ABSTRACT

I will present a general framework along with a computational platform for estimating human exposure to environmental hazards, focusing particularly on air pollution. The abstract framework, one that embraces a number of specific approaches that have been taken in the past, consists of modules that represent the determinants of exposure and these can to some extent be developed independently of one another in implementations of the framework. I will describe one particularly important implementation, a large computer model that accounts for the random behavior of individuals. That model, an adaption of the well-known pNEM model developed by the EPA, accounts for behaviour by sampling from 24-hour recall time-activity databases. On approval, individuals can access the model through their web-browsers. They can then construct online, their own versions of this model for an environmental hazard on interest, by specifying its parameters and uploading to the UBC site, any relevant datasets. I will show some outputs from that model for predicting PM10 exposures in Vancouver and London, and how it may be used by regulators in scenario analysis. I will also describe how it can be used in connection with other modules from the abstract framework, to enable the local levels of a hazard to be predicted using data from remote monitoring sites.