(
(
(
(
(
(
4
500000
470000
medium to large dunes
(
(
(
(
!
(
(
468000
bedforms (Deleu et al., 2004)
bedforms (Deleu et al., 2004)
medium to large dunes
(
( very-large dunes with lee slope to the NE
4
4 very-large dunes with lee slope to the SW
! ! very-large symmetric dune
aspect (orientation of slope)
Flat (-1)
North (0-22.5)
Northeast (22.5-67.5)
East (67.5-112.5)
Southeast (112.5-157.5)
South (157.5-202.5)
Southwest (202.5-247.5)
West(247.5-292.5)
Northwest (292.5-337.5)
North (337.5-360)
Easting (m)
480000
Figure 2: Classification decision tree developed for Westhinder study area of a B-BPI and a F-BPI
classification, resulting in 5 classes: depressions, crests, flats, slopes and small-scale crests
466000
(
(
(
very-large symmetric dune
!
very-large dunes with lee slope to the SW
(
!
(
!
very-large dunes with lee slope to the NE
(
(
4
Slopes
470000
4
medium to large dunes
(
Small scale crests
468000
4
bedforms (Deleu et al., 2004)
4
Flats
466000
4
470000
< 1 stdev
4
Small scale crests
468000
>= 1 stdev
4
466000
< 1 stdev
460000
4
>= 1 stdev
Raster grid derived from bathymetry
Resulting classes
Bathymetry data compiled from the Ministry
of the Flemish Community AWZ - WWK
added with data from the Hydrographic Office
of the Netherlands and the UKHO
Broersbank
(
!
(
!
Trapegeer
UTM31N - WGS84 coordinates
0
1
2
km
4
Figure 7: BPI classification applied on multibeam bathymetric dataset (resolution
2 m) of larger area in Hinderbanken: clear distinction between bedforms on
sandbanks and swales. Patchy pattern in swales possible indicator of gravel
REFERENCES
• Ashley, G.M., 1990. Classification of large-scale subaqueous bedforms: a new look at an old problem. Journal of Sedimentary Petrology, 60 (1), 160-172.
• Deleu, S., V. Van Lancker, D. Van den Eynde, and G. Moerkerke, 2004. Morphodynamic evolution of the kink of an offshore tidal sandbank: the Westhinder Bank (Southern North Sea). Continental Shelf Research, 24 (15), 1587-1610.
• Iampietro, P., and R. Kvitek, 2002. Quantitative seafloor habitat classification using GIS terrain analysis: Effects of data density, resolution, and scale. In Proceedings of the 22nd Annual ESRI User Conference. San Diego, CA, July 8-12.
http://gis.esri.com/library/userconf/proc02
• Lundblad, E., D. J. Wright, J. Miller, E. M. Larkin, R. Rinehart, S. M. Anderson, T. Battista, D. F. Naar, and B. T. Donahue, in press, A Benthic Terrain Classification Scheme for American Samoa. Submitted to Marine Geodesy.
• Weiss, A. D. 2001. Topographic Positions and Landforms Analysis (Conference Poster). ESRI International User Conference. San Diego, CA, July 9-13.
• Wright, D.J., Lundblad, E.R., Larkin, E.M., Rinehart, R.W., Oregon State University Davey Jones Locker Seafloor Mapping/Marine GIS Lab (project background at http://dusk.geo.orst.edu/esri04/p1433_ron.html) Joshua Murphy, Lori Cary-Kothera, Kyle Draganov NOAA Coastal
Services Center