Mapping intertidal vegetation in the wash estuary using remote sensing techniques

Abstract

The mapping and monitoring of the intertidal zone of the East Coast of England is of considerable interest to conservationists and coastal managers. Intertidal vegetation offers natural protection against coastal erosion and considerably reduces the cost of man made sea defences. Monitoring of intertidal vegetation may also be of value in providing an early warning of sea-level change. This thesis considers the most effective way of classifying and monitoring the intertidal zone using remote sensing techniques and incorporating the results into a coastal monitoring Geographic Information System (GIS). The coastal monitoring GIS is used to model the advantages and disadvantages of different classification strategies. The Wash Estuary forms the principal study area. This study uses multi-temporal data which requires atmospheric and radiometric correction. All Landsat 5 TM images used in the present study were referenced to a common image, based on the techniques of F.G.Hall of NASA, to allow meaningful comparisons to be made. The data sets were geometrically corrected to allow incorporation into the coastal monitoring GIS. Two strategies were used to classify the intertidal zone: a conventional maximum likelihood classifier and a fully constrained mixture model using the least squares technique. The results of the classifications were incorporated into the coastal monitoring GIS along with information acquired from previous ground based surveys. Statistical analysis of the classifications was carried out by cross tabulation with the ground based surveys in order to determine the accuracy of the methods. Detailed, reliable information arrived at cheaply, objectively and at regular intervals would provide a valuable resource for the management and monitoring of the coastal environment.