Towards quantifying changes in forest cover in the Araucaria forest-grassland mosaic in southern Brazil

Quantitative estimates of past vegetation cover are needed both regionally so human-landscape interactions can be better understood, and globally to evaluate the effects of changing vegetation cover on the climate system. Models reducing the bias in the pollen representation of vegetation cover have been developed in the temperate regions of the northern hemisphere, while experience applying them in other parts of the world is limited. The Araucaria forest-grassland mosaic of southern Brazil is an area where open and forested plant communities exist naturally and have changed their cover over time due to changing climate and human activities. Therefore, this area is of particular interest for such studies. Modern pollen and vegetation comparisons were carried out in two protected areas: Vila Velha State Park in Paraná state and Aparados da Serra National Park in Rio Grande do Sul. Vegetation mapping and analysis focused on Araucaria angustifolia as the characteristic tree of this vegetation type. In the Araucaria forest-grassland mosaic open vegetation and woodland change at a scale of hundreds of metres. These changes are difficult to capture based on pollen proportions in surface samples. While the abundance of Poaceae pollen is not a good indicator of locally open conditions, several pollen taxa can be used as indicators of the local vegetation cover. Pollen vegetation ratios (R-values) compare well between the two study regions, indicating that pollen production of individual species within the large plant families of Poaceae and Asteraceae are similar within the overall region. Araucaria angustifolia pollen is underrepresented with regards to its vegetation cover, while Poaceae are among the highest pollen producers in the region. Diverse woodland species other than A. angustifolia were grouped as one forest taxon and as the species composition of woodlands differed between the two study areas, so did the estimated pollen productivity of this group. It would be rewarding in future investigations to estimate pollen productivity for groups of trees with the same pollen dispersal syndrome. The application of pollen dispersal models designed for closed canopy in the protected areas was challenging. Further model development is required to deal with pollen released at different levels in semi-open vegetation types