Coastal Erosion Modeling using Remotely Sensed Data
Maged m.m. and s.b. Mansor
Faculty of Engineering
University Putra Malaysia
43400 Serdang, Selangor
Fax : + 6-03-948 8939
shattri@eng.upm.edu.my
Abstract
Coastal erosion is caused by the input of the wave energy to the coastal zone. This wave energy could sedimentation or erosion along the coastline. One of the areas reported to experience this problem is the coastline of Kuala Terengganu. The aim of this study is to develop system to predict the coastal erosion. Mathematical model was carried out to investigate and predict the erosion. This model utilized several types of data, including remotely sensed data, ship observations and ground truth data. In this project, aerial photos and airborne AIRSAR/TOPSAR data were utilized to predict the coastal erosion. The main parameter extracted from the remotely sensed data were the remotely sensed data is the wave spectra density along the shoreling of Kuala Terengganu.
The results show that the shoreline of Kuala Terengganu is exposed to sedimentation and erosion at various locations. The probability distribution function model has been adapted to predict the shoreline changed utilizing ship observation data from 1971-1983. By integrating the wave spectra pattern derived from TOPSAR data with the wave spectra change derived from ship observation, the sedimentation and erosion in coastal water of Kuala Terengganu could be predicted. Statistical analysis shows that shoreline changes are in Equilibrium State with nature. The changes are usually temporary and would normally recover back to its original state.
The study indicates that coastal erosion occurred seasonally during the northeast monsoon period due to wave actions. However, the sedimentation was occurred during south-west monsoon period and during internal monsoon period.
Introduction
The applications of remote sensing data for coastal processes has a more interest between researchers. Most of the works used the historical data of aerial photo and satellite imageries to detect shoreline changes (Mazalan et al., 1989; frithy et al., 1995). The interest subject has been taken into account between researchers, which is the coastal sedimentation and erosion. A using remote sensing techniques to study coastal erosion will
produce a wide concept for erosion problem. Because of the fact that remote sensing techniques able to figure out the interaction between sea and shoreline change. Furthermore, remote sensing techniques able to cover a large area over than 300 km. Recently radar imageries such as ERS-1 show ability for coastal studies. Ibrahim and Samsudin 1996; Maged et al., 1998 have done such studies. Most of studies conducted for coastal erosion problem either by classic methods (Lukman et al., 1995) or remote sensing data (Raj, 1982; Mazalan et al., 1989; Maged et al, 1997) could not predict coastal erosion over long time. The aim of this study is to predict the coastal erosion especially at long term. This will be done over one hundreds years period.
Methodology
Study Area
One of the area reported to experience rapid erosion is the coastline of Kula Terenggarnu. This area is exposed to highest wave during the north-east monsoon compare to south-west and transitional period (Wong 1981). The study area is located in the South China Sea between 5
° 21'N to 5° 21'N and 103° 10' E. The highest rate of shoreline change reported in the study area was 30m/yr in the study of Mazlan (1989).
Wave Spectra Model
Wave spectra derived from the C-band AIRSAR by applying two dimensional Fourier Transform from PCI EASI/PACE image -processing system this. The wave spectra derived from C-band AIRSAR were map into the real wave spectra by using quasi-linear model.
S(Q) = H(Kx; Kc) S(L) S (k) (1)
Whee S (Q) is a quasi-linear transform function, Kx is wave number azimuth direction; Kc is the cut-off wavenumber, which function of wind speed. S(k) is AIRSAR wave spectra while S(L) is real wave spectra measured in situ.
Mathematical Model of shoreline Change
Mathematical model of shoreline changes is based on the rate changes of the sediment volume. Mathematical model utilized several types of data, including AIRSAR wave spectra; quasi-linear wave spectra; average ship observation data (1971-1983) and ground truth. This has been done by using ACESS 1.07 software.
Statistical Model
In order to predict the shoreline changes over one hundreds years or more, the probability distribution function of the different sources data was used. This was done under the following assumptions:
-
Human activities was neglected
- The time interval between erosion and sedimentation events begin equaled or exceeded
- The shoreline change occurred seasonally.
- The shoreline change rate for different period are estimated from the following relation:
R = Ay + B (2)
Where y is given by Weibull distrubiton
Y = {1n (S T)}1/k
Where S is number of erosion and sedimentation events per year, T is return period (years) and k is the length of records by years. The probability of erosion and sedimentation occurrence could be expressed as a percentage change of occurrence. This could be given by the following
P = 100[1-(1-1/t)L] (3)
Where p is the chance of erosion and sedimentation occurrences and L is time. The model has been used to estimate return period of the wave percentage occurrence. In its study we have used the above model to predict shoreline change over one hundreds year.
