Spatial patterns of soil phytoliths in a wet vs. dry neotropical forest: Implications for paleoecology

Abstract

To improve paleoenvironmental reconstruction based on fossil phytoliths, it is vital to establish modern calibrations that explicitly consider important functional aspects of vegetation, such as structure, composition, diversity, and spatial heterogeneity. No such studies currently exist for Central- and South America. To begin to solve this problem, we address two questions: 1) do phytolith assemblages reflect the overall differences in structure, composition, and diversity between rainforest and dry forest? and 2) do phytolith assemblages from lateral transects capture variation in structure and composition typical of these vegetation types? At La Selva (rainforest) and Palo Verde (dry forest) biological stations (Costa Rica), we collected soil phytolith samples from, respectively, fourteen and eleven quadrats along vegetation transects. Along each transect, we measured gradients in canopy openness (LAI), elevation, and moisture. Statistical analyses were used to characterize and compare vegetation structure, composition, and diversity to soil phytolith assemblage composition, and to test for correlation with environmental gradients at each site. The two sites were then compared using multivariate statistical methods. Results indicate that rainforest and dry forest vegetation can be distinguished by analyzing multiple phytolith samples collected within sites. Within each vegetation type, the comparison of phytolith assemblages along transects allows reconstruction of important structural (canopy openness and patchiness) and compositional (dominant plant functional types) aspects of the vegetation. Thus, the analysis of phytolith samples collected along temporally constrained stratigraphic levels is a viable and vital tool for inferring heterogeneity in canopy cover and gross vegetation composition in the deep-time fossil record.