Monitoring of Caspian Sea Coastline Changes Using Object-Oriented Techniques

Abstract

Coastal zone monitoring is an important task in national development and environmental protection, in which, extraction of shorelines should be regarded a fundamental research of necessity. Very dynamic coastlines such as Caspian Sea coasts pose considerable hazards to human use and future development. Therefore several rapid reliable techniques are required to monitor and update coastline maps of these areas to explore rates of environmental retreats. In the current study, various semi-automated methods like NDWI, NDSI and Tasseled Cap have been accordingly applied and results were integrated with some objectoriented classification methods. Landsat MSS, TM and ETM imageries of the past three decades were consequently processed by an object-oriented approach performed with an eCognition software package. By comparing three classified maps of the south Caspian Sea coasts (Babolsar Port to Feridonkenar) in 1977, 1984 and 2002 with a unique region growing image segmentation technique (multi-resolution segmentation), areas of rapid change were progressively identified. Revealed models demonstrate that several yearly persuaded fluctuations and considerable periodical changes on the study area coastlines particularly during the last decade observed by TOPEX/Jason satellites. These great variations have occurred as the result of 2.6 meters increases in height of water in the sea from 1984 to 1995. This has successively caused a diminishing of coastal lands, about 185 km2 mainly on the Babolsar Port, changing landcover and landuse types by depletion of significant agricultural and residential areas. Implementations of such significant changes signify that the majority of local biotic and biotic components, all over the surrounding areas, would be in crucial threat in the near future.