The geosimulation of West Nile virus propagation: a multi-agent and climate sensitive tool for risk management in public health
1 Department of Computer Sciences and Software Engineering, Laval University, Quebec, G1V 0A6, Canada
2 Centre for Research in Geomatics, Laval University, Quebec, G1V 0A6, Canada
3 Institut national de santé publique du Québec (INSPQ), 945 avenue Wolfe, Quebec, G1V 5B3, Canada
4 Centre hospitalier universitaire de Québec (CHUQ), 2705, boulevard Laurier, Quebec, G1V 4G2, Canada
International Journal of Health Geographics 2008, 7:35 doi:10.1186/1476-072X-7-35Published: 7 July 2008
Since 1999, the expansion of the West Nile virus (WNV) epizooty has led public health authorities to build and operate surveillance systems in North America. These systems are very useful to collect data, but cannot be used to forecast the probable spread of the virus in coming years. Such forecasts, if proven reliable, would permit preventive measures to be put into place at the appropriate level of expected risk and at the appropriate time. It is within this context that the Multi-Agent GeoSimulation approach has been selected to develop a system that simulates the interactions of populations of mosquitoes and birds over space and time in relation to the spread and transmission of WNV. This simulation takes place in a virtual mapping environment representing a large administrative territory (e.g. province, state) and carried out under various climate scenarios in order to simulate the effects of vector control measures such as larviciding at scales of 1/20 000 or smaller.
After setting some hypotheses, a conceptual model and system architecture were developed to describe the population dynamics and interactions of mosquitoes (genus Culex) and American crows, which were chosen as the main actors in the simulation. Based on a mathematical compartment model used to simulate the population dynamics, an operational prototype was developed for the Southern part of Quebec (Canada). The system allows users to modify the parameters of the model, to select various climate and larviciding scenarios, to visualize on a digital map the progression (on a weekly or daily basis) of the infection in and around the crows' roosts and to generate graphs showing the evolution of the populations. The basic units for visualisation are municipalities.
In all likelihood this system might be used to support short term decision-making related to WNV vector control measures, including the use of larvicides, according to climatic scenarios. Once fully calibrated in several real-life contexts, this promising approach opens the door to the study and management of other zoonotic diseases such as Lyme disease.