Year

2011

Degree Name

Doctor of Philosophy

Department

School of Earth and Environmental Sciences

Abstract

The vulnerability of a coast is related to its geomorphic condition, slope status, past shoreline change, rate of relative sea-level rise (SLR), tidal range and exposure to waves. A Coastal Vulnerability Index (CVI) was developed by Gornitz and Kanciruk in 1989 which has been adopted and modified by subsequent researchers and organizations. The coast of Bangladesh is considered to be one of the most vulnerable in the world, but a thorough assessment of its vulnerability has not been undertaken previously. This study has developed a vulnerability index for the entire coast of the country based on these six physical variables. A total of 957 cells of the coast (excluding river mouths) have been divided into five vulnerability classes for all of the variables except SLR. The SLR variable has been divided into four vulnerability classes.

The geomorphology variable has been mapped from satellite imagery to classify the coast into categories such as mud flat, sandy beach, mangrove or cliff. Elevation obtained from SRTM, ASTER and GTOPO datasets has shown contradictory results, indicating overestimated elevation in dense forested areas, especially in the Sundarbans coastal zone. Therefore, slope has been used in preference. The shoreline of the coast has been detected using a band ratio approach to discriminate the water line on Landsat images over a 20- year period from 1989 to 2009. Rates of shoreline change have been calculated using the End Point Rate (EPR) method in the Digital Shoreline Analysis System (DSAS) extension in ArcGIS®. The coast has been observed to be very dynamic with recorded erosion of 285 m/yr and accretion of 633 m/yr. Sea-level rise has been calculated using tide-gauge data indicating high variability along the coast. However, relative sea-level change along the Bangladesh coast is still unclear. Similarly, tidal range obtained from secondary sources has been found to be variable throughout the coast, with the highest tidal range around the Feni River estuary, including Sandwip Island. Although most of the studies have considered wave heights as a variable in the CVI calculation, this study has considered surge height because of the frequent cyclones hitting the coast.

Coastal vulnerability indices for the Bangladesh coast have been calculated using four formulae: i) product mean ii) average sum of squares, iii) square means of product mean, and iv) sum of product methods. Additionally, the fourth formula has been modified to explore a more reliable vulnerability measure for the coast. Different formulae have generated similar patterns of vulnerability along the coast. In general, this assessment has identified the Meghna River estuary area, including the northern part of Hatiya Island, as very highly vulnerable. Shorelines along the Bhola coastal zone, part of the Barguna Patuakhali coastal zone, and parts of Sandwip and Hatiya Islands have been indicated as highly vulnerable. Variable vulnerabilities have been observed along the Sundarbans and part of Barguna Patuakhali coast. The vulnerability of the southeast coast is very low along the Cox’s Bazar coastal zone and very low, low or moderate along the Chittagong coastal zone.

The physical susceptibility of the coast is only one aspect that needs to be integrated with population for a fuller assessment of vulnerability. Coastal population data at a spatial resolution to fit into this study is not available. Therefore, large scale preliminary population data has been integrated into the CVI to discuss a coastal community vulnerability index (CCVI). Instead of identifying the length of shoreline in different vulnerability classes, this study has indicated a zone-based distribution of vulnerability indicating the most or the least vulnerable parts of the coast. This CVI has been undertaken at national scale with a focus on six coastal zones that requires follow-up local scale studies.

Share

COinS
 

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.