1 Data-ontsluiting en analyse: radiofrequente elektromagnetische veldenJohn Bolte (RIVM), Tessa Eikelboom (IVM, VU) Jos Kamer and Gerard van der Zande (Agentschap Telecom)

2 Elektromagnetisch spectrumRadiofrequente velden: 300 Hz – 300 GHz Belangrijkste bekende kortetermijn gezondheidseffecten hangen samen met opwarming Ook andere niet-thermische effecten? John Bolte RIVM & THUAS 25 January 2017

3 Exposure in cohorts: Questionnaires & Prediction modelstime-activity patterns building (materials) electrical equipment mobile phone / internet use self-reported non-specific physical symptoms / adverse health effects John Bolte RIVM & THUAS 25 January 2017

4 measurement set gps ELF magnetic radiofrequencyJohn Bolte RIVM & THUAS 25 January 2017

5 Electric E-field activities in city centreTrain Bike Inside shop Walk Home Railway station John Bolte RIVM & THUAS 25 January 2017

6 Ontsluiting: Biases en nauwkeurigheidBolte, JFB, Lessons learnt on biases and uncertainties in personal exposure measurement surveys of radiofrequency electromagnetic fields with exposimeters. Environ. Int.. 94: Review.    Bolte JFB, Maslanyj M, Addison D, Mee T, Kamer J, Colussi L, Do car-mounted mobile measurements used for radio-frequency spectrum regulation have an application for exposure assessments in epidemiological studies? Environ. Int. 86: 75–83. Bolte J.F.B., Van der Zande G., Kamer J., Calibration and uncertainties in personal exposure measurements of radiofrequency electromagnetic fields. Bioelectromagnetics 32(8): 6 John Bolte RIVM & THUAS 25 January 2017

7 Coupling: format - Date:59: ,6 0,05 0,05 0,05 0,05 0,05 0,05 0,05 :59: ,5 0,05 0,05 0,05 0,05 0,05 0,05 0,05 /13/09 0:00: ,6 0,05 0,05 0,05 0,05 0,05 0,05 0,05 06/13/09 0:00: ,5 0,05 0,05 0,05 0,05 0,05 0,05 0,05 John Bolte RIVM & THUAS 25 January 2017

8 Electric E-field activities in city centreTrain Bike Inside shop Walk Home Railway station Train Bike Inside shop Walk Home Railway station John Bolte RIVM & THUAS 25 January 2017

9 Sources of biases and uncertaintiesMechanical errors (shock or fall) Measurement process due to hardware and software filters (manufacturer calibration, multiple signals detection in one band, modulation/duty factor sensitivity, non-detects through detection limit, resolution, out of band response aka cross talk, flatness of response within a frequency band, linearity in response, temperature and humidity sensitivity, sampling interval, repeatability of indicator, temporal drift) Anisotropy (hor. and vert, angle of incidence, and polarization) Influence of the body Temporal and spatial variability Repeatability John Bolte RIVM & THUAS 25 January 2017

10 Mechanical error: clippingJohn Bolte RIVM & THUAS 25 January 2017

11 Uncertainties in measurement valuesWhen do we meet compliance? Directive 2013/35/EU mentions: John Bolte RIVM & THUAS 25 January 2017

12 Do all devices measure the same field?Calibration: remove systematic biases GHz Transverse ElectroMagnetic (GTEM) cell Do all devices measure the same field? John Bolte RIVM & THUAS 25 January 2017

13 Variability per unit and frequency bandInput = 2.5 V/m Correction factors (input/response) typically 0.7 – 1.5 unit 1: broke, repaired unit 2: tends to give higher registrations than rest John Bolte RIVM & THUAS 25 January 2017

14 Linearity (trendlines)John Bolte RIVM & THUAS 25 January 2017

15 Anisotropy: Elevation angle of arrivalcorrection: John Bolte - RIVM E-Monument 4th meeting PHE, Chilton, UK July 30, 2014

16 Open Area Test Site (OATS): influence of the bodyJohn Bolte RIVM & THUAS 25 January 2017

17 Influence of the body (linear V/m scale)Horizontal Polarisation Input 2.5 V/m V/m 1 2 3 4 5 6 0º 45º 90º 135º 180º 225º 270º 315º input 2,5 V/m FM (V/m) TV3 (V/m) TV4&5 (V/m) GSMrx (DL) (V/m) DCSrx (DL) (V/m) UMTStx (UL) (V/m) UMTSrx (DL) (V/m) WIFI (V/m) Vertical Polarisation Input 2.5 V/m 1 2 3 4 5 0º 45º 90º 135º 180º 225º 270º 315º input 2,5 V/m FM (V/m) TV3 (V/m) TV4&5 (V/m) GSMrx (DL) (V/m) DCSrx (DL) (V/m) UMTStx (UL) (V/m) UMTSrx (DL) (V/m) WIFI (V/m) Influence of the body (linear V/m scale) unit 7 John Bolte RIVM & THUAS 25 January 2017

18 body shielding and polarisationBolte et al. , 2011, Bioelectromagnetics John Bolte RIVM & THUAS 25 January 2017

19 Spy 7 Spy 6 Spy 4511 Spy 1207 GPS 121 Spy 9 RFeyes MaschekJohn Bolte - RIVM E-Monument 4th meeting PHE, Chilton, UK July 30, 2014

20 Other measurement devices, other measurementsDepending on position on body Nearest neighbour projection on route for GSM in Industrial Area, Amersfoort, NL Cumulative distance along route John Bolte RIVM & THUAS 25 January 2017

21 Reproducibility, Temporal and Spatial variationBeekhuizen et al., 2013 Cumulative Distance (m) John Bolte RIVM & THUAS 25 January 2017

22 Analyse Bolte J, Eikelboom T, Personal radiofrequency electromagnetic field measurements in the Netherlands: exposure level and variability for everyday activities, times of day and types of area. Env. Internat 48: – DOI: /j.envint 22 John Bolte RIVM & THUAS 25 January 2017

23 TWA Power density (mW/m2): Transporthr/Nses/pax N=98, balanced: sex, age, ses, work, housing type, degree of urbanisation John Bolte RIVM & THUAS 25 January 2017

24 TWA Power density: At home - insidehr/Nses/pax N=98, balanced: sex, age, ses, work, housing type, degree of urbanisation John Bolte RIVM & THUAS 25 January 2017

25 Temporal exposure N=98, balanced: sex, age, ses, employment, type of house, degree of urbanisation John Bolte RIVM & THUAS 25 January 2017

26 John Bolte RIVM & THUAS 25 January 2017

27 Distribution of session means per activity Transport - DCS downlinkJohn Bolte RIVM & THUAS 25 January 2017

28 Measured vs. Prediction (model +Questionnaire) DCS downlinkmeasured log(power density) rSpearman = .23 <50% 50-90% >90% predicted log(power density) John Bolte RIVM & THUAS 25 January 2017

29 Exposure distribution per demographic featureJohn Bolte RIVM & THUAS 25 January 2017

30 Concluding remarks Characterisation of everyday activities based on temporal and spatial analysis is possible Contrasting everyday activities can be found Correction factors for different response of 11 RF exposimeters necessary, based on: Out of band response test software filter / Precision within band fin=unknown error margin  Influence of body motion average? / Polarisation reflections average? / Multiple signals / sources no detect unknown error   Response & linearity constant factor input/response  Do these features change over time??? John Bolte RIVM & THUAS 25 January 2017

31 Which exposure metric is relevant for effect?example: magnetic fluc density (B field) Essence of exposure: emphasising the principal components of the exposure registration shape time weighted average? sudden field changes? windows above a threshold? John Bolte RIVM & THUAS 25 January 2017

32 Two different AEMs - DECT (1.8 GHz)hr/Nses/pax Powerdensity RCM N=98, balanced: sex, age, ses, work, housing type, degree of urbanisation John Bolte RIVM & THUAS 25 January 2017

33 Individual exposure – effect: personal medicineSelf-declared electrosensitives Ecological Momentary Assessment real time data no recall bias natural context alarm cues at 6 hr intervals (eg. 8:00, 14:00, 20:00) 3 weeks = 63 blocks John Bolte RIVM & THUAS 25 January 2017

34 GSM, DCS and UMTS basestationstrain station walk bike John Bolte RIVM & THUAS 25 January 2017

35 Which exposure metric is relevant for effect?example: magnetic fluc density (B field) Essence of exposure: emphasising the principal components of the exposure registration shape time weighted average? sudden field changes? windows above a threshold? No Dosis – effect relation exists for fieldstrength level in everyday activities John Bolte RIVM & THUAS 25 January 2017

36 Conclusies Blootstelling aan elektromagnetische velden hangt af van:ondernomen activiteiten (gedrag) belgedrag van omgeving en eigen belgedrag percentage stedelijke bebouwing tijd van de dag en kan door eigen ingreep mogelijk tot de helft verminderd worden - Modellering van individuele blootstelling is complex door sterke overlap en variantie Bolte JFB and Eikelboom T, Personal radiofrequency electromagnetic field measurements in the Netherlands: exposure level and variability for everyday activities, times of day and types of area. Env. Int. 48:133:142. John Bolte RIVM & THUAS 25 January 2017

37 Personal environmental sensing: electromagnetic fieldsgps Personal environmental sensing: electromagnetic fields radio frequency ELF magnetic John Bolte RIVM & THUAS 25 January 2017

38 Elektromagnetische velden in UtrechtTrain Bike Inside shop Walk Home Railway station John Bolte RIVM & THUAS 25 January 2017

39 Standards and bodies For measurements and modelling of electromagnetic fields from devices: e.g. distance, height, duration, methods, types of devices, etc. BUT that is not the same as exposure: where is a person for how long in what field strength? Exposure over action values does not mean adverse health effects NOR does under action values mean no adverse health effects: long term health effects, non-specific health effects And: John Bolte RIVM & THUAS 25 January 2017

40 From internal to external field limits:Same: - Exposure limit values Different: Models Safety factors John Bolte RIVM & THUAS 25 January 2017

41 Exposimeters: body shielding and polarisationBolte et al. , 2011, Bioelectromagnetics John Bolte RIVM & THUAS 25 January 2017

42 Calibration of EME Spy 121 John Bolte - RIVM E-Monument 4th meeting PHE, Chilton, UK July 30, 2014