Viel JF, Cardis E, Moissonnier M, de Seze R, Hours M. Radiofrequency exposure in the French general population: Band, time, location and activity variability. Environ Int. Aug 3, 2009 Ahead of print.
Information on radiofrequency (RF) exposure to individuals is scarce, and less is known about the relative importance of different sources of exposure. Existing RF sources operate in different frequency bands which can be subdivided in two broad categories: external sources (broadcast transmitters (radio, TV), or mobile phone base stations) and internal sources (mobile phones, in-house bases for cordless phones, or microwave ovens). The relative contribution of these sources to overall exposure depends on individual home and workplace circumstances. Up to now, crude proxies for exposure and possibly unrepresentative spot measurements have been used.
The goal of this survey was to use personal exposure meters to: assess RF exposure on a population basis; check its variability with time, location and activity; and clarify the relative contribution of different sources to total exposure.
The study took place in two French cities with their suburban areas and rural surroundings: Besançon (120,000 inhabitants), with a population dispersed widely in the surroundings, and a low density of RF sources; Lyon (1,200,000 inhabitants), second largest urban area in France with a high population density and numerous RF sources. A total of 398 people were enrolled into the study between December 2005 and September 2006. For exposure assessment, commercial personal exposure meters (PEM) were used, configured for a 13-second measurement period with a data collection period of 24h, yielding up to 6,643 data points. Subjects were asked to perform their routine tasks while wearing the PEM. Regression on order statistics were used to handle difficulties with measures below detection limits.
The highest proportion of non-detects were observed for total field (46.6%), cordless phones (17.2%), WiFi-microwave sources (14.1%), and FM transmitters (11.0%). Regarding peak exposures (assessed by the proportions of measurements above 1 V/m), the only sources of RF exposure were the subject's own mobile phone (GSM and DCS uplink), DECT phones, and microwave ovens. The total field mean value was 0.201 V/m, though this was higher in urban areas, during daytime, among adults, and when moving. Most of the time, recorded field strengths were not detectable with the exposure meter. When focusing on specific channels, the highest mean exposure resulted from FM sources (0.044 V/m), followed by WiFi-microwave (0.038 V/m), cordless phones (0.037 V/m), and mobile phones (UMTS uplink: 0.036 V/m, UMTS downlink: 0.037 V/m).
Discussion and Conclusion
This study is one of the first to assess personal band selective exposure in a large dataset, sampled from the general population, and informing about the most relevant exposure contributions in the everyday environment. Some limitations applied, namely that the presence of the human body can alter the patterns of wave propagation, and that mobile phone measurements depend heavily on the way the device is used. For these reasons, the authors did not attempt to determine the whole body specific absorption rates (SAR) from the electric fields measured. The strength of the statistical method used is its resistance against error due to the distribution of the data. However, the large proportion of censored values observed in the study warrants cautious interpretation of the results. Regarding peak exposures, individual devices (mobile phones, cordless phones, and microwave ovens) dominated the exposure, while the highest mean value was found for FM transmitters, confirming some preliminary results. Among mobile phone bands, UMTS signals had unexpectedly high values. Various exposure factors contributed to variability, namely region, place of residence (some of which can be explained by density of base stations), day versus night, adult versus youth, and outdoor versus indoor. It remains to be confirmed whether these observed differences are due to true exposure-relevant factors.