GISdevelopment.net ---> AARS ---> ACRS 1992 ---> Poster Session P Management of Wetlands – A Remote Sensing Approach S. Sudhakar, V. Jayasree, M. Prithviraj Regional Remote Sensing Service Centre, IIT Campus Kharagpur A. K. Raha State forest department, Calcutta Abstract Sunderbans, one of the world’s major mangrove ecosystem needs proper planning and timely execution of conservational measures to control further degradation. Satellite Remote Sensing, with its repeatability and synoptic coverage in separate spectral bands, has opened new vistas in the survey and monitoring of the inaccessible areas. Realizing the importance of data base on the condition and areal extent of mangroves for a scientific forest management strategy, an attempt has been made in this work, to segregate three density classes within complex mangrove ecosystem in the district 24 Paraganas (South), West Bengal using IRS-1A, LISS-II data pertaining to November 1988. The overall classification accuracy was more than 85 percent. Introduction The terms “Wetland” is quite often used to cover wide range of aquatic habitats like marshes, swamps and bogs. It is defined as “the lands transitional between terrestrial and aquatic systems where the water table is usually at or nears the surface or the land is covered by shallow water” (Cowardian et. al. 1979). The mangrove, a distinct type of vegetation dispersed on relief lying under constant influence of tidal and fresh water, constitute a part of wetland ecosystem. The mangroves comprises of physiologically specialized and ecologically adapted plants to survive in salt / blackish water conditions. Most of the species are characterized by high osmotic and suction pressure in the leaves, vivipary and presence of specialized roots of stilt/pneumatophers with buttress nature of stem to withstand the adverse conditions of the costal wetland environment. The general distribution of mangroves is restricted to the tropical and temperature zones of the world and their exuberant growth and development requires primarily the following conditions: Air temperature near about 200C during coldest months with the seasonal temperature range not exceeding 4°C; Shallow shores free form strong wage and tidal action; A wide horizontal tidal range with substratum containing fine grained alluvium. India has got an extensive coverage of total wetlands where characteristic vegetation – the mangroves, grow. They cover an area of 6740 sq. kms, which is about 7 percent of the world’s mangrove population. About three mangrove population. About three fourth occur in Sundarbans, in the 24 Paraganas (south) district of West Bengal and the remaining major patch is found in the Andaman and Nicobar group of islands (Mang Sanjoy, 1991; Thothathri, 1981). Mangrove being an important component of the complex wetland ecosystem needs to be protected and conserved. Therefore. Periodic monitoring of this coastal environment is of paramount importance to obtain the knowledge about their areal extent and condition. The acquisition of data pertaining to these changes in the wetland condition through conventional means is costly, time consuming and difficult due to inaccessibility. But with aid of satellite Remote Sensing methods, having the advantages like synoptic view and repetitive coverage, the task of regular monitoring of this coastal environment has become simpler. In this paper the successful accomplishment of segregation of the different density classes of the canopy cover of the Sunderbans by following the digital classification scheme (maximum likelihood algorithm) in VAX-11/780 computer with VIPS – 32 Software on the IRS- 1A LISS-II data ahs been highlighted. Study Area The district is, situated between latitudes 21° 30’N – 22°37’ N and longitudes 88° 03’ E – 89° 07’E. It is bounded on the north by Calcutta and 24 Parganas (North) districts, on the east by 24 Parganas (North) district and Khulna district of Bangladesh, on the south by the Bay of Bengal, and on the west by the river Hugli separating it from the districts of Medinapur (East) and Haora (fig. 1). Data Used IRS-1A, LISS-II data path/row 18-52 A1, A2, B1, B2; 17 – 52 A2; 18 – 53 A1, B1 and 17 – 53 A1 pertaining to November 1988 were used for the study. Survey of India topographical map nos. 79B, 79C, 79F and 79G on 1:250,000 scale an aerial photographs on 1:50,000 scale were also used as collaberal data. Working plan maps pertaining to sunderbans were consulted wherever necessary. Methodology The different scenes understudy were first merged on their paths and geometrically corrected with respect to the topographical maps. Standard False Colour Composites (std. FCC) were generated by assigning red, green and blue colour to bands 4, 3 and 2 respectively. Different stretching were applied to highlights the desired information for easy identification of earth features. Based on tone/texture along with the associated features, various “ground truth” points were identified and marked on the Std. FCC. In view of inaccessibility to enter the vegetated swampy zones of Sunerbans, emphasis was given to the fringes of the mangrove Islands during the ground truth collection. Aerial photographs were also consulted wherever necessary while assigning training areas for classification. Representative areas in Sunderbans were subjected to Normalized Difference Vegetation (NDVI) and Tasseled Cap Transformation based on Kaunth – Thomas algorithms to understand the density classes within the complex mangroves vegetation. The total scene was then assigned different training sets based on the gathered ground truth information and by consulting outputs from Normalized Differences Vegetation Index, Kaunth and Thomas algorithm and aerial photographs. To avoid misclassification, the total scene was classified twice, once, on the major area pertaining to mangrove zone of the Sunderbans and secondly on the area without was done using maximum likelihood algorithm. After classification, respective rows of the scenes were mosaiced to represent the entire area understudy. The district, road, rails range and Sunderban Tiger Project area boundaries were digitized District mask file and the remaining boundaries as overlaying file were generated. Using the district mask file the district was extracted and later other boundaries were superimposed from the overlaying files. The various steps involved in the present study is given in Fig. 2. Statistical data on different mangrove density classes along with other land use / land cover classes were also generated (Table 1). Results and Discussion Sunderbans, perhaps the largest block of mangrove in the world has got nearly 20 species of flora. However, only two species namely, Sundri (Heritiera formes) and Gewa (Excocaria agallocha) constitute the major part in the mangrove forests of Sunderbans. The mangrove vegetation in Sunderbans can broadly be divided into tow main soil-vegetation types. They are Euestuarine and Prestuarine. The later can be further divided into three sub-types viz. Tidal Mangroves, Prohaline and Euhaline. (Ananda Rao & Sastry, 1982). The euestuarine type in Sunderbans is composed of gregarious growth of Nypa fruticans and Phoenix paludosa. Often, they form pure strands along the elevated fringes and drier border lands of protected tidal waves. The tidal mangroves zone where the salinity is more or less same as that of open sea is found to be covered with a single species – Avicenna marina. But with increase of elevation and consequent sea-water there is formation of mixed nature of vegetation mainly of Sonneratia, Bruguiera, Aegialitis and often Excoecaria species. Due to complexity and mixed nature of different mangrove species, it is rather difficult to segregate on species basis until and unless one is fully conversant with the phonological conditions of different species. Efforts are on the make use of temporal.