Identifying Dutch elm disease ‘danger-spots’ on the Isle of Man with an agent-based model Bruce Mitchell a , Joana Barros b , Daniel Wendel c Department of Geography, Environment and Development Studies, Birkbeck, University of London, Malet Street, London, WC1E 7HX Telephone: (+440) 207 079 0644 Fax: (+44) 207 6316 498 Email a: bruce.birkbeck@ntlworld.com b: j.barros@bbk.ac.uk c: djwendel@gmail.com Keywords: Dutch elm disease control, agent-based modelling, local vulnerability 1. Introduction The paper presents an agent-based model (ABM) of the spread of Dutch Elm Disease (DED), applied to the Isle of Man (IoM). The objective is to provide the Manx forestry authority with a tool to support their DED control campaign. IoM has an estimated population of 250,000 elm trees, and is unusual in that DED – transmitted by the bark beetle Scolytus – is still being successfully fought. A strict control program was established soon after the disease arrived in 1992. So far, just over 1,000 trees have been lost, compared with around 30 million on the British mainland. DED has been studied since 1918, and its epidemiology and lifecycle are by now well known. However, most DED studies are essentially aspatial. Models of DED have been developed, but focus on either the biological aspects of the disease – (Castro and Bolke (2004)) or on the spread of the disease (Swinton and Gilligan, 1996). The present study proposes a three-dimensional spatial analytical agent-based-model (ABM) approach to identifying ‘danger-spots’ – locations on the IoM where outbreaks might lead to the greatest mortality among the Isle’s elm population. 2. The DED Model A prototype model was built in StarLogo TNG (MIT Media Laboratory, 2008), which supports modelling in three dimensions. This feature, together with the ease with which a model may be designed and implemented, makes StarLogo TNG particularly suitable for the project. Agent-based modelling was suitable for this project as it provided the opportunity to study the agents’ behaviours independently as well as collectively.