NL FR EN
www.belgium.be

Monitoring of Coral Reef in View of Sustainable Island Development; Case Study using Hyperspectral Remote Sensing: Fordata, Tanimbar, Indonesia (KABAR)

Projet de recherche S0/00/60b (Action de recherche S0)

Personnes :

  • Prof. dr.  GOOSSENS Rudi - Universiteit Gent (UGent)
    Partenaire financé belge
    Durée: 1/3/2005-31/12/2006
  • Dhr.  DEBRUYN Walter - Vlaamse Instelling voor Technologisch Onderzoek (VITO)
    Partenaire financé belge
    Durée: 1/3/2005-31/12/2006

Description :

Context and objectives

The biologically diverse coral reefs surrounding Pulau Fordata (Tanimbar Archipelago, Indonesia) support an important part of the livelihoods of the local people. The unsustainable use of this natural resource, however, has led to its deterioration, especially around the main island. The general objective of this project, therefore, is the development of an efficient monitoring system for coral reef ecosystems based on hyperspectral remote sensing information in order to enhance the sustainable development of small island communities like on Pulau Fordata. More specifically, it aims to assess the condition of the coral reef benthic cover in the Pulau Fordata region using CASI hyperspectral data, to create a bathymetric model of the reef structures, and to develop generic data processing techniques and algorithms which can be applied to coral reefs worldwide.

Methodology

• A bathymetric model has been derived by applying the (semi-)analytical model by Lee et al. (1998, 1999) to the hyperspectral CASI dataset in an iterative optimisation set-up implemented in ENVI IDL.
• A new algorithm to classify coral reef benthic cover without prior water column correction has been developed consisting of two separate steps. First, a geomorphological classification has been made so that different benthic cover types with nearly identical reflectance spectra but located on different reef parts, and thus depth, can be segregated. Consequently, each geomorphological unit is separately classified using an algorithm encompassing a semi-unsupervised hierarchical clustering followed by class aggregation based on spatial and spectral analyses and class labelling using ground truth data.

Lee, Z., Carder, K.L., Mobley, C.D., Steward, R.G. and Patch, J.S., 1998. Hyperspectral remote sensing for shallow waters. 1. A semianalytical model. Applied Optics, 37: 6329-6338.
Lee, Z., Carder, K.L., Mobley, C.D., Steward, R.G. and Patch, J.S., 1999., Hyperspectral remote sensing for shallow waters. 2. Deriving bottom depths and water properties by optimization. Applied Optics, 38: 3831-3843.


Results

• bathymetric map of the Pulau Nukaha reef system (East of Pulau Fordata): visual analysis indicated a reasonably good representation (mean residual error of -0.9m) of the topographic structure of the reef system up to a depth of 35m;
• benthic cover map of the Pulau Nukaha reef system: a detailed view of the distribution of the benthic cover types in the study area has been obtained, discriminating 11 coral and 3 algae classes;
• algorithms have been developed and/or optimised to support hyperspectral coral reef mapping, including: bore sight calibration based on inherent image information, sunglint correction, bathymetric mapping based on radiative transfer modelling, and benthic cover classification without the need of water column correction;
• synthesis workflow describing the most important steps and procedures to be taken when deriving bathymetric and benthic cover information on coral reefs using hyperspectral remote sensing data; tailored for the use of radiative transfer optimisation approaches.

Documentation :