School of BioSciences - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemFrom little things big things grow - savanna burning, suppressed trees and escape from the fire trap in Australian mesic savannas( 2017)This thesis explores growth responses and strategies of fire-suppressed trees in mesic savannas. Frequent fires are common to savannas globally, and cause most savanna trees to remain trapped as resprouts in the understorey by a cycle of topkill, where all above-ground parts of the plant are killed, followed by resprouting. Escape of suppressed resprouts from this fire trap is reflected in savanna tree community structure and composition. In this thesis, I contribute to the growing body of work from across the savanna biome that seeks to unravel the different effects of fire, competition and species growth strategies as mechanisms driving savanna tree communities. This question is fundamental to understanding what limits tree biomass in savannas and to predicting effects of different fire regimes in both the short term and in future climate scenarios. Despite much argument and modelling, mechanistic drivers of mesic savannas remain topics of conjecture, in part due to historical, environmental and species trait differences between continents. I collected the data used throughout this thesis within the Tiwi Carbon Study, a nine-year long fire experiment that aimed to provide accounts of carbon stored in soils, live vegetation and dead biomass under different fire regimes. The carbon economy is becoming a significant economic contributor to Aboriginal communities across remote northern Australia, with associated human benefits of social empowerment, wellbeing and connection to traditional practices. The increased focus on active management of northern Australian savannas for carbon sequestration and emissions abatement within a carbon market provides a human perspective to the ecological focus of my thesis. Within this context, there is a renewed imperative to understand what limits trees in savannas to anticipate effects of changes to fire regimes on carbon stocks and biodiversity. Using individual-level data I collected for 11 common resprouting savanna tree species subjected to different fire regimes on the Tiwi Islands in monsoonal northern Australia, I: (1) develop a theoretic framework that describes persistence and escape of suppressed resprouts subjected to frequent fire; (2) develop novel methods for estimating species and fire-specific escape heights; (3) model resprout growth and escape from the fire trap as mediated by fire and competition; (4) define sapling growth strategies based on functional and architectural traits that may influence escape potential; and, (5) demonstrate the effects of varying fire frequencies on savanna structure and composition. I found that the likelihood of escape from the fire trap is context-specific and related to differences in fire intensity, species traits and topkill-avoidance. Fire promotes fast growth of trees compared to fire exclusion, which may promote higher escape rates over shorter timeframes. However, less frequent fire leads to increased midstorey densities overall, thus affecting stand structure. In Australian savannas, eucalypts receive particular attention because of their canopy dominance, but I found minimal evidence of distinctly different growth responses between eucalypt and non-eucalypt resprouts that might explain this. Fine-scale environmental variation and individual species characteristics must be considered for robust estimates of escape from the fire trap. My research further implicates non-fire disturbances and different reproductive strategies as potentially illuminating drivers of different species responses – important topics for future research.
-
ItemAssessing and managing interacting species at risk of coextinction( 2016)Interactions between organisms are ubiquitous: predators hunt prey, plants compete for light, and pollinators visit flowers to forage on nectar. Through their interactions species influence each other's population dynamics and ultimately their persistence: Darwin was already convinced that if bumblebees became extinct their food plants would follow quickly. Despite their importance, interactions are commonly ignored when we assess species' extinction risk or plan for their conservation management. My thesis is divided into six chapters, addressing two important components of conserving interdependent species. First, I assess if and how we can use a common type of data - observed interaction networks - to assess the coextinction risk of interacting species in networks, and to predict how interactions influence cascading extinctions when interdependent species are lost. Secondly, I investigate how interacting species can be protected in combined management approaches, focussing on the increasingly common method of translocating species for conservation. To answer this questions, I develop a range of statistical and mathematical modelling approaches and apply these to theoretical simulations and empirical data. In chapter 2, I investigate how quantitative methods can help to identify those species in interaction networks that are at risk of coextinction, while incorporating important factors such as uncertainty and imperfect detection of species in the field. I develop a hierarchical $N$-mixture model that accounts for imperfect detection and allows one to disentangle two factors that influence interaction frequencies between species: the probability that two species interact, and the abundances of species. This enables one to estimate with uncertainty the number of interaction partners of a species and the community size of dependents. I fit the model to data that from simulations of different parameter scenarios and to empirical networks of flower-visiting insects found on a threatened ecological community of plants from the Stirling Ranges National Park in Western Australia. In chapter 3, I extend this modelling approach to investigate how imperfect detection and uncertainty influence the progression of extinction through mutualistic networks. Therefore, I apply the modelling approach from chapter 2 to observed networks to correct these networks for sampling bias. Then, I sequentially remove plant species from the networks to investigate how extinction cascades differ between observed and corrected networks. I show that networks corrected for sampling bias, are more densely connected and the interactions between species are more diffusely distributed throughout the networks. This causes corrected networks to be less specialised, and plant species to be more redundant, leading to increased network robustness. The results of chapter 2 and 3 indicate that imperfect detection strongly affects observed interaction networks and suggests that it is unwise to draw strong inferences for the conservation status of species and the robustness of ecosystems without acknowledging imperfect detection and uncertainty. In the second part of this thesis, I investigate management actions for improving the persistence of cothreatened interacting species, with a particular focus on conservation translocations. The fourth chapter investigates how useful current single-species translocation guidelines are for conserving cothreatened species and the interactions between them. I first classify potential systems of cothreatened species and devise appropriate management options for each system. Secondly, I extend current single-species guidelines to incorporate interactions in the assessment, planning and implementation phase for the conservation of multiple interacting species. For each phase of a translocation, I present case studies of threatened interacting species where a combined translocation could save the species. In chapter 5, I examine in detail how different types of interactions influence the optimal size of founder populations and the order in which interacting species should be translocated. I use mathematical models for coupled two-species systems, in which species interact in consumer-resource, competitive or mutualistic interactions. While some common rules in translocating interacting species emerge, most decisions about necessary founder sizes and translocation order are interaction-type specific. In the two chapters about combined translocations of cothreatened species, I show that interspecific interactions are important processes that shape population dynamics, and should therefore be incorporated into the quantitative planning of multi-species translocations. Finally in chapter 6, I synthesise the findings of my work and highlight future research avenues.