Calibration and testing of a portable NaI(Tl) gamma-ray spectrometer-dosimeter for evaluation of terrestrial radionuclides and 137Cs contributions to ambient dose equivalent rate outdoors

Direct measurement of gamma dose rate in air using a gamma-ray dosimeter is one of the common procedures to assess the external exposure of humans in the case of an environmental contamination with technogenic gamma-ray emitting radionuclides [(Generic procedures for monitoring in a nuclear or radiological emergency. IAEA-TECDOC-1092, IAEA – International Atomic Energy Agency, 1999); (Radiation monitoring of the exposure doses of population at the territories radioactively contaminated due to the accident at the Chernobyl NPP. Recommended Practices. Adopted 27.12.07, implemented 27.12.07. Federal Center of Hygiene and Epidemiology of Federal Service for Surveillance on Consumer Rights Protection and Human Well-Being, Moscow, 2007)]. The dosimeter reading in terms of ambient dose equivalent rate, Ḣ*(10), (ADER reading ) is the sum of the following major components: 1) dose rate due to the terrestrial radionuclides of U series, Th series and K (ADER TRN ); 2) dose rate due to the directly ionizing and photon component of cosmic radiation; 3) intrinsic noise of the dosimeter; 4) dose rate due to technogenic radionuclides [1, 2, 3]. Additional small contribution to ADER reading can be associated with some other terrestrial radionuclides, including those of the U series and La, the so called “cosmogenic” natural gamma-emitting radionuclides (Be and Na) and Rn progenies in the atmosphere [3, 4, 5, 6]. For correct estimation of external exposure from artificial sources, it is necessary to subtract all other components from the dosimeter reading. Several approaches for separation of the ADER reading components have been proposed and practically implemented. Specifically, for assessing the sum of DOI: 10.21514/1998-426Х-2017-10-1-18-29 УДК:539.16.08:52-732


Introduction
Direct measurement of gamma dose rate in air using a gamma-ray dosimeter is one of the common procedures to assess the external exposure of humans in the case of an environmental contamination with technogenic gamma-ray emitting radionuclides [(Generic procedures for monitoring in a nuclear or radiological emergency.IAEA-TECDOC-1092, IAEA -International Atomic Energy Agency, 1999); (Radiation monitoring of the exposure doses of population at the territories radioactively contaminated due to the accident at the Chernobyl NPP.Recommended Practices.Adopted 27.12.07,implemented 27.12.07.Federal Center of Hygiene and Epidemiology of Federal Service for Surveillance on Consumer Rights Protection and Human Well-Being, Moscow, 2007)].The dosimeter reading in terms of ambient dose equivalent rate, Ḣ* (10), (ADER reading ) is the sum of the following major components: 1) dose rate due to the terrestrial radionuclides of 238 U series, 232 Th series and 40 K (ADER TRN ); 2) dose rate due to the directly ionizing and photon component of cosmic radiation; 3) intrinsic noise of the dosimeter; 4) dose rate due to technogenic radionuclides [1,2,3].Additional small contribution to ADER reading can be associated with some other terrestrial radionuclides, including those of the 235 U series and 138 La, the so called "cosmogenic" natural gamma-emitting radionuclides ( 7 Be and 22 Na) and Rn progenies in the atmosphere [3,4,5,6].For correct estimation of external exposure from artificial sources, it is necessary to subtract all other components from the dosimeter reading.
Several approaches for separation of the ADER reading components have been proposed and practically implemented.Specifically, for assessing the sum of Радиационная гигиена Том 10 № 1, 2017 contributions of cosmic radiation and intrinsic noise of a dosimeter, additional measurements with the dosimeter can be carried out on a surface of a large water body (e.g., lake or sea) [(Radiation monitoring…, 2007), 2 , 7, 8].The contribution from terrestrial radionuclides to the total ADER may be evaluated by determination of the 238 U, 232 Th and 40 K activity concentrations in soil using stationary or/and portable gamma-ray spectrometers.It can be done with: 1) soil sampling and subsequent laboratory analysis or/and 2) direct measurements in field (in situ measurements), e.g., [2,5,9,10,11].
In the 1960-70s, Beck et al. [5,12] developed the theoretical principles of in situ gamma-ray spectrometry and practically demonstrated that in situ measurements of soil activity could provide more representative data for evaluation of gamma dose rate in air than the data obtained by soil sample collection and laboratory analysis.Since these pioneering studies the field has rapidly expanded (for review see [13,14,15]).The assessment of the external dose components was one of the applications of in situ gamma-ray spectrometry after the Chernobyl accident (1986).In particular, a portable germanium detector was used to differentiate absorbed dose rates in air due to uranium series, thorium series, 40 K and 137 Cs during an extended survey of indoor and outdoor terrestrial gamma radiation in Greece [16].
The present investigation is a study on the application of a commercially available portable gamma-ray spectrometry system for in situ analysis of the activity concentration in soil and ADER in the remote period after the Chernobyl accident.We have used a NaI(Tl) based spectrometer-dosimeter from ATOMTEX (Belarus).The AT6101D spectrometer [17] has the pattern approval certificates of the Republic of Belarus, the Russian Federation, Ukraine and Kazakhstan.The device is calibrated by the manufacturer to determine: 1) activity concentrations (Bq kg -1 ) of the terrestrial radionuclides of 226 Ra, 232 Th and 40 K in soil, agricultural raw materials, forestry products and construction materials; 2) ground contamination density by 137 Cs, A Cs (Bq m -2 ); 3) total ambient dose equivalent rate, ADER (nSv h -1 ).
Potentially, the device can be used to separate contributions of terrestrial radionuclides and 137 Cs to the total ADER.Unfortunately, the spectrometer was not factory calibrated in such a manner.Therefore, the main aim of the present study was to deduce the calibration coefficient for conversion of the measured AC eff (Bq kg -1 ) to the terrestrial component of ambient dose equivalent rate, ADER TRN (nSv h -1 ).It has allowed quantifying the dose rate associated with terrestrial radionuclides in the presence of 137 Cs contamination.Another aim was to test the spectrometry system in field at the radioactively contaminated areas of the Bryansk region (Russia).
The laboratory and field measurements were carried out in the period 2014-2016.

Instruments and the factory calibration procedure
The portable spectrometer-dosimeter MKS AT 6101D [17] consists of two blocks: detection and processing units that are connected by a water-proof cable (Fig. 1).The detection unit is a NaI(Tl) cylindrical scintillation detector (63 mm in diameter and 63 mm in length).The energy resolution (FWHM at 662 keV of 137 Cs) of the detector is 7.3%.The detector is placed in a temperature-and-shock-resistant dust-proof and moisture-proof container (121 mm in diameter and 477 mm in length).The weight of the unit is 4 kg.The available geometries for measurements are 2π (on a surface) and 4π (in a well).Spectrometric information from the detection unit is transferred into a processing unit (mass = 0.8 kg) and is displayed on LCD screen.The multichannel analyzer is adjusted to 512 channels at the energy range from 40 keV to 3 MeV.Up to 300 spectra can be stored in the energy independent memory of the spectrometer.The instrumental spectra processing algorithm allows data display in the form of activity concentration of the terrestrial radionuclides, ground contamination density by 137 Cs and its activity concentration.For energy calibration of the spectrometer, point sources of 241 Am, 109 Cd, 57 Co, 139 Ce, 113 Sn, 54 Mn, 22 Na, 60 Co, 152 Eu, 137 Cs, 228 Th and 88 Y were used.The 3 window matrix method (see e.g., [18]) was used to calculate the 226 Ra ( 238 U series), 232 Th and 40 K activities of the soil samples.The energy windows were centered on the 1461 keV ( 40 K), 1764 keV ( 214 Bi) and 2615 keV ( 208 Tl) full-energy peaks (Fig. 2) for the estimation of 40 K, 238 U, and 232 Th activity concentrations, respectively.The uranium 238 U and thorium 232 Th content was calculated under the assumption that secular equilibrium exists between all of the radionuclides within the decay series.High volume standards containing 40 K, 226 Ra and 232 Th have been used to calibrate the spectrometer.The field spectrum processing includes subtraction of the so called "background spectrum" recorded by the manufacturer inside a 10 cm thick lead shield.Intrinsic relative uncertainty of the monitored radionuclide concentration measurements is ±20% for the terrestrial radionuclides and ±30% for 137 Cs (maximum).Th by a single quantity, which takes into account the total external exposure associated with them, 'effective' activity concentration of the natural radionuclides in soil, AC eff (Bq kg -1 ), is calculated automatically within the AT6101D dosimeterspectrometer using the formula: The definition of AC eff , which is in use in the Russian Federation, is closely related to so called "radium equivalent activity, Ra eq ", a commonly used radiological hazard index (e.g., [19,20,21]).For the calculation of Ra eq , it is assumed that 370 Bq kg −1 of 226 Ra, 259 Bq kg −1 of 232 Th or 4810 Bq kg −1 of 40 K produce the same γ-ray dose rates.Therefore, the correction coefficients are 1.43 for 232 Th and 0.077 for 40 K.These are slightly (by about 10%) different from those values that have been used by the AT6101D producer for derivation of AC eff (eq.1).
The ambient gamma-radiation dose equivalent rate value in an inspection point is determined by the instrument spectrum analysis with the "spectrum-dose" operational functions.The measured pulse height distribution is converted automatically into the physical quantity of dose rate using the response func-tions that are recorded in the memory unit of the spectrometerdosimeter.The correction functions (coefficients) are energy dependent.The validity of the spectrum-to-dose conversion function has been verified by the manufacturer using a strong standard 137 Cs source.For the reference dose rate values of 0.7, 7.00 and 70.0 µSv h -1 , the readings were 0.681, 7.03 and 70.8 µSv h -1 , respectively [AT6101D Spectrometer.Operation manual.ATOMTEX, Minsk, Belarus, 2014].According to the manufacturer, the intrinsic relative uncertainty of the ADER measurements in field is within ±20% (maximum).
The manufacturer provides a reference KCl volume source (mass = 0.2 kg) for periodical energy calibration of the spectrometer in situ (the 1461 keV full-energy peak of 40 K).The energy calibration can also be done using a 137 Cs reference source (the 662 keV full-energy peak of 137 Cs-137m Ba).Additionally, the spectrometer has a built-in automatic stabilization which is achieved by using light emitting diode.Its instability during a continuous survey is estimated as ±1.5% (maximum).

Additional in-house calibration and field testing
To estimate contribution of cosmic radiation and intrinsic noise of the device into the primary reading of the spectrometer, ADER reading , three gamma-ray spectra were recorded on the frozen surface of the Finnish Gulf in the Leningradskaya region (Russia) in December 2016, shortly after formation of an ice cover.The measurements were performed at a distance of about 2.5 km from the shore-line.The site (60.113°N, 29.898° E) has an average water depth of 6 m.
To determine a relationship between AC eff and ADER, spectrometric measurements were made outdoors at the territory of St.-Petersburg and at two settlements in the Leningradskaya region (Table 1).The areas can be considered as a "background" because they did not receive any substantial amount of Chernobyl fallout (Table 1).In these areas, the current ground contamination level by 137 Cs (from global fallout and the Chernobyl accident) is estimated below 5 kBq m -2 .A total of 27 sites (20 ground plots and 7 paved areas) were selected for the measurements (Table 2).The sites characteristics can be found in Tables 3 and 4. At 12 plots (Table 3), in situ measurements were performed with a downward facing detector at heights of 0.1 m and 1.0 m above the ground (or paved surface).The detection and processing units were mounted on an aluminum tripod (mass = 4.0 kg) (Fig. 1).Data from this series of measurement were used to evaluate: 1) a relationship between AC eff and ADER for our spectrometric system and 2) a relationship between the results obtained at the two heights above the ground.The validity of the AC eff to ADER conversion coefficient was checked in a second series of measurements carried out at 15 different plots at a height of 0.1 m above the ground (Table 4).The first practical testing of the spectrometer and empirically derived calibration factors was performed in Novozybkov, the Bryansk region.The town and its surroundings had been heavily contaminated following the Chernobyl accident (Table 1).The average residual ground contamination by 137 Cs was estimated as 396 kBq m -2 in 2014 [22].The spectrometric measurements were performed at the typical locations [2]: grassland, kitchengarden, yard, street and forest (Table 5).For comparison, measurements were also made at a less contaminated site in the Bryansk region, Zatishie, where the average level of ground contamination in 2014 was only 41 kBq m -2 [22].At each surveyed plot, gamma-ray spectra were recorded at a height of 0.1 m and 1.0 m, respectively, above the ground.b -the ADER is calculated using eq.( 2) (ADER = 11 + AC eff × 0.51).
c -the primary reading of the AT1601D spectrometer-dosimeter.A statistical uncertainty (at the 2 sigma level) of the ADER measurements is below ±2%.SD -standard deviation.

The end of table 5
All outdoor measurements (excluding the measurements on ice) were conducted in dry weather in summer time.Counting times ranged from 300 s to 1800 s.
Statistical analysis of data was performed using Excel for MS Windows and the tools provided by free on-line calculators, Free Statistics Calculator [23] and Centr sovremennykh psykhotekhnologiy [24].

Additional in-house calibration
Three repeated measurements performed on the frozen surface of the Finnish Gulf showed the same value of 7 nSv h -1 .This value includes response to cosmic radiation and contribution from intrinsic noise of the spectrometer.The measurements revealed a rather weak response of the spectrometer to the directly ionizing and photon component of cosmic radiation because the population-weighted average dose rate from this source at sea level corresponds to 31 nGy h -1 (absorbed dose in air) or 31 nSv h -1 (effective dose) [6].One should note that the spectrometer-dosimeter AT6101D is not designed to measure the cosmic radiation.
A summary of the dose rate and activity concentration of radionuclides at 27 background plots is presented in Table 2.The measurements were performed close to the soil or paved surface i.e. in the position recommended by the AT6101D manufacturer for evaluation of radionuclides activity in the 2π geometry [AT6101D Spectrometer.Operation manual, 2014].
No 662 keV peak can be seen in the pulse height distributions recorded at any of the background plots.An example of a spectrum obtained at a background area is shown in Fig. 2. A spectrum recorded at a contaminated area is also shown for comparison to demonstrate a prominent 662 keV peak from 137 Cs-137m Ba.The ground contamination density by 137 Cs at the background plots was found to be below the detection limit (1.3 kBq m -2 ).This value should be taken with care because the detection limit, which is recommended by the manufacturer of the AT6101D spectrometer, is equal to 4 kBq m -2 .But in any case, the residual 137 Cs contamination at our background plots is rather low (less than 5 kBq m -2 ).
Activity concentrations of terrestrial radionuclides were in the ranges: 458-1870 Bq kg -1 for 40 K, 11-102 Bq kg -1 for 226 Ra and 22-161 Bq kg -1 for 232 Th.Values of AC eff varied from 80 Bq kg -1 to 442 Bq kg -1 .The minimum values were registered at ground plots while the highest activities were recorded at sites paved with granite.Asphalted plots occupied the intermediate position in terms of recorded activity.A similar pattern can be seen with respect to dose rates registered at plots covered by soil, asphalt and granite material (Table 2).
To derive the ADER to AC eff ratio for the two heights 0.1 m and 1 m, a linear regression analysis was applied to the data obtained at 12 plots.The characteristics and radiometric data for the plots are presented in Table 3, while the constructed regression lines and regression equations are shown at Fig. 3 (the top and middle panels).The median dose rate at 1 m height is slightly smaller (by 3%) than that at a height of 0.1 m (Table 3).A similar difference (by 5% on average) between the two heights can be seen with respect to AC eff .Although small, the differences between detector heights over ground for both radiometric parameters are statistically significant (the non-parametric Wilcoxon signedrank test, P < 0.01, n=12).Beck and de Planque [4] calculated exposure rates in air from natural gamma emitters homogeneously distributed in soil.A 2% reduction in the exposure rates was found with increasing the height of measurement point from 0 m to 1 m above the ground (see Table 3
The slope coefficient of the regression line (Fig. 3) was calculated as 0.496 [s.e.= 0.013] nSv h -1 per Bq kg -1 for the 0.1 m height and 0.529 [s.e.= 0.011] nSv h -1 per Bq kg -1 for the 1 m height.The slope coefficients can be interpreted as conversion coefficients from AC eff to the ambient dose equivalent rates due to terrestrial radionuclides of 238 U series, 232 Th series and 40 K (ADER TRN ).The Student's t-test showed that the two slopes did not significantly differ from each other (t-value = 1.938,P > 0.05).Therefore we have decided to pool data from the two heights together to derive a single coefficient for further analysis of experimental data.The scatter plot and regression equation for the pooled data are presented in the lower panel of Fig. 3.As one can see from this Figure, the unified ADER TRN to AC eff ratio (conversion coefficient) is approximately equal to 0.51(rounded from 0.509) nSv h -1 per Bq kg -1 .
The intercept of the regression line of the pooled data (Fig. 3, the lower panel) was calculated as 11 nSv h -1 [s.e.= 2 nSv h -1 ; 95% confidence interval = (8-14) nSv h -1 ].The intercept can be interpreted as the sum of contributions from intrinsic noise of the spectrometer, cosmic radiation, the natural radionuclides 7 Be and 22 Na, and a residual 137 Cs contamination.Cosmic radiation and intrinsic noise of the spectrometer are the major contributors to ADER reading at zero value of AC eff .These sources give 7 nSv h -1 .The 7 Be and 22 Na contributions to the ambient gamma dose rate are very small, far below 1 Радиационная гигиена Том 10 № 1, 2017 Cs contamination of 4 kBq m -2 and the dose rate conversion factor DCF 137Cs = 1 nSv h -1 per 1 kBq m -2 [3], we can estimate the 137 Cs input to ADER at the background plots by a value of 4 nSv h -1 .The actual 137 Cs input may be substantially lower than the theoretical one because the DCF 137Cs value of 1 (nSv h -1 )/(kBq m -2 ) was deduced for undisturbed meadows before 2011.The soils at the majority of our plots had been tilled, which may reduce the normalized dose rate due to 137 Cs with a factor of ~2.For paved surfaces, the reduction factor for the DCF 137Cs may be even greater: ~ 6 [2].

Validation of the additional calibration
The validity of the derived relationship between ADER TRN and AC eff was checked at 15 different plots (Table 4).Expected ADER reading (nSv h -1 ) was calculated from measured AC eff (Bq kg -1 ) using the formula: where DR 0 is the spectrometer-dosimeter reading at zero value of AC eff (11 nSv h -1 ) and CF is the conversion coefficient from AC eff to ADER TRN [0.51 (nSv h -1 )/(Bq kg -1 )].
A very good consistency between the ADER calculated with AC eff and the measured ADER has been found (the last column in Table 4).The deviation from unity for the ratio between calculated and measured values did not exceed 8%; the mean and median ratio was 0.99.To some extent, the good agreement is associated with the fact that the same spectrum was used for derivation of AC eff and ADER, although different calibration procedures were used to derive values of AC eff and ADER.
It is also interesting to compare the experimental value of 0.51 (nSv h -1 )/(Bq kg -1 ) for AC eff to ADER TRN conversion coefficient with theoretical expectations.Bossew et al. [3] summarized available literature data and concluded that dose conversion coefficients differ remarkably between authors (eight references).For example, dose conversion coefficient for 238 U homogeneously distributed in soil varied from 0.357 to 0.463 (nGy h -1 )/(Bq kg -1 ).The Sv/Gy conversion coefficient for U series is estimated to be 1.253 (see [3] and references therein).Therefore, the theoretically expected conversion coefficient from AC eff to ADER TRN may range from 0.45 (nSv h -1 )/(Bq kg -1 ) to 0.58 (nSv h -1 )/(Bq kg -1 ) with a mean value of 0.54 (nSv h -1 )/(Bq kg -1 ) and median value of 0.56 (nSv h -1 )/(Bq kg -1 ).Our experimental value of 0.51 (nSv h -1 )/(Bq kg -1 ) lies within the range and deviates less than 10% from the theoretical mean and median values.
Our calibration coefficients must be used with caution and only for the devices similar to the one employed in this study.We also suppose that each spectrometerdosimeter should be additionally calibrated before using it for decomposition of ADER because systematic uncertainty of the AC eff and ADER measurements is not negligible.It is declared by the manufacturer that this uncertainty may be as large as ±20% for determinations of activity concentration of terrestrial radionuclides and measurements of ADER [AT6101D Spectrometer.Operation manual, 2014].Additional uncertainties of the ADER evaluation using the AC eff data may be associated with variations in vertical distributions of radionuclides in soil or in its cover (e.g., asphalt, flagstones, granite gravel).It can be reasonably assumed that terrestrial radionuclides are distributed uniformly throughout a soil profile [2].However, this assumption is hardly applicable for paved areas.The asphalted pavement has a complex structure consisting of asphalt layer, base and sub-base.Activity concentrations of natural radionuclides may vary between building materials used in construction of roads and pavements [25].Nearby buildings may also influence the outdoor dose rate expected only from soil [10].

In field measurements in the Bryansk region
Table 5 shows the ambient dose equivalent rate, surface ground contamination with 137 Cs, activity concentrations of 40 K, 226 Ra, 232 Th and their 'effective' activity concentrations determined at a height of 0.1 m above the ground at seven typical plots in contaminated areas in the Bryansk region, as measured in 2015-2016.
Activity concentrations of terrestrial radionuclides were in the ranges: 156-582 Bq kg -1 for 40 K, 10-30 Bq kg -1 for 226 Ra and 10-31 Bq kg -1 for 232 Th.Values of AC eff varied from 32 Bq kg -1 to 119 Bq kg -1 .The lowest values were registered in the forest at Zatishie, while the highest activities were recorded at a street covered by asphalt.The median AC eff in soils from the Bryansk region (67 Bq kg -1 , n = 6) was about 1.6 times lower in comparison with the level registered in soils from St.-Petersburg and the Leningradskaya region (111 Bq kg -1 , n = 20).The difference between the groups is statistically significant (the Mann-Whitney U test, P < 0.01).
Measured ground deposition densities (inventories) of 137 Cs (Table 5) ranged widely from 3.1 kBq m -2 to 447 kBq m -2 .The minimum inventory was found at a street in Novozybkov.The street had been covered with new asphalt several years after the Chernobyl accident.The low value of the 137 Cs inventory at the treated street shows that no significant recontamination of the new asphalt surface has occurred.The maximum inventory was registered at the undisturbed grassland and forest sites located at the Novozybkov's suburbs.The measured ground deposition density of 137 Cs at the forest and grassland (410-450 kBq m -2 ) corresponds well to the contamination level officially reported for Novozybkov in 2014 (~400 kBq m -2 ) [22].The measured ground deposition density of 137 Cs in the forest at Zatishie (45 kBq m -2 ) also correlates very well with the officially established level of 137 Cs contamination: 41 kBq m -2 [22].The measured inventories of 137 Cs at the kitchengarden Les-13gar, yard Les-13ya and grassland Les-13gr in Novozybkov appeared to be 2-3 folds less than the expected 137 Cs inventory of ~400 kBq m -2 .The observed difference can be explained by a deep penetration of 137 Cs at these three plots.The ground areas had been dug by the owner many times after the Chernobyl accident.These observations indicate the necessity of further calibration and testing of AT6101D spectrometer with respect to the quantitative determination of 137 Cs in undisturbed and disturbed soils.
Table 6 shows measured ADER, 'effective' activity concentration of natural radionuclides and estimations of dose rates due to: 1) terrestrial radionuclides (ADER TRN ), and 2) 137 Cs (ADER Cs-tot ) for the height of 1 m above the ground at contaminated areas in the Bryansk region.Ambient dose equivalent rate at a height of 1 m above the ground ranged from 78 nSv h -1 to 546 nSv h -1 .The median value of ADER at the Bryansk region (241 nSv h -1 ; n = 7) is about three times higher than the one in St.-Petersburg and the Leningradskaya region (74 nSv h -1 ; n = 12).The difference between the contaminated and background regions is statistically significant (the Mann-Whitney U test, P < 0.01).No correlation was found between AC eff and ADER at the contaminated plots.Spearman's coefficient of correlation is -0.143 (Р > 0.05; n = 7).At the same time, ADER and 137 Cs ground deposition densities at the plots (n = 7) were strongly positively correlated; Spearman's rank correlation coefficient has been calculated as 0.929 (P < 0.05).
Total dose rate due to 137 Cs, ADER Cs-tot (nSv h -1 ), was calculated using the following formula: where ADER reading is the dosimeter reading; DR 0 is the expected reading of the dosimeter at zero value of AC eff and no 137 Cs contamination (8 nSv h -1 ); AC eff is 'effective' activity concentration of terrestrial radionuclides (Bq kg -1 ); CF is the conversion coefficient from AC eff to ADER TRN [0.51 (nSv h -1 )/ (Bq kg -1 )].
Note that the DR 0 value of 8 nSv h -1 includes the value of 7 nSv h -1 obtained during measurements above the surface of the Finish Gulf.Additionally, 1 nSv h -1 may be attributed to: 1) contribution from natural radionuclides different from 238 U series, 232 Th series and 40 K, and 2) a small difference in altitudes between the surface of the Finnish Gulf (0 m a.s.l.) and the terrains in the Bryansk region (150-170 m a.s.l.).
Calculated values of ADER Cs-tot are shown in column 7 at Table 6, while the contribution of the man-made source to ADER reading is presented in column 8 at Table 6.In Novozybkov, the contribution of 137 Cs to the total ambient dose rate is over 90% at the undisturbed grassland and forest, about 75% at disturbed ground plots and about 40% at an asphalted street.This can be explained by the high residual 137 Cs contamination of soils in Novozybkov: ~ 400 kBq m -2 .The level of the 137 Cs contamination in Zatishie (~ 40 kBq m -2 ) was about ten times lower than that in Novozybkov.Nevertheless, the artificial source dominated the ambient dose rate in the Zatishie's forest where the total measured ADER (78 nSv h -1 ) was comparable with the average ADER (68 nSv h -1 ) determined at nine ground plots from the Leningradskaya region.To a great extent, the strong contribution (70%) of 137 Cs to the total dose in the Zatishie's forest is attributable to the relatively low activity concentrations of natural radionuclides in soil (AC eff = 31.6Bq kg -1 ; Table 5) and the relevant gamma dose rate in air at a height of 1 m above the ground (ADER TRN = 15 nSv h -1 ; Table 6).
The tested spectrometer-dosimeter and technology to separate the natural and 137 Cs components of the ambient dose equivalent rate can be applied not only for the Chernobylaffected territories but also for other sites where radioactive contamination is fully dominated by 137 Cs.For example, these may be areas contaminated as a result of the accident happened in Elektrostal (Russia) in April 2013, or sites of the peaceful underground nuclear explosions which were carried out in the USSR in the last century [26,27].Potentially, the territories contaminated by radiocaesium ( 137 Cs + 134 Cs) after Fukushima accident (2011) can be surveyed using the AT6101D device, although some correction for the presence of 134 Cs might be required.

Conclusions
The results of the measurements confirm the applicability of in situ gamma-ray spectrometry for decomposition of the ambient dose equivalent rate outdoors.After additional calibration it is possible to use a commercially available NaI(Tl) gamma-ray spectrometer-dosimeter to separate the natural and 137 Cs components of the ambient dose equivalent rate.The additional calibration requires performing in situ measurements at the environment which has negligible contamination by 137 Cs (at the level of a few kBq m -2 ).The conversion coefficient from AC eff to ADER TRN of 0.51 (nSv h -1 )/(Bq kg -1 ) has been obtained using a regression analysis of experimental data.The intrinsic noise of the spectrometer and its response to cosmic radiation at sea level has been estimated to 7 nSv h -1 .The dosimeter-spectrometer and experimentally derived calibration coefficients have been tested in field in the Bryansk region that had been heavily contaminated by Chernobyl fallout.The contribution of 137 Cs to the total ADER varies between 40% and 95%.Currently further studies are conducted to: 1) obtain new data from areas contaminated by 137 Cs, and 2) derive calibration coefficients for indoor locations and for backpack based radiation detection systems.a -the plot is located inside settlement.b -the plot is located outside settlement.c -the primary reading of the AT1601D spectrometer-dosimeter.A statistical uncertainty (at the 2 sigma level) of the ADER measurements is less than ±2%.Statistical uncertainties (±%, at the 2 sigma level) for determination of AC eff , ADER TRN and ADER Cs-tot are given in brackets.

Fig. 1 .
Fig. 1.In situ gamma-ray spectroscopic measurements with MKS AT6101D at a kitchengarden in Khittolovo (plot Hit-2), the Leningradskaya region in 2015.The detection unit and the data analysis unit (DAU) are set up on an aluminum tripod.The white arrows indicate approximate position of the effective center of the detector: at a height of about 1 m (the top panel A) and 0.1 m (the lower panel B) above the ground

Fig. 2 .
Fig. 2. In situ NaI(Tl) gamma-ray spectra recorded with MKS AT6101D at the background (plot Hit-2, Khittolovo, the Leningradskaya region) and contaminated (plot Les-13kit, Novozybkov, the Bryansk region) kitchengardens in 2015.Positions of peaks from the natural radionuclides 214 Bi ( 238 U series), 208 Tl ( 232 Th series) and 40 K, and man-made radionuclide 137 Cs are indicated by arrows.Counting time is 925 s for the background plot and it is 695 s for the contaminated plot

Fig. 3 .
Fig. 3. Relationship between total ambient dose equivalent rate (ADER) and 'effective' activity concentration of terrestrial radionuclides (AC eff ) determined with in situ NaI(Tl) measurements (MKS AT1601D) at a height of 0.1 m and 1 m above the ground at background areas in St.-Petersburg and the Leningradskaya region in 2014-2016

Table 1 List of background and contaminated settlements surveyed in 2014-2016
[28,29]initial official 137 Cs inventory is given for 1986 according to[28,29].The values in brackets are the137Cs inventory determined in this study (on the year of the survey) with in situ measurements at a height of 0.1 m above the ground.

Table 3 Values of ambient dose equivalent rate (ADER reading ) and 'effective' activity concentration of terrestrial radionuclides (AC eff ) determined at heights of 0.1 m and 1 m above the ground at 12 background plots in 2014-2016
Радиационная гигиена Том 10 № 1, 2017

Table 4 A comparison between calculated (based on determination of the 'effective' activity concentration of terrestrial radionuclides, AC eff ) and measured values of ambient dose equivalent rate (ADER) at 15 background plots in 2014-2015
a -a statistical uncertainty (±%, at the 2 sigma level) for determination of AC eff is given in brackets.

Table 5 Ambient dose equivalent rate (ADER reading ), surface ground contamination with 137 Cs (A Cs ), activity concentration of terrestrial radionuclides and 'effective' activity concentration of terrestrial radionuclides (AC eff ) determined with in situ measurements (AT1601D) at a height of 0.1 m above the ground at contaminated areas of the Bryansk region in 2015-2016
reading of the AT1601D spectrometer-dosimeter.A statistical uncertainty (at the 2 sigma level) of the ADER measurements is below ±2%.Statistical uncertainties (±%, at the 2 sigma level) of determination of A Cs , AC 40K , AC 226Ra , AC 232Th and AC eff are given in brackets.
b-the plot is located outside settlement.c -the primary