Journal of Vegetation Science 25 (2014) 481–490 Representativeness of tree diversity in the modern pollen rain of Andean montane forests Nele Jantz, J€ urgen Homeier & Hermann Behling Keywords Ecuador; Palynology; Pollen diversity; Pollen trap; Taxonomic surrogacy; Tropical Andean forests; Tropical mountain vegetation Nomenclature APG (The Angiosperm Phylogeny Group) III (2009) Received 30 May 2012 Accepted 18 June 2013 Co-ordinating Editor: Val erio Pillar Jantz, N. (corresponding author, nele.jantz@ biologie.uni-goettingen.de) & Behling, H. (hbehlin@gwdg.de): Department of Palynology and Climate Dynamics, Albrecht-von-Haller Institute for Plant Sciences, University of G€ ottingen, Untere Karsp€ ule 2, 37073 G€ ottingen, Germany Homeier, J. (jhomeie@gwdg.de): Department of Plant Ecology, Albrecht-von-Haller Institute for Plant Sciences, University of G€ ottingen, Untere Karsp€ ule 2, 37073 G€ ottingen, Germany Abstract Questions: To assess the relationship between modern pollen rain and Andean montane forest vegetation for diversity, and provide a basis for interpretations of palaeoecological data in the northern Andes, we asked: (1) can the reduction of plant and pollen data to family level preserve information about diversity in both data sets; (2) how precisely do tree pollen and spore types represent richness pat- terns along an altitudinal gradient on tropical mountains; and (3) how similar are tree pollen and spore family richness in relationship to tree family richness? Location: Tropical montane rain forests, Podocarpus National Park in the Andes of South Ecuador (3°S, 79°W, 1000–3000 m a.s.l.). Methods: We analysed tree diversity and species composition in three different rain forest types: Premontane (PMF), lower montane (LMF) and upper montane (UMF). We investigated modern pollen rain using pollen traps. After testing the reliability of a taxonomic surrogacy on the plant data, we compared abundance and representation, as well as diversity of the two data sets at family level. This was done using rarefaction and Sørensen indices. Results: The correlation between tree species and families was high (r = 0.81, P < 0.001). Sample rarefaction on tree pollen and plant family data revealed highest pollen diversity on sites of the UMF, but highest tree diversity on LMF and PMF sites. The Sørensen indices indicate down-drift of pollen from higher altitudes in PMF and LMF and up-drift in UMF. Between 1% and 50% of pollen taxa of each sample originate from outside the plot. Conclusions: Taxonomic surrogacy at family level is a good tool for comparing presence–absence patterns of plant and pollen data in tropical regions with high tree diversity. On a family basis, pollen presence–absence data represent the cor- responding tree vegetation data, but uncertainties increase with decreasing alti- tude. The higher diversity in pollen data of the UMF, but slightly lower diversity in the LMF and PMF, can at least partly be explained by wind patterns, local abundance of shrubs and herbs and differences in evenness. Introduction Modern pollen rain has long been recognized as a valuable reference for paleoecological studies (Giesecke et al. 2010) as it provides the necessary basis for more quantitative views on past environmental conditions. Studies of mod- ern pollen rain in tropical South and Central America are, however, still scarce, and concentrated on the northern neotropics (e.g. Bush & Rivera 1998, 2001; Bush 2000; Correa-Metrio et al. 2011), the central Andes region of Peru and Bolivia (e.g. Kuentz et al. 2007; Ortu~ no et al. 2011) Amazonia (e.g. Bush et al. 2001; Berrio et al. 2003; Gosling et al. 2005, 2009; Urrego et al. 2011) and coastal regions (Collins & Bush 2010). Few studies have been car- ried out in the northern Andes region (Salgado-Labouriau 1979; Grabandt 1980; Witte 1994; Hansen et al. 2003; Rangel 2005; Olivera et al. 2009), and only one in the spe- cies-rich region of the Andean depression between Gir on- Cuenca in South Ecuador and Huancabamba in North Peru (Niemann et al. 2010). The focus of these studies has mostly been on pollen representation in comparison to vegetation composition, often accompanied by assessments of the pollen spectra of a specific ecosystem type. Multivariate analyses, Journal of Vegetation Science Doi: 10.1111/jvs.12105 © 2013 International Association for Vegetation Science 481