Field performance of a nephelometer in rural kitchens: effects of high humidity excursions and correlations to gravimetric analyses

Rural kitchens of solid-fuel burning households constitute the microenvironment responsible for the majority of human exposures to health-damaging air pollutants, particularly respirable particles and carbon monoxide. Portable nephelometers facilitate cheaper, more precise, time-resolved characterization of particles in rural homes than are attainable by gravitational methods alone. However, field performance of nephelometers must contend with aerosols that are highly variable in terms of chemical content, size, and relative humidity. Previous field validations of nephelometer performance in residential settings explore relatively low particle concentrations, with the vast majority of 24-h average gravitational PM2.5 concentrations falling below 40 mu g/m(3). We investigate relationships between 24-h gravitational particle measurements and nephelometric data logged by the personal DataRAM(pDR) in highly polluted rural Chinese kitchens, where gravitationally determined 24-h average respirable particle concentrations were as high as 700 mu g/m(3). We find that where relative humidity remained below 95%, nephelometric response was strongly linear despite complex mixtures of aerosols and variable ambient conditions. Where 95% relative humidity was exceeded for even a brief duration, nephelometrically determined 24-h mean particle concentrations were nonsystematically distorted relative to gravitational data, and neither concurrent relative humidity measurements nor use of robust statistical measures of central tendency offered means of correction. This nonsystematic distortion is particularly problematic for rural exposure assessment studies, which emphasize upper quantiles of time-resolved particle measurements within 24-h samples. Precise, accurate interpretation of nephelometrically resolved short-term particle concentrations requires calibration based on short-term gravitational sampling.