Atmospheric stability effects on potential radiological releases at a [608818]

Atmospheric stability effects on potential radiological releases at a
nuclear research facility in Romania: Characterising the atmosphericmixing state
Scott D. Chambersa,*, Dan Galeriub, Alastair G. Williamsa, Anca Melintescub,
Alan D. Grif fithsa, Jagoda Crawforda, Leisa Dyera, Marin Dumab, Bogdan Zorilab,c
aAustralian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia
b“Horia Hulubei ”National Institute for Physics and Nuclear Engineering, 30 Reactorului St., POB MG-6, 077125 Bucharest, Magurele, Romania
cDepartment of Electricity, Solid Physics and Biophysics, Faculty of Physics, University of Bucharest, Magurele, Romania
article info
Article history:
Received 10 November 2015Received in revised form13 January 2016Accepted 17 January 2016Available online xxx
Keywords:
222Rn
Atmospheric stability
Mixing depth
Radon flux
Radioactive releasesTritiumabstract
A radon-based nocturnal stability classi fication scheme is developed for a flat inland site near Bucharest,
Romania, characterised by signi ficant local surface roughness heterogeneity, and compared with tradi-
tional meteorologically-based techniques. Eight months of hourly meteorological and atmospheric radon
observations from a 60 m tower at the IFIN-HH nuclear research facility are analysed. Heterogeneous
surface roughness conditions in the 1 km radius exclusion zone around the site hinder accurate char-acterisation of nocturnal atmospheric mixing conditions using conventional meteorological techniques,so a radon-based scheme is trialled. When the nocturnal boundary layer is very stable, the Pasquill
eGifford “radiation ”scheme overestimates the atmosphere's capacity to dilute pollutants with near-
surface sources (such as tritiated water vapour) by 20% compared to the radon-based scheme. Underthese conditions, near-surface wind speeds drop well below 1 m s
/C01and nocturnal mixing depths vary
from ~25 m to less than 10 m above ground level (a.g.l.). Combining nocturnal radon with daytime
ceilometer data, we were able to reconstruct the full diurnal cycle of mixing depths. Average daytimemixing depths at this flat inland site range from 1200 to 1800 m a.g.l. in summer, and 500 e900 m a.g.l. in
winter. Using tower observations to constrain the nocturnal radon-derived effective mixing depth, we
were able to estimate the seasonal range in the Bucharest regional radon flux as: 12 mBq m
/C02s/C01in
winter to 14 mBq m/C02s/C01in summer.
Crown Copyright ©2016 Published by Elsevier Ltd. All rights reserved.
1. Introduction
Nuclear facilities are commonly required to monitor their
emissions of radioactive gases and aerosols to the environment, in
order to gauge the integrated environmental impacts of routine
releases of pollutants, and to help in the forecasting of potential
health risks associated with accidental releases ( Galeriu et al., 2014;
IAEA, 2011a,b ; EURATOM TREATY, http://www.euratom.org/ ). An
important component of regulatory monitoring programs is the
routine measurement of meteorological quantities that can be used
to characterise the state of the atmosphere and its ability to dilute
the emitted pollutants and transport them away from the area.
Near-surface concentrations of hazardous air-borne pollutantsto which workers, residents, wildlife or crops may be exposed, are
primarily a function of the source strength, the volume of the at-
mosphere into which they mix, and entrainment processes (e.g. Pal,
2014 ). For pollutants with sources typically below the height of the
nocturnal inversion layer, concentrations will usually peak in calm
pre-dawn conditions, when the lower atmosphere is most poorly
mixed (e.g. Avino et al., 2003; Baciu, 2005; Galmarini, 2006; Pearce
et al., 2011; Crawford et al., 2016; Grundstr €om, 2015; Chambers
et al., 2015a,b ). On the other hand, emissions from elevated sour-
ces (e.g. tall stacks) frequently result in peak concentrations shortly
after dawn (so-called “fumigation events ”;Oke, 1987 ), when tur-
bulence erodes the nocturnal inversion and incorporates the
overlying air into the developing convective boundary layer.
Since establishing and maintaining dense regional monitoring
networks is logistically and economically prohibitive, efforts tounderstand the impact of radioactive releases into the atmosphere*Corresponding author.
E-mail address: szc@ansto.gov.au (S.D. Chambers).
Contents lists available at ScienceDirect
Journal of Environmental Radioactivity
journal homepage: www.elsevier.com/locate/jenvrad
http://dx.doi.org/10.1016/j.jenvrad.2016.01.010
0265-931X/Crown Copyright ©2016 Published by Elsevier Ltd. All rights reserved.Journal of Environmental Radioactivity 154 (2016) 68 e82

tend to rely mainly upon plume dispersion modelling and our
ability to accurately assess the “stability ”(degree of mixing) of the
lower atmosphere. The most accurate conventional techniques to
characterise atmospheric stability (e.g., Richardson number, Obu-
khov length; Foken, 2008 ) are based on surface and boundary layer
similarity theory. However, since these require sophisticated
(research quality) instruments that are expensive and hard to
maintain routinely, they are not widely adopted. Consequently, less
accurate categorical techniques derived from routine meteorolog-
ical measurements, such as the Pasquil-Gifford turbulence and ra-
diation schemes ( Pasquil, 1961; Turner, 1964; Pasquill and Smith,
1983 ), tend to be more commonly adopted.
A new approach to stability classi fication that is gaining
acceptance relies upon observations of the naturally-occurring
radioactive trace gas Radon-222 (e.g. Perrino et al., 2001; Perrino,
2012; Chambers et al., 2015a,b , and references therein). Because
radon is emitted steadily by all soils and rocks, it has a surface
source function that varies slowly in both space and time. In
particular, diurnal radon source variations can be considered to be
constant at most inland sites. Once in the atmosphere, radon is
unreactive, poorly soluble and decays at a rate that is small incomparison to vertical mixing time scales. As a result, near-surface
radon concentrations represent a direct measure of the intensity
and vertical extent of nocturnal atmospheric mixing, and perform
better than meteorological proxies in characterising the outcomes
of vertical mixing on surface-emitted passive scalar quantities
(IAEA, 2012; Williams et al., 2013; Chambers et al., 2015b ).
Horia Hulubei National Institute for Research and Development
in Physics and Nuclear Engineering (IFIN-HH) is the largest research
institute in Romania. Close to Bucharest, the site is located in a
topographically-complex mixed urban-rural landscape. In order to
better assess the fate of the radioactive gases and aerosols that are
exhausted from this facility via a 40 m stack, a comprehensive suite
of meteorological measurements have been made since 1995
(when the reactor and radiopharmaceutical production facilities
were operational) from a nearby 60 m tower ( Galeriu et al., 2014 ).
The effective release height at IFIN-HH site is similar to that at the
nearby Cernavoda Nuclear Power Plant (CNPP), which currently
produces around 20% of Romania's electricity. CNPP is also situated
in relatively complex terrain, and is the subject of ongoing research
on the in fluence of atmospheric stability on pollutant concentra-
tions using conventional meteorological techniques.
Since 1997, when reactor operations at IFIN-HH ceased, the
meteorological tower has served as a real-time data provider for
the RODOS (Real-time On-line DecisiOn Support) decision support
system for nuclear emergencies ( Galeriu et al., 2011; Ehrhardt,
1997 ). In particular, FDMH (Food and Dose Module Hydrogen), a
component of RODOS, was used to investigate the radiological
impact of accidental releases of tritiated water and transfer from
the atmosphere to the food chain or inhalation pathways ( IAEA,
2014; Galeriu et al., 2000 ). Since the start of the reactor decom-
missioning process in 2007 it has become necessary to increase the
levels of mandatory meteorological surveying, and tower obser-
vations also began to be made available for use in land eatmosphere
interaction studies.
At night, when atmospheric conditions are stable, near-surface
concentrations of tritiated water (HTO) increase, thereby
increasing the radiological hazard. Under these conditions HTO
more readily makes its way into soil water or enters the crops in the
surrounding rural areas and is converted to organic forms. If tritium
from HTO binds to carbon within the crops, organically bound
tritium (OBT) is formed. Since OBT is a very stable organic form the
radiological hazard (through risk of subsequent ingestion and
incorporation into DNA) is increased. At present crop uptake of HTO
at night and subsequent formation of OBT is not well understood(Melintescu et al., 2015; Galeriu and Melintescu, 2011, 2015 ). Of the
few models used to estimate the OBT:HTO ratio that consider
nocturnal uptake, parameterisations of the process are very
simplistic ( Melintescu and Galeriu, 2015; IAEA, 2014; Melintescu
et al., 2015; Galeriu and Melintescu, 2015 ).
Nocturnal atmospheric stability classi fication by conventional
meteorological approaches has proven to be problematic at sites
with signi ficant surface roughness heterogeneity such as IFIN-HH
(Galeriu et al., 2014 ). In light of this issue, the current study aims
to: (i) develop a versatile, continuous (i.e., not categorical) classi-
fication scheme for the atmospheric mixing state at the IFIN-HH
site based on near-surface atmospheric radon measurements, (ii)
characterise the behaviour of key meteorological quantities within
the de fined mixing states, (iii) compare the ef ficacy and consistency
of the Pasquil-Gifford and radon-based approaches for nocturnal
atmospheric stability classi fication by assessing their ability to
distinguish changes in key parameters (including near-surface
wind speed and gradients, temperature gradients, and net radia-
tion) between identi fied stability categories, and (iv) characterise
changes in atmospheric mixing depth across the diurnal cycle for a
range of atmospheric stabilities in summer and winter. Detailedinvestigation of identi fied stability in fluences on nocturnal tritium
concentrations, as well as improvement of parameterisations of
tritium uptake by crops, will be the subject of a subsequent
investigation.
2. Materials and methods2.1. Site and measurements
IFIN-HH is situated approximately 10 km SSW of the Bucharest
central business district (44
/C142102.7200N, 26/C1402038.4200E) in a mixed
urban-rural landscape ( Fig. 1 ). Locally, the surface is quite inho-
mogeneous and, in places, roughness elements (including trees and
buildings) reach 10 e15 m above ground level, a.g.l. When the
reactor and radiopharmaceutical production at this site were in
operation, radioactive gases and aerosols were exhausted from this
facility via a 40 m stack, with the effective source height ranging
from 44 to 55 m. However, since operations at this facility ceased,
radioactive emissions from this site have been minor and often
hard to distinguish from background.
A 60 m instrumented tower is located on the premises of this
research facility. Wind speed, wind direction, temperature and
relative humidity are monitored at 30 and 60 m a.g.l., solar radia-
tion is monitored at 10 m, net radiation and rainfall at 30 m agl. All
observations are logged as 10-min averages of 10-s readings and
subsequently integrated to hourly averages for analysis. In addition
to the meteorological observations, atmospheric radon concentra-
tion is monitored at 10 m a.g.l., using an “AlphaGUARD ”(PQ2000
PRO, Saphymo, Germany; which also monitors atmospheric pres-
sure, temperature and relative humidity). The AlphaGUARD was
situated in a Stevenson's Screen (well-ventilated, weatherproof
enclosure) to protect it from precipitation, operated in diffusion
mode, and set for hourly integration.
Information regarding cloud cover, cloud height, atmospheric
mixing depth and aerosol properties is recorded nearby using a
vertically-oriented modi fied CHM 15k Nimbus ceilometer (Nimbus
Jenoptik, Germany). The ceilometer's measurement range is
5me15 km, with a 15 m resolution and full overlap range of
300e400 m. Its principle of measurement is similar to that of a
lidar, using light detection and ranging to target aerosols and cloud.
For the purpose of this investigation the ceilometer's 30 s output
was averaged to hourly resolution, and mixing depth was taken to
be the height of the first identi fied inversion (PBL_0). Further de-
tails regarding the tower instrumentation are provided in GaleriuS.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 69

et al. (2014) .
Due to a range of technical dif ficulties, primarily with operation
of the AlphaGUARD unit, the longest, most consistent period of
overlap between all of the abovementioned observations available
at the time of this study was 8 months (May eDecember 2012).
2.2. Radon as an atmospheric tracer
Radon (222Rn) is a naturally-occurring, unreactive, poorly solu-
ble, radioactive gas with a relatively well-de fined source function
from ice-free, unsaturated terrestrial surfaces that varies little with
space or time compared to that of most anthropogenic emissions.
The half-life of radon (T 1/2¼3.82d) is suf ficiently long for it to be
considered a conservative tracer over the course of a single night,
but short enough for radon not to accumulate in the atmosphere,
and typically exhibit an order of magnitude gradient between the
atmospheric boundary layer (ABL) and the lower troposphere. This
combination of physical characteristics enables radon to be used as
a quantitative proxy for the effects of near-surface vertical mixing
on scalar quantities ( Perrino et al., 2001; Sesana et al., 2003; Baciu,
2005; Galmarini, 2006; Galeriu et al., 2011; Williams et al., 2013;
Chambers et al., 2015a,b ).
2.3. Atmospheric stability measures
A range of stability typing methods have been tested at the IFIN-
HH site. The generally-preferred climatological technique classi fies
atmospheric stability based on net radiation and estimated 10 m
wind speed, and results in six stability states over the diurnal cycle,
very similar to those of the Pasquill Gifford radiation scheme (PG_R;
Turner, 1964 ). The same stability classi fication technique was
adopted by Baciu (2005) for their 5-year study of radon in
Bucharest sampling at 2 m a.g.l. Since 10 m wind speed observa-
tions are not available on the IFIN-HH tower, for routine operations
they are estimated from the 30 and 60 m wind speed observations,
based on the roughness effect on the wind pro file assuming open,flat terrain.
Of the other conventional measures available, here we choose to
adopt the bulk Richardson number ( RiB; Eq. (1)) for our mostly-
nocturnal stability classi fication requirements, since the Obukhov
length can be dominated by errors when turbulence levels are very
low (Mahrt et al., 2006; Banta et al., 2007 ). Negative RiBvalues are
indicative of convective mixing, whereas values in excess of 0.25
(the “critical ”Richardson number, RiC;Stull, 1988 ) indicate strong
suppression of vertical turbulence by thermal strati fication.
RiB¼
g
qv!/C18Dqv
DU2/C19
Dz (1)
where gis the acceleration due to gravity (m s/C02),qvis virtual po-
tential temperature (/C14K),Uis wind speed (m s/C01) and zis the
measurement height (m). The Doperator here indicates a differ-
ence between the 30 m and 60 m levels on the IFIN-HH tower.
It is worth noting that since the PG-R and RiBclassi fications for
the INIF-HH tower site are derived from measurements between 30
and 60 m agl., both of which can be above the height of the
nocturnal inversion under conditions of moderate to strong sta-
bility (see Section 3.3), they are unlikely to be equally representa-
tive of the true atmospheric state of the nocturnal boundary layer
under all conditions.
Unlike the PG-R and RiBschemes, which classify atmospheric
stability on an hourly basis, the radon-based approach employed in
this study classi fies entire days (24-h periods), based on near-
surface radon concentration observations performed in a 10-h
nocturnal window. A slightly modi fied version of the approach
originally introduced by Chambers et al. (2015b) is brie fly outlined
below.
The range of radon concentrations recorded at the IFIN-HH site
(2e80 Bq m/C03;Fig. 2 a) was similar to that reported in Bucharest by
Baciu (2005) and in northwest Romania by Florea and Duliu (2012) .
The observed variability in the hourly radon time-series is a
Fig. 1. Location of the IFIN-HH site relative to Bucharest. Also shown is the 1 km buffer zone around the site and the relative proportion of urban and agricultu ral areas. Inset shows
location of IFIN-HH site in Romania and distance from the Black Sea coast.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 70

superposition of variability on a range of temporal scales (hourly,
diurnal, synoptic and seasonal). While contributions on greater
than diurnal timescales are most closely related to the air mass'
recent fetch history (over the past 2 e3 weeks, due to radon's half-
life), only the diurnal variability can be directly related to daily
changes in the atmospheric mixing state (or “atmospheric stabil-
ity”;Chambers et al., 2015a,b ).
As also noted for Bucharest by Baciu (2005) , the diurnal radon
cycle at IFIN-HH ( Fig. 3 a) is characterised by a morning maximum,
coinciding with sunrise ( Fig. 3 b), and afternoon minimum
(1400e1600 h) when the daytime convective boundary layer is
deepest. On average, morning radon concentrations first begin to
decrease at 0700 h when net radiation becomes positive and con-
vection starts to erode the nocturnal boundary layer. Only when net
radiation drops to less than 50% of its peak daytime value in the late
afternoon does radon accumulation typically recommence.
Once the diurnal component of radon variability has been iso-
lated (i.e., difference between the black and red curves in Fig. 2 b),the mean nocturnal increase of the ‘diurnal ’radon signal from
1900h e0500 h ( Fig. 3 a) can be calculated (Eq. (2)) and directly
related to the mean state of atmospheric mixing for that night. The
chosen diurnal window is site speci fic. Ideally, the start of the
window should correspond to sunset in mid-winter, and the end of
the window should be the hour before sunrise in mid-summer.
Furthermore, the nocturnal radon increase should be referenced
to the radon concentration at the start of the window. In this way,
one stability classi fication scheme can be uniformly applied to all
seasons. The only exception to this rule is at sites where there is a
large seasonality in the strength or homogeneity of the radon
source function, as discussed by Williams et al. (2015) .
RnNInc¼ Pt¼0500
t¼1900ðRnt/C0Rn1900/C17
10!
(2)
Atmospheric stability categories can then be assigned based on
chosen thresholds of the nocturnal radon increase ( RnNInc). In140 160 180 200 220 240 260 280 300 320 340020406080Radon (Bq m-3)(a)
300 305 310 315 320 325 330 335 340020406080
(b)
Day of 2012 Hourly radon
Fetch contribution
Fig. 2. (a) Hourly 10 m radon concentrations observed at INIF-HH, and (b) a subset of (a) with the synoptic time-scale “fetch ”contributions to the radon concentration variability
indicated in red. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
16 18 20 22 0 2 4 6 8 10 12 146912151821Radon (Bq m-3)
Hour of composite day(a)
16 18 20 22 0 2 4 6 8 10 12 140100200300400500600
accumulation
window(b)Radiation (W m-2) Solar
Net
Fig. 3. Typical diurnal cycles at IFIN-HH of (a) radon concentration at 10 m, and (b) global solar and net radiation. Note that the time axes have been adjusted t o focus on the
nocturnal portion of each day. All reported times are local time (LST ¼UTCț2 h).S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 71

Fig. 4 a, four stability categories have been arbitrarily assigned at the
IFIN-HH site based on the quartile ranges of nocturnal radon
accumulation observed over the 8-month period (see also Table 1 ).
The number of possible categories will depend upon the size of the
available dataset, on the understanding that each category should
contain enough events to generate robust statistics. Since there was
approximately 220 days of valid observations in this study, subse-
quent statistics are based on ~55 events (whole days) per stability
category (or ~27 events per category when separated by quasi-
season, as described below).
Once a stability category has been assigned to each 10-h
nocturnal window, the same category is subsequently assigned to
the entire 24-h period containing that night (1500 e1500 h). The
mid-afternoon (1500 h) delineation point was chosen since this is
when the atmosphere is generally most well mixed, and concen-
trations of radon (as well as other primary pollutants of near sur-
face origin) reach their minimum values (e.g. Fig. 3 a).
We grouped days of the whole dataset according to this radon-
based classi fication scheme. Diurnal composites of the 10 m radon
concentration in each of the four stability categories are shown in
Fig. 4 b. For stable nocturnal conditions, the diurnal radon cycle had
the largest amplitude. As will be seen in Section 3.3(Fig. 16 fig), this
corresponds to the largest diurnal change in mixing depth. For
near-neutral (well-mixed) conditions, on the other hand, theconverse was true. We then separated the dataset into the four
warmest (Jun eSep) and four coolest (May, Oct eDec) months of the
8-month study period (see Table 2 ), which, for the sake of
simplicity, we henceforth refer to as “summer ”and “winter ”
(although they don't correspond to actual seasons). Diurnal radon
composites for our quasi-seasons are shown in Fig. 5 . Importantly,
no large changes in the rates of radon accumulation from sunset
and midnight is evident between summer and winter in any of the
stability categories, indicating that any seasonality that exists in the
characteristics of the regional radon source function is likely to be
small.
2.4. Characterising tower in fluences and the fetch environment at
IFIN-HH
The predominant wind directions at the tower site were north
through east, and southwest ( Fig. 6 ; see also Galeriu et al., 2014 ).
Only for about 10 e15% of the year did winds track from the research
facility over the Bucharest central business district. As shown inFig. 1 , a heavily vegetated “exclusion zone ”with a radius of
approximately 1 km exists around the site. The combination of
vegetation and research facility infrastructure result in a hetero-
geneous distribution of roughness elements about the measure-
ment tower as indicated by the roughness length map ( Fig. 7 ). The
least obstructed approaches to the tower are from the west and
north. Since the lowest height for routine wind and temperature
observations is 30 m (twice the height of the tallest local roughness
elements), it is dif ficult to accurately assess nocturnal atmospheric
stability conditions using conventional meteorological techniques.
Analysing the 60 e30 m wind speed gradient as a function of at-
mospheric stability and wind direction (
Fig. 8 ), a clear tower in-
fluence was observed in the SSE quadrant due to the location of the
instruments on a short (1.5 m) boom at ~346/C14. Further information
regarding the potential tower in fluence on IFIN-HH meteorological
observations is given by Galeriu et al. (2014) .
3. Results and discussion
3.1. Comparison of stability classi fication techniques at the IFIN-HH
tall-tower site
3.1.1. Pasquill eGifford scheme
In addition to the standard hourly PG-R classi fication scheme
mentioned in Section 2.3, for the purposes of more direct com-
parison with the radon-based classi fication method we also
devised a “whole day ”PG-R scheme, whereby nights were classi-
fied in their entirety according to the most commonly assigned
hourly PG-R value between 2000 and 0500 h (inclusive). In an
analogous way to the radon-based classi fication scheme, this
classi fication was subsequently applied to a 24-h period
(1500e1500 h).0 3 6 9 12 15 18050100150200
Near
neutralWeakly
stableModerately stableCumulative frequency
Mean nocturnal radon (Bq m-3)Stable(a)
16 18 20 22 0 2 4 6 8 10 12 1404812162024(b)Radon (Bq m-3)
Hour of composite day Near-neutral
Weakly stable
Moderately stable
Stable
Fig. 4. (a) Cumulative frequency distribution of mean nocturnal radon concentration from 1900 h to 0500 h (referenced to 1900 h value and shifted by ț3B qm/C03to avoid negative
concentrations), and (b) mean diurnal radon cycles for each of the stability categories.
Table 1Atmospheric stability categories, speci fic to the IFIN-HH tower site, as de fined by
nocturnal mean radon accumulation.
Quartile Nocturnal mean
diurnal radonNocturnal mean
stability CategoryVertical mixing
Q1 <1.37 Bq m
/C03Near neutral Strong
Q2 1.37 e4.24 Bq m/C03Weakly stable Moderate
Q3 4.24 e7.34 Bq m/C03Moderately stable Weak
Q4 >7.34 Bq m/C03Stable Very weakS.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 72

Diurnal radon composites over the whole dataset for the three
nocturnal PG-R classi fications (4 ¼neutral, 5 ¼moderately stable
and 6 ¼stable) are shown in Fig. 9 . By this scheme, 18% of nights
were classi fied as neutral, 44% as moderately stable and 38% as
stable (compared to the even quartile split between four categories
for the radon-based method). Direct comparison of these results
with Fig. 4 b shows that the PG-R and radon-based “neutral ”clas-
sification schemes result in a similar amplitude diurnal radon cycle.
The moderately stable PG-R category appears to yield results be-
tween the moderately mixed and weakly mixed radon-based cat-
egories, and the stable PG-R category falls between the weakly
mixed and stable radon-based categories.
An alternative means of comparing the radon-based and PG-R
classi fication schemes is to group hourly PG-R classi fications by
radon-based stability classes. Fig. 10 a,b show diurnal composites of
median hourly PG-R classi fications grouped by radon-derived sta-
bility category for summer and winter data separately. Summer
nights classi fied as stable according to the radon-based scheme
(Fig. 10 a; blue inverted triangles) were also usually classi fied as
stable (iPG ¼6) according to the PG-R scheme. Most other nights in
summer, however, were classi fied as moderately stable according
to the PG-R scheme ( Fig. 10 a), despite a wide range of conditions
that are well resolved by the radon-based scheme ( Fig. 5 a). The PG-
R scheme attributed a wider range of stability classi fications to the
first half of winter nights ( Fig.10 b). Shortly after midnight in winter,
however, the PG-R scheme appeared less capable of consistentlyresolving differences in stability categories. This period coincided
with a “flattening ”of the winter nocturnal radon accumulation
under “stable ”conditions ( Fig. 5 b; see Section 3.3for further dis-
cussion), as well as more variability in the “moderately stable ”
nocturnal radon accumulation than was the case in summer. This
seasonal change in nocturnal performance of the PG-R classi fica-
tion scheme is likely attributable to a seasonal change in the rela-
tive contributions of thermal strati fication and wind-induced
mechanical mixing to the atmospheric state of the nocturnal
boundary layer at IFIN-HH ( Fig. 11 ; see also Section 3.2and Fig. 13
eeh). To demonstrate this point, Fig. 11 shows composite plots
the diurnal component of the 10 m radon concentration as a
function 60 e30 m virtual potential temperature and wind speed
gradients between 11pm and 3am each night (to avoid cases when
mixing depth drops below 10 m a.g.l., see Section 3.3). In winter, the
observed radon concentration is a stronger function of thermal
strati fication than in summer. Conversely, the broader range of
60e30 m wind speed gradients observed in summer indicate
substantially reduced mechanical mixing at night.
Despite the radon-based classi fication scheme being derived
purely using nocturnal data and subsequently applied to whole15 17 19 21 23 1 3 5 7 9 11 13051015202530Diurnal radon (Bq m-3)
Hour of composite day Near neutral
Weakly stable
Moderately stable
Stable(a) Summer
15 17 19 21 23 1 3 5 7 9 11 13051015202530
(b) Winter
Fig. 5. Diurnal composite 10 m radon concentrations (diurnal-component only) for the (a) warmest, and (b) coolest, 4 months of the 8-month dataset.
0-45 45-90 90-135 135-180 180-225 225-270 270-315 315-3600102030% of total hourly observations
Wind direction sector
Fig. 6. Relative frequency of hourly wind direction (in 45/C14bins) at the IFIN-HH tower
site.
Fig. 7. Roughness length (z o[m]) map of a 12 /C212 km region centred on the INIF-HH
site.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 73

days, there is a close correspondence between the hourly PG-R and
radon classi fications over the entire diurnal cycle. Days that expe-
rience the most stable nocturnal conditions are frequently associ-ated with clear-sky conditions (which enable rapid surface cooling
and strong thermal strati fication). Clear-sky conditions are also
usually associated with the strongest convective mixing during the
day. Fig. 10 a shows that days categorised as “stable ”by the radon-
based scheme show the most rapid transitions between stable
nocturnal conditions (iPG ¼6) and unstable daytime conditions
(iPG¼1). Furthermore, the slowest transition between these states
occurs for the days classi fied as near-neutral by the radon-based
scheme ( Fig. 10 a; black squares). The apparent success with
which the radon-based classi fication scheme can be applied to a
whole 24-h period is attributable to the short-term persistence of
synoptic atmospheric conditions. Based on almost 3 years of ob-
servations in eastern Australia, Chambers et al. (2011) demon-
strated that on almost 75% of occasions (i.e., whenever fronts or
other severe weather systems are not actually moving through the
area), a given atmospheric state (i.e. cloudy/clear, windy/calm, etc.)will usually persist for most of a day.
While both the PG-R and radon-based classi fication schemes
predict a similar range of radon concentrations under near-neutral
conditions, the radon-based scheme estimates a diurnal amplitude
almost 25% greater than that of the PG-R scheme under stable
conditions. Given that the behaviour of radon (a passive atmo-
spheric tracer with a surface source) would be fairly representative
of the behaviour of tritiated water vapour (or other gaseous pol-
lutants with sources beneath the nocturnal inversion layer), the
choice of stability classi fication technique could have serious im-
plications for public or ecosystem exposure estimates in the event
of a release from the IFIN-HH site.
To provide an indication of the atmosphere's capacity for
dispersing pollutants under stable atmospheric conditions as
determined by the radon-based and PG-R classi fication schemes,
distributions (10th, 50th, 90th percentiles) of hourly radon con-
centrations over the diurnal cycle for days classi fied as having
“stable ”nocturnal conditions are shown in Fig. 12 , for both summer
and winter cases. Peak median hourly concentrations are under-
estimated by the PG-R scheme by around 20%, and in summer, the
10th percentile concentrations are low by more than 50%. Underthe most stable conditions, however, it is likely that the stack height
(40 m a.g.l.) would be well above the height of the nocturnal
inversion (see Section 3.3), which, as discussed by Crawford et al.
(2016) , would reduce the magnitude of near-surface
concentrations.
3.1.2. Bulk Richardson number
In both summer and winter, the median hourly nocturnal Ri
B
values based on all days classi fied as “near-neutral ”by the radon-
based scheme were below the critical Richardson number,
RiC¼0.25 ( Fig. 10 ), indicating mixing and the lack of large thermal
gradients in the lowest 60 m of the nocturnal boundary layer. In
summer, the median hourly nocturnal RiBvalues exceeded the
critical value by progressively larger amounts for weakly stable,
moderately stable and stable radon-based categories ( Fig. 10 c).
While winter median hourly RiBvalues also exceeded the critical
value for all stable conditions as classi fied by the radon-based
scheme, their range was greatly reduced, and the distinction be-
tween RiBvalues for the weakly and moderately stable categories
was less clear ( Fig. 10 d). As evident in Figs. 11 and 13 (e-h) , and
further discussed in Section 3.3, this change in RiBbehaviour with
radon-based stability classi fication can be largely attributed to
changing relative contributions of wind and temperature gradients
to the nocturnal stability regime.
3.2. Characterisation of meteorological parameters for the radon-
based stability categories
Meteorological observations from the IFIN-HH tower were
grouped into radon-based stability category and separated by
summer/winter quasi-seasons (see Section 2.3). Diurnal cycles of
their hourly mean values are shown in Fig. 13 (standard deviations
have been omitted to avoid crowding the figures). Near-neutral
conditions, often associated with stronger winds and cloud cover,
had lower values of incoming solar radiation in both seasons,whereas “stable ”and “moderately stable ”days in summer exhibi-
ted the highest values, consistent with clear-sky conditions ( Figs. 13
a,b). The mean diurnal cycles of incoming solar radiation in winter
under weakly and moderately mixed conditions were more vari-
able. This is likely attributable to re flection off the sides of patchy
cloud-cover at this time of lower solar elevation.
Near-neutral conditions typically exhibited the lowest rates of
surface cooling (net radiation; Fig. 13 c,d), due to the insulating
effect of increased nocturnal cloud cover. Of the other evenings, the0 45 90 135 180 225 270 315 360-0.50.00.51.01.52.0U60m – U30m (m s-1)
Wind direction (oN) near neutral
weakly stable
moderately stable
stable
Fig. 8. Median 60 e30 m wind speed gradient between 0000 and 0700 h each night as
a function of wind direction and radon-derived atmospheric stability classi fication.
15 17 19 21 23 1 3 5 7 9 11 13048121620Radon (Bq m-3)
Hour of composite day Neutral
Moderately stable
Stable
Fig. 9. Diurnal composite 10 m radon concentrations for the three daily PG-R
classi fications.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 74

stable nights (clearest sky conditions) exhibited the highest rates of
change in surface cooling over the night. Weakly and moderately
mixed nights maintained slightly more consistent rates of cooling
throughout the night as a result of higher near-surface wind speeds
than for the most stable nights (as shown in Section 3.3). Near-
neutral days were characterised by the smallest virtual potential
temperature ( qv) lapse rates between 30 and 60 m ( Fig. 13 e,f).
Nightly-average qvlapse rates progressively increased with
increasing strength of atmospheric stability as indicated by the
radon-based classi fication scheme, although overall magnitudes
were lower in winter than summer.
In summer, the 30 m wind speeds reduced fairly consistently
with increasing nocturnal stability category ( Fig. 13 g). This is
consistent with, and builds upon, the findings of Baciu (2005) , who
also reported signi ficant anti-correlations between wind speed and
near-surface radon concentration in the Bucharest region within
the neutral to stable Pasquill eGifford stability regimes. Of partic-
ular interest under stable conditions are the wind speed minima
observed at 1900 h and 0700 h. At a flat inland site, the near-surface
wind speed under stable conditions is expected to be <1.5 m s/C01
(Sesana et al., 2003 ), and on average as low as 0.5 m s/C01(Chambers
et al., 2015b ). It is therefore most likely that the wind speed minima
at 1700 h in Fig. 13 g corresponds to the time at which the height of
the nocturnal inversion drops below the 30 m observation height.
Between 2000 and 0600 h, the observed 30 m wind speeds are
more representative of the lower residual layer, with evidence ofnocturnal jets being more common around 0400 e0500 h. At
around 0700 e0800 h the slower-moving near-surface air is mixed
up to the 30 m level as the nocturnal inversion is burned off. Further
evidence supporting this hypothesis is provided in Fig. 14 b, which
indicates the drop in median hourly wind speed at the 60 m level
occurring on average an hour earlier in the evening, and an hour
later in the morning. Furthermore, throughout the night when both
the 30 and 60 m wind observations are above the nocturnal
inversion (disconnected from the surface), wind speeds are more
similar than is the case under near-neutral conditions ( Fig. 14 a),
when both observation heights are within a well-mixed nocturnal
boundary layer that is well-connected to the surface roughness
elements.
In winter, however, the distinction between weakly and
moderately mixed categories in the 30 m wind speed data is much
less consistent and the broad minima in wind speed under stable
conditions are less pronounced than in summer. It is likely that the
atmospheric state of the near-surface air on winter nights at the
INIF-HH tower is most signi ficantly controlled by large near-surface
temperature gradients occurring well below the lowest tower
measurement height (see Section 3.3and Fig. 16 ). Since the radon
observations are performed much nearer the surface (10 m a.g.l.),
the accumulation of this tracer would better re flect the in fluence of
this thin layer of thermally strati fied air. Fig. 15 a shows the qlapse
rate difference between the 30 m observation level and tempera-
ture measured by the AlphaGUARD radon detector at 10 m within a15 17 19 21 23 1 3 5 7 9 11 1301234567Winter SummerPasquill-Gifford (radiation) iPG near neutral
iPG weakly stable
iPG moderately stable
iPG stable(a)
15 17 19 21 23 1 3 5 7 9 11 1301234567
(b)
iPG near neutral
iPG weakly stable
iPG moderately stable
iPG stable
15 17 19 21 23 1 3 5 7 9 11 13-1.5-1.0-0.50.00.51.01.5
(c)Bulk Richardson Number
Hour of composite day RiB near neutral
RiB weakly stable
RiB moderately stable
RiB stable
1 5 1 7 1 9 2 1 2 3 13579 1 1 1 3-1.5-1.0-0.50.00.51.01.5
RiC(d)
near neutral
weakly stable
moderately stable
stable
Fig. 10. Composite median hourly Pasquill eGifford and bulk Richardson number stability classi fications as a function of whole-day radon-based stability classi fication.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 75

weather-proof enclosure. The drop in 10 e30 m qlapse rate 0200 h
(Fig. 15 a), and flattening of the radon accumulation curve ( Fig. 5 b),
indicates that the winter nocturnal inversion drops even below the
10 m level under the most stable atmospheric conditions (see also
Fig. 16 ).
3.3. Characterisation of mixing depth as a function of nocturnal
stability category
On average, the ceilometer-reported mixing depths were char-
acterised by minimum values mid-morning (~400 m a.g.l.) and
maximum values in the late afternoon (~1400 m a.g.l.) ( Fig. 16 ).
While more recent models of ceilometer have been successfullyused to monitor nocturnal mixing depths (e.g. Pal et al., 2012; Pal,
2014 ), due to the 300 e400 m full overlap range of the ceilometer
used in this study mixing depths below 400 m could not be reliably
estimated. Throughout the night, reported mixing depths were
often quite variable. At times when there was a fresh aerosol source
after sunset, and the nocturnal mixing depth was greater than
~150 m, the ceilometer appeared to report representative nocturnal
boundary layer depths. In other cases, however, the reported
mixing depths were variable, usually corresponding to the top of
the residual layer from the previous day's CBL or decoupled aerosol
layers within the residual layer. Average hourly mixing depthsreported by the ceilometer stabilized again mid-morning when the
nocturnal inversion had begun to erode and mixing heights had
grown past 300 e500 m.
In addition to the ceilometer observations, we used the 10 m
radon concentrations and a the box model shown in Eq. (3)(e.g.,
Griffiths et al., 2013; Sesana et al., 2003 ), to calculate the “effective
mixing height ”(h
e); essentially a near-surface nocturnal mixing
length scale that can be estimated for varying degrees of atmo-spheric stability (except near-neutral conditions). The combination
of radon and active remote sensing techniques to investigate
diurnal mixing patterns has been proposed or applied in a number
of previous studies ( Fontan et al., 1979; Guedalia et al., 1980;
Griffiths et al., 2013; Pal et al., 2015 ).
In brief, assuming a constant radon flux within the nocturnal
measurement footprint, neglecting advective effects, and ac-
counting for entrainment/detrainment of air between the stable
nocturnal boundary layer and the overlying free atmosphere, a
simple budget equation can be proposed for the change in radon
concentration (C) within the lowest layer of the atmosphere:
dC
dt¼F
he/C0lC/C0D (3)
where l¼2.098 /C210/C06s/C01is the radon decay constant, F is theFig. 11. Variation in the diurnal component of the observed nocturnal (2200 e0300 h) radon concentration as a function of the virtual potential temperature lapse rate and wind
speed gradient between 30 and 60 m in summer and winter.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 76

radon source term and D is the dilution term (non-zero when the
boundary layer is growing). Eq. (3)can be solved iteratively to
obtain a time-series of he, the depth that the nocturnal boundary
layer would be if it was uniformly mixed ( Griffiths et al., 2013 ).
This model is only valid for the portion of the diurnal cycle
where the in fluence of local radon sources on observed concen-
trations dominates that of remote (larger-scale) sources; effectively,
from around sunset until several hours after sunrise. In this context
“local ”is understood to represent a region with radius of order
10 s km.
Before combining the radon-derived mixing depths with the
ceilometer values the estimated heit was necessary to “calibrate ”
Eq.(3)to the local radon source function. Results from the ERA-
Interim land model reported by Karstens et al. (2015) indicated a
seasonality of the radon source function in the vicinity of Bucharest
of 6e18 mBq m/C02s/C01. These values overlapped with, but were
lower than, the seasonal (summer/winter) variability of
14e34 mBq m/C02s/C01derived from the GLDAS-Noah LSM reported in
the same study. Both the ERA-Interim and GLDAS-Noah results
indicated that the radon source function in the vicinity of Bucharest
was considerably lower than that in the mountainous regions to the
northwest (Cluj: 22 e40 mBq m/C02s/C01,Cosma et al., 2005; Simion
et al., 2013 ).
Based on the information above we initially employed a radon
flux of 14 mBq m/C02s/C01. Diurnal cycles of mixing depth based on
combined ceilometer and radon-based mixing depth for weaklystable to stable conditions in summer are shown in Fig. 16 b,d,f.
Immediately apparent is the large discrepancy between mixing
heights reported by the two methods in the morning overlap
period (0700 e0900 h). Based on the 10 m radon observations, Eq.
(3)appears to signi ficantly underestimate mixing heights when the
NBL is growing into the residual layer to form the new convective
boundary layer. This is likely attributable to the slow response time
of the radon detector. The morning transition period is a time of
rapidly decreasing radon concentration and, as described by
Griffiths et al. (2015) , radon detectors that base their measurement
on a combination of218Po and214Po alpha-particle decay can
require a substantial response-time correction. As a result, the
AlphaGUARD would be reporting radon concentrations throughout
the morning transition considerably higher than the actual values.
Throughout the night, however, when radon concentrations are
varying much more slowly, the reported radon concentrations and
corresponding mixing depths are expected to be more represen-
tative of real-time values.
On average, nocturnal mixing depths in summer under weakly
stable conditions were 40 e60 m but reduced to around 10 e20 m
under stable conditions. To better justify our choice of radon source
function for summer we examined the diurnal cycle of 30 m wind
speed ( Fig. 14 b, red circles). The wind speed minima observed, on
average, at 1900 h and 0800 h, most likely represent the times at
which the nocturnal inversion dropped below, and rose above,
respectively, the 30 m measurement height. To con firm this, we16 18 20 22 0 2 4 6 8 10 1201020304050
WinterRadon (Bq m-3)
Summer(a) Radon "stab le" category
16 18 20 22 0 2 4 6 8 10 1201020304050
(b) PG-R "stable" category
10th percentile
50th percentile
90th percentile
16 18 20 22 0 2 4 6 8 10 1201020304050
(c) Radon "stable" category
Hour of composite day16 18 20 22 0 2 4 6 8 10 1201020304050
(d) PG-R "stable" category
Fig. 12. Comparison of potential “extreme pollution/exposure events ”(distributions of hourly radon concentrations under “stable ”atmospheric conditions) in summer and winter
as predicted by the (a,c) radon-based scheme, and (b,d) PG-R scheme.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 77

also checked the diurnal cycle of 60 m wind speed on stable nights
in summer ( Fig. 14 b, black squares), and compared these obser-
vations to neutral (well-mixed) conditions ( Fig. 14 a). Based onFig. 16 f, the radon-derived nocturnal mixing depth at 1900 h was
also 30 m, indicating that our choice of radon source function was
appropriate.Fig. 13. Diurnal hourly-mean composites of meteorological parameters from the IFIN-HH tower as a function of radon-derived nocturnal stability classi fication in summer (a,c,e,g)
and winter (b,d,f,h).S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 78

Compared to the 0300 h radon-derived mixing depth in Fig. 16 f
(10 m), the predicted values at 0400 e0500 h were higher even
though the sun had yet to rise. However, looking at Fig. 5 a, the 10 m
radon concentration ceases to rise between 0500 h and 0700 h on
stable nights. The most likely explanation of these observations is
that on stable nights in summer the nocturnal mixing depth
dropped below 10 m between 0400 and 0500 h.
Fig. 5 b shows that under stable conditions in winter, on average,
the radon concentration at 10 m plateaus between 0200 and0800 h. Furthermore, Fig. 16 e shows that radon-derived mixing
depths between 0200 h and 0700 h are considerably higher, and
more variable, than earlier in the night. As for the summer case, this
likely indicates that the nocturnal mixing depth under stable con-
ditions in winter also drops below 10 m a.g.l. Assuming a winter
radon source function of 12 mBq m
/C02s/C01, the decreasing trend of
radon-derived mixing depth between 1700 h and 0000 h (blue
dotted line in Fig. 16 e) crosses 10 m at 0200 h, consistent with the
observations in Fig. 5 b. For weakly stable nights in winter the
average nocturnal mixing depth was typically 30 e50 m ( Fig. 16 a).
Afternoon mixing depths in summer ( Fig. 16 a,c,e) were highest
on clear-sky days ( “stable ”nocturnal classi fication), reachingaverage depths of almost 1800 m agl. During other conditions ABL
depths were typically 1200 e1400 m. It is interesting to note that, in
winter, daytime mixing depths for the most stable classi fied days
were generally lower than for other conditions. In these cases, the
clear-sky conditions usually occurred as the result of synoptic-scale
subsidence under anti-cyclonic conditions ( Fig. 15 b), resulting in
relatively low synoptic inversion heights.
In summary, the average seasonal (summer/winter) variability
in the local Bucharest radon source function is likely to be from 12to 14 mBq m
/C02s/C01from winter to summer. This is consistent with
our assessment of the nocturnal build-up of radon ( Fig. 5 ) that, on
average, there is little seasonal variability in the radon source
function at this site. The results in Fig. 16 indicate that for moder-
ately stable to stable nocturnal conditions the IFIN-HH release
height would usually be above the height of the nocturnal inver-
sion, resulting in relatively small concentrations of tritiated water
within the nocturnal boundary layer. It should be noted, however,
that the standard deviation of mixing height estimates under stable
conditions in winter was a factor of 2 e3 times higher than in
summer. The most consistently enhanced nocturnal concentrations
of tritiated water vapour near the surface in the vicinity of IFIN-HHFig. 14. Comparison of diurnal composite hourly median 30 and 60 m wind speeds under (a) near-neutral and (b) stable atmospheric conditions in summer.
Fig. 15. Composite hourly mean (a) winter potential temperature difference between 30 m and the AlphaGUARD located at 10 m a.g.l., and (b) atmospheric pressur e at 10 m a.g.l.S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 79

are likely to occur under weakly stable conditions, when the
nocturnal inversion is often above the release height, and meanwinds are not suf ficient to well ventilate the area.
4. Conclusions
We analysed 8 months of continuous hourly meteorological and
atmospheric radon observations from a 60 m tower on thepremises of the “Horia Hulubei ”National Institute for Research and
Development in Physics and Nuclear Engineering (IFIN-HH) inBucharest, Romania. Heterogeneous roughness elements up to
15 m tall within the 2 km diameter exclusion zone of this facility
have hitherto hampered efforts to characterise the atmospheric
mixing state by conventional meteorological techniques. We have
developed and applied a continuous (i.e. not categorical) radon-
based stability classi fication scheme and demonstrated its
Fig. 16. Diurnal variability in atmospheric mixing depth at the IFIN-HH site as derived by a combination of ceilometer and radon-derived mixing heights for we akly stable to stable
conditions in summer and winter.
Table 2Comparison of mean monthly temperature (C), relative humidity (%), wind speed (m s
/C01), wind direction (all at 30 m a.g.l.) and precipitation for the IFIN-HH site between 2012
and the 3-year (2010 e2012) average.
Month Temperature Rel. Humidity Wind speed Wind direction Precipitation
3-yr Mean 2012 3-yr Mean 2012 3-yr Mean 2012 3-yr Mean 2012 3-yr Mean 2012
Jan /C01.7 /C01.7 81.9 74.8 2.5 3.4 314 278 41.6 48.4
Feb /C02.6 /C06.2 74.5 73.0 2.9 2.2 48 45 36.1 33.6
Mar 6.0 6.9 60.6 55.5 3.0 2.7 262 298 15.8 11
Apr 12.8 14.7 55.2 56.6 3.0 3.0 98 220 30.5 44May 17.4 18.3 69.3 66.8 2.6 2.4 29 20 105 146Jun 22.1 23.9 61.8 60.3 2.2 2.0 214 328 49.6 36Jul 24.8 27.4 55.4 46.6 2.5 2.7 359 53 27.7 22Aug 25.0 25.4 48.2 45.2 2.3 2.2 41 35 24.4 25.4Sep 20.5 20.7 51.9 53.7 2.6 2.5 53 62 29 40Oct 11.9 14.8 59.1 65.3 2.6 2.5 46 43 28.6 25.2
Nov 7.4 7.4 75.2 79.8 2.3 2.1 188 73 6.6 3
Dec 0.4 /C01.1 81.4 79.3 2.3 2.5 228 203 39 74S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 80

superior performance compared to more commonly-used meteo-
rological techniques (Pasquill eGifford “radiation ”stability typing
and the bulk Richardson number technique). Despite the fact that
the IFIN-HH reactor is no longer operational, seasonal characteri-
sation of the atmospheric mixing state will be useful throughout
the decommissioning process, and “lessons learned ”will be readily
transferrable to numerous other sites including the nearby, oper-
ational Cernavoda Nuclear Power Plant.
Under stable nocturnal conditions the Pasquill eGifford “radia-
tion ”scheme was found to overestimate the atmosphere's capacity
to dilute pollutants with near-surface sources by 20% compared to
the radon-based scheme. Under these conditions, near-surface
wind speeds within the “rough ”exclusion zone drop well below
1ms/C01and nocturnal mixing depths vary from ~25 m to less than
10 m agl. The most consistently enhanced nocturnal concentrations
of tritiated water vapour in the vicinity of IFIN-HH are likely to
occur under weakly stable conditions, when the nocturnal inver-
sion is often above the release height, and mean winds are not
sufficient to well ventilate the area. For moderately stable to stable
conditions the release height would typically be above the
nocturnal inversion.
By applying a combination of ceilometer observations and
radon-derived effective mixing heights, we were able to charac-
terise the variability of mixing depths across the whole diurnal
cycle both for the “summer ”/”winter ”quasi-seasonal extremes
(based on the four warmest and four coldest months of the avail-
able 8-month dataset), and as a function of atmospheric stability.
Average daytime mixing depths at this flat inland site were
observed to vary from 1200 to 1800 m agl., in “summer ”,t o
500e900 m agl., in “winter ”.
By constraining radon box model estimates of nocturnal effec-
tive mixing depth with tower observations we were able to well-
constrain the seasonal variability in the Bucharest regional radon
flux: 12 mBq m/C02s/C01in winter to 14 mBq m/C02s/C01in summer;
values which were in close agreement with recently published
European radon flux maps.
Acknowledgements
This study is part of the project EXPLORATORY IDEAS 65/2011
financed by the Romanian Authority for Scienti fic Research (Grant
Number: PN-II-ID-PCE-2011-3-0396 / IDEI65/2011 and name of thegrant applied for “Interdisciplinary approach for dynamic model-
ling of tritium transfer in crops ”).
References
Avino, P., Brocco, D., Lepore, L., Pareti, S., 2003. Interpretation of atmospheric
pollution phenomena in relationship with the vertical atmospheric remixing by
means of natural radioactivity measurements (radon) of particulate matter.Ann. Chim. 93 (5 e6), 589 e594.
Baciu, A.C., 2005. Radon and thoron progeny concentration variability in relation to
meteorological conditions at Bucharest (Romania). J. Environ. Radioact. 83,
171e189.
Banta, R.M., Mahrt, L., Vickers, D., Sun, J., Balsley, B.B., Pichugina, Y.L., Williams, E.L.,
2007. The very stable boundary layer on nights with weak low-level jets.
J. Atmos. Sci. 64, 3068 e3090 .
Chambers, S., Williams, A.G., Zahorowski, W., Grif fiths, A., Crawford, J., 2011.
Separating remote fetch and local mixing in fluences on vertical radon mea-
surements in the lower atmosphere. Tellus B 63, 843 e859. http://dx.doi.org/
10.1111/j.1600-0889.2011.00565.x .
Chambers, S.D., Wang, F., Williams, A.G., Xiaodong, D., Zhang, H., Lonati, G.,
Crawford, J., Grif fiths, A.D., Ianniello, A., Allegrini, I., 2015a. Quantifying the
influences of atmospheric stability on air pollution in Lanzhou, China, using a
radon-based stability monitor. Atmos. Environ. 107, 233 e243.
Chambers, S.D., Williams, A.G., Crawford, J., Grif fiths, A.D., 2015b. On the use of
radon for quantifying the effects of atmospheric stability on urban emissions.
Atmos. Chem. Phys. 15, 1175 e1190 .
Cosma, C., Cozar, O., Jurcut, T., Baciu, C., Pop, I., 2005. Simultaneous measurement of
radon and thoron exhalation rate from soil and building materials. Radioact.Environ. 7, 699 e705. ISBN: 978-0-08-044137-5 .
Crawford, J., Chambers, S.D., Cohen, D.D., Williams, A.G., Grif fiths, A.D., Stelcer, E.,
2016. Assessing the impact of atmospheric stability on primary and secondary
aerosols at Richmond, Australia, using Radon-222. Atmos. Environ. 127,
107e117.
Ehrhardt, J., 1997. The RODOS system: decision support for off-site emergency
management in Europe. Radiat. Prot. Dosim. 73 (1 e4), 35e40.
Florea, N., Duliu, O.G., 2012. Eighteen years of continuous observation of radon and
thoron progenies atmospheric activity. J. Environ. Radioact. 104, 14 e23.
Foken, T., 2008. Micrometeorology. Springer-Verlag, Berlin Heidelberg, ISBN 978-3-
540-74665-2, p. 306 .
Fontan, J., Guedalia, D., Druilhet, A., Lopez, A., 1979. Une methode de mesure de la
stabilite verticale de l ’atmosphere pres du sol. Bound. Layer Meteorol. 17 (1),
3e14.http://dx.doi.org/10.1007/BF00121933 .
Galeriu, D., Melintesu, A., Turcanu, C., Raskob, W., 2000. FDMH -The Tritium Module
in RODOS – Actual Status and Perspectives Workshop on Tritium Behavior in the
Environment, Kurri, Kumatori, Osaka, Japan, 8-9 May, 2000 .
Galeriu, D., Melintescu, A., 2011. Research and development of environmental
tritium modelling. Fusion Sci. Technol. 60 (4), 1232 e1237 .
Galeriu, D., Melintescu, A., 2015. Progresses in tritium accident modelling in the
frame of IAEA EMRAS II. Fusion Sci. Technol. 67 (2), 343
e348.
Galeriu, D., Melintescu, A., Duma, M., Zorila, B., Gheorghiu, A., 2014. Nuclear
meteorology at IFIN-HH. Romanian J. Phys. 59 (9 e10), 999 e1011 .
Galeriu, D., Melintescu, A., Stochioiu, A., Nicolae, D., Balin, I., 2011. Radon, as a tracer
for mixing height dynamics – an overview and RADO perspectives. Romanian
Rep. Phys. 63 (1), 115 e127.
Galmarini, S., 2006. One year of222Rn concentration in the atmospheric surface
layer. Atmos. Chem. Phys. 6, 2865 e2886. http://dx.doi.org/10.5194/acp-6-2865-
2006 .
Griffiths, A.D., Chambers, S.D., Williams, A.G., Werczynski, S., 2015. Increasing the
accuracy and temporal resolution of two- filter Radon-222 measurements by
correcting for the instrument response. In: Atmospheric Composition &
Chemistry Observations &Modelling Conference Incorporating the Cape Grim
Annual Science Meeting, Murramarang NSW Australia, 11 e13th November,
2015 .
Griffiths, A.D., Parkes, S.D., Chambers, S.D., McCabe, M.F., Williams, A.G., 2013.
Improved mixing height monitoring through a combination of lidar and radon
measurements. Atmos. Meas. Tech. 6, 207 e218. http://dx.doi.org/10.5194/amt-
6-207-2013 .
Grundstrom, M., 2015. The In fluence of Atmospheric Circulation and Meteorology
on Urban Air Pollution and Pollen Exposure (Doctoral Thesis). University of
Gothenburgh, Faculty of Science, Department of Biological and EnvironmentalSciences, ISBN 978-91-85529-79-7. http://hdl.handle.net/2077/38728 .
Guedalia, D., Ntsila, A., Druilhet, A., Fontan, J., 1980. Monitoring of the atmospheric
stability above an urban and suburban site using sodar and radon measure-
ments. J. Appl. Meteorol. 19 (7), 839 e848. http://dx.doi.org/10.1175/1520-
0450(1980)019 <0839:MOTASA >2.0.CO;2 .
International Atomic Energy Agency, 2012. Sources and Measurements of Radon
and Radon Progeny Applied to Climate and Air Quality Studies, ISBN 978-92-0-
123610-4. Proceedings Series, STI/PUB/1541 .
International Atomic Energy Agency, 2014. Transfer of tritium in the environment
after accidental releases from nuclear facilities. Report of Working Group 7
“Tritium Accidents ”of EMRAS II Topical Heading, Approaches for Assessing
Emergency Situations Environmental Modelling for RAdiation Safety (EMRAS II)Program, IAEA_TECDOC-1738. International Atomic Energy Agency, ISBN 978-
92-0-102814-3 .
International Atomic Energy Agency, 2011a. Safety Aspects in Siting for Nuclear
Installations. Draft Safety guide, Revision No. SG-S9, DS433. last accessed 5 Nov
2015. http://regelwerk.grs.de/downloads/ds433_rev0016-2013-03-18draft_
clean.pdf .
International Atomic Energy Agency, 2011b. Meteorological and Hydrological
Hazards in Site Evaluation for Nuclear Installations. IAEA Safety Standards No.
SSG-18. last accessed 5 Nov 2015. http://www-pub.iaea.org/MTCD/
publications/PDF/Pub1506_web.pdf .
Karstens, U., Schwingshackl, C., Schmithüsen, D., Levin, I., 2015. A process-based
222Rnflux map for Europe and its comparison to long-term observations.
Atmos. Chem. Phys. Discuss. 15, 17397 e17448. http://dx.doi.org/10.5194/acpd-
15-17397-2015 .
Melintescu, A., Galeriu, D., Diabat /C19e, S., Strack, S., 2015. Preparatory steps for a robust
dynamic model for organically bound tritium dynamics in agricultural crops.
Fusion Sci. Technol. 67 (3), 479 e482.
Melintescu, A., Galeriu, D., 2015. Uncertainty of OBT Models. 4th Organically Bound
Tritium (OBT) Workshop, August 31 eSeptember 2, 2015, Bucharest, Romaia .
Oke, T.R., 1987. Boundary Layer Climates, second ed. Methuen &Co. Ltd., University
Press, Cambridge, ISBN 0-415-04319-0 .
Pal, S., Xueref-Remy, I., Ammoura, L., Chazette, P., Gibert, F., Royera, P., Dieudonn /C19e, E.,
Dupont, J.-C., Haeffelin, M., Lac, C., Lopez, M., Morille, Y., Ravetta, F., 2012.
Spatio-temporal variability of the atmospheric boundary layer depth over the
Paris agglomeration: an assessment of the impact of the urban heat island in-
tensity. Atmos. Environ. 63, 261 e275.
Pal, S., 2014. Monitoring depth of shallow atmospheric boundary layer to comple-
ment LiDAR measurements affected by partial overlap. Remote Sens. 6,
8468e8493. http://dx.doi.org/10.3390/rs6098468 .
Pal, S., Lopez, M., Schmidt, M., Ramonet, M., Gibert, F., Xueref-Remy, I., Ciais, P., 2015.
Investigation of the atmospheric boundary layer depth variability and itsS.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 81

impact on the222Rn concentration at a rural site in France. J. Geophys. Res.
Atmos. 120, 623 e643. http://dx.doi.org/10.1002/2014JD022322 .
Pasquil, D., 1961. The estimation of the dispersion of windborne material. Meteorol.
Mag. 90, 33 e49.
Pasquill, F., Smith, F.B., 1983. Atmospheric Diffusion, third ed. Ellis Horwood Ltd.,
John Wiley &Sons, Chichester, p. 437 .
Pearce, J.L., Beringer, J., Nicholls, N., Hyndman, R.J., Tapper, N.J., 2011. Quantifying
the in fluence of local meteorology on air quality using generalized additive
models. Atmos. Environ. 45, 1328 e1336 .
Perrino, C., Pietrodangelo, A., Febo, A., 2001. An atmospheric stability index based
on radon progeny measurements for the evaluation of primary urban pollution.
Atmos. Environ. 35, 5235 e5244 .
Perrino, C., 2012. Natural Radioactivity from Radon Progeny as a Tool for the
Interpretation of Atmospheric Pollution Events, Sources and Measurements of
Radon and Radon Progeny Applied to Climate and Air Quality Studies. Inter-
national Atomic Energy Agency, Vienna, Austria, pp. 151 e159. ISBN 92-0-
123610-4 .
Sesana, L., Caprioli, E., Marcazzan, G.M., 2003. Long period study of outdoor radonconcentration in Milan and correlation between its temporal variations and
dispersion properties of atmosphere. J. Environ. Radioact. 65, 147 e160. http://
dx.doi.org/10.1016/S0265-931X(02)00093-0 .
Simion, E., Mihalcea, I., Cuculeanu, V., Simion, F., 2013. Dynamics of outdoor radon
and thoron progeny concentrations in some geographical areas of Romania.
Rom. Journ. Phys. 58 (Supplement), S251 eS261. Bucharest .
Stull, R.B., 1988. An Introduction to Boundary Layer Meteorology. Kluwer Academic
Publishers, The Netherlands, ISBN 90-277-2769-4 .
Turner, B., 1964. A diffusion model for an urban area. J. Appl. Meteorol. 3, 83 e91.
Williams, A.G., Chambers, S.D., Conen, F., Reimann, S., Hill, M., Grif fiths, A.D.,
Crawford, J., 2015. Radon as a Tracer of Atmospheric In fluences on Traf fic-
related Air Pollution in a Small Inland City. Tellus B, to be submitted January
2016 .
Williams, A.G., Chambers, S.D., Grif fiths, A., 2013. Bulk mixing and decoupling of the
nocturnal stable boundary layer characterized using a ubiquitous natural tracer.
Bound. Lay. Meteorol. 149, 381 e402. http://dx.doi.org/10.1007/s10546-013-
9849-3 .S.D. Chambers et al. / Journal of Environmental Radioactivity 154 (2016) 68 e82 82