MANGROVE VEGETATION ASSESSMENT IN THE SANTIAGO RIVER MOUTH, MEXICO, BY MEANS OF SUPERVISED CLASSIFICATION USING LANDSAT TM IMAGERY

This paper presents a mangrove vegetation assessment from 1970 to 1993 of the Santiago River Mouth, Nayarit, West of Mexico. The aims of thi s work are to describe the plant composition and structure of mangrove in the study area, and to evaluate the deforestation level and its am plitude by means of a retrospective analysis of the cover and distribu tion area of mangrove species using a LandsatTM image, aerial photogra phs and oblique video. Mangrove of the study area is dominated by Lagu ncularia racemosa with the average importance value of 158.18 and 400 ha of plant cover, followed by Avicennia germinans, with an average im portance value of 138.52 and 324 ha of plant cover. Mangrove showed se ven height and five diametrical classes that include the two dominant species. L. racemosa was the dominant species in six of the eight comp ass lines. The highest absolute frequencies for both dominant species were found in the second height class frequency, and the first diametr ic class frequency. Cover area and distribution of mangrove in the stu dy area were mapped using a LandsatTM5 image(April 1993). A supervised classification was applied using the maximum likelihood algorithm, co nsidering ten initial classes. This classification was evaluated by ob taining a classification error matrix and by assessing its accuracy. T he mangrove vegetation area reported before, considering the same area for image analysis, resulted to be overestimated in 56% regarding the value obtained in our photointerpretation (1065 ha). From the latter mangrove area, the current cover is 724 ha, which represents a decreas e of 32% in a 23-yr period. (C) 1998 Elsevier Science B.V.