Integrating GIS approaches with geographic profiling as a novel conservation tool
Abstract
Geographic profiling (GP) was originally developed to solve the problem of
information overload when dealing with cases of serial crime. In criminology, the
model uses spatial data relating to the locations of connected crimes to prioritise the
search for the criminal’s anchor point (usually a home or workplace), and is extremely
successful in this field. Previous work has shown how the same approach can be
adapted to biological data, but to date the model has assumed a spatially homogenous
landscape, and has made no attempt to integrate more complex spatial information (eg,
altitude, land use). It is this issue that I address here. In addition, I show for the first
time how the model can be applied to conservation data and – taking the model back
to its origins in criminology – to wildlife crime. In Chapter 2, I use the Dirichlet
Process Mixture (DPM) model of geographic profiling to locate sleep trees for tarsiers
in dense jungle in Indonesia, using as input the locations at which calls were recorded,
demonstrating how the model can be applied to locating the nests, dens or roosts of
other elusive animals and potentially improving estimates of population size, with
important implications for management of both species and habitats. In Chapter 3, I
show how spatial information in the form of citizen science could be used to improve
a study of invasive mink in the Hebrides. In Chapter 4, I turn to the issue of ‘commuter
crime’ in a study of poaching in Savé Valley Conservancy (SVC) in Zimbabwe, in
which although poaching occurs inside SVC the majority of poachers live outside,
showing how the model can be adjusted to reflect a simple binary classification of the
landscape (inside or outside SVC). Finally, in Chapter 5, I combine more complex
land use information (estimates of farm density) with the GP model to improve
predictions of human-wildlife conflict.
Authors
Faulkner, SallyCollections
- Theses [4121]