Bouwe Reijenga

@breijenga.bsky.social

437 Followers 481 Following

(Macro)evolutionary biologist • postdoc @OxUniEarthSci from fossils, phylogenies and theory to community assembly and diversification trends

Relationship between phylogenetic metrics and assemblage species richness for real bird clades and model simulations. Colours and rows denote the empirical relationships for all clades (orange) and relationships simulated under the ‘historical variable rate’ (yellow), ‘historical global rate’ (blue) colonisation rate and ‘non-historical’ (dark blue) model. For the simulation scenarios the mean proportion of co-occurring species across 2500 replicated simulations are plotted. Columns represent the five phylogenetic metrics capturing different aspects of clade evolutionary history: The crown age of the clade, phylogenetic imbalance (Colless' index), mean branch length of the extant species, ρ a measure of temporal change in diversification rates through time where positive (negative) values indicate an increase (decrease) and log-transformed clade species richness. Shaded areas represent 95% confidence intervals of fitted generalised linear models with quasibinomial error distribution and logit link function weighted for clade richness. The proportion of co-occurring species correlates positively with clade age, shows no correlation with phylogenetic imbalance, a positive correlation with the mean age of species, no correlation with changes in diversification rate, and a slight negative relationship with clade species richness. The two historical models predict these patterns well, but the non-historical model predicts none.
Observed variation in assemblage richness of passerine clades compared to model predictions is shown between three model predictions: A historical model with variable colonisation rates between clades, a historical model with a global rate across clades, and a non-historical model in which colonisation is assumed to be near instantaneous after speciation. The three plots follow a standard structure where the proportion of co-occuring species across the clades is plotted between the model prediction and the empriical values, i.e. as per the figure legend: Assemblage richness is plotted as a % of the total number of species in each clade. Model predictions are shown for (a) the ‘historical variable rate’, (b) ‘historical global rate’, and (c) the ‘non-historical’ model. A 1:1 line shows where the empirical the predicted richness matches. Bars represent the 95% confidence intervals of predicted assemblage richness from 2500 replicate simulations of each model. Colours indicate if empirical assemblages fall within (blue) or outside (red) of model expectations. The results show that the historical variable rate model captures all clades, the historical global model predicts 21/25 and the non-historical model predicts 16/25 clades but shows very little variation between predicted proportional richness.
Visualisation of how speciation history influence assemblage richness. (a) Conceptualisation of the link between allopatric speciation, range expansion and assemblage richness. Upward-pointing arrows show timing of events (speciation, colonisation, and local extinction) on a hypothetical phylogentic tree that shows two speciation events. When speciation occurs the ancestral geographic range (red rectangle) is divided into two (yellow and green) and at most one species will be present (λ1) in the local assemblage (circle). Upon the second speciation event, the ancestor has gone locally extinct. Neither of the descendants (blue and purple) will therefore be present (λ0) in the local assemblage (open circles). The build-up of local assemblage thus depends on the colonisation of species (γ), as well as local extinction (μ), as species undergo expansion (green) or contractions (yellow) of their geographic range. (b) Two hypothetical clades, each with four species (similar design as above), with the top clade having older species on average than the bottom clade. Filled circles highlight the presence of a lineage in the local assemblage and arrows indicate colonisation events. For both clade the ancestor was present in the assemblage, but the dynamics of allopatric speciation and colonisation result in a higher assemblage richness at present in the clade with older species.
Species richness patterns of exemplary clades. (a, c, e) Assemblage species richness (as a proportion of clade richness) of three exemplar avian families (Passerellidae, Dendrocolaptidae & Onychorhynchidae) showing low, intermediate and high levels of co-occurrence across the Americas. Pie charts show clade richness (decreasing in size in order) and the black lines point toward the grid cell of highest richness all in different locations in the Americas. (b) The relationship between clade richness and maximum local species richness (n = 40 families). An ordinary least squares regression estimates an intercept of −0.257 and slope of 0.761. (d) Distribution of maximum (proportional) local species richness across clades, showing large variation between clades.
a, Reproductive isolation is greater for populations that have been evolving in different environments (between environments, red) compared with populations that have been evolving in the same environment (within environments, yellow). k indicates the number of effects sizes extracted with the number of studies in parentheses. The thick black lines are the 95% CIs and the thin black lines are the prediction intervals. b, However, the magnitude of reproductive isolation is independent of the number of generations of the experimental evolution study. Reproductive isolation is z-transformed. Each point is an effect size (yellow diamonds for within-population crosses and red circles for between-population crosses). The size of the point indicates the precision (1/s.e.m.) of the effect size, as shown in the legend in a. The lines represent the posterior predicted mean of reproductive isolation, with the shaded area indicating the 95% Cl (dotted yellow lines are within-population crosses, and solid red lines are between-population crosses).
Figure 2 from Jouault et al. in Current Biology. The figure consists of 4 x 3 graphs in which each row shows the diversification (origination, extinction, net diversification, and diversity) history of a group of hymenoptera. Emphasis is put on the effect that the rise of angiosperms and the KPg mass extinction event have had on the complex histories of hymenoptera, showing sharp increases in diversity after the K/Pg
The final figure in the set, showing what happens from the Danian to the Selandian-Thanetian and Ypressian. Only millions of years later in the Selandian and Thanetian, the intercept and local diversity increase and remain stable throughout and in the Ypressian.
Similar to the previous post, we add how the SAR has changed from the Maastrichtian to the Danian, post-K/Pg. Immediately Post-K/Pg in the Danian, and perhaps contrasting ideas of broadly ranging disaster taxa, the SAR slope and therefore turnover and regionalisation of biota appears to increase dramatically.
The figure shows the first results in which we show how the species-area relationship has changed from the Campanian to the Maastrichtian. The top row of graphs shows the data point plotted between geographical area size, measured as the distance covered by a minimum spannign tree connecting all localities, and the species richness corrected for sampling intensity. The bottom row shows the intercept and slope values plotted through time, caculated by fitting a ordinary least squares regression. The figures show that Pre-K/Pg in the Maastrichtian, we find that terrestrial tetrapods in general show a decline in regional diversity and turnover across space. The decrease in regional diversity is driven by a decrease in intercept for nonavian dinosaurs, but all groups such as mammals show a decrease in slope.
Methods description in which we highlight the usage of the paleobiology database from which we compiled a dataset of 14887 North American terrestrial tetrapod fossil occurrences, covering 3766 localities and 1661 species. We focus on Dinosaurs, mammals, squamates, turtles and crocodiles, across six geological stages spanning roughly 40 milion years. We validate our approach with spatial explicit neutral models, described in the supplementary material of the paper.
Description of species area relationships where increases in the amount of sampled area increase the number of species found. We highlight that changes in the intercept of this relationship can be roughly interpreted as changes in the number of species found locally, or the smallest measured site. Secondly, changes in slope can be roughly interpreted as a change in beta diversity, or how the composition of communities changes with increasing distance between sampled sites.
Figure from Close et al. (2020) Proc B. The figure shows how species richness of terrestrial tetrapods changes from the Permian till the present. Specifically, while diversity appears constant from the Permian until the end of the Paleogene, diversity appears to reach a new, higher equilibrium after the K/Pg mass extinction
(A) and (C) show mean log-transformed genus-level origination rates (blue), and (B) and (D) show mean extinction (red) rates plotted against log clade duration for 245 marine orders sampled across the Phanerozoic. Rates in (A) and (B) are calculated via Foote’s per capita metric, and those for (C) and (D) are calculated by Alroy’s second-for-third metric that correct for sampling. Rates are shown as the mean number of events per interval, per lineage and are not normalized for interval length. Colored lines represent ordinary least-squares regression fits between log rate and log duration, and shaded areas show 95% confidence intervals. β corresponds to the standardized regression coefficient that is equal to the slope of the relationship between rate and duration when their variance has been standardized to 1. Overall the figure shows that A and B show significant, strong negative relationships between rate and duration. However, when corrected for uneven sampling in C and D, only origination shows a very weak significant trend