Andrea Paz recent started her postdoc at ETH Zürich in Switzerland. She is an evolutionary biologist interested in unveiling the processes generating the patterns of species distributions. Here, Andrea shares her recent work investigating the environmental correlates of diversity for multiple clades and diversity measures in the Atlantic Forest.
Andrea visiting the Atlantic Forest during a trip she had with part of the US and Brazil team. The photo is in Boracéia Biological Station – a field station from the Universidade de São Paulo.
Personal links. Personal website | Twitter
Institute. Research conducted as a PhD student at the Graduate Center, City University of New York | Currently a postdoc at ETH Zürich.
Academic life stage. Starting a postdoc.
Major research themes. Biogeography, species distributions, amphibians, environmental drivers of species and community distributions.
Current study system. I just finished my PhD studying several taxa in the Atlantic Forest of Brazil. The Atlantic Forest is considered a biodiversity hotspot because of its high diversity and endemism levels and its high level of threat (less than 10% of the original forest persists). This forest is a super interesting system that includes broad latitudinal and altitudinal gradients (not a very usual combination) and thus has huge environmental variation and heterogeneity. All this variation makes this system perfect for testing many ecological and evolutionary questions related to the effect of environmental differences in biodiversity.
Boracéia Biological Station – field station from the Universidade de São Paulo.
Recent JBI paper. Paz, A., Brown, J.L., Cordeiro, C.L.O., Aguirre-Santoro, J., Assis, C., Amaro, R.C., Raposo do Amaral, F., Bochorny, T., Bacci, L.F., Caddah, M.K., d’Horta, F., Kaehler, M., Lyra, M., Grohmann, C.H., Reginato, M., Silva-Brandão, K.L., Freitas, A.V.L., Goldenberg, R., Lohmann, L.G., Michelangeli, F.A., Miyaki, C., Rodrigues, M.T., Silva, T.S. and Carnaval, A.C. (2021). Environmental correlates of taxonomic and phylogenetic diversity in the Atlantic Forest. Journal of Biogeography, 48(6), 1377-1391 https://doi.org/10.1111/jbi.14083
Motivation behind this paper. This study results from a huge collaborative effort between scientists in several countries, including the USA and Brazil. The Atlantic Forest is a big and diverse place in terms of its biology but also by the heterogeneity of landscapes it presents. For these reasons, it is hard to monitor and study this hotspot of diversity everywhere, because many factors may play a role in explaining its diversity patterns. We wanted to have a better understanding of what are the environmental correlates of different diversity dimensions in the Atlantic Forest and test whether those could apply to several taxonomic groups, including both plants and animals.
Key methodologies. Here, we used a machine learning approach where an ensemble of models (including random forest, neural networks among others) was created to better predict observed patterns of biodiversity based on abiotic variables. This allowed us to understand the correlates of diversity in the forest for several taxonomic groups. The literature shows incongruent results between different taxa, but using multiple taxa in a single biome helped us find some more general conclusions about the organisms in the forest, including how precipitation is a main predictor of diversity. In contrast, topography had a very small contribution.
A lot of the work for this publication was computer based and I did much of it working with Dr. Thiago Silva at UNESP Rio Claro in Brazil (left) and at University of Stirling in Scotland (right).
Unexpected challenges. It was very interesting getting data together from different research groups, universities and taxonomic groups. Standardizing it was definitely a challenge. For example, everyone has a different way of identifying their specimens, some use numbers from fieldwork, others use numbers from laboratory work and others from the museums. Also, the precision in naming taxa is different, with some specialties using just binomials (genus & species) and others an extra layer of clustering, such as tribes, subfamilies, etc. For the molecular portion, scientists in different disciplines also use different genes to understand the evolutionary history of their groups of interest and even different techniques to reconstruct those histories. On the other hand, bringing diverse perspectives together helped us better understand the potential processes driving diversity in this forest.
Major results. The major result is that even though the Atlantic Forest is huge and heterogeneous, we can indeed use environments to predict diversity patterns (at least for phylogenetic and taxonomic diversity) irrespective of the taxonomic group. Even more surprising is that a single driver – precipitation – was of particular importance to all groups and different measures of diversity. The model applied in our study shows a lot of promise to predict changes in diversity with a changing climate.
Next steps for this research. We are building a model that allows for predicting trends in biodiversity change in near real-time for the forest. This tool will allow us to flag areas that are either gaining or losing diversity for different groups of plants and animals because of environmental change. We hope this will become a tool easily applied in conservation actions in the near future.
Presenting the results of this work in São Paulo (Brazil), at the FAPESP International Symposium: Dimensions US and Biota São Paulo in 2019.
If you could study any organism on Earth, what would it be? I started my career studying amphibians and hope to keep going back to them :). They are amazing models to study environmental impacts at the population, species, and community levels. Also, they are beautiful!!
Anything else to add? This was the first chapter of my PhD and the first time I led a paper with so many collaborators. It was a really amazing experience getting to work with so many cool scientists!