Environment, Succession and Land-Use Potential in the Semi-Evergreen Forests of Southern Sri Lanka

Abstract

Relationships between mature forest, successional stands and habitat variation in the dry zone of southern Sri Lanka are investigated by multivariate methods, including an attempt to use vegetation and habitat data to predict yield for teak and coconut. A common sampling frame is used for ecological and yield prediction studies. Crop sites are matched by geomorphological variables with mature forest samples and, if available, one or more stands of successional vegetation regenerating after the abandonment of chena cultivation. Habitat data are simplified by principal components analysis to remove collinearity and component scores are used to seek relationships with vegetation data simplified by ordination and classification. Stepwise multiple regression of component scores on axis coordinates from ordinations provides hypotheses of additive habitat relationships. Floristic patterns on ordination axes are summarized as seriation arrays usingX2 tests to isolate species restricted to axis segments. Successional floristic patterns, with parallel diversity and structural trends, are dominant in vegetation data, including gap regeneration in mature forest samples. Separate soil moisture (involving topography and textural effects) and soil type trends are also interpretable in mature forest vegetation. Successional samples also contain regional patterns which might reflect rainfall differences. Trend surface analysis, seriation and yield prediction models are used to relate crop yield to vegetation data. Regression is used to correlate habitat data with crop yield. However, no strong relationship with crop yield variables is found. Weak prediction of teak volume increment per tree is possible from vegetation data, and both teak and coconut yields increase weakly with increasingly fine soil texture. Further understanding of dry zone vegetation in Sri Lanka, including its conservation value and indicator potential, is currently threatened by widespread and unplanned forest clearance.