PJB-2006-63
SPATIAL DISTRIBUTION OF ECOLOGICAL COMMUNITIES USING REMOTELY SENSED DATA
RIFFAT NASEEM MALIK* AND SYED ZAHOOR HUSAIN**
Abstract
In Pakistan in spite of few attempts for mapping land-cover types, satellite remotely sensed data has not been used extensively; and its potential is not being explored for providing information on mapping vegetation cover in general and ecological communities in particular. In this study, we used SPOT(Systeme Pour l'Observation de la Terre) multispectral (XS) satellite sensor data in visible and near infrared portion of the light spectrum as a surrogate for distribution of ecological vegetation groups defined by the classification and ordination methods (the most commonly used multivariate techniques used in floristic composition classification in vegetation ecology) and non-vegetation classes. The results indicate that classification of vegetation groups based on species composition identified using classification and ordination techniques to some extent resemble to those groups classified using SPOT XS data with least accuracy in comparison to non-vegetation classes which were more homogenous and spectrally separable and were classified more accurately in comparison. Two classification models i.e. supervised maximum likelihood and fuzzy supervised classification showed similar overall level of accuracies. The possibilities of lower classification accuracies and difficulties of classifying ecological communities based on the species composition using remotely sensor data are discussed.
To Cite this article:
Download
