ABSTRACT
Vector-borne diseases have been and are currently a considerable concern for the public and the military as they have significant health and economic impacts. The ease of global travel and trade serve as means by which vectors can be introduced into novel areas, increasing disease risk. Predictive tools including mathematical models are essential in managing vector-borne disease risk but vary in their predictive power, ease of use, and the extent to which they have been compared to field data. To date, considerable effort has been expended on developing very detailed, computationally intensive, and spatially explicit computer simulation models to predict, for example, the abundance of mosquito vectors such as Aedes aegypti, a common vector for dengue fever. However, these highly complex models can be difficult to implement and parameterize requiring a great deal of data. While there is considerable value in continuing to develop highly detailed computer simulation models, here we explain and explore the use of simpler, spatially explicit rule-based models that can be used to quickly develop relative risk maps of vector abundance and disease risk. With additional parameterizations, these models may yield outcomes that accurately track mosquito abundance data. The modeling framework is implemented using ArcGIS software and is adaptable to the level and quality of data available. The rule-based approach posits that a generalized understanding of the ecological requirements for mosquitoes can be used to develop a spatially explicit model focused on suitable mosquito habitat. Combined with human census data, the highest relative risk of vector-borne diseases would lie where mosquito habitat suitability and human density were highest. This approach is quasi-mechanistic in that the ecological information incorporated into the model relates to the reproductive biology of the vector, although detailed data are not required. As more and better data become available, however, this information can be added to improve model predictions and confidence. Although we think detailed, computationally intensive vector models are relevant, in some cases a streamlined, rule-based approach may be more appropriate especially when data and time are limited. We provide an example of the application of this spatially explicit modeling approach to the mosquito vector Culex tarsalis in West Texas. Our approach can be broadly applied to a range of vectors or other risk factors and may be of use to municipalities or can have applications relevant to the military in foreign lands.
