A Wide Brownian Land: Drought Adaptation and Niche Diversification in an Australian Conifer

Abstract

Plants cover most of earth’s landmasses. Their ability to adapt to changing climates through evolutionary time has formed the distinctive biomes we see today. In plants, maintaining water supply from the roots to the leaves is critical because leaves are the sites of photosynthesis, where the sugars necessary for growth are produced. The failure of a plant’s hydraulic system can ultimately lead to death, so drought imposes a high selection pressure on plant traits. Species adapted to dry and arid climates have evolved traits that maintain the water supply critical to life. If species cannot survive changing climatic conditions they risk extinction, and drought therefore also poses a serious threat to species diversity. Plant lineages that have evolved drought-tolerance often have higher diversification rates because they can ‘radiate’ into new arid niches that most other species cannot exist in. Callitris is an ecologically diverse, cupressoid conifer genus native to Australia and New Caledonia. Drought-adapted Callitris species are the world’s most drought-tolerant trees. No other trees can operate under such water-tensions without suffering significant injury or mortality. Additionally, drought-adapted Callitris species do not experience a hydraulic conductance-saftey trade-off. The traits related to an absence of the hydraulic trade-off has not been identified. Callitris therefore presents an ideal study group to investigate trait-related radiations into droughted environments. Callitris species possess ‘callitroid thickenings’ in their water-conducting tracheids, but drought-adapted Callitris have a much higher frequency of thickenings than Callitris from wet habitats, suggesting that callitroid thickenings are a trait critical to drought-tolerance. This thesis investigates the evolution of drought-tolerance in Callitris from a macroevolutionary perspective. I used data from R.D. Heady’s (1997) SEM study of callitroid thickening and a new phylogeny of the southern hemisphere component of the Cupressaceae, the Callitroideae, to investigate the evolution of climatic niches and drought-tolerance traits in Callitris. The trait evolution studies emphasised a weakness in phylogenetic comparative methods (PCMs), leading to a simulation-based study that investigated the effect of extinctions on tree reconstruction and subsequently on PCMs. The first chapter in this thesis investigates the evolution of hydraulic drought-tolerance traits in Callitris by comparing the evolution of key hydraulic traits. I expected the frequency of callitroid thickenings (FCT) to be evolving under selection because of its strong ecological signal. Instead I found that FCT fitted a Brownian Motion (BM) model, suggesting evolution via drift. However, I also found that FCT was convergent with arid habitats, revealing that FCT appears to be under selection. The seemingly contradictory pattern of a trait under selection fitting a model of evolutionary drift suggests that FCT may be critical to drought-adaptation and speciation. FCT might enable high conductance and high safety by reinforcing pit borders, rather than by reducing pit apertures. Chapter two investigates climatic niche evolution in the Callitroideae to ask whether the Callitris has evolved into a drier niche than its closest relatives and if the evolution of arid niches has led to higher diversification in Callitris. I found that Callitris has a drier, hotter and flatter niche than the rest of the Callitroideae (RoC). Sister-pair comparisons showed that drought-adapted Callitris had divergent niches as a result of geographic fragmentation. Despite extraordinary drought-tolerance Callitris had a lower diversification rate than the RoC, suggesting high selection and recent extinctions. Range fragmentation appears to have been the result of intensifying aridification, leading to local adaptation and geographic speciation. Finally, chapter three investigates the effect of tree reconstructions on PCMs. I simulated phylogenies under four random extinction scenarios and traits under BM and Ornstein–Uhlenbeck (OU) models. Using extant only taxa from the simulations, I reconstructed phylogenies in BEAST. I found that the ‘slowdown’ in lineages-through-time plots was only observed in the reconstructed trees, suggesting it is an artefact of phylogenetic reconstruction methods. Reconstructed trees produced higher phylogenetic signal, erroneous ancestral trait reconstructions and incorrect inferences of the evolutionary model. In particular, traits simulated under a BM model were inferred as evolving under Early Burst models and traits simulated under an OU model were inferred as fitting a BM model. These results show that PCMs are heavily influenced by tree shape, which can lead to the erroneous interpretation of evolutionary histories. They also explain how traits under selection and associated with geographic speciation could fit a BM model. Show more