THE USE OF INDIVIDUAL-BASED MODELS FOR STUDYING THE RESPONSE OF BIRD POPULATIONS TO LANDSCAPE CHANGE
Hazel R. Parry*1, Christopher J. Topping2, Nigel Boatman1, Marc Kennedy1 & Alistair Murray1
1 Central Science Laboratory, Sand Hutton, York, YO41 1LZ, UK
2 Department of Wildlife Ecology & Biodiversity, NERI, University of Aarhus, Kalø, Grenåvej 14, DK-8410 Rønde, Denmark
* Email: h.parry@csl.gov.uk
Predicting the response of birds to changes in their environment is an ongoing challenge. Individual-based models (IBMs) have the potential to provide useful insights but have not been widely used to date. Advantages of IBMs over more ‘traditional’ population-based or autoecological approaches include the ability to simulate the effects of new land use types with given properties and landscape structure (as model drivers operate at a lower level), examine impacts at multiple landscape scales, incorporate temporal variability in habitat suitability and relate population density fluctuations to underlying population dynamics.
In order for these advantages to be realised, and for IBMs to achieve their potential for informing policy development, IBM approaches need to be recognised and accepted by policy makers. The application of IBMs therefore needs to address policy needs for simplicity, clarity and a rapid response. A scoping study has been undertaken to investigate the key factors influencing the confidence of policy makers and those responsible for commissioning policy-orientated research in such techniques, and to provide relevant information to address the issues raised.
A case study addressing the implications of the abolition of set-aside for Skylarks in the UK is used to demonstrate the application of the model to policy-relevant issues, building on an existing model constructed by Topping et al. (2005). The validity of the model was tested by comparing the effects of changing land use as predicted by the model with data on Skylark population dynamics, to determine whether the population trends under different land uses in the model reflect what is known about the observed population response. We demonstrate methods with which to ‘validate’ such models, in particular comparisons with regional statistical data available from the Breeding Bird Survey (BTO), and use a novel methodology, BACCO (Bayesian Analysis of Computer Code Output), to analyse model uncertainty. We indicate ways in which such models may be better understood, analysed and made available (for example through a semantic grid service) for greater scientific and policy gain.
Once confidence in the model’s plausibility, sensitivity and validity was gained, model experiments were used to test the response of Skylarks to the removal of set-aside, as well as the within-year effects of spraying with non-selective herbicide in spring in preparation for a return to cropping, a practice commonly applied to set-aside in the UK.
Future work on the model will improve the suitability of the model for use in the UK by incorporating more UK specific parameters (at present a number of Denmark-specific parameters are used, e.g. for farm management, landscape etc.). We define which parameters need to be UK specific and show how this might then lead to the use of the model as a powerful tool for policy decision-making.
Topping, C.J., R.M. Sibly, H.R. Akçakaya, G.C. Smith & D.R. Crocker. 2005. Risk Assessment of UK Skylark Populations Using Life-History and Individual-Based Landscape Models. Ecotoxicology 14, 925–936
Hazel Parry works at the Central Science Laboratory, York, which is an executive agency of Defra (from 1 April this will be a new agency: ‘Fera’). She completed her PhD from the University of Leeds in 2006 and now has a permanent research position at this institution. Her research interests include: agent-based and individual-based modelling, computer simulation of crop pest population dynamics and dispersal, grid computing, geographical information systems, environmentally sensitive farming, and predicting the implications of policy and environmental change for agricultural landscapes and farmland ecology.
Hazel’s webpage: www.scispace.net/hazelp.
