Methods for Conservation and Ecology
I am an ecologist, interested in developing
methodology and theory to inform conservation decisions in a changing world
methodology and theory to inform conservation decisions in a changing world
Nothing in the world truly stands still. My research broadly focuses on measuring how the world is changing to understand where and how we should conserve species. I use a mixture of approaches in machine learning and AI to answer questions in relation to conservation decision-making, connectivity, migration and population ecology.
|
Machine Learning and AIThe increases in data availability and computer processing power have opened up opportunities for machine learning to revolutionise both ecology and conservation. I have worked on the use of Convolutional Neural Nets to classify meerkat calls into types (e.g. alarm call vs close call) to infer collective behaviour. I have also worked on segmenting seed x-rays (left image) to identify different parts of the seed for trait extraction. Finally, I have also worked on the use of reinforcement learning for protected area prioritisation and see a lot of potential for this work in adaptive conservation planning.
|
Conservation decision-makingBudgets are often restricted in conservation, and achieving the greatest benefit with limited funds is a crucial first step in minimising biodiversity loss. Managers must therefore make decisions about where, when, what and how frequently to carry out different conservation measures, despite uncertainties surrounding population size, where a species might be, and how effective the conservation intervention might be. I have therefore developped methods for optimising ranger patrols and optimising conservation in connected systems and identifying where and what species to collect for ex situ conservation.
|
|
|
Connectivity Conservation
Natural systems rarely stand still. Most species fluctuate not only in their distributions, but also in their movements - birds can fly, but also plants can have their seeds moved by the wind. Considering the connectivity of populations is essential to their conservation. Site loss will not only impact that site, but all the sites that are connected to it. Reserve systems are prone to failure where connectivity between sites is not accounted for. I have developped code for modelling networks and predicting population change under habitat loss and investigated this in a shorebird case study.
|
Animal migration and trackingRoughly a quarter of animal species migrate, but we still know remarkably little about where they go. I have done extensive work mapping migration and behaviour using multi-sensor gelocation methods, monitoring group dynamics during migration and developping R packages for tracking behaviour during the migratory cycle.
|
|
|
Population EcologyAnimal populations fluctuate both in time and space in response not only to each other, but also in response to their environment. I have been involved in theoretical modelling of invasive species spread, coral reef resilience and in modelling population trends in unmarked populations of shorebirds and dragonflies.
|
