Evaluation of the use of empirical ambient ozone pollutant modeling and subject activity logs as an indirect measurement of ozone exposure

The personal ozone exposure of women on the track team at Rice University was monitored using Harvard passive samplers over a period of six weeks during August and September of 1997. Each subject logged their location and activity in and around campus during the exposure period. A three-dimensional kriging model of ozone was developed to estimate the ozone exposure for each subject from ambient fixed site monitoring data. The ozone predictions from the model were combined with the activity and location information, adjusted for indoor environment when applicable, to estimate personal exposure concentrations for each subject. Using two independent approaches, the kriging model was proven to provide accurate spatial and temporal estimates of ozone at subject exposure points. The results from this work show that the method developed to estimate exposure through kriging over (x,y,t) to predict ozone concentrations at exposure points combined with subject activity/time logs produces exposure estimates within the error bounds of the analytical methods for personal monitoring.