Modern pollen–vegetation relationship of plant communities in the Uruguayan Atlantic coast

Abstract

Abstract Modern pollen–vegetation relationships are the basis for any paleo-study and are especially needed to understand fossil pollen assemblages and their ecological inferences. Some authors have proposed that lakes and ponds represent the surrounding vegetation through pollen grains captured by the lake. The aim of this work was to establish the modern pollen–vegetation relationships of coastal plant communities and pollen spectra represented in the surface sediments of a coastal shallow lake, and to evaluate whether spatial heterogeneity was captured by the Lake Chaparral in Perla de Rocha, Rocha, Uruguay. Pollen grains of five surface sediment samples were analyzed and related with a stratified sampling of vegetation communities surrounding the lake. Correspondence analysis and t test were used to determine plant communities and analyze the diversity of pollen assemblages. Several plant species relevant to conservation were registered, and plant communities were differentiated: coastal forest, dunes, prairies, and wetlands. However, due to limitations in taxonomic identification level for the grains of graminoids (Poaceae, Cyperaceae) in pollen assemblages, herbaceous communities highly represented by these botanical families were inferred as open environments. Pollen spectra registered at Chaparral Lake represented local and nearby vegetation well, according to both old and new theoretical models of lakes as a catchment pollen system. Several pollen grains were good indicators of the coastal forest (Myrsine sp., Ephedra tweediana Fisch. & C. A. Mey., Lithraea sp. and Tripodanthus acutifolius (Ruiz & Pav.) Tiegh.), and dunes (Chenopodiaceae, and Ambrosia sp. and Senecio sp.). The modern pollen–vegetation relationship established in this work is similar to those registered for the southern coast of Brazil, due to a shared forest component in both vegetation and pollen spectra. These results constitute primary data for the area, and we think this system is very appropriate for the historic reconstruction of coastal vegetation, particularly the coastal forest.