Document: DIE-3-40-23

Forest types in Southern Yucatan, Mexico.

PEREZ-SALICRUP, D.R.* and D.R.FOSTER

Harvard Forest, Harvard University, Petersham, MA 01366 USA ¹

Abstract:
The Southern Yucatan Peninsula (SYP) contains the largest and most rapidly disappearing track of continuous tropical vegetation in Mexico. The region is covered by a mosaic of two major forest types, which can be readily distinguished in terms of their structural features (i.e.: canopy height, tree density, epiphyte abundance, etc.). We evaluated whether these two forest types could be recognized based on their tree species composition on a large geographical scale. We established a network of 110 permanent 500 m² circular plots, concentrated in six regions within a 30,867 km² area in SYP. The six regions where plots were concentrated captured local gradients of environmental conditions (i.e.: precipitation, elevation, topography, etc.), and land use history (i.e.: intensity of timber extraction, shifting cultivation, etc.). The two forest types were readily identified with DCA and Discriminant Function Analysis. Both forest types were present in all but one of the six study regions. This result supports the notion that these two major forest types are associated with local topographic features, rather than with regional gradients of natural conditions or land use history. Within clusters of each of the forest types, the six geographical regions were distinguished, which indicates that there are slight tree composition differences in both forest types across the six regions where plots were established. These differences could be the result of natural conditions and human land use. A Mantel's test between matrices of Sorensen's index of similarity and geographic distance between all plots showed a significant positive correlation, which indicates that geographically closer plots share more tree species than distant plots. However, plots within a forest type consistently shared more tree species than plots on different forest types irrespective of geographic distance. Hence, a forest plot is more likely to share tree species with a distant plot of the same forest type, than with a closer plot of the other forest type. These results suggest that the difference in tree species assemblage between the two major forest types in SYP is not related to regional gradients of natural conditions and land use history.

Keywords: Forest types, DCA, Discriminant Function Analysis, Mantel's test, Yucatan, Tropical Forests, Land use history, Gradients in Natural Conditions

This abstract is being presented at: 3:30 PM in session:
Poster Session #17: Vegetative Analysis.