Pink-footed and Greylag geese winter in Britain and can cause damage to crops,
resulting in a conflict with agriculture. An understanding of where geese are likely to
feed would help to target suitable areas for goose management plans, aimed at
relieving such conflict. The aim of this project was to create models to predict the
feeding distribution of both Pink-footed and Greylag geese. Two separate approaches
were taken to model goose feeding distribution from landscape characteristics. The
first was a standard approach, logistic regression, which predicted the probability of a
field being used by geese from the field’s landscape characteristics. Models were
based on goose distribution data from field surveys. The main factors affecting field
choice by both species were distance from the nearest building and distance from the
roost. The inclusion of autologistic terms did not improve the fit of the models. A
second, more novel approach to predicting goose distribution was taken to see if more
accurate predictions could be produced. This modelling technique involved
simulating the movements of Greylag geese throughout the day. The rules
constraining goose movement in the model were derived from analysis of radiotracked
geese. Flight direction was constrained by altitude or distance from the river
while the probability of landing was dependent on the distance from buildings. The
accuracy of the models in predicting goose distribution was tested both within the
study area, Strathearn and Strathallan, and in another area, Loch Leven. Models
based on animal movements have the theoretical advantage of incorporating barriers
to movement, but the simulation model did not out-perform the logistic regression
model. The models can be applied to other goose feeding areas relatively easily and
can be used to identify areas where management plans for both Pink-footed and
Greylag geese should be targeted.