Background
Air quality standards are typically based on long term averages – whereas a person may encounter exposure peaks throughout the day. Exposure peaks may contribute meaningfully to health impacts beyond their contribution to long term averages, and therefore should be considered alongside longer-term exposures. We aim to define and explain peak exposure to black carbon air pollution and look at the relationship between short peak exposures and longer term personal exposure.

Methods
A peak detection algorithm was applied to pooled data from two independent studies. High-resolution personal black carbon monitoring was performed in 175 healthy adult volunteers for a minimum of two 24-h periods per person. At the same time, we retrieved information on the time-activity pattern. Data covered Belgium, Spain, and the United Kingdom. In total, 2053 monitoring days were included.

Results
Exposure profiles revealed 2.8 ± 1.6 (avg ± SD) peaks per person per day. The average black carbon concentration during a peak was 4206 ng/m³. On 5.5% of the time participants were exposed to peak concentrations, but this contributed to 21.0% of their total exposure. The short time in transport (8%), was responsible for 32.7% of the peaks. 24.1% of the measurements in transport were categorized as peak exposure; while sleeping this was only 0.9%. When considering transport modes, participants were most likely to encounter peaks while cycling (34.0%). Most peaks were encountered at rush hour, from Monday through Friday, and in the cold season. Gender and age had no impact on the presence of peaks. Daily average black carbon exposure showed only a moderate correlation with peak frequency (r = 0.44). This correlation coefficient increased when considering longer term exposure to r > 0.60 from 10 days onward.

Conclusions
The occurrence of peaks varied substantially over time, across microenvironments and transport modes. Daily average exposure was moderately correlated with peak frequency. Real-time air pollution alerting systems may use the peak detection algorithm to support citizens in self-management of air pollution health effects.