Diapositiva 1

1
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
http//www.ual.es/GruposInv/ProyectoCostas/index.h
tm
Preliminary Results on High Accuracy Estimation
of Shoreline Change Rate Based on Coastal
Elevation Models
F.J. Aguilara, I. Fernándeza, J.L. Pérezb, A.
Lópeza, M.A. Aguilara A. Mozasb, J. Cardenalb
a Dept. of Agricultural Engineering, Almería
University, Spain b Dept. of Cartographic
Engineering, Geodesy and Photogrammetry, Jaén
University, Spain
Corresponding Author F.J. Aguilar
(faguilar_at_ual.es)
Kyoto, Japan. 10 August 2010
2
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Detailed coastal topographic information will be
the key variable in understanding the likely
impacts of global anthropogenic and natural
hazards (including SLR due to Climate Change).
The shoreline is one of the most important and
critical indicators of coastal evolution and
vulnerability
Kyoto, Japan. 10 August 2010
1
3
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Urban development on the coastal area and
resource use conflicts spawn environmental
degradation and increasing hazard vulnerability.
In Spain, more than 44 population are now living
in coastal areas (7 of territory)
Anthropic Pressure

• Increase of urban areas (soil sealing and change
  of ISA)
• Shortage of planning
• Sand extraction to attend the demand from
  greenhouse crops

Kyoto, Japan. 10 August 2010
2
4
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Yes but, How to extract the shoreline? Some of
the methods are based on the intersection between
the Coastal Elevation Model (CEM) and the plane
corresponding to the chosen tidal datum
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
An open coast tide station very close to our
working coastal area is needed to accurately
estimate its MHW
Sometimes it is recommendable to use as a more
reliable vertical reference the Mean Sea Level
(MSL) tidal datum
Kyoto, Japan. 10 August 2010
3
5
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
1. A classic approach the Cross-Shore Profile
method (CSP) (e.g. Stockdon et al., 2002)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
R2 0.98
Stockdon, H.F., Sallenger, A.H., List, J.H.,
Holman, R.A., 2002. Estimation of shoreline
position and change using airborne topographic
Lidar data. Journal of Coastal Research, 18(3),
pp. 502-513.
Kyoto, Japan. 10 August 2010
4
6
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
1. A classic approach the Cross-Shore Profile
method (CSP)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
But most of times the results are not as we wish
R2 0.33
Kyoto, Japan. 10 August 2010
5
7
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
2. Our approach the Elevation Gradient Trend
Propagation method (EGTP)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Kyoto, Japan. 10 August 2010
6
8
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
2. Our approach the Elevation Gradient Trend
Propagation method (EGTP)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Kyoto, Japan. 10 August 2010
7
9
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Kyoto, Japan. 10 August 2010
8
10
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
ALMANZORA RIVER MOUTH
Main wave direction East-West
Average wave height 1 m
Maximum wave height 5 m
Maximum tide range (microtidal coast) 0.5 m
Kyoto, Japan. 10 August 2010
9
11
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Data corresponding to 2001 come from an analogic
RGB photogrammetric flight at an approximated
scale of 15000 taken on 9 April 2001. 1 m
grid-spacing CEM was carried out by means of
stereo matching techniques ranking over
previously digitized images and subsequent
exhaustive and careful edition by one operator
(SOCET SET environment). The estimated vertical
accuracy of the photogrammetrically-derived CEM
was around 30 cm
Kyoto, Japan. 10 August 2010
10
12
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Data corresponding to 2009 come from a combined
flight
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
General information Height above ground 1000
m Number of strips 4 Number of photographs
86 Digital camera DMC (Digital Mapping Camera)
Intergraph GSD (cm) 10 RGBNir (12 bits) Forward
overlap () 65 Side overlap () 60 LiDAR
sensor ALS60 LEICA FOV (º) 35 Max. laser pulse
Rate (Hz) 96100 Max. point spacing across
track(m) 1,33 Max. point spacing along track (m)
1,46 Average point density (points/m2)
1,61 Average point space (m) 0,79 Average point
area (m2) 0,62 Estimated height accuracy (m) 0,08
Kyoto, Japan. 10 August 2010
11
13
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Methodology to compute the corresponding
shoreline change rate between 2001 and 2009
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
End Point Rate (EPR) Computation (m/year) erosion
(-) or accretion ()
95 confidence level
Kyoto, Japan. 10 August 2010
12
14
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Performance of the Two Tested Shoreline Mapping
Methods
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
CSP method seems to be quite sensitive to noise
due to an incorrect separation between water and
land, unexpected artifacts along foreshore
profile or actually non-straight profiles. EGTP
can be deemed as much more robust than CSP,
maintaining a greater number of useful transects
Year Transects lost Transects lost Uncertainty (sxs) Uncertainty (sxs)
CSP EGTP CSP EGTP
2001 12.84 3.19 2.95 4.10 m
2009 12.18 1.54 1.05 m 1.48 m
Kyoto, Japan. 10 August 2010
13
15
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Quantitative Analysis (ANOVA where EPR is the
explained variable)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Source of variation Degrees of freedom Sum of squares F Signific. (plt0.05)
Method (A) 1 0.659 0.853 0.355
Tidal datum (B) 1 0.037 0.048 0.826
Transect spacing (C) 2 1.073 0.694 0.499
Zone (D) 28 33809.29 1562.34 lt0.001
AB 1 0.134 0.174 0.676
AC 2 0.834 0.539 0.582
BC 2 0.085 0.055 0.946
ABC 2 0.0292 0.037 0.962
AD 25 66.756 3.455 lt0.001
BD 28 49.099 2.268 lt0.001
ABD 25 9.991 0.517 0.977
CD 56 41.945 0.969 0.540
ACD 47 5.744 0.158 1
BCD 56 1.802 0.041 1
ABCD 47 1.003 0.0.027 1
Error 12412 9592.72
A (method) CSP and EGTP B (tidal datum) MSL
and MHW C (transect spacing) 5 m, 10 m and 20
m D (Zone) 29 homogeneous areas
Kyoto, Japan. 10 August 2010
14
16
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
Zone EPR (m/year) Useful transects EPR (m/year) EPR (m/year) Useful transects
Zone Method CSP Method CSP Method EGTP Method EGTP Method EGTP
1 0 3 -0.95 16 16
2 ----- ---- -0.03 9 9
3 3.77 9 2.91 10 10
4 0 1 -0.19 9 9
5 -0.40 17 -0.34 15 15
6 0 14 0 15 15
7 0.28 8 0.09 8 8
8 ---- ---- 0.27 3 3
9 -0.10 15 -0.28 21 21
10 2.50 8 2.39 14 14
11 ---- ---- 0 5 5
12 ---- ---- -3.97 4 4
13 3.46 4 3.85 24 24
14 0.30 13 0.12 18 18
15 1.72 2 1.46 7 7
16 -0.09 16 -0.13 29 29
17 -1.07 20 -1.01 26 26
18 3.76 94 3.64 122 122
19 2.84 52 2.92 56 56
20 -2.16 82 -2.12 82 82
21 0 23 0 23 23
22 2.02 96 1.90 84 84
23 0.27 24 0.30 20 20
24 -0.58 102 -0.30 102 102
25 -0.07 58 -0.16 60 60
26 -1.38 43 -1.10 42 42
27 2.05 25 1.98 29 29
28 -2.17 75 -2.17 59 59
29 0 47 -0.04 47 47
Average 0.39 851 0.55 959 959
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Kyoto, Japan. 10 August 2010
15
17
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
ORTHOPHOTO 2009
2009
2001
1989
PUNTA DE LOS HORNICOS
Kyoto, Japan. 10 August 2010
16
18
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
PALOMARES BEACH
1989
2009
2001
VERA BEACH
Kyoto, Japan. 10 August 2010
17
19
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
ACCRETION
2009
1989
2001
VERA BEACH
EROSION
Kyoto, Japan. 10 August 2010
18
20
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
QUITAPELLEJOS BEACH
Kyoto, Japan. 10 August 2010
19
21
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS

• The new grid-based approach can be strongly
  recommended because its precision, local slope
  acquisition, robustness regarding the presence of
  noise and outliers, and capability to deal with
  very curved and even closed coastal features.
• The preliminary results also indicate that,
  though the global rate-of-change for the whole
  coastline between 2001 and 2009 may be catalogued
  as relatively low (0.55 0.50 m/year of net
  accretion), the local results for every one of
  the 29 homogeneous units considered have been
  extremely variable and statistically significant.
• Many local phenomena, registered in a short-term
  period and mainly due to human activities such as
  the presence of new engineered structures and
  artificial beach regeneration, may strongly
  affect the shoreline evolution in certain and
  localized areas

Kyoto, Japan. 10 August 2010
20
22
Research Project RNM 3575 Multisource Geospatial
Data Integration and Mining for the Monitoring
and Modelling of Coastal Areas Evolution and
Vulnerability
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Thank you very much for your kind attention
Kyoto, Japan. 10 August 2010
21
23

• 2009 LiDAR data. Vertical Accuracy

Average dz 0.029 Minimum dz -0.284 Maximum dz
0.180 Std deviation 0.089 N 62
24

• 2009 LiDAR data. Vertical Accuracy

25
MDE
ez
elc
MSL
Ground truth
tga
HIGH WATER LINE
26
1. A classic approach the Cross-Shore Profile
method (CSP)
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Theoretical framework to estimate the point
shoreline error (xs) propagated from CSP method
FUNCTIONAL MODEL
LEAST SQUARES ESTIMATION
Kyoto, Japan. 10 August 2010
27
2. Our approach the Elevation Gradient Trend
Propagation method (EGTP) Computed uncertainty
for the shoreline determination
INTRODUCTION
SHORELINE EXTRACTION METHODS
STUDY SITE, DATASETS METHODOLOGY
RESULTS DISCUSSION
CONCLUSIONS
Kyoto, Japan. 10 August 2010
9