Local hydrological conditions influence tree diversity and composition across the Amazon basin

www.ecography.org ECOGRAPHY Page 1 of 17 © 2022 The Authors. Ecography published by John Wiley & Sons Ltd on behalf of Nordic Society Oikos This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 45 1–16 Ecography Subject Editor: Kenneth Feeley Editor-in-Chief: Miguel Araújo Accepted 21 June 2022 doi: 10.1111/ecog.06125 2022: e06125 Local hydrological conditions influence tree diversity and composition across the Amazon basin Manuel J. Marca-Zevallos, Gabriel M. Moulatlet, Thaiane R. Sousa, Juliana Schietti, Luiz de Souza Coelho, José Ferreira Ramos, Diogenes de Andrade Lima Filho, Iêda Leão Amaral, Francisca Dionízia de Almeida Matos, Lorena M. Rincón, Juan David Cardenas Revilla, Marcelo Petratti Pansonato, Rogério Gribel, Edelcilio Marques Barbosa, Ires Paula de Andrade Miranda, Luiz Carlos de Matos Bonates, Juan Ernesto Guevara, Rafael P. Salomão, Leandro Valle Ferreira, Dário Dantas do Amaral, Nigel C. A. Pitman, Corine Vriesendorp, Tim R. Baker, Roel Brienen, Marcelo de Jesus Veiga Carim, José Renan da Silva Guimarães, Percy Núñez Vargas, Isau Huamantupa-Chuquimaco, William F. Laurance, Susan G. W. Laurance, Ana Andrade, José Luís Camargo, Abel Monteagudo Mendoza, Rodolfo Vasquez, Luis Valenzuela Gamarra, Hugo F. Mogollón, Ben Hur Marimon-Junior, Beatriz S. Marimon, Timothy J. Killeen, Emanuelle de Sousa Farias, David Neill, Marcelo Brilhante de Medeiros, Marcelo Fragomeni Simon, John Terborgh, Juan Carlos Montero, Juan Carlos Licona, Bonifacio Mostacedo, Roosevelt García-Villacorta, Alejandro Araujo-Murakami, Luzmila Arroyo, Daniel Villarroel, Nállarett Dávila, Fernanda Coelho de Souza, Fernanda Antunes Carvalho, James A. Comiskey, Alfonso Alonso, Francisco Dallmeier, Alexandre A. Oliveira, Carolina V. Castilho, Jon Lloyd, Ted R. Feldpausch, Marcos Ríos Paredes, Nicolás Castaño Arboleda, Dairon Cárdenas López, Gerardo A. Aymard Corredor, Anthony Di Fiore, Agustín Rudas, Adriana Prieto, Flávia Rodrigues Barbosa, Janaína Costa Noronha, Domingos de Jesus Rodrigues, Rainiellen de Sá Carpanedo, Eurídice N. Honorio Coronado, Carlos A. Peres, William Milliken, Alfredo Fuentes, J. Sebastián Tello, Carlos Cerón, Bente Klitgaard, Milton Tirado, Rodrigo Sierra, Kenneth R. Young, Gonzalo Francisco Rivas-Torres, Pablo R. Stevenson, Angela Cano, Ophelia Wang, Cláudia Baider, Research article Jos Barlow, Joice Ferreira, Erika Berenguer, Juliana Stropp, Henrik Balslev, Manuel Augusto Ahuite Reategui, Italo Mesones, Elvis H. Valderrama Sandoval, Therany Gonzales, Susamar Pansini, Neidiane Farias Costa Reis, Adeilza Felipe Sampaio, Vincent Antoine Vos, Walter Palacios Cuenca, Angelo Gilberto Manzatto, William Farfan-Rios, Miles R. Silman, Karina Garcia-Cabrera, Patricio von Hildebrand, Marcelino Carneiro Guedes, Janaina Barbosa Pedrosa Costa, Juan Fernando Phillips, César I. A. Vela, José Julio de Toledo, Daniela Pauletto, Fernando Cornejo Valverde, Maria Natalia Umaña, Oliver L. Phillips, William E. Magnusson, Hans ter Steege and Flávia R. C. Costa 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 2 of 17 17 M. J. Marca-Zevallos (https://orcid.org/0000-0001-9428-7441) and T. R. Sousa, Programa de Pós-Graduação em Ecologia, Inst. Nacional de Pesquisas da Amazônia, Manaus, Brazil. MJ.M-Z also at: Herbario Vargas, Univ. Nacional de San Antonio Abad del Cusco, Cusco, Cuzco, Peru. – G. M. Moulatlet (https://orcid.org/0000-0003-2571-1207), Facultad de Ciencias de la Tierra y Agua, Univ. Regional Amazónica Ikiam, Tena, Ecuador and Inst. de Ecología A.C., Xalapa, Mexico. – J. Schietti, L. d. S. Coelho, J. F. Ramos, D. d. A. Lima Filho, Iêda Leão Amaral, F. D. de Almeida Matos, L. M. Rincón, J. D. Cardenas Revilla, M. P. Pansonato, R. Gribel, E. M. Barbosa, I. P. d. A. Miranda, L. C. d. M. Bonates, J. Carlos Montero and F. R. C. Costa (https://orcid. org/0000-0002-9600-4625) ✉ ([email protected]), Coordenação de Biodiversidade, Inst. Nacional de Pesquisas da Amazônia – INPA, Petrópolis, Manaus, AM, Brazil. – J. E. Guevara, Grupo de Investigación en Biodiversidad, Medio Ambiente y Salud-BIOMAS, Univ. de las Américas, Quito, Ecuador and Keller Science Action Center, The Field Museum, Chicago, IL, USA. – R. P. Salomão, Programa Professor Visitante Nacional Sênior na Amazônia – CAPES, Univ. Federal Rural da Amazônia, Belém, PA, Brazil. – RPS, L. V. Ferreira and D. Dantas do Amaral, Coordenação de Botânica, Museu Paraense Emílio Goeldi, Belém, PA, Brazil. – N. C. A. Pitman and C. Vriesendorp, Science and Education, The Field Museum, Chicago, IL, USA. – T. R. Baker, R. Brienen, F. Coelho de Souza (https://orcid.org/0000-0001-8893-9052), T. R. Feldpausch, E. N. Honorio Coronado and O. L. Phillips, School of Geography, Univ. of Leeds, Leeds, UK. – M. d. J. V. Carim and J. R. d. S. Guimarães, Depto de Botânica, Inst. de Pesquisas Científicas e Tecnológicas do Amapá – IEPA, Campus do IEPA da Fazendinha, Amapá, Brazil. – P. Núñez Vargas, I. Huamantupa-Chuquimaco and W. Farfan-Rios, Herbario Vargas, Univ. Nacional de San Antonio Abad del Cusco, Cusco, Cuzco, Peru. – W. F. Laurance, S. G. W. Laurance and J. Terborgh, Centre for Tropical Environmental and Sustainability Science and College of Science and Engineering, James Cook Univ., Cairns, Queensland, Australia. – A. Andrade and J. L. Camargo, Projeto Dinâmica Biológica de Fragmentos Florestais, Inst. Nacional de Pesquisas da Amazônia – INPA, Petrópolis, Manaus, AM, Brazil. – A. Monteagudo Mendoza, R. Vasquez and L. Valenzuela Gamarra, Jardín Botánico de Missouri, Oxapampa, Pasco, Peru. – H. F. Mogollón, Endangered Species Coalition, Silver Spring, MD, USA. – B. H. Marimon-Junior and B. S. Marimon, Programa de Pós-Graduação em Ecologia e Conservação, Univ. do Estado de Mato Grosso, Nova Xavantina, MT, Brazil. – T. J. Killeen, Agteca-Amazonica, Santa Cruz, Bolivia. – E. d. S. Farias, Laboratório de Ecologia de Doenças Transmissíveis da Amazônia (EDTA), Inst. Leônidas e Maria Deane, Fiocruz, Adrianópolis, Manaus, AM, Brazil and Programa de Pós-graduação em Biodiversidade e Saúde, Inst. Oswaldo Cruz – IOC/FIOCRUZ, Rio de Janeiro, RJ, Brazil. – D. Neill, Ecosistemas, Biodiversidad y Conservación de Especies, Univ. Estatal Amazónica, Puyo, Pastaza, Ecuador. – M. B. de Medeiros and M. F. Simon, Prédio da Botânica e Ecologia, Embrapa Recursos Genéticos e Biotecnologia, Parque Estação Biológica, Brasilia, DF, Brazil. JT also at: Dept of Biology and Florida Museum of Natural History, Univ. of Florida, Gainesville, FL, USA. – JCM and J. C. Licona, Inst. Boliviano de Investigacion Forestal, Doble via La Guardia, Santa Cruz, Santa Cruz, Bolivia. – B. Mostacedo, Facultad de Ciencias Agrícolas, Univ. Autónoma Gabriel René Moreno, Santa Cruz, Santa Cruz, Bolivia. – R. García-Villacorta, Dept of Ecology and Evolutionary Biology, Cornell Univ., Corson Hall, Ithaca, NY, USA and Peruvian Center for Biodiversity and Conservation (PCBC), Iquitos, Peru. – A. Araujo-Murakami, L. Arroyo and D. Villarroel, Museo de Historia Natural Noel Kempff Mercado, Univ. Autónoma Gabriel Rene Moreno, Santa Cruz, Santa Cruz, Bolivia. – N. Dávila, Biologia Vegetal, Univ. Estadual de Campinas, Campinas, SP, Brazil. – FCDS, F. A. Carvalho and W. E. Magnusson (https://orcid.org/0000-0003-19883950), Coordenação de Pesquisas em Ecologia, Inst. Nacional de Pesquisas da Amazônia – INPA, Petrópolis, Manaus, AM, Brazil. – J. A. Comiskey, Inventory and Monitoring Program, National Park Service, Fredericksburg, VA, USA. – JAC, A. Alonso and F. Dallmeier, Center for Conservation and Sustainability, Smithsonian Conservation Biology Inst., Washington, DC, USA. – MPP, A. A. Oliveira and C. Baider, Inst. de Biociências – Dept Ecologia, Univ. de Sao Paulo – USP, Cidade Universitária, São Paulo, SP, Brazil. – C. V. Castilho, EMBRAPA – Centro de Pesquisa Agroflorestal de Roraima, Boa Vista, RR, Brazil. FAC also at: Univ. Federal de Minas Gerais, Inst. de Ciências Biológicas, Depto de Genética, Ecologia e Evolução, Belo Horizonte, MG, Brazil. – J. Lloyd (https://orcid.org/0000-0003-2314-590X), Faculty of Natural Sciences, Dept of Life Sciences, Imperial College London, London, UK. TRF also at: Geography, College of Life and Environmental Sciences, Univ. of Exeter, Exeter, UK. – M. Ríos Paredes, Servicios de Biodiversidad EIRL, Iquitos, Loreto, Peru. – N. Castaño Arboleda and D. Cárdenas López, Herbario Amazónico Colombiano, Inst. SINCHI, Bogotá, DC, Colombia. – G. A. Aymard Corredor, Programa de Ciencias del Agro y el Mar, Herbario Universitario (PORT), UNELLEZ-Guanare, Guanare, Portuguesa, Venezuela. – A. Di Fiore, Dept of Anthropology, Univ. of Texas at Austin, Austin, TX, USA. – A. Rudas and A. Prieto, Inst. de Ciencias Naturales, Univ. Nacional de Colombia, Bogotá, DC, Colombia. – F. R. Barbosa, J. C. Noronha, D. d. J. Rodrigues and R. d. S. Carpanedo, ICNHS, Federal Univ. of Mato Grosso, Setor Industrial, Sinop, MT, Brazil. ENHC also at: Inst. de Investigaciones de la Amazonía Peruana (IIAP), Iquitos, Loreto, Peru. – C. A. Peres, School of Environmental Sciences, Univ. of East Anglia, Norwich, UK. – W. Milliken, Natural Capital and Plant Health, Royal Botanic Gardens, Kew, Richmond, Surrey, UK. – A. Fuentes, Herbario Nacional de Bolivia, Univ. UMSA, La Paz, La Paz, Bolivia. – AF, J. S. Tello and WF-R also at: Center for Conservation and Sustainable Development, Missouri Botanical Garden, St. Louis, MO, USA. – C. Cerón, Escuela de Biología Herbario Alfredo Paredes, Univ. Central, Quito, Pichincha, Ecuador. – B. Klitgaard, Dept for Identification & Naming, Royal Botanic Gardens, Kew, Richmond, Surrey, UK. – M. Tirado and R. Sierra, GeoIS, Quito, Pichincha, Ecuador. – K. R. Young, Geography and the Environment, Univ. of Texas at Austin, Austin, TX, USA. – G. F. Rivas-Torres, Colegio de Ciencias Biológicas y Ambientales-COCIBA & Galapagos Inst. for the Arts and Sciences-GAIAS, Univ. San Francisco de Quito-USFQ, Quito, Pichincha, Ecuador and Dept of Wildlife Ecology and Conservation, Univ. of Florida, Gainesville, FL, USA. – P. R. Stevenson and A. Cano (https://orcid.org/0000-0002-2831-4066), Laboratorio de Ecología de Bosques Tropicales y Primatología, Univ. de los Andes, Bogotá, DC, Colombia. – O. Wang, Environmental Science and Policy, Northern Arizona Univ., Flagstaff, AZ, USA. CB also at: The Mauritius Herbarium, Agricultural Services, Ministry of Agro-Industry and Food Security, Reduit, Mauritius. – J. Barlow and E. Berenguer, Lancaster Environment Centre, Lancaster Univ., Lancaster, Lancashire, UK. – J. Ferreira, Embrapa Amazônia Oriental, Belém, PA, Brazil. EB also at: Environmental Change Inst., Univ. of Oxford, Oxford, Oxfordshire, UK. – J. Stropp, Dept of Biogeography and Global Change, Museo Nacional de Ciencias Naturales (MNCN-CSIC), Madrid, Spain. – H. Balslev, Dept of Biology, Aarhus Univ., Aarhus C, Aarhus, Denmark. – M. A. Ahuite Reategui, Medio Ambiente, PLUSPRETOL, Iquitos, Loreto, Peru. – I. Mesones, Dept of Integrative Biology, Univ. of California, Berkeley, CA, USA. – E. H. Valderrama Sandoval, Dept of Biology, Univ. of Missouri, St. Louis, MO, USA and Facultad de Biologia, Univ. Nacional de la Amazonia Peruana, Iquitos, Loreto, Peru. – T. Gonzales, ACEER Foundation, Puerto Maldonado, Madre de Dios, Peru. – S. Pansini, N. F. C. Reis and A. F. Sampaio, Programa de Pós-Graduação em Biodiversidade e Biotecnologia PPG-Bionorte, Univ. Federal de Rondônia, Porto Velho, RO, Brazil. – V. A. Vos, Univ. Autónoma del Beni José Ballivián, Campus Univ. Final, Beni, Bolivia. – W. Palacios Cuenca, Herbario Nacional del Ecuador, Univ. Técnica del Norte, Quito, Pichincha, Ecuador. AC also at: Cambridge Univ. Botanic Garden, Cambridge, UK. – A. G. Manzatto, Depto de Biologia, Univ. Federal de Rondônia, Porto Velho, RO, Brazil. WF-R, Living Earth Collaborative, Washington Univ. in Saint Louis, St. Louis, MO, USA. – M. R. Silman and K. Garcia-Cabrera, Biology Dept and Center for Energy, Environment and Sustainability, Wake Forest Univ., Winston Salem, NC, USA. – P. von Hildebrand, Fundación Estación de Biología, Bogotá, DC, Colombia. – M. C. Guedes and J. B. P. Costa, Empresa Brasileira de Pesquisa Agropecuária, Embrapa Amapá, Macapá, Amapá, Brazil. – J. F. Phillips, Fundación Puerto Rastrojo, Bogotá, DC, Colombia. – C. I. A. Vela, Escuela Profesional de Ingeniería Forestal, Univ. Nacional de San Antonio Abad del Cusco, Puerto Maldonado, Madre de Dios, Peru. – J. J. de Toledo, Univ. Federal do Amapá, Ciências Ambientais, Macapá, AP, Brazil. – D. Pauletto, Inst. de Biodiversidade e Floresta, Univ. Federal do Oeste do Pará, Santarém, PA, Brazil. – F. C. Valverde, Andes to Amazon Biodiversity Program, Madre de Dios, Madre de Dios, Peru. – M. N. Umaña (https://orcid.org/0000-0001-5876-7720), Dept of Ecology and Evolutionary Biology, Univ. of Michigan, Ann Arbor, MI, USA. – H. ter Steege (https://orcid.org/0000-0002-8738-2659), Naturalis Biodiversity Center, Leiden, the Netherlands and Systems Ecology, Vrije Univ. Amsterdam, Amsterdam, the Netherlands. 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 3 of 17 Tree diversity and composition in Amazonia are known to be strongly determined by the water supplied by precipitation. Nevertheless, within the same climatic regime, water availability is modulated by local topography and soil characteristics (hereafter referred to as local hydrological conditions), varying from saturated and poorly drained to well-drained and potentially dry areas. While these conditions may be expected to influence species distribution, the impacts of local hydrological conditions on tree diversity and composition remain poorly understood at the whole Amazon basin scale. Using a dataset of 443 1-ha non-flooded forest plots distributed across the basin, we investigate how local hydrological conditions influence 1) tree alpha diversity, 2) the community-weighted wood density mean (CWM-wd) – a proxy for hydraulic resistance and 3) tree species composition. We find that the effect of local hydrological conditions on tree diversity depends on climate, being more evident in wetter forests, where diversity increases towards locations with well-drained soils. CWM-wd increased towards better drained soils in Southern and Western Amazonia. Tree species composition changed along local soil hydrological gradients in Central-Eastern, Western and Southern Amazonia, and those changes were correlated with changes in the mean wood density of plots. Our results suggest that local hydrological gradients filter species, influencing the diversity and composition of Amazonian forests. Overall, this study shows that the effect of local hydrological conditions is pervasive, extending over wide Amazonian regions, and reinforces the importance of accounting for local topography and hydrology to better understand the likely response and resilience of forests to increased frequency of extreme climate events and rising temperatures. Keywords: Amazon basin, HAND, species composition, tree diversity, wood density Introduction Large-scale patterns of floristic diversity and composition across tropical forests have been well explained by annual precipitation and seasonality (Clinebell et al. 1995, ter Steege et al. 2003, Stropp et al. 2009, Esquivel-Muelbert et al. 2017), besides soil fertility (ter Steege et al. 2006) and neutral processes (Condit et al. 2002, Tuomisto et al. 2003, Coronado et al. 2009, Emilio et al. 2010). Nevertheless, within any given climate, at local scale, water availability for plants is modulated by local topographic and edaphic conditions, giving rise to the local hydrological conditions (Daws et al. 2002, Moeslund et al. 2013, Costa et al. 2022), which in turn affect variations in plant species diversity and composition (Svenning 2001, ter Steege et al. 2003, Jirka et al. 2007, Schietti et al. 2013, Moulatlet et al. 2014, Zuleta et al. 2020). In Amazonia, where the world’s largest tropical forest is located, we now understand how local hydrological conditions can modulate the effects of droughts (Sousa et al. 2020, Esteban et al. 2021, Costa et al. 2022), with locally wetter areas constituting potential refuges for diversity (McLaughlin et al. 2017). Global warming is leading to an increase in the frequency and intensity of droughts (Marengo et al. 2018). As a consequence, Amazonian forests are experiencing a reduction in water availability for plants (Pascolini-Campbell et al. 2021), and higher tree mortality and biomass loss (Phillips et al. 2009, Brienen et al. 2015, Berenguer et al. 2021). Therefore, it is crucial to assess the influence of local hydrological conditions on forest diversity, and on taxonomic and functional composition. The local hydrological conditions of non-flooded forests are mainly determined by elevation (relief), terrain aspect and slope (Moeslund et al. 2013). These characteristics control the land’s drainage capacity, the local redistribution of precipitation runoff and the vertical distance to groundwater (Rennó et al. 2008, Nobre et al. 2011, Moeslund et al. 2013). Along a topographic profile, lower areas are moist and often water-saturated, and have soils with low drainage capacity due to the proximity to the water table (Rennó et al. 2008, Nobre et al. 2011, Moeslund et al. 2013). In contrast, the upper areas such as hilltops are typically drier and have well-drained soils due to increasing distance to the water table (Rennó et al. 2008, Nobre et al. 2011, Moeslund et al. 2013, Fan et al. 2017). Besides topography, soil characteristics – especially soil texture – also affect the water availability for plants by regulating water infiltration and the capillary rise (Hacke et al. 2000, Fan et al. 2017). Thus, soil characteristics and topography together determine the local hydrological conditions, and consequently the water availability for plants at the local scale. However, the effect of local hydrological conditions on plants across large regions should also depend on climate, as the variation in precipitation determines how much water enters the system. Therefore, we can expect an interaction between local hydrological conditions and precipitation affecting the plant community properties (Blanchard et al. 2019, Muscarella et al. 2019). The local hydrological conditions can be expected to affect species richness through physiological or forest dynamics processes. Considering that water deficit imposes a physiological limitation to many plant species (Nepstad et al. 2007, Meir et al. 2015, Giardina et al. 2018), drier local conditions given by the combination of dry climate and deep water tables at high topographic positions should limit the number of species able to colonize and survive. Conversely, even within dry climates, low topographic positions with shallow water tables should provide higher soil moisture, alleviating seasonal water deficits and promoting higher richness (Segura et al. 2003). This scenario should shift under wet climates, where shallow water tables lead to seasonal or permanent waterlogging and anoxia, stressful conditions that can reduce the number of species (Féret and Asner 2014). Shallow water tables also limit root development (Fan et al. 2017), and shallow rooting systems are more prone to uprooting, increasing tree mortality rates (Gale and Hall 2001, Ferry et al. 2010, Toledo et al. 2011). Higher mortality rates could lead to increased richness if the disturbance is moderate or decreased richness if it is intense (Connell 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 4 of 17 1978, Bongers et al. 2009). This expectation could be reversed under dry climates, where waterlogging is less common, and tree mortality tends to be higher in the upper topographic positions exposed to wind (Segura et al. 2003). A functional ecology approach provides insights into environmental filtering along environmental gradients through the evaluation of morpho-physio-phenological traits that affect an individual’s growth, reproduction and survival (Violle et al. 2007). Earlier studies of non-flooded Amazonian forests at local scales indicate that local hydrological conditions filter functional traits such as wood density (Ferry et al. 2010, Cosme et al. 2017); hydraulic resistance (Oliveira et al. 2019, Fontes et al. 2020, Garcia et al. 2022); leaf traits – e.g. specific leaf area (SLA), leaf nitrogen concentration, leaf size, leaf mass per area, leaf dry matter content, leaf thickness and chlorophyll content (Kraft et al. 2008, Cosme et al. 2017, Schmitt et al. 2020); seed mass (Kraft et al. 2008); and canopy chemical traits (Asner et al. 2015). However, no study has so far determined if hydrological filtering of traits is a widespread phenomenon over the Amazon basin and how it interacts with other environmental factors. A key trait to understand plant hydraulics is wood density, given that higher wood density is associated to higher xylem resistance to embolism (Hacke et al. 2001, Santiago et al. 2004, Hoeber et al. 2014), a trait that is selected in drier climates (Chave et al. 2006, Ibanez et al. 2017, Blanchard et al. 2019) and drier local hydrological conditions (Oliveira et al. 2019). Although other functional traits besides wood density can be relevant to the environmental filtering of species distributions by hydrology, we focused on the trait that tends to have lower intraspecific variation (compared to leaf traits) and broader data coverage. Filtering of functional traits by the local hydrological conditions can lead to species sorting across hydrological environments, with consequences for species composition and alpha diversity (Valencia et al. 2004, Baldeck et al. 2013, Schietti et al. 2013, Féret and Asner 2014, Moulatlet et al. 2014, Zuleta et al. 2020). This can be evaluated through the correlation between the community average of any trait (e.g. wood density) and the multivariate axes representing species composition (ter Steege et al. 2006). Given that species pools differ among Amazonian regions (ter Steege et al. 2006, 2013) owing to biogeographical and evolutionary causes, it can be expected that the effects of local hydrological conditions on species composition would be detected within regions (as defined by geography and substrate origin in Feldpausch et al. 2011). This is because which species groups are associated with locally wetter or drier conditions along hydrological gradients within each region should depend on which species are present in the region. Furthermore, larger changes in species composition along the gradient of local hydrological conditions may be expected in terrains with ancient well-dissected relief, since these produce stronger hydrological contrast along topography. The aim of this study was to understand the effects of local hydrological conditions on the properties of tree communities at the Amazon basin scale. We addressed the following specific questions: 1) How do local hydrological conditions influence tree alpha diversity? 2) How do local hydrological conditions influence wood density? 3) How do local hydrological conditions affect tree species compositional change along environmental gradients? The general hypothesis is that local hydrological conditions will influence the environmental filtering of species distributions, thus generating spatial patterns of alpha diversity, mean wood density and species composition. Specifically, we predict that: 1) Tree alpha diversity will decrease in low topographic positions (due to soil–water excess) in climatically wetter forests, but increase in those positions (due to higher local moisture) in climatically drier forests. 2) Forests growing on soils with greater drainage capacity and lower water availability will have higher wood density as it provides higher hydraulic resistance against droughts. 3) Species composition will be sorted along gradients of local hydrological conditions within regions, following changes in community-weighted means of wood density (CWM-wd). Material and methods Vegetation data We used the Amazon Tree Diversity Network (ATDN) dataset, with plots distributed throughout the Amazon basin (as delimited by Mayorga et al. 2012) (Fig. 1). Our analyses were restricted to 1-ha lowland terra-firme forest plots below 500 a.s.l. (excluding plots on white sand and inundated forests) and to individuals with a diameter ≥ 10 cm, excluding all lianas. Plots varied in dimensions and shapes, with most being square or rectangular 1 ha, while 11.7% were 250 × 40 m and following altitudinal contours (design described in Magnusson et al. 2005). In addition, we considered only plots with at least 80% of individuals identified to species level. As species identification was done by different taxonomists, we excluded 18 634 individuals (8.45% of the total number of individuals; mean = 42, min = 0, max = 173 individuals per plot) that were not identified to the species level to avoid confusion with morphospecies synonymy. This introduced no bias in the analyses, as there was no association between the proportion of morphospecies per plot with the main variables of interest (i.e. local hydrological conditions; Supporting information). By including only those individuals identified at the species level, more robust patterns of alpha diversity and composition are expected (Pos et al. 2014). We also excluded plots with georeferencing problems, such as those with coordinates displaced from terra-firme towards rivers or lakes. Finally, we excluded 18 plots from areas without height above nearest drainage (HAND) data. Thus, we carried out the analyses using 443 plots, which total 210 801 individuals of 3527 species, distributed in 619 genera and 104 families. 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 5 of 17 Vegetation metrics Alpha diversity We quantified species, genus and family diversities for each plot using Fisher’s alpha (logarithmic series model), which is based on the number of individuals and species in each plot and is relatively insensitive to the difference in the number of individuals among plots (Fisher et al. 1943). Wood density The wood density of each plot was represented by the CWM of the species present in that plot. We obtained wood density data from the global wood density database (Chave et al. 2009, Zanne et al. 2009). When the wood density of a species was not available, we used the mean wood density at the genus, family or plot level (Baker et al. 2004, Rozendaal et al. 2020). We used genus level wood density data for 1381 species out of the 2417 species, and family and plot level data for 170 and 10 species, respectively. Then, we calculated the CWM-wd, which is the estimated mean wood density of the individuals in each plot (Garnier et al. 2004). Species composition We excluded species occurring in fewer than two plots to reduce the potential noise generated by the inclusion of rare species in the analysis (Cao et al. 2001, McCune and Grace 2002). The final number of species and individuals in the full dataset for species composition analyses was 241 and 206 459, respectively. This database was the same used in the analyses of wood density. We first tested if the expectation of species composition differences among regions hold true for the specific dataset used here. This was implemented with a PERMANOVA (Anderson 2001), where the response variable was the matrix of Bray–Curtis dissimilarities in species abundances among plots, and the Amazon regions the independent variables. Figure 1. Location of 443 1-ha plots used in this study. The map shows the distribution of plots (coloured dots) across the Amazon basin. The limits of the Amazon basin (black outline) were defined according to Mayorga et al. (2012). Background information shows the annual precipitation layer obtained from CHELSA ver. 1.2 (1979–2013, 30 arcsec of resolution, <http://chelsa-climate.org/> [Karger et al. 2018]). Dot colours indicate the four geomorphological regions (from Feldpausch et al. 2011): Central–East (blue, CA–EA = 122 plots), Guiana Shield (green, GS = 52 plots), South (red, SA = 102 plots) and West (yellow, WA = 167 plots). 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 6 of 17 This method tests whether plots from the same region are floristically more similar to each other than would be expected by chance. The differences among regions were confirmed (PERMANOVA: p = 001, R2 = 0.10). Given the confirmed differences among regions, we ran one non-metric multidimensional scaling (NMDS) ordination within each of the four Amazonian regions, and the resulting axes were the response variables representing species composition in the generalised least square (GLS) regression analyses (see Data analyses section). Ordinations were run in two dimensions to keep stress close to 0.2 (McCune and Grace 2002), and were based on Bray–Curtis distance matrices calculated over quantitative data – species abundances – standardized by plot (i.e. the abundance of each species divided by the total abundance in each plot). Ordinations based on quantitative data mostly capture the patterns of more abundant species, which tend to have larger contributions to the difference or similarity between plots (McCune and Grace 2002). The variation of the original data captured by the NMDS axes was obtained by regression of the matrix of original floristic distances among samples calculated with the same dissimilarity index used in the ordination method, and the matrix of distances among samples obtained from the final NMDS solution (McCune and Grace 2002). Environmental variables The environmental variables considered in the present study were proxies for the hydrological conditions at the local and at the regional scales, and for the soil fertility. Local hydrological conditions were represented by the HAND, which captures the influence of topography on soil hydrology (Rennó et al. 2008, Nobre et al. 2011), and by one variable capturing the water retention capacity of the soil– soil texture (Hacke et al. 2000). HAND is calculated using information from digital elevation models (DEM) and drainage networks by normalizing the topography (elevation) as a function of the relative height over the nearest drainage. This variable is directly related to the gravitational potential of soil water and, therefore, its values are correlated with the water table depth, providing a representation of the local hydrological conditions. High HAND values mean high drainage potential; low HAND values indicate proximity to the water table, where water accumulates and may create waterlogged conditions (Nobre et al. 2011). We extracted HAND data from layers elaborated by Banon and Novo (2018) with ~30 × 30 m spatial resolution (available at: <www.dsr.inpe.br/ amazondrainage/home.php?content=hand>). To capture hydrological variability within each 1 ha plot, we reprojected the layers to ~100 × 100 m resolution using bilinear interpolation, which assigns the output cell value by taking the weighted average of the neighbouring cells. This method is recommended to assign raster values to points in continuous datasets without distinct boundaries. The HAND layer was based on Strahler’s sixth order drainages, as it provides more stable information about the drainage capacity and the water condition of the environment across varying geomorphologies (Banon and Novo 2018). Finally, for all analysis, HAND values were log-transformed (base 2) to meet normality assumptions. The soil texture was represented by the percentages of sand or clay. We extracted soil texture data for a depth of 15 cm for each plot from the SoilGrids platform (Hengl et al. 2017) with ~250 × 250 m spatial resolution. Hydrological conditions at the regional scale were represented by the historical average of the maximum cumulative water deficit (MCWD) for each plot. MCWD is the measure of the annual water deficit that considers both the duration and the intensity of the dry season (Aragão et al. 2007). MCWD represents the most negative value of water deficit (WD), given by the difference between precipitation (P) and evapotranspiration (E) within each year. For each month (n), WD is quantified as: if WD EP then WD WD EP else WD nnn nn nn n - - -+< =-+ = 1 1 0 0 ; ; . For this calculation, we considered the hydrological year, so the starting point for the calculation was not necessarily the first month of the calendar year, but the wettest month (n − 1) of the first year of the time series, according to EsquivelMuelbert et al. (2019). The precipitation data were obtained from the TerraClimate platform (Abatzoglou et al. 2018) with ~ 4 × 4 km spatial resolution from 1959 to 2018. We assumed the evapotranspiration to be fixed at 100 mm per month, being an approximation (mean) from soil evapotranspiration obtained in different locations in Amazonia (Aragão et al. 2007). We estimated soil fertility using soil cation concentration (SCC, i.e. the sum of Mg+2, Ca+2 and K+1). SCC is an indicator of soil fertility correlated with the most limiting nutrient in the tropical regions – phosphorus availability (Quesada et al. 2009, Moulatlet et al. 2017, Figueiredo et al. 2018). We obtained SCC values from the layer produced by Zuquim et al. (2019), with ~11 × 11 km spatial resolution, which was reprojected to ~ 5 × 5 km resolution using the bilinear method. Four Amazonian regions (Central and Eastern Amazonia, Guiana Shield, Southern Amazonia and Western Amazonia), defined by geography and substrate origin in Feldpausch et al. (2011), were used to constrain the analyses of species composition, given their expected effect on the species pool (ter Steege et al. 2013). Central and Eastern Amazonia are characterized by having nutrient-poor soils derived from reworked rocks and late Cretaceous sediments that experienced weathering for more than 20 million years (Irion 1978, Quesada et al. 2010), and the relief is either flat or undulated (Sombroek 2000). The Guiana Shield and Southern Amazonia have weathered, nutrient-poor soils formed on ancient Cretaceous crystalline substrates (Irion 1978, Quesada et al. 2010), and the relief is characterized by hilly dissected lands and rounded hills (Sombroek 2000). Western Amazonia is formed on more recent pre-Andean sediments from the Cretaceous–Tertiary 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 7 of 17 and Pliocene periods, where large areas dominated by shallow soils that can contribute to the fertility of the underlying soils through the weathering of the source material (Irion 1978, Quesada et al. 2010), the relief is characterized by undulating to rolling land (Sombroek 2000). Data analyses We modelled each of the response variables – Fisher’s alpha at species, genus and family levels, the CWM-wd of the entire Amazon basin, and the NMDS axes 1 and 2 of the ordinations of species composition for each Amazon region – as a function of the proxies for the hydrological conditions at the local scale (HAND and soil texture) and at the regional scale (MCWD), plus soil fertility (SCC), using GLS models. We also included in the models the interaction between HAND and MCWD, expected to affect the water conditions for plants (Blanchard et al. 2019, Muscarella et al. 2019). Sand or clay were included in alternative GLS models to avoid multicollinearity, and the textural class in the best performing models was retained. We included a term to account for the spatial autocorrelation, specifying the most appropriate autocorrelation structure for each model, selected from the lowest values of the Akaike information criterion (AIC) (Supporting information). We examined the multicollinearity among variables, using the variance inflation factor (VIF) available in the vif function of the car package (Fox et al. 2021). A VIF greater than 5 is a signal of collinearity (Dormann et al. 2013). In each of our GLS models, all variables had VIF < 4 (Supporting information), indicating that collinearity was unlikely to affect our results. Furthermore, the pairwise correlation among variables was low (rho < 0.31) (Supporting information). We report the partial effects of variables from multiple models, i.e. the effect of a variable when the effects of others are held constant (Allen 1997). To visualize and report the individual or interaction effects of the explanatory variables on the response variables we used partial plots of the significant effects (p > 0.05) identified by each GLS model, and the classes were defined based on the standard deviation around the mean of each of these variables. To compare the effect size of the explanatory variables on the response variables, all variables were standardized to have mean of zero and standard deviation of one. We applied a logarithmic transformation to Fisher’s alpha values at the species level to ensure normal distribution of residuals; this transformation was not necessary for genus and family levels. In the models for Fisher’s alpha we included an exponential variance structure for MCWD to reduce the effect of heteroscedasticity in the models, i.e. a decrease in residual variance for the response variable (Fisher’s alpha) along the explanatory variable (MCWD) (Zuur et al. 2009). The selection of the variance structure was based on the lowest values of AIC (Zuur et al. 2009) (Supporting information). The models for species composition were run within each Amazonian region, using as response variables the NMDS-1 and NMDS-2 axes of the ordination carried on the plots belonging to each region. We first tested for the correlation between the first two NMDS ordination axes, to determine if both could be used as separate response variables. The independence among axes was confirmed (Supporting information), thus the GLS models including all the factors described above were run for each axis within each region. We conducted another test of the effect of hydrological and soil predictors on species composition by running an ordination based on all plots (following the same standards used before), and modelling the two resulting axes as a function of MCWD, HAND, %Clay and SCC, and including regions in interaction with HAND. This provides an assessment of how general the effects detected in the first approach are. Finally, to determine if species composition and wood density are correlated we performed Spearman’s correlation tests between species composition axes (NMDS-1 and NMDS-2) and the CWM-wd of the plots of each region and adjusted probability values with the Bonferroni correction (Armstrong 2014). All analyses were carried out in R (R ver. 4.0.2). The nlme package (Pinheiro et al. 2021) was used for GLS models, the vegan package (Oksanen et al. 2020) for NMDS ordinations, the adonis function from the vegan R package for PERMANOVA and the corr.test function from the psych package (Revelle 2021) for correlation tests. Results Tree alpha diversity Fisher’s alpha at species and genus levels was mainly influenced by maximum climatological water deficit (MCWD) and by the interaction between the HAND and MCWD, whereas at family level it was influenced by MCWD and the soil cation concentration (SCC) (Table 1). MCWD had the largest relative contribution in the models at all taxonomic levels (Table 1, Fig. 2a–c). The Fisher’s alpha diversity at all taxonomic levels increased towards wetter forests, with less negative MCWD values (Fig. 3a–c). The partial effect of the interaction between MCWD and HAND (Fig. 3d, e) indicated that in climatically wetter forests (MCWD between -136.7 and −1.60 mm, corresponding approximately to 1–2 dry months in the year, 222 plots) the Fisher’s alpha of species and genera were higher in plots with high HAND values, i.e. well-drained areas with deep water tables, such as the plateaus and hilltops (Fig. 3d, e, blue). However, in climatically drier forests (MCWD between −429.6 and −138.0 mm, > 2 dry months, 221 plots), there was a subtle tendency towards greater diversity in sites with low HAND values, i.e. the zones with higher water availability and water table close to the surface, corresponding to bottomlands and riparian forests (Fig. 3d, e, red). Family alpha diversity increased towards forests with higher SCC (Fig. 3f). There was a marginally significant interaction between MCWD and HAND at the family level (Table 1), with a similar trend to that observed at species and genus levels (Supporting information). 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 8 of 17 Wood density The CWM-wd differed among the Amazonian regions, and was affected by SCC and the interaction between HAND and the regions (Table 1, Fig. 2d). CWM-wd strongly decreased with SCC (Fig. 3g). CWM-wd was higher in Central, Eastern Amazonia and Guiana Shield forests, and lower in Southern and Western Amazonian forests (Fig. 3h). The effect of HAND can be seen in Southern and Western Amazonian forests, where the CWM-wd increased with HAND, i.e. towards uplands and well-drained areas (Fig. 3i). Furthermore, there was a tendency (marginally significant, Table 1) for an average increase in the CWM-wd towards forests with higher HAND values all over the basin (Supporting information). Species composition The two-dimensional NMDS solution produced high congruence between species composition distances observed in the original space and those obtained in the reduced space, low stress and > 50% of the variation of the original composition was captured in all the regions (Supporting information). The results of the GLS models (Table 2) showed Figure 2. Standardized effect of the variables used in the generalized least square (GLS) models for alpha diversity and community weighted mean of wood density (CWM-wd) in the Amazon basin. Species (a), genus (b) and family (c) alpha diversity were mainly influenced by the maximum cumulative water deficit (MCWD). CWM-wd was mainly influenced by SCC (soil fertility) and the Amazonian regions (d). For each term in the model, the points represent the standardized effect and the lines represent standard deviation 1. “:” = interaction. Table 1. Standardized coefficients and their respective probabilities, resulting from the generalized least square (GLS) models for tree alpha diversity (i.e. Fisher’s alpha (FA) for species: FA-species, FA-genus and FA-family) and the community weighted mean of wood density (CWM-wd) for 443 1-ha plots across the Amazon basin. MCWD – maximum cumulative water deficit, HAND – height above nearest drainage, SCC – soil cation concentration and soil texture (% Clay). ‘:’ = interaction. *, ** and *** indicate significant coefficients with p < 0.05, p < 0.01 and p < 0.001, respectively. MCWD HAND SCC %Clay HAND:MCWD Region HAND:Region R2 FA-Species 0.495*** 0.035 0.097 0.012 0.086* – – 0.64 FA-Genus 0.507*** 0.06 0.133 0.078 0.107** – – 0.63 FA-Family 0.457*** −0.006 0.247** 0.037 0.063 – – 0.57 CWM-wd −0.055 0.06 −0.4*** 0.034 0.14 −0.369** 4.58* 0.76 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 9 of 17 Figure 3. Partial effect of maximum cumulative water deficit (MCWD) on species (a), genus (b) and family (c). Partial effect of the interaction between height above nearest drainage (HAND) and maximum cumulative water deficit (MCWD) on the species (d) and genus (e) Fisher alpha diversity. Panels (d) and (e) show the HAND effect in wetter (blue) and drier (red) forests. In order to visualize the interaction between MCWD and HAND, the number of plots was divided into two groups based on the median. Red is used for plots with MCWD values with higher water deficit (between −429.6 and −138.0, 221 plots); and blue is for plots with MCWD values with less water deficit (between −136.7 and −1.60, 222 plots). Partial effect of soil cation concentration (SCC), a proxy of soil fertility, on family Fisher alpha diversity (f). Partial effect of SCC (g) and Amazon regions (h) on the community weighted mean of wood density (CWM-wd). Partial effect of the interaction between HAND and Amazon regions on CWM-wd (i). Different colours represent the four Amazon regions: Central and Eastern Amazonia (blue, CA–EA = 122 plots), Guiana Shield (green, GS = 52 plots), Southern Amazonia (red, AS = 102 plots) and West Amazonia (yellow, WA = 167 plots). *, ** and *** indicate significant effects with p < 0.01 and p < 0.001, respectively. 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 16 of 17 Magnusson, W. E. et al. 2005. RAPELD: a modification of the Gentry method for biodiversity surveys in long-term ecological research sites. – Biota Neotrop. 5: 19–24. Marca-Zevallos, M. J. et al. 2022. Data from: Local hydrological conditions influence tree diversity and composition across the Amazon basin. – Dryad Digital Repository, <https://doi. org/10.5061/dryad.qnk98sfkg>. Marengo, J. A. et al. 2018. Changes in climate and land use over the Amazon Region: current and future variability and trends. – Front. Earth Sci. 6: 1–21. Marimon, B. S. et al. 2014. Disequilibrium and hyperdynamic tree turnover at the forest-cerrado transition zone in southern Amazonia. – Plant Ecol. Divers. 7: 281–292. Marra, D. M. et al. 2014. Large-scale wind disturbances promote tree diversity in a Central Amazon Forest. – PLoS One 9: e103711. Mayorga, E. M. G. et al. 2012. LBA-ECO CD-06 Amazon River Basin Land and stream drainage direction maps. – ORNL DAAC, <https://doi.org/10.3334/ORNLDAAC/1086>. McCune, B. and Grace, J. B. 2002. Analysis of ecological communities. – MjMSoftware. McLaughlin, B. C. et al. 2017. Hydrologic refugia, plants and climate change. – Global Change Biol. 23: 2941–2961. Meir, P. et al. 2015. Threshold responses to soil moisture deficit by trees and soil in tropical rain forests: insights from field experiments. – Bioscience 65: 882–892. Moeslund, J. E. et al. 2013. Topography as a driver of local terrestrial vascular plant diversity patterns. – Nord. J. Bot. 31: 129–144. Moulatlet, G. M. et al. 2014. Local hydrological conditions explain floristic composition in lowland Amazonian forests. – Biotropica 46: 395–403. Moulatlet, G. M. et al. 2017. Using digital soil maps to infer edaphic affinities of plant species in Amazonia: problems and prospects. – Ecol. Evol. 7: 8463–8477. Muller-Landau, H. C. 2004. Interspecific and inter-site variation in wood specific gravity of tropical trees. – Biotropica 36: 20–32. Muscarella, R. et al. 2019. Effects of topography on tropical forest structure depend on climate context. – J. Ecol. 108: 1–15. Nepstad, D. C. et al. 2007. Mortality of large trees and lianas following experimental drought in an Amazon forest. – Ecology 88: 2259–2269. Nobre, A. D. et al. 2011. Height above the nearest drainage – a hydrologically relevant new terrain model. – J. Hydrol. 404: 13–29. Oksanen, A. J. et al. 2020. vegan: community ecology package, ver. 2.5-7. – <https://CRAN.R-project.org/package=vegan>. Oliveira, R. S. et al. 2019. Embolism resistance drives the distribution of Amazonian rainforest tree species along hydro-topographic gradients. – New Phytol. 221: 1457–1465. Oliveira-Filho, A. T. et al. 1998. Effects of canopy gaps, topography and soils on the distribution of woody species in a central Brazilian deciduous dry forest. – Biotropica 30: 362–375. Parolin, P. 2001. Morphological and physiological adjustments to waterlogging and drought in seedlings of Amazonian floodplain trees. – Oecologia 128: 326–335. Parolin, P. et al. 2004. Central Amazonian floodplain forests: tree adaptations in a pulsing system. – Bot. Rev. 70: 357–380. Pascolini-Campbell, M. et al. 2021. A 10 per cent increase in global land evapotranspiration from 2003 to 2019. – Nature 593: 543–547. Phillips, O. L. et al. 1994. Dynamics and species richness of tropical rain forests. – Proc. Natl Acad. Sci. USA 91: 2805–2809. Phillips, O. L. et al. 2009. Drought sensitivity of the Amazon rainforest. – Science 323: 1344–1347. Pinheiro, J. et al. 2021. nlme: linear and nonlinear mixed effects models. – R package ver. 3.1-152, <https://CRAN.R-project. org/package=nlme>. Pos, E. T. et al. 2014. Are all species necessary to reveal ecologically important patterns? – Ecol. Evol. 4: 4626–4636. Quesada, C. A. et al. 2009. Regional and large-scale patterns in Amazon forest structure and function are mediated by variations in soil physical and chemical properties. – Biogeosciences 6: 3993–4057. Quesada, C. A. et al. 2010. Variations in chemical and physical properties of Amazon forest soils in relation to their genesis. – Biogeosciences 7: 1515–1541. Quesada, C. A. et al. 2012. Basin-wide variations in Amazon forest structure and function are mediated by both soils and climate. – Biogeosciences 9: 2203–2246. Rennó, C. D. et al. 2008. HAND, a new terrain descriptor using SRTM-DEM : Mapping terra-firme rainforest environments in Amazonia. – Remote Sens. Environ. 112: 3469–3481. Revelle, W. 2021. psych: procedures for personality and psychological research, ver. 3.0-12. – <https://cran.r-project.org/ package=psych>. Roebroek, C. T. J. et al. 2020. Global distribution of hydrologic controls on forest growth. – Hydrol. Earth Syst. Sci. 24: 4625–4639. Rozendaal, D. M. A. et al. 2020. Competition influences tree growth, but not mortality, across environmental gradients in Amazonia and tropical Africa. – Ecology 101:e03052. Santiago, L. S. et al. 2004. Leaf photosynthetic traits scale with hydraulic conductivity and wood density in Panamanian forest canopy trees. – Oecologia 140: 543–550. Schietti, J. et al. 2013. Vertical distance from drainage drives floristic composition changes in an Amazonian rainforest. – Plant Ecol. Divers. 7: 37–41. Schmitt, S. et al. 2020. Topography consistently drives intraand inter-specific leaf trait variation within tree species complexes in a Neotropical forest. – Oikos 129: 1521–1530. Segura, G. et al. 2003. Tree community structure and stem mortality along a water availability gradient in a Mexican tropical dry forest. – Plant Ecol. 169: 259–271. Silvertown, J. et al. 2015. Hydrological niches in terrestrial plant communities: a review. – J. Ecol. 103: 93–108. Sombroek, W. 2000. Amazon landforms and soils in relation to biological diversity. – Acta Amaz. 30: 81–100. Sousa, T. R. et al. 2020. Palms and trees resist extreme drought in Amazon forests with shallow water tables. – J. Ecol. 108: 2070–2082. Stropp, J. et al. 2009. Disentangling regional and local tree diversity in the Amazon. – Ecography 32: 46–54. Svenning, J. 2001. On the role of microenvironmental heterogeneity in the ecology and diversification of neotropical rain-forest palms (Arecaceae). – Bot. Rev. 67: 1–53. ter Steege, H. and Hammond, D. S. 2001. Character convergence, diversity and disturbance in tropical rain forest in Guyana. – Ecology 82: 3197–3212. ter Steege, H. et al. 2003. A spatial model of tree α-diversity and tree density for the Amazon. – Biodivers. Conserv. 12: 2255–2277. 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Page 17 of 17 ter Steege, H. et al. 2006. Continental-scale patterns of canopy tree composition and function across Amazonia. – Nature 443: 444–447. ter Steege, H. et al. 2013. Hyperdominance in the Amazonian tree flora. – Science 342: 1243092. Toledo, J. J. et al. 2011. How much variation in tree mortality is predicted by soil and topography in Central Amazonia? – For. Ecol. Manage. 262: 331–338. Tuomisto, H. et al. 2002. Distribution and diversity of pteridophytes and Melastomataceae along edaphic gradients in Yasuní National Park, Ecuadorian Amazonia. – Biotropica 34: 516–533. Tuomisto, H. et al. 2003. Dispersal, environment and floristic variation of western Amazonian forests. – Science 299: 241–244. Tuomisto, H. et al. 2014. Species richness and diversity along edaphic and climatic gradients in Amazonia. – Ecography 37: 1034–1046. Umaña, M. N. et al. 2021. Shifts in taxonomic and functional composition of trees along rainfall and phosphorus gradients in central Panama. – J. Ecol. 109: 51–61. Valencia, R. et al. 2004. Tree species distributions and local habitat variation in the Amazon: large forest plot in eastern Ecuador. – J. Ecol. 92: 214–229. Violle, C. et al. 2007. Let the concept of trait be functional! – Oikos 116: 882–892. Wittmann, F. and Parolin, P. 2005. Aboveground roots in Amazonian floodplain trees. – Biotropica 37: 609–619. Wittmann, F. et al. 2006. Tree species composition and diversity gradients in white-water forests across the Amazon Basin. – J. Biogeogr. 33: 1334–1347. Wittmann, F. et al. 2010. Phytogeography, species diversity, community structure and dynamics of Central Amazonian floodplain forests. – In: Junk, W. J et al. (eds), Amazonian floodplain forests. Ecological Studies 210. Springer, pp. 61–102. Zanne, A. E. et al. 2009. Data from: Towards a worldwide wood economics spectrum. – Dryad Digital Repository, <https://doi. org/10.5061/dryad.234>. Zuleta, D. et al. 2020. Importance of topography for tree species habitat distributions in a terra firme forest in the Colombian Amazon. – Plant Soil 450: 133–149. Zuquim, G. et al. 2019. Making the most of scarce data : Mapping soil gradients in data-poor areas using species occurrence records. – Methods Ecol. Evol. 10: 788–801. Zuur, A. F. et al. 2009. Mixed effects models and extensions in ecology with R. – Springer. 16000587, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/ecog.06125 by Spanish Cochrane National Provision (Ministerio de Sanidad), Wiley Online Library on [01/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License