Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Australie → Optimal fire management for biodiversity conservation

University of Melbourne (2016)

Optimal fire management for biodiversity conservation

Giljohann, Katherine M

Titre : Optimal fire management for biodiversity conservation

Auteur : Giljohann, Katherine M

Université de soutenance : University of Melbourne

Grade : Doctor of Philosophy (PhD) 2016

Stochastic disturbances are important natural events shaping the structure and dynamics of landscapes, ecological communities and populations. Despite being integral to many ecosystems, they present a significant source of uncertainty for the conservation of biodiversity. Species are adapted to certain disturbance regimes, yet these are often not well-known. Additionally, many large-scale disturbance regimes are changing across the world due to anthropogenic impacts on ecological systems and climatic changes. Modified disturbance regimes are likely to alter species responses ; posing an additional challenge to management. Methods and tools for accounting and dealing with uncertainties are needed to support management decisions in disturbance-prone systems. Using fire as a model disturbance type and the semi-arid mallee region of southern Australia as a study system, this thesis investigates methods for managing disturbance-prone ecosystems for biodiversity conservation. The research focuses on uncertainties that arise from stochasticity in future disturbances, technical complexities in the setting and application of conservation-based management objectives, and the level of detail required for realistic predictions of population dynamics in the presence of multiple disturbance types and for disturbance-sensitive species.
Five major findings of this thesis apply to different aspects of managing disturbance-prone ecosystems for conservation. Three of the findings relate to the management of multiple species at the landscape scale. First, a systematic comparison of four management objectives, all with an over-arching aim to maximise multiple species persistence, highlighted that the optimal management action was sensitive to how the objective was specified. This result demonstrated the need to consider the underlying biases of different quantitative objectives carefully when specifying conservation-based objectives. Management objectives based on biodiversity indices enable the requirements of multiple species to be represented in a coherent form. The second major finding was through exploring the sensitivity of a biodiversity index, the geometric mean of species’ relative abundance, to different types of species data used to calculate the index, to how successional changes in habitat are represented and how different species are valued. The exploration revealed a high sensitivity of the index to the choice of data that underpins it ; which in turn influenced conservation decisions. Awareness of the sensitivities of biodiversity indices will improve environmental decision making. Third, an evaluation of future differences in species extinction risk under a range of fixed targets for hazard reduction burning provided a robust basis for discriminating between alternative policies. This demonstrated that carefully considering potential trade-offs between conservation objectives and other values can assist more effective and efficient management plans.
The last two main findings of this thesis relate to the management of individual species. A life-cycle analysis of a perennial resprouting hummock grass, Triodia scariosa, revealed that although highly resilient to naturally occurring disturbances and variable climatic conditions, an interaction between below-average rainfall, prescribed fire and herbivory significantly reduced post-fire survival in both seedlings and resprouting individuals, whereas herbivory following wildfire had a minor effect on survival despite variation in post-fire rainfall. As interactions between disturbances and other ecosystem drivers can induce unexpected outcomes, effective management requires an integrated approach that simultaneously considers multiple ecosystem drivers. Lastly, evaluating the utility of a demographically equivalent population ratio in a case study of the serotinous fire-killed tree Callitris verrucosa, revealed that monitoring only the intervals between disturbances will not necessarily be sufficient to ensure population persistence in disturbance-sensitive species ; as persistence depends on the state of the population at the time of the disturbance. This demonstrated that accounting for population structure can lead to greater realism when predicting future population trends and more effective conservation decisions for disturbance-sensitive species.
Stochastic disturbances, like fire, have inherent uncertainty that can never be eliminated. Accounting for stochasticity in ecosystem dynamics assists informed decision making and the conservation of biodiversity. The work presented in this thesis offers new ecological insights and decision support tools for environments subject to stochastic large-scale disturbances. Four general principles for understanding disturbance-prone ecosystems emerge : the need to understand the state of the system, to account for multiple uncertainties, explore trade-offs in decision making and take an adaptive approach to management. Careful consideration of these principles will contribute to more realistic and well-informed conservation decisions in disturbance-prone environments.


Page publiée le 12 juillet 2017