György Thuróczy

National Research Institute for Radiobiology and Radiohygiene

Dept. of Non-Ionizing Radiation

Address: 1221 Budapest, Anna u 5. HungaryTel: +36 1 482 2019 fax: +36 1 482 2020

thuroczy@hp.osski.hu, www.osski.hu

PUBLIC EXPOSURE TO RF FROM INSTALLED SOURCES:

SITE MEASUREMENTS AND PERSONAL EXPOSIMETRY

• Characteristics of EMF environment– Sources– Exposure levels– Exposure variations

• RF measurements • in-situ survey around mobile base stations• Survey in underground (metro) stations by

broadband measurements• personal RF exposimetry

Outline

• Main features of RF exposure:– wide range of exposure levels (environmental vs.

occupational)– Spatial variations in space – Inhomogeneous and partial body exposure – Variations of exposure level in time– Rapid development of new technologies with

appearing new frequencies and signals– Rapid changes of exposure situations

Characteristics of EMF environment and human exposure

Characterization of human exposure to RF

Exposure Source

Long term- low level

whole body

Intermittent whole body

Intermittent partial body

Intermittent Highly local

Short term Transient whole or

partial body

Broadcast +++ +

Telecom ++ ++ +++

Medicine + ++

Industry ++ ++ + +

Home ++

Military + + ++ +++

Emerging

technology

+ + +++

(Valberg et al, 2007)

Changing the environmental exposure to RF• Environmental RF exposure always changes in time:

– rapid proliferation of RF sources (i.e. base station, BS)

– raise of ambient RF radiation: 1980 (Tell, USA): ~ 50 W/m2 1999 (Hamnerius, Sweden): ~ 500 W/m2

– Increasing contributions of GSM sources in the RF range (up to 39-61 %)

– Increasing the indoor exposure due to the new wireless devices

System City area%

Town%

Rural%

Residental%

Total%

Radio 13 1 11 20 15

Television 13 1 48 24 23

NMT 450 2 1 2 0 1

NMT 900 1 2 0 0 1

GSM 900 61 53 39 41 47

GSM 1800 4 9 0 5 3

Various 6 33 0 10 10

Mean contributions from different RF sources in Sweden (mean percent part of the total ratio from RF environment given in

percent, 30 MHz-2100 MHz), Hamnerius 1999

RF environmental exposure in the frequency range 80 MHz-1850 MHz

55,0

65,0

75,0

85,0

95,0

105,0

115,0

80,0 280,0 480,0 680,0 880,0 1080,0 1280,0 1480,0 1680,0

Frequency (MHz)

Elec

tric

Fie

ld S

tren

ght (

dbuV

/m)

Vert (dbuV/m)

Hor(dbuV/m)

Z (dBuV/m)

GSM 900

GSM 1800

Human exposure in the environment and workplaces: wide range of level

Medium frequency range: 0.3-3 MHz Mantiply, 1997

Occupational

Public

Human RF exposure in the environment and workplaces: wide range of level

Very High Frequency range: 30-300 MHzMantiply, 1997

Occupational

Public

Human RF exposure in the environment and workplaces: wide range of level

Ultra High Frequency range: 0.3-3GHz Mantiply, 1997

Public

Occupational

RF exposure levels - outdoor

(Valberg et al, 2007)

RF exposure levels - indoor

(Valberg et al, 2007)

Exposure and sources compared to EU Recommendation of public exposure limits

For example: 100%= 100 T@50 Hz ; 4,5 W/m2@900 MHz

Exposure to EMF from different sources compared to public limit of EU Recommendation (1999/519/EC)

0,0001

0,001

0,01

0,1

1

10

100

1000

10000

100000

GS

M B

S

3G B

S

TV

-rad

io B

C

mob

ile p

hone

indu

stri

al d

ev.

Acc

ess

cont

r.

Pow

er li

ne M

F

Hom

e (5

0 H

z)

App

lianc

es

In p

erce

nt o

f pu

blic

lim

it (%

)

Exposure and home sources compared to base stations

in percent of the EU Recommendation for public limit

Exposure to EMF from different sources compared to public limit of EU Recommendation (1999/519/EC)

0,00010,0010,010,1

110

1001000

10000100000

GSM

BS

3G B

S

mic

row

ave

oven

was

hing

m.

elec

tric

ove

n

ligh

t tub

e

hair

dry

er

TV

dev

ice

In p

erce

nt o

f pu

blic

lim

it (

%)

For example: 100%= 100 T@50 Hz ; 4,5 W/m2@900 MHz

Measurements – Why?

Typical cases when RF exposure prediction or measurement are required:

• Compliance testing of human exposure (workers, general public) according to compliance levels of ICNIRP or EU reference levels (typically required by public, local authority)

• Source characterisation (typically required by operators)• Pre-installation measurement (requested typically by local

governments)• New antenna installation (in case of many existing antennas)• Compliance testing of electromagnetic compatibility (EMC)• Scientific request for epidemiological study (i.e. to define

cohorts within the population)

Aims and corresponding methods: Environmental RF exposure

Goal of RF exposure measurement

Broadband Frequency selective

SAR

Compliance testing for general public exposure

yes occasionally no

Compliance testing, occupational

yes occasionally occasionally

Source characterisation no yes no

Pre-installation measurement occasionally yes no

New source installation yes yes no

Compliance testing, EMC no yes no

Scientific request for epidemiology

yes occasionally occasionally

• Measured value: field strength and variations

– advantages: • We have a long term experiences of the methods• Well developed measurement devices are available• Provide a relevant characterisation of the real

exposure of the given site• Possibility to validate by numerical methods

– Disadvantage and limitations: • Large time requested (i.e. spatial mapping,

monitoring, frequency selective measurement)• Expert’s work is needed (expensive)• May have large variations in field strength • May not relevant to the real individuals’ exposure

In situ (on-site) EMF measurement

Site measurements of RF exposure: variations and uncertainties

– The exposure of population show a very large variation. Basically there are four different sources of such variations and uncertainties:

• large scale variation because of variations of exposure between different places and between different times at a given location (influenced by how the measurement sites were selected, may extend to few orders of magnitude).

• signal variations induced by propagation path and technology

• temporal variations in traffic density at a given location (2-3 fold)

• uncertainty in measurements due to the measurements techniques (up to 30 %)

from J.Wiart

from J.Wiart

– Survey on mobile base stations by frequency selective measurements

– Survey in underground (metro) stations by broadband measurements

– Personal RF exposimetry by frequency selective exposimeter

Measurements

Survey on mobile base stations by frequency selective measurements

Issues and questions:

What levels of exposure to radiofrequency fields are in the environment to the vicinity of base stations?

How does the increased deployment of antennas relate to exposure levels?

Do these exposures to electromagnetic fields from base station antenna comply with standards and regulations?

Is any relation between the public exposure and the distance from the antenna?

Methods/1

•According to the COST-STM frequency selective measurements were performed.

•The selected frequencies effectively covered the frequencies used for GSM 900 and GSM 1800 down-links.

•The assumption was made that far field conditions applied. Therefore the measured electric fields could be converted to power densities.

•Measurements were performed using wide band antennas connected to a spectrum analyser.

•The antenna was always mounted on a tripod at slightly varied heights around 1.5 m.

General outline of frequency specific measurements, comprising an antenna, a spectrum analyser and data storage facilities.

Methods/2Spectral measurements

• Antenna and spectrumanalyzer– PBA 10200 (ARCS) 80 MHz-2200 MHz– Advantest U4941

• Measurements– Sweep: 40 MHz, (925-965 MHz)– RBW: 100 kHz, VBW: 100 kHz– x ,y, z orthogonal, max-hold– data storage (700 points)– distance measurements with laser beam

• Evaluation– From SRAM to Excel spreadsheet– dbV/m, V/m, mW/m2

– According to COST STM

Methods/2

•The effective power density was obtained by the vectorial summation of the orthogonal electric field (V/m) components. Using the formulae P = E2/377, the result is expressed as the effective power density in mW/m2.

•All data were obtained from spot measurements (N=292), and in most cases no information concerning the variations of the field strengths versus time was available.

•During the sample (scanning) time, the maximum field strength in each direction could be obtained by using the “peak hold” function of the analyser.

•The data analysis was generally performed off-line. The measured exposures were expressed in mW/m2. The data were expressed as:

– Ssum (mW/m²): The sum of all power densities in the respective GSM down-link

band–Si (mW/m²): The highest power density measured at a single frequency in the

respective GSM down-link band

Field strenght of GSM base station

55,00

65,00

75,00

85,00

95,00

105,00

115,00

925,0 930,0 935,0 940,0 945,0 950,0 955,0 960,0 965,0

Frequency (MHz)

Fiel

d st

reng

ht (d

buV

/m)

Vert (dbuV/m)

Hor(dbuV/m)

Z (dBuV/m)

Frequency spectrum of GSM 900 MHz base station: down-link

Power density of GSM base station

0,0000

0,0100

0,0200

0,0300

0,0400

0,0500

0,0600

925,0 935,0 945,0 955,0 965,0

Frequency (MHz)

Po

wer

den

sity

(m

W/m

2)

Res (mW/m2)

Power density spectrum of GSM base station: down-link

RF exposure around base stations: Data from all measurement sites

0,001

0,010

0,100

1,000

10,000

100,000

SS

um

po

we

r d

en

sit

y (m

W/m

2)

Ssum Median: 0,156 mW/m2 (CI 95%: 0,0209 - 4,922, N = 292)

Ssum (mW/m²): The sum of all power densities in the respective GSM band

Distribution histogram of measured power density in different ranges of all measurement sites

-15,0

10,0

35,0

60,0

85,0

SSum power density range (mW/m2)

Per

cent

age

of a

ll da

ta (%

)

SSum (mW/m2) % (N=292) S>10 1,9

1<S<10 14,20.1<S<1 50,9

0.01<S<0.1 32,60.001<S<0.01 0,4

Ssum (mW/m²): The sum of all power densities in the respective GSM band

Ssum (mW/m2) Indoor Outdoor Street Country side

No access available

Number of sites 106 161 101 20 23

Median 0,141 0,187 0,182 0,177 0,278

CI (95%) 0,021 – 5,499 0,020–4,985 0,018–4,983 0,024–1,590 0,022-3,801

Si (mW/m2) Indoor Outdoor Street Country

side No access available

Median 0,0006 0,0133 0,0106 0,0150 0,0167

CI (95%) 0,0002–0,360 0,0001–0,245 0,0001–0,237 0,001-0,127 0,001-0,122

Power density in different Type of environment of all measurements (mW/m2)

Ssum (mW/m²): The sum of all power densities in the respective GSM band

Si (mW/m²):The highest power density measured at a single frequency in the respective GSM band

0,001

0,010

0,100

1,000

10,000

100,000

All Outdoor Indoor Street Fields

Ssu

m (m

W/m

²)

median

Power density in different type of environment

Ssum (mW/m²): The sum of all power densities in the respective GSM band

ICNIRP/EU limit on 900 MHz @ 4500 mW/m2

on 1800 MHz @ 9000 mW/m2

Power density vs. distance from the base station

0,00

0,01

0,10

1,00

10,00

100,00

1 46 91 136 181 226 271

distance (m)

po

wer

den

sity

(S

sum

, mw

/m2)

Ssum (mW/m²): The sum of all power densities in the respective GSM band

EU limit at 900 MHz @ 4500 mW/m2

at 1800 MHz @ 9000 mW/m2

Base station survey:Results and Discussion

•According to the results of present spectral measurements at more than 290 sites, the exposure levels from the base stations were many times below the ICNIRP/EU limits (4500 - 9000 mW/m2) respectively.

•The exposure did not exceed the tens of mW/m2 (a few microwatt/cm2) at locations accessible to public.

•Within 300 m of the base station no clear expression could be found between the exposure levels and distances similarly to other studies.

– Survey on mobile base stations by frequency selective measurements

– Survey in underground (metro) stations by broadband measurements

– Personal RF exposimetry by frequency selective exposimeter

Measurement #2

Peron B

Peron A

Metro tunel

GSM panel antennatoward peron

Radiatedcables

GSM panel antenna

toward the tunel

2-14,6 V/m

x

x1-8,1 V/m

x

0,3-0,4 V/m

GSM Base station in METRO station(n=83)

GSM Base station in METRO station:broadband measurements (N=83)

Electric Field strength close to Mobile Base Station (~ 1m)in the area accessible to public (1,7 m height)

0

2

4

6

8

10

12

14

16

Ele

ctr

ic f

ield

(V

/m)

Average: 3,3 V/m, SD: 2,07 V/m; Median: 2,9 V/m (CI 95%: 1,4-7,19)

Maximum measured values in METRO stations and in METRO cabins

0

2

4

6

8

10

station cabin

Ele

ctri

c fi

eld

(V

/m)

GSM Base station in METRO station:broadband measurements

• – Survey on mobile base stations by frequency

selective measurements– Survey in underground (metro) stations by

broadband measurements– Personal RF exposimetry by frequency

selective exposimeter

Measurement #3

Personal RF exposimetry

• According to our previous site measurements:

– The main RF field variations comes from the location

therefore

– Assess the personal exposure = assess the exposure where the person is sleeping, working, walking….

Personal RF exposimetry: aims

• Record the RF exposure coming from main wireless systems and broadcast.

• Avoid interference with the person’s activity

• Long term frequency selective recording of human exposure to RF

Personal RF exposimetry Evaluation of RF general population exposure

• Dosimetric problems: – Difficulties in retrospective exposure assessment for epi studies

– Exposure misclassification due to different RF sources

– Long term exposure variations in time, exposure variation in space

• Objectives – Characterise RF exposure levels of individuals– Evaluate the importance of different exposure sources in the

general and personal environment– Identify, if possible, the main factors which may predict exposure

levels

• Further Aim– support for any future epidemiological study

Personal RF exposimetry: the device

• Personal exposure meter (PEM): – Antennessa DSP090 (France)

• Frequency bands recording selectively:– FM (88 to 108 MHz)

– TV (174 to 223 MHz) & (470 to 830 MHz)

– GSM 900 Tx (mobile phone: 875 to 915 MHz) & Rx (base station: 935 to 960 MHz)

– GSM1800 Tx(1710 to 1795 MHz) & Rx(1805 to1880 MHz)

– UMTS Tx (1920 to 1980 MHz) & Rx (2110 to 2170 MHz).

• dynamic range 40 dB within the E-field range:– from 0.05 V/m to 5 V/m at each band

Personal RF exposimetry: sources of variations and uncertainties

– Spatial variations in signal strengths induced by various propagation path and technology (few orders),

– Isotropy (i.e. close to the body)

– Frequency selectivity (co-channel error)

– sensitivity and accuracy

– Personal variation due to position (proximity) to the human body

• Isotropy (free space): 2 dB at 95% confidence (1 dB at 66%)

• Vertical dosimeter

Personal RF exposimetry: sources of uncertainties: influences of the body

• The body has an influence but in each frequency band of interest the total exposure is the sum of different sources and reflections therefore the exposure is often coming from everywhere.

• The mean values as obtained with the PEMs tend to underestimate the free field measurements (64% - 72%).

• Taking into account this underestimation of the free field conditions, the simple free space model might be a valid approximation for the real exposure.

dosimeter

Measurement in the SwissCom laboratory, Lehmann et al, 2007

Personal RF exposimetry Personal frequency selective exposimeter for RF range

Antennessa DSP 090

• The purpose of the current study was to evaluate the usefulness of an RF personal exposimeter (dosimeter) for assessing individual radiofrequency (RF) exposure in an urban environment.

• Measurements taken by RF personal dosimeter (PEM) were also compared to preliminary site measurements taken around mobile base stations.

Measurements with personal exposimeter in Budapest (Hungary)

• Subjects– n=21 participants (plus 4 pilot study, 2 eliminated by technical reasons)

– Residency in Budapest (capital, 2,5 million inhabitants)

– Most of them were our colleagues from the institute or their relatives

• Protocol– time-activity diary by the subjects following the form designed for the

study (we used similarly forms for 50 Hz personal exposimetry)

– Time of the survey: 24 h data recording

• Device, recording and location– personal exposure meter (PEM): Antennessa DSP090– Recording: sample rate 15 sec, ~1440 minutes (24 h)– Location: mounted on the body and /or in a carry bag. Fix location

near the bed at night and close to subject indoors (i.e. on the desk, table etc.)

Measurements with personal exposimeter in Budapest (Hungary)

• Selected exposure metrics:– Duration of (exposure, activity) time (min)– Field intensity (V/m): max., arithmetic mean, S.D.– Time Weighted Average (TWA)

TWA: Average field intensity x duration of time (Vm-1 x min)

• Exposure metrics analyzed according to:– Frequency bands (channels)– Different time periods (activity)

• Mainly the GSM(rx)/DCS(rx)/TV4&5 were analysed so far (fixed installed sources: base stations, broadcasts antennas).

Methods: exposure metrics and analysis

• Total measuring time (1440 min – 24 h)

• Periods according to activity types: – (1)home (2)bed (3)travel (4)work (5)other/else

• Periods when the measured field exceeded the detection threshold of the meter:(>0,05 V/m)

• Periods when the measured data fall into different bins according to field level ranges: – (a-low) 0,05-0,1 V/m (b-medium) 0,1-1,0 V/m (c-high) 1,0-5,0 V/m

• Periods when the measured field was equal to the detection limit (=0,05 V/m, or may below)

Exposure metrics and analysis: time periods

Exposure time above 0,05 V/m

0123456789

10F

M

TV

3

TV

4&

5

GS

Mtx

GS

Mrx

DC

Stx

DC

SR

x

UM

TS

tx

UM

TS

rx

pe

rce

nt

of

tim

e (

%)

Home

Travel

Work

Else

Mean percent of total time (24h), when subjects were surely exposed above 0,05 V/m by channels and activity types

Averages of periods (minutes) within 24 h (~1440 min) when the data of measured field fall into different bins of exposure ranges according to selected

channels (base stations and TV4&5) by different activity types.

0,01

0,1

1

10

100

1000

ho

me

bed

trav

el

wo

rk

oth

er

ho

me

bed

trav

el

wo

rk

oth

er

ho

me

bed

trav

el

wo

rk

oth

er

GSMrx DCSrx tv4&5

Ave

rag

e ti

me

[min

]

E>1 V/m 0,1<E<1 V/m E<0,1 V/m

(900 MHz) (1800 MHz)

Average of maximum measured electric field strengths during the whole recording periods (24 h)

Average of the maximum values during 24 h (N=21)

0,05

0,25

0,45

0,65

0,85

1,05

1,25

1,45

1,65

1,85

home bed travel work other

V/m

TV4&5

GSMrx

DCSrx

UMTSrx

Average of the mean values during 24 h (N=21)

0,05

0,1

0,15

0,2

0,25

home bed travel work other

V/m

TV4&5

GSMrx

DCSrx

UMTSrx

Average of mean measured electric field strengths during the whole recording periods (24 h)

Mean RF exposure according to frequency bands during total measuring time, when E>0,05 Vm-1 by participants

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

0 200 400 600 800 1000 1200

time when E>0,05 Vm-1(minutes)

RF

fie

ld (

V/m

)

TV4&5 GSMtx

GSMrx DCStx

DCSrx

Base stations

Mobile phones

Personal RF exposimetry:Conclusions

• The results from personal exposure showed that one third of the participants spent 40-70 % percent of 24h recording time above the detection limits of PEM (0,05 V/m) and half of subjects spent less than 10 %.

• The highest exposure was detected during the traveling period and the lowest in the bed at home.

• We also concluded that duration of time exposed to RF levels above the detection limit of the PEM is a useful exposure metric to compare and contrast individual RF exposure.

Personal RF exposimetry:possible further studies

• Further data must allow for more interpretation of results:

– different sites (urban-rural)– different population categories (adults ;

mobile phone users or not; children ; …)– diary form, daily and place : (residential

room ; office ; restraint space – underground etc - ; foot or car moving, indoor or outdoor).

– data collection on 24 h or 1 week (at 2 different seasons ?)

– sampling interval to define : 14 s or 1’27 minutes or others ?

Personal RF exposimetry:Conclusions: advantages and disadvantages

– Measured values: field strength and time duration in exposure (data logging)

• advantages:

– Time profile of individuals’ exposure

– Long term recording

– Easy to perform without expert

– Possible exposure data for epidemiological studies (i.e. cohort study)

• disadvantage and limitations:

– Uncertainties and variations due to the placement of exposimeter on the body (especially at RF range)

– Many devices requested for a large study (expensive)

– Good organisation and logistics are requested

– Data evaluation needs a long time and expertise

– Sensitivity may be limited

Personal exposimetry and RF dose-response considerations: questions and remarks

– The current established RF dose-response concept based on the threshold principle.

– The current concept of dose based on, that the RF exposure as such can not be accumulated with time.

– Otherwise the time scale in dose concept may be relevant. (i.e. the biological effect may fade with time because of physiological (heating-cooling) and other (repair) mechanisms).

– The “dose” should be defined as a transformation of exposure that corresponds to biological efficiency.

General conclusions of site and personal exposure measurements

– All studies support that in all examined positions RF exposure at the sites accessible to public were many times below (ten thousands to millions of ICNIRP reference levels) the exposure limits of the ICNIRP/EU guidelines.

– Otherwise the continuous rise of ambient RF radiation levels have been detected.

– Because of the proliferation of base station antennas the contribution of RF exposure due to GSM and other wireless technology is increasing.

Forthcoming issues in RF exposure assessment and dosimetry

• More detailed environmental and personal RF exposure assessment because of rapid proliferation of GSM, 3G, wireless and new emerging technology

• More accurate local SAR models and measurement of non-uniform partial body and highly localized RF exposure

• Moving toward to microscopic (cellular, sub-cellular) RF dosimetry

• More accurate thermal modeling relevant to dosimetric evaluation

• Efforts to move from physical quantities toward a biologically relevant quantity for implementation of the dose concepts

Permanent debate:measurements vs. calculation?

Everyone believes a measurement except the person who did it.

No one believes a calculation except the person who did it.

from Peter Zollmann (Vodafone)

Thank you for your attention!

Current dose response concepts of RF exposure:threshold of health effects

SAR (W/kg)0.08 0.4 4.0

Thermal related health effects

Est

abli

shed

hea

lth

eff

ects

Threshold of health effects

other measures safety factors (ICNIRP) Guidelinesconcepts

PRECAUTIONARY PRINCIPLEPRECAUTIONARY PRINCIPLE ??

Emitted RF power levels

(Valberg et al, 2007)

Time above 0.05 V/m or below by each participant

0%

20%

40%

60%

80%

100%

Per

cent

of t

ime

(%)

19 9 10 4 17 16 3 18 14 21 13 20 5 11 15 8 12 2 6 7 1

Participant No.

Percent distribution of time (min) when measured RF field was above the detection limit (0.05 V/m) during total measuring time

E=0,05 V/m

E>0,05 V/m

Exposition above 0,05 V/m: by 1/3 of participants 40 - 70 % of total time by 1/2 of participants 10 % of total time

Duration of exposed time (E>0,05 V/m) vs. TWA exposure according to activity types

(a trial of calculation of cumulative exposures by participants)

Duration of exposed time (E>0,05 V/m) vs. TWA exposure according to activity types by participants

0,1

1

10

100

1000

0,01 1 100 10000

TWA exposure (Vm-1 x min)

ex

po

se

d t

ime

(m

in) home

bed

travel

work

else