Spatiotemporal Analysis of Rapidly Evolving Coastal Topography using Open Source GRASS GIS

1
Spatio-temporal Analysis of Rapidly Evolving
Coastal Topography using Open Source GRASS GIS
Helena Mitasova Department
of Marine, Earth and Atmospheric Sciences, NCSU,
Raleigh, T. Drake, MEAS NCSU, R. Harmon, US Army
Res. Office, J. McNinch Virginia Inst. of Marine
Science, D. Bernstein CMWS, Coastal Carolina
Univ., C. Freeman, UNC-CH http//www.skagit.meas.n
csu.edu/helena
2
Goal Explore the possibilities to gain new
insight into short-term coastal topography
evolution by combining modern mapping
technologies with Open source GRASS GIS and
provide methodology and valuable information for
coastal management. Study areas - Jockey's
Ridge sand dune migration - Bald Head Island pre
and post-nourishment beach evolution - Bogue
Island
3
Mapping technologies
Beach topography Real Time Kinematic GPS
Nearshore bathymetry multibeam and conventional
sonar
Coastal topography LIDAR USGS/NOAA ATM 1996-2000
USACE FRF, Duck, NC
www.csc.noaa.gov/crs/tcm/about_ldart.html
4
Modern digital coastal data and GIS

• Challenges
• massive data sets, oversampling,
• noisy data and complex surfaces
• anisotropy and heterogeneous coverage
• almost all data are spatio-temporal
• Robust and flexible gridding technique is
  crucial
• tunable spatial approximation such as RST
• (Regularized Spline with Smoothing and Tension)

H. Mitasova
5
Jockeys Ridge evolution
Combining 98 IR-DOQQ, 99 LIDAR and 02 RTK GPS to
assess the change decreasing elevation, migration
C
Height Year m 1950 43 1974 34 1995 27 1999
26 2002 24
A
B
N
H. Mitasova
6
Jockeys Ridge Migration
location horizontal
change vertical change elevation
direction
98-99 99-02 99-02 99
02 main dune peak -
- -2.1 25.7
23.6 main dune crest A NA 44.7
3.9 14.3 18.2
S main dune crest B 10 18.4
4.0 6.5 10.5
S south dune peak - -
-3.0 10.8 7.8 south
dune crest 7.6 21.4
4.7 1.4 6.1 SW south
dune fence C - -
1.7 3.6 5.3 south dune
fence D - -
2.8 7.2 10.0 east dune near peak
- - -1.7
15.3 13.6 east dune crest
16.2 25.0 3.0
1.7 3.7 SW

H. Mitasova
7
Dune crests movement based on1998(IR-DOQQ) -
1999(LIDAR) - 2002(RTKGPS)
A South dune B
East dune
Horizontal movement 40m W Change in
elevation - 2 m
Horizontal movement 40m SW 40m remains to
reach road!
D Central section
CMain ridge
Horizontal movement 40m SW Change in
elevation - 2.6 m! 25.9 gt 23.3m
Horizontal movement 40m SW Change in
elevation - 2.6 m! 25.9 gt 23.3m
Horizontal movement no crest Change in
elevation 3m
8
Jockeys Ridge 2000 - 2002
A
2000 2002
Minigolf castle
9
Bald Head Island
Human impact on evolution of shore topography and
nearshore bathymetry channel deepening,
nourishment Natural forces hurricanes
Cape Fear Rv
LIDAR 1997-2000 RTK GPS 2001-02
elevation m
Single beam sonar
2000, 2002 Multibeam sonar 2001
Integrated 10m resolution model from multiple
sources
10
Off-shore mound interferometric sonar
2m res. model interpolated by RST
Original gridded data 1.3x1.3km 65m swaths,
10-20m gaps
change elev.-0.5m slope-3deg volume8400m3
April 2001 Jan. 2002
Overlayed 0.5m res surfaces
Original gridded data 250x300
More results at skagit.meas.ncsu.edu/helena/measw
ork/mound/mound.html
11
Bald Head Island shore change
Recent change
Historical change 1850 - 1962
1998 LIDAR 0m 2000 LIDAR 0m
2001 Dec. RTK GPS after nourishment
after Cleary et al. 1989
800m
Shoreline rotates around pivot area
12
South Beach evolution 1997-2000
Overlayed 1997 and 2000 LIDAR surfaces central
section is relatively stable, rest erodes while
changing its shape
West Center East
convex -gt concave
stable pivot area
2000 zgt0m zlt0m
stable
concave -gt convex
13
Beach slope change 1997-2000
14
Slope and curvature change
Severely eroding area approximated and analyzed
by RST
1998
Slope
Profile curvature
concave convex
2000
15
South Beach change 1997-2000
Elevation change 1997-2000
Volume change loss 376,000 m3 gain 30,000 m3
m
loss gain
detail draped over 1997 surface
detail draped over 1997 surface
acceleration
Second order change 1997-1998-2000
H. Mitasova
16
RTK GPS 2001-2001
RTK GPS data are oversampled in the direction of
the vehicle movement. Anisotropic interpolation
is needed when distance between profiles is
significantly greater than resolution.
RST default parameters
RST with anisotropic tension
Impact on volume change estimate
120000m3/160000m3
H. Mitasova
17
Pre-nourishment 2000 LIDAR
LIDAR 2000
H. Mitasova
18
Change after nourishment
LIDAR 00 RTK GPS Dec. 01
H. Mitasova
19
Change after nourishment
LIDAR 00 RTK GPS May 02
H. Mitasova
20
Change after nourishment
LIDAR 00 RTK GPS Sep. 02
21
Change after nourishment
LIDAR 00 RTK GPS Dec 02
22
Change after nourishment profiles
LIDAR 2000 RTKS Dec. 2001 May 2002
September 2002
23
Change after nourishment
LIDAR 2000 September 2002 December 2002
24
Moved volumes after nourishment
Sep. 2002 - Dec. 2001
RTK GPS
Aug. 2000 Fall 1997 LIDAR
Sep. 2002-Aug. 2000 RTK GPS - LIDAR
H. Mitasova
25
Moved volumes
South Beach
Cape Fear
H. Mitasova
26
Cape Fear evolution
LIDAR 1997 1998
1999, grey is 2000
RTK GPS Dec 2001 May 2002
Sep 2002
27
Cape Fear elevation change
LIDAR 1997 RTK GPS Dec. 2001
Dec. 2002
1997-2000 2000-Dec. 2001
Dec. 2001-Dec. 2002
28
Zalewski June 2001
Bernstein December 2002
29
Bogue Island collaboration with Chris Freeman
and Dave Bernstein
Slope
Profile curvature
Sand bars
30
Bogue Island May - August 2002 change
31
Conclusions I
Combination of modern mapping techniques with
Open source GIS provides unique insight into 3D
coastal topography evolution at high spatial and
temporal resolution. GIS based analysis and
visualization allows us to quantify the observed
changes (elevation, shoreline, volume, slope and
shape) and evaluate effectiveness of
stabilization measures. The developed methodology
is being further enhanced and applied to other
areas.
H. Mitasova
32
Conclusions II
Jockey's Ridge LIDAR and RTK GPS data confirmed
that Jockeys Ridge is flattening and GIS
analysis allowed us to quantify the dune
migration and identify the critical areas Future
Continue monitoring, LIDAR (ATM or EAARL) would
help to assess the impact of fences.
Modeling? Compare with historical elevation and
land use data collaboration with Dr. Overton. 3D
subsurface data? Museum exhibition evolution of
Jockey's Ridge, explain why it is not the largest
dune anymore
H. Mitasova
33
Conclusions III
Bald Head Island Analysis based on LIDAR and RTK
GPS data showed systematic, spatially variable
erosion of the beach accompanied with beach shape
change. After renourishment the rates increased
in the west section and more beach become more
stable in the east. Future Analysis of the
entire area as a single system - bathymetry
(fate of eroded sand back to channel, CF shoal,
sandbars ?) - new LIDAR survey - modeling
(DELFT3D? MIKE?)
H. Mitasova
34
Acknowledgment
This project is funded by the NRC/ARO
fellowship. In addition to observations acquired
by co-authors, data from NOAA-USGS (LDART), USACE
FRF Duck NC (FRF web site) and UNC Wilmigton
(Anders et al. 1990, Clearly et al. 1989) were
used.
1997 1998
1999 2000
12/2001 05/2002
09/2002 1997-2002
Cape Fear change 1997 - 2002
H. Mitasova
35
Open source GIS GRASS
grass.itc.it
General purpose GIS for raster, vector, site and
image data processing Links to other OS
projects MapServer, GSTATS? Developed at US
Army CERL (1982-1995). General Public Licence
(GPL) in 1999 free to run, modify, distribute,
and release (but it cannot be improved and
released as a proprietary system)
36
Regularized Spline with Tension
RST (Mitasova and Mitas 1993) implemented in
GRASS to support spatial approximation. Simulates
behavior of flexible plate forced to pass close
to the data points.

• Properties
• 2D, 3D and 4D implementation,
• flexible properties through tension and
  smoothing,
• simultaneous computation of slope, aspect,
  curvatures,
• segmented processing for large data sets
• formally equivalent to universal kriging
  (covariance function determined by
  smoothness seminorm