Environmental correlates of phylogenetic endemism in amphibians and the conservation of refugia in the Coastal Forests of Eastern Africa

Abstract

Aims To quantify the spatial distribution of amphibian phylogenetic endemism (PE), an indicator of potential refugia, to test PE for correlations with current and historical environmental predictors, and to evaluate the effectiveness of current protected areas at conserving evolutionary history. Location Coastal Forests of Eastern Africa (CFEA) and the adjacent low‐elevation Eastern Afromontane (EA). Methods We integrated new and existing spatial and phylogenetic data to map PE for almost the full amphibian assemblage (41 of 55 species), including 35 intraspecific lineages from several species and complexes showing high phylogeographic structure. Using spatial and non‐spatial regressive models, we tested whether PE can be predicted by measures of Quaternary climate change, forest stability, topographic heterogeneity and current climate. PE results were intersected with the protected area network to evaluate current conservation effectiveness. Results We detect refugia in Tanzania and coastal Kenya previously identified as CFEA centres of endemism but also new areas (lowland Tanga region and Pangani river, Zaraninge forest, Mafia island, Matumbi hills). Results show that refugia for amphibians (high PE) are located in areas with long‐term Quaternary climate stability and benign current climate (high precipitation of driest quarter, high annual precipitation), with climatically unstable areas demonstrating low PE. Conservation analyses revealed that ten PE hotspots account for over 25% of the total PE, but only small parts of these areas are under conservation protection. Main Conclusions Utilizing cryptic diversity from novel phylogeographic data and distribution modelling improves our understanding of endemism patterns, with climate stability being strongly correlated with the distribution of PE. Our analyses point towards high PE areas being refugia, which require an urgent need to consolidate protected areas within centres of endemism in this highly threatened biodiversity hotspot.

Species sorting and mass effect along forest succession: Evidence from taxonomic, functional, and phylogenetic diversity of amphibian communities

“…While these taxonomic, functional, and phylogenetic results support a strong environmental signal in the structure of amphibian communities along environmental gradients (Hernández‐Ordóñez et al, ; Russildi, Arroyo‐Rodríguez, Hernandez‐Ordoñez, Pineda, & Reynoso, ; Suazo‐Ortuño et al, ), they also highlight the need for more integrated assessments of amphibian diversity that go beyond counting species and individuals. A clear phylogeny including over 2,800 frogs, salamanders, and caecilians of the world is available since 2011 (Pyron & Wiens, ), but to date, only four studies have evaluated the phylogenetic patterns of amphibian diversity along environmental gradients (Barratt et al, ; Martins et al, ; Nowakowski et al, ; Ribeiro et al, ). Similarly, ecological, physiological, and life‐history trait data are increasingly available (Oliveira, Sao‐Pedro, Santos‐Barrera, Penone, & Costa, ; Ribeiro et al, ), but their use in amphibian community ecology still relies on a few traditional, nonintegrative approaches.…”

Vanishing refuge? Testing the forest refuge hypothesis in coastal East Africa using genome‐wide sequence data for seven amphibians

“…Finally, we summed models over multiple time periods to visualize the potential spatio‐temporal distributions of each taxon. We collected all available locality information per taxon using the Global Biodiversity Information Facility (GBIF), published data (Barratt et al., 2017a; Burgess & Clarke, ; Ohler & Frétey, ) and our own and collaborator's fieldwork. Data were filtered to remove any imprecise or ambiguous localities, or points that could not be accurately matched to each taxon with certainty.…”

“…Previous species richness and endemism studies in the CFEA have recognized the existence of a biogeographic division situated between the northern (Zanzibar) and southern (Inhambane) Zanzibar-Inhambane ecoregions (Azeria et al, 2007;Burgess et al, 1992Burgess et al, , 1998Burgess et al, , 2004 Barratt et al, 2017aBarratt et al, , 2017bBwong et al, 2017). In East Africa, such patterns are often associated with vicariant diversification through a forest refuge model of speciation (Endler, 1982;Haffer, 1969Haffer, , 1997Mayr & O'Hara, 1986;Moreau, 1954;Moritz et al, 2000;Plana, 2004) or attributed to ecological change (Lorenzen et al, 2012).…”

“…Historical environmental change across tropical East Africa has been frequent since the Miocene, and the current CFEA are considered to be the remnants of a once continuous forest that has expanded and contracted for the past 40 million years (Axelrod & Raven, ; Demenocal, ; Maslin et al., ; Mumbi, Marchant, Hooghiemstra, & Wooller, ). Combined knowledge of endemism patterns and environmental change have led to the assumption that current CFEA biodiversity mainly originated from the isolation and persistence of ancient lineages in forest refugia, with local extinctions and in some cases adaptation to non‐forest habitats across the rest of the region (Azeria, Sanmartín, Ås, Carlson, & Burgess, ; Barratt et al., 2017a; Burgess et al., ). The coastal forests have thus been described as a “vanishing refuge” (Burgess et al., ), although to date forest refugial processes have not been thoroughly tested against alternative modes of diversification (Damasceno, Strangas, Carnaval, Rodrigues, & Moritz, ; Kirschel et al., ; Schneider et al., ; Zhen et al., ).…”

“…Here, we use nuclear DNA (nucDNA) single nucleotide polymorphism (SNP) data acquired from restriction site‐associated DNA sequencing (RAD‐seq), mitochondrial DNA (mtDNA) and georeferenced occurrences for seven widespread amphibians found in forest and non‐forest habitats across the CFEA. Although these seven taxa represent only a small proportion of the 55 known amphibians in this region (Barratt et al., 2017a), they cover a variety of life histories and ecological associations (from generalists to forest specialists), which should reflect the evolutionary processes that have occurred in this region. Our sampling represents almost the complete range of each of the study taxa, allowing us to test the forest refuge hypothesis against alternative modes of diversification, including allopatric divergence across landscape barriers (rivers and topographic) and parapatric divergence across ecological gradients.…”