Supervisor

 S. A. Sisson (Statistics) and Richard Kingsford (BEES)

 Availability

 Now, until taken.

The Macquarie river is a major river of the Murray-Darling Basin. Its catchment begins in the Great Dividing Range, around Mudgee, Bathurst and Orange, and then the river flows northwest with most of it ending in one of the most important wetland systems in Australia - the Macquarie Marshes. About 80% of river flows originate in the Great Dividing Range. Major changes have occurred to river flows as a result of river regulation and diversions of water upstream from the Macquarie Marshes.

 

This project would adopt practical Bayesian methods (e.g. using WinBUGS) with the aim of determining how the river has changed as a result of river regulation and in particular the effects on river flows at different points in the river. The project would investigate relationships between rainfall and river flows before and after river regulation.

 

Objective: Determine Bayesian statistical models relating annual rainfall to river flows before and after regulation for two different parts of the river: the mid-section and before the Macquarie Marshes. Question: WHat is the nature of the effect of river regulation?

 

Data: Rainfall data - more than 100 years available! River flow data - gauge data; available for a number of different places on the river; usually about 15-30 years before regulation and 70-90 uyears after river regulation. Key sites: Dubbo, Warren and Oxley.

 

 

 Title

 Supervisor

 David Warton (Statistics) and Tracey Rogers (Australian Marine Mammal Research Centre)

 Availability

 Now, until taken.

Data are available via satellite of daily movements of eight seals, over 1-3 month periods. It is of interest to explore how seal movements relate to the locations of penguin colonies, and whether different seals express personal preferences for foraging near different penguin colonies. This project will involve modelling seal movement data, probably using a Markov modelling approach, and using the model to test specific hypotheses about seal movement in relation to penguin colonies. This project offers the opportunity to interact with applied scientists working in a non-UNSW institution.

 

 

 Title

 Supervisor

 S. A. Sisson and D. I. Warton

 Availability

 Session II 2006

It has recently become School policy that undergraduate courses are assessed by a number of class tests rather than take-home assignments to avoid problems of plagiarism. However, the effects of this choice as compared to alternative tutorial assessment schemes in terms of student learning and engagement in the subject is unclear.

 

Under the auspices of the University's Learning and Teaching Awards, this project will analyse data collected from all undergraduate statistics service courses over Session II, 2005 to Session I 2006 in order to determine the effect of assessment choice in tutorials in terms of student engagement with the material, and their understanding of the subject.

 

Given the official status of this project, this project has a small monetary stipend available. The goal is to analyse these data and provide evidence to support or contradict  current School practice in tutorial assessment.

 

 Supervisor

 David Warton

 Availability

 Now, until taken.

The study of species abundance distributions is a fundamental topic in ecology, and the identification of appropriate models for species abundance distributions have the potential to be very useful in modelling abundances in community ecology. This project will explore some possible approaches to species abundance modelling using hierarchical models based on the lognormal distribution, using an empirical Bayes approach.

 

This project offers the opportunity to work on an applied problem, learn new statistical techniques and develop skills using new statistical software.

 

 Supervisor

 David Warton (Statistics) and Dan Ramp (BEES)

 Availability

 Now, until taken.

 Description

 GPS data are available on the location of roadkills (deaths of wild animals due to collisions with road vehicles) along a 42km stretch of road on the Snowy Mountains Highway over a 2 year period. It is of interest to understand what factors increase the incidence of roadkill, and to identify "hotspots" along the road where there is a high risk of roadkill.

 

This project offers the opportunity to work on an applied problem, interact with biological researchers, learn new statistical techniques and develop skills using new statistical software.