Journal of Environmental Radioactivity 102 (2011) 551e558
lable at ScienceDirectContents lists avaiJournal of Environmental Radioactivity
journal homepage: www.elsevier .com/locate/ jenvradBiotic, temporal and spatial variability of tritium concentrations in transpirate
samples collected in the vicinity of a near-surface low-level nuclear waste
disposal site and nearby research reactor
J.R. Twining*, C.E. Hughes, J.J. Harrison, S. Hankin, J. Crawford, M. Johansen, L. Dyer
Institute for Environmental Research, ANSTO, Locked Bag 2001, Kirrawee DC, NSW 2232, Australiaa r t i c l e i n f o
Article history:
Received 10 November 2010
Received in revised form
18 February 2011
Accepted 18 February 2011
Available online 12 March 2011
Keywords:
Tritium
Transpiration
Correlation modelling
Groundwater
Leachate* Corresponding author. Tel.: þ612 97173060; fax:
E-mail address: jrt@ansto.gov.au (J.R. Twining).
0265-931X/$ e see front matter Crown Copyright 

doi:10.1016/j.jenvrad.2011.02.012a b s t r a c t
The results of a 21 month sampling program measuring tritium in tree transpirate with respect to local
sources are reported. The aim was to assess the potential of tree transpirate to indicate the presence of
sub-surface seepage plumes.
Transpirate gathered from trees near low-level nuclear waste disposal trenches contained activity
concentrations of 3H that were significantly higher (up to w700 Bq L
1) than local background levels
(0e10 Bq L
1). The effects of the waste source declined rapidly with distance to be at background levels
within 10s of metres. A research reactor 1.6 km south of the site contributed significant (p < 0.01) local
fallout 3H but its influence did not reach as far as the disposal trenches.
The elevated 3H levels in transpirate were, however, substantially lower than groundwater concen-
trations measured across the site (ranging from 0 to 91% with a median of 2%). Temporal patterns of tree
transpirate 3H, together with local meteorological observations, indicate that soil water within the active
root zones comprised a mixture of seepage and rainfall infiltration. The degree of mixing was variable
given that the soil water activity concentrations were heterogeneous at a scale equivalent to the effective
rooting volume of the trees. In addition, water taken up by roots was not well mixed within the trees.
Based on correlation modelling, net rainfall less evaporation (a surrogate for infiltration) over a period of
from 2 to 3 weeks prior to sampling seems to be the optimum predictor of transpirate 3H variability for
any sampled tree at this site.
The results demonstrate successful use of 3H in transpirate from trees to indicate the presence and
general extent of sub-surface contamination at a low-level nuclear waste site.
Crown Copyright 
 2011 Published by Elsevier Ltd. All rights reserved.1. Introduction and aim
The Little Forest Burial Ground (LFBG) is a legacy, near-surface,
waste disposal site located at Lucas Heights near the old HIgh Flux
Australian Reactor (HIFAR, Fig. 1) in southern Sydney. Both the LFBG
(via groundwater) and HIFAR (via atmospheric dispersal) are
potential sources of tritium (3H) to the nearby environment.
Disposal operations at the LFBG took place from 1960 to 1968
(Isaacs and Mears, 1977) and 3H comprised part of the waste. Since
then, the site has been regularly maintained and routine moni-
toring of groundwater, grass and high volume air sampling, as
reported in annual ANSTO Environmental Monitoring reports (e.g.
Hoffmann et al., 2008), has continued to the present day. Recentlyþ612 97173599.
2011 Published by Elsevier Ltd. Allinitiated studies at the site have been more intensive than the
routine work and have been directed towards better under-
standing, and modelling of, the behaviour of radioactive materials
under Australian conditions as well as to assess management
options for the site. Aspects of the improved modelling will hope-
fully be transferrable to other Australian sites, such as the proposed
national nuclear waste storage facility under consideration.
At any waste site, one of the key issues is the potential hazard to
humans or other species that may arise due to the movement of
radionuclides through the environment to potential receptors. It is
therefore necessary to understand the environmental pathways
which may lead to such exposures, which usually include the
possibility of movement of radionuclides in groundwaters through
the vicinity of the site. It was believed that trees at the LFBG had the
potential to intersect any sub-surface dispersion plumes from the
buried waste, thereby providing a more expansive spatial coverage
than the existing set of bores at the site and allowing sampling overrights reserved.
552 J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558
Fig. 1. The Little Forest Burial Ground and nearby HIFAR research reactor. A, Location
map and sampling sites for long, NS transect; B, Aerial view of LFBG showing locations
of sampled trees comprising standard set and nearby transects.the waste without otherwise disturbing it. Tritiated water, being
a conservative tracer of such a plume, would be taken up by plants
and transpired. The soilerooteleaf pathway was identified as the
major transport mechanism for 3H from contaminated ground-
water by Evenden et al. (1998). Thus, this radioecological study was
instigated in 2007 to assess that potential.
Since that time, there has also been an increasing emphasis
placed on the need to undertake environmental dose assessment of
nuclear activities (e.g. ICRP 108, 2008). This paper will report on the
findings of a 21 month survey of transpirate 3H over, and adjacent
to, the trenches and with respect to the other local 3H source,
HIFAR. Whilst 3H itself is not of undue concern from a radiological
perspective, particularly at the levels measured in this investiga-
tion, these results will be used to identify the trees which have
substantial access to any waste seepage plume. Those trees will be
targeted for additional tissue analyses for other radionuclides to
assess their relative transport rates in the plume as well as their
bioaccumulation rates for dose assessment modelling purposes.2. Materials and methods
2.1. Site description
At the LFBG, a variety of low-level nuclear waste materials were
disposed of in a series of 77 unlined trenchesw3m deep, 25m long
and spaced 2.7 m apart. As each trench was filled, it was re-covered
with topsoil to a depth of w1.0 m. There was a total of 1675 m3 of
waste, comprising w150 GBq of activity and including 1730 kg of
beryllium (Isaacs and Mears, 1977). The locations of trenches are
still evident from minor surface subsidence at the present day
(Fig. 1b).
The local geology, described by Hughes et al. (2010), consists of
topsoil and clay to a depth of 1.5e2 m, which is derived from
weathered shale to a depth of approximately 6e8 m. The shale
overlies the Hawkesbury Sandstone which extends to a depth of
around 200 m. Surrounding the trenches, a perched saturated zone
exists near the top of the weathered shale at w3 m. This saturated
zone is hypothesised to be fed by leachate from the LFBG trenches.
Below this level other perched saturated zones have been observed,
with the main aquifer occurring in the Hawkesbury Sandstone.
2.2. Sampling
A variety of approaches were assessed to determine the best
methods for sampling and analysis of 3H in transpirate. More
detailed descriptions of those techniques are available in Twining
et al. (2009). The following is a synopsis of the optimal method-
ology applied over the period.
2.2.1. Biota selection
Seven accessible trees and shrubs over, or adjacent to, the
trenches were identified as being most likely to intersect any sub-
surface seepage plumes. These plants, comprising a gum tree
(Eucalyptus paniculata; Eu 1), two similar Acacia sp (Ac 1, Acacia
obtusifolia and Ac 2, Acacia longifolia longifolia), two turpentine
trees (Syncarpia glomulifera; Sy 1 and 2), a Leptospermum poly-
galifolium shrub (Le 1) and a Banksia serrata tree (Ba 1) are identi-
fied in Fig. 1b and are hereafter referred to as the standard set.
Typical rooting depths for these species are not well established,
particularly for this site; however the following values are esti-
mates provided from the scientific literature. In schlerophyllous
scrub and forest in south eastern Australia the range of maximum
rooting depths for Banksia spp and a Leptospermum sp. is 2.3e2.5 m
(Canadell et al., 1996, citing Specht and Rayson, 1957). McColl
(1969) reported an average rooting depth of just 1.2 m for
E. paniculata based on observations of soil profiles in a coastal forest
south of Sydney. No information on the rooting depths of the Acacia
spp nor of the Syncarpia sp. in this study is otherwise available,
although Canadell et al. (1996, Fig. 2 therein) provided an estimated
mean maximum rooting depth for trees of 6.9  1.2 m and for
shrubs of 5.0  0.9. Average rooting depths will be somewhat less
than the maxima and it should be recognised that local soil
conditions will also constrain plant growth.
Initial samples of transpirate were collected in Jul 2007. Subse-
quent sets of samples using the same trees and as far as possible the
same branches, were collected approximately every three months
until Mar 2009. In Oct 2007, samples were collected from more
branches on the trees identified as Ba 1, Sy 1 and Ac 2 (Fig. 1b) to
better estimate variability within these trees. In addition, three
transects of approximately 400 m length located adjacent to the
trenches were sampled on or about 13 Mar 2008 (West); 24 Jun
2008 (North) and 30 Sep 2008 (South East) (Fig. 1b). A fourth, much
longer transect (w15 km) beginning at the HIFAR reactor which is
approximately 1.6 km south of the LFBG, then running northward
J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558 553
800 
700 
600 
500 
400 
300 
200 
100 
0 
1 2 3 4 5 
Branch no. 
Fig. 2. Tritium activity concentrations in transpirate collected from 5 branches on each
of 3 trees [Ac 1 (dotted), Ba 1 (dashed) and Sy 1 (solid line) refer to Section 2.2.1 for
sample details] on 4 October 2007. Analytical errors (1 s.d.) are also shown.
3 
T ra n  sp  i ra t e  H (B q  / L ) through the LFBG and beyond (Fig. 1a) was sampled on 30 Sep 2008
and 23 Mar 2009. Plants sampled in these transects comprised
Eucalyptus; Syncarpia and Acacia spp.
2.2.2. Collection of transpirate
The leaves at the distal end of a limb (in the case of trees or large
shrubs) or limbs (in the case of smaller shrubs) were enclosed in
a large, dry, clean, clear PVC bag with the open end being taped
closed around the branch (or branches) using adhesive cloth tape.
The bag was typically placed on a north-facing (sun-exposed) limb,
on sunny days after dew evaporation, for four or more hours.
The condensate in the bag was collected with new, 50 mL
disposable syringes. The samples were filtered in situ using dual
stage, 1.0 and 0.45 mm polypropylene membrane cartridges, into
clean, dry, HDPE plastic bottles. Volumes from 0 to >250 mL
(typical range 50e150 mL) were collected over the entire period
with smaller volumes collected on cooler, more overcast days. The
filtered samples were returned to ANSTO and placed into a cool
room to await further processing.
2.2.3. Sample processing
From the filtered transpirates, 25mL aliquots were processed for
3H analysis using Eichrom tritium columns (Eichrom, 1996) to
remove dissolved organics (including organically-bound 3H) and
ions. After treatment, 5 mL of each sample was added to trans-
lucent, HDPE scintillation vials with 11 mL of Ultima Gold uLLT
liquid scintillant cocktail and the vials were sealed for counting. A
similar procedure was used to prepare standards and blanks to be
included in each counting run.
2.3. Analysis
Samples, with blanks (distilled ‘dead’ water acquired from an
aquifer with no bomb pulse 3H) and standards (prepared using
Amersham TRY64, batch 133, 1 Nov 2002), were counted on
a Packard Tricarb 2900TR LSC using up to 31  20 min cycles over
an energy range from 2 to 18.6 keV. Detection efficiencies were
approximately 25% with a detection limit of w10 Bq L
1. Quench
correction used instrument TSIE values and a correction curve
derived from a set of 10 standard samples quenched with
increasing amounts of food dye. The 3H activity in each sample was
determined using total counts, after background subtraction and
accounting for within-run efficiency (standards) and quenching
(TSIE). Measurement uncertainties quoted in this study include
counting, weighing and standard activity uncertainties.2.4. Additional data
Meteorology observations form part of the routine monitoring
program carried out at ANSTO. A range of variables are collected at
frequencies up to every 15 min from the Lucas Heights meteoro-
logical station (Clark, 2003) approximately 1.6 km south from the
LFBG. Daily averages of temperature, rainfall and pan evaporation
were used to help interpret the variability in transpirate 3H over
time.
Similarly, routine water samples have been regularly collected
and analysed for a range of contaminants, including 3H, from
boreholes across the LFBG and rainfall from the meteorological
station (Hughes et al., 2010). In addition, water samples were
gathered from series of shallow (3e5 m) core holes immediately
after drilling near the trenches in Aug 2009.These results were used
to assist interpretation of the spatial variability observed in the
transpirate 3H data.
3. Results and discussion
The full set of results for 3H in transpirate at LFBG is available from
the records at ANSTO via the first author. The results reported here-
after are subsets or averages relevant to the hypotheses being
assessed ineachcase. All further references to 3H in transpirate, soil or
groundwater at this site refer to non-organically-bound 3H inwater.
3.1. Within-tree variability
Results for 3H in transpirate collected on 4 Oct 2007 from 5
different branches on each of 3 trees (Ac 2; Ba 1 and Sy 1) are shown
in Fig. 2. The measurement uncertainties are generally within the
symbol of each datum. Hence, it can readily be seen that there were
significant (p < 0.01) differences in 3H being measured from
different branches within any tree. The absolute differences
between trees will be discussed below. Variability was greater in Ac
2 than in Ba 1whichwas similar to Sy 1. The trees increased in size in
that same order, so it may be that this patternwas related to rooting
depth. However, rooting depth was not recorded to confirm this.
It is clear that the trees were all exhibiting different 3H levels in
different branches. This observation points to two related factors.
Firstly, the roots were accessing different concentrations of 3H in
the soil (i.e. heterogeneous sources) and secondly, there was
limited mixing occurring within the trees once water was taken up
by the roots.
Regarding small scale variability at field sites, Amano and Garten
(1991) encountered very heterogeneous soil water 3H concentra-
tions at a site adjacent to the Oak River National Laboratory in the
USA. They relied on integration within the plant to assess average
soil water concentrations for their modelling. Combining several
leaf samples to provide a soil water value for their modelling may
have been sufficient to overcome the variability identified in that
study. Modellers within the IAEA BIOMASS Program also agreed
that 3H concentrations in soils at sites with sub-surface contami-
nation can vary sharply over short distances (IAEA, 2003). From
that, good predictions of 3H transport required sophisticated
models and detailed information on meteorological conditions and
soil characteristics unless the models were generalised over larger
areas.
The limited mixing of 3H within tree water streams was not
unexpected and similar disparities had been observed for tritiated
water uptake in another Australian species (Twining, 1994). Xylem
tissues are themain conductive vessels carryingwater from roots to
leaves for photosynthesis. Studies carried out in the rainforests of
the Amazon (Jordan and Kline, 1977) and in coniferous areas of
north-western USA (Kline et al., 1976) conclusively showed that the
554 J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558rateof transpiration ineither environmentwasdirectlyproportional
to the sapwood cross-sectional area, and hence to the amount of
xylem tissue present. These vascular tissues need to be highly
conductive and yet with reduced lateral porosity to allow for water
to be drawn vertically over substantial distances. In support of
reduced horizontal mixing, Mathew et al. (1988) injected tritiated
water into a number of holes bored into the base of a 2.5mpalm tree
and thenobserved the appearance of 3H in the leaves at the crownby
sampling leaflets over time (the water was collected by vacuum
freeze-drying and hence did not include any organically-bound 3H).
They found that the 3H peak arrival varied from w5.5e9 days in
different leaves, almost twice the time fromone branch to the other.
This indicated that some flowpathsweremore tortuous than others
but also that those paths were essentially discrete.
Evidence for some horizontal mixing does exist, however. Kline
et al. (1976) undertook a similar study in which tritiated water was
injected into holes near the base of Douglas fir trees. The vertical
and horizontal movement of 3H after 10 days was assessed by
taking cores into the pith of the tree at 20, 30, 40 and 50 m heights
above the injection point. As expected, these showed the most
rapid vertical movement occurred towards the inner parts of the
cortex (sapwood) where the xylem predominates. Most of the 3H
was at 40 m but a peak was starting to become apparent at 50 m.
However some was also detected in the pith (heartwood). The
pattern in this case suggested that 3Hmoved horizontally out of the
xylem and into the pith during the passage of the peak activity but
then moved back into the xylem after the peak had passed. There
was little or no evidence of vertical movement within the pith.
None of these studies measured organically-bound 3H.
3.2. Temporal variability in transpirate tritium activity
concentrations
Variability in average transpirate 3H over the entire sampling
period for the standard set of trees is shown in Fig. 3. The values for
each time period weremostly significantly different (p< 0.01) from
the previous datum for each branch, although the degree of change
was variable. The marked differences on 4 Oct 2007 were, to some
degree, due to the sampling of multiple branches on some trees
that day, as discussed earlier. Apart from that excursion, most of the
trees retained a similar degree of activity concentration in tran-
spirate over time despite the significant differences between each
sample. For example, Sy 1 always had higher average 3H activity
concentrations than did Le 1, which in turn always had more than
Ba 1 and similarly for Ac 1. This pointed to heterogeneity in the
spatial distribution of 3H in the soil profile, as indicated previously
by the heterogeneity observed between branches and as will be
discussed further below. It also indicated that the soil 3H levels atAc 1 Ac 2 Ba
Fig. 3. Average 3H activity concentration in transpirate from each tree (see Section 2.2.1 fo
indicates that a different branch was selected and sampled after July 2007.
Jul-07
Oct-07
Jan-08
Apr-08any one location remained generally consistent over time. At
a similar site to the LFBG, near the Oak Ridge National Laboratory
Waste Storage Area, Amano and Garten (1991, citing Amano et al.
1987) reported that the change in soil 3H concentrations was fairly
slow, as reflected in our data.
There was also a tendency for the trees to follow a similar trend
from one date to the next, with most trees either increasing or
decreasing their average activity concentration simultaneously. This
suggested that environmental factors were having some influence
across the site, the most obvious being rainfall which will tend to
dilute the 3H activity concentrations in the surface soils as well as in
the trenches (assuming that the trenches are the major source of 3H
in this location). Amano and Garten (1992) showed surface dilution
leading to reduced 3Hvalues in the top10 cmof the soil profile at the
Oak Ridge site over both summer and winter compared to consis-
tently higher activity concentrations for the next 70 cm. At LFBG,
infiltration into the clay soils below the topsoil zone is believed to be
slow. However, evidence suggests that the poorly capped trenches
provide a preferentialflowpathwayallowing rainfall to dilute the 3H
in the trenches. The resulting saturated layer, observed at a depth of
2.5e3.5 m, may therefore have 3H activity concentrations that also
vary in response to rainfall.
The effect of rainfall is not only to dilute the surface soil 3H. Plants
can also respond to changingwater potential by altering their active
root zone, i.e. the depth at which they obtain their water, to access
water at energetically more favourable levels (Guswa, 2008 and
citations within). This is a physiological response to access water
that requires less effort to adsorb and transpire. Over an annual
cycle, Stringer et al. (1989) studiedoaks onahillside inKentucky that
was also affected by lateral groundwater 3flow of H from the nearby
low-levelwaste repositoryatMaxeyFlats. In spring, the surface soils,
from the top of the hill slope to lower down, had a good water
content of about 0.4 m3 m
3 due to snow melt; xylem water
potentials in trees were all low (around 
0.2 MPa); and foliage 3H
was also relatively low (at about 5  103 Bq L
1). By well into
summer, the surface soils had dried by about a factor of 2 and the
trees at the top of the hill slope had increased their water potential
up to about 
1.75 MPa, indicating greater stress. These plants had
also retained their earlier foliage 3H concentrations. The trees at the
bottom of the slope, however, had onlymildly increased their water
potential (to about 
0.35 MPa) but their foliage 3H had increased
dramatically tomore than 1106 Bq L
1. The trees had shifted their
water source to the deeper, but relatively more accessible, seepage
from theMaxey Flats sitewhichwas unavailable to the plants higher
up the hill slope. The plants at LFBGmay be similarly switching their
water source in response to rainfall inputs and thereby temporarily
changing their access to any potential sub-surface seepage plumes
from the trenches.600
 1 Eu 1 Le 1 Sy 1 Sy 2
500
400
300
200
100
0
r tree details) from the standard set on each sampling date. The dotted line for Eu 1
Jul-08
Oct-08
Jan-09
Bq/l
J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558 555To test this hypothesis, daily data of rainfall less pan evaporation
measured at Lucas Heights were used to estimate a local value for
the cumulative rainfall deficit as a surrogate for potential infiltra-
tion. The effects of daily temperature ranges over the period of the
study were also assessed. These were compared with the
percentage shift in transpirate 3H activity concentration from one
sample to the next for each branch over the entire sampling period.
Percentages were used rather than the absolute differences to
normalise samples to a common baseline. It should be noted that
some data were excluded from these analyses, as follows: Two
branches (Le 1.2 and Sy 2.2) had too few points with successive 3H
measures to allow for accurate regression analysis; branch Eu 1.1
had consistently low activity concentrations at less than twice the
background and hence was prone to excessive variability due to
counting uncertainty; and one branch (Ac 2.1) died soon after final
sampling and hence could be considered to have potentially
abnormal water use patterns.
Initial observations showed no direct relationship between the
shift in transpirate 3H and daily temperatures, rainfall events orA 
-5 -4 -3 -2
y = -0.15x + 0.01 
R2   = 0.18 
B 
-5 -4 -3 -2
y = -0.26x - 0.33 
R2   = 0.46 
RaiRnfaailnl fdaell fdiceitfi c–i t1 144 dda ley
Fig. 4. Correlations of the % shift in transpirate tritium activity concentration in individual
evaporation at ANSTO (1.6 km S) for the preceding 14 days. A, larger trees: Eucalyptus panic
and Acacia obtusifolia, Banksia serrata and Leptospermum polygalifolium.
Tritium shift % T ri  ti  u m S h i ft  % cumulative rainfall deficit, however there was some indication that
a leading average rainfall deficit (representing potential infiltra-
tion), estimated over some days or weeks, may be having an effect.
Modellers within the IAEA BIOMASS Program Tritium Working
Group also agreed that the balance between rainfall and evapora-
tion, together with mechanical dispersion processes in soil, were
key factors in the transport of 3H through the unsaturated layer
above a contaminated zone (IAEA, 2003). Fig. 4 a and b show the
trends for shifts in transpirate 3H concentrations for larger and
smaller trees, respectively, using an arbitrary lag in the potential
infiltration of 14 days. The correlations are not signi cant (r2fi ¼ 0.18
and 0.46 respectively), however a consistent trend was observable.
To determine an optimum lag period, the transpirate 3H activity
concentrations were correlated with net rainfall deficit over daily
time steps from 1 to 90 days using a modelling approach described
in Venables and Ripley (2002). Fig. 5 shows the results of the
correlation modelling. Values around zero imply poor or no
correlation between the number of days over which approximate
infiltration was calculated and the measured shift in transpirate 3H2.5 
2 
1.5 
1 
0.5 
0 
-1 0 1 2
-0.5 
-1 
2 
1.5 
1 
0.5 
0 
-1 0 1 2 
-0.5 
-1 
-1.5 
-2
a ledaindgi nagv earvaegrea g(me m(m) m) 
branches between successive samples at LFBG with the cumulative rainfall minus pan
ulata and Syncarpia glomulifera; B, smaller trees and shrubs; Acacia longifolia longifolia
556 J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558
1.2
Ac1.1
0.8 Le1.1
0.4 Ac2.2
Ac2.3
0 Ac2.4
-0.4 Ac2.5
Ba1.1
-0.8 Ba1.2
-1.2 Sy1.1
1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 Sy2.1
Preceeding number of days
Fig. 5. Correlations for the % shift in tritium activity concentrations between
sequential samples for individual branches at LFBG with the rainfall deficit at ANSTO
(1.6 km S) over the number of days prior to sampling.
350.0 
300.0 
250.0 
200.0 
150.0 
Correlation coefficient
q  . L activity concentration. Negative values imply an inverse relation-
ship (i.e. 3H activity concentrations decrease with increasing rain-
fall) and the converse is true for positive values.
It is apparent from Fig. 5 that therewas considerable noise in the
data. A simple correlation using rainfall deficit alone did not fully
explain temporal shifts in 3H activity concentration in transpirate at
LFBG, remembering also that the meteorological station was
approximately 1.6 km distant from the sampled trees. Measuring
additional parameters at the time of sampling, such as soil moisture
content, air pressure, relative humidity etc. may have provided
more modelling power to assist in evaluation. However, some
points of consistency were discerned.
Firstly, most of the correlations were negative. This supports
the hypothesis that rainfall infiltration into surface soil was dilu-
ting the activity concentration of 3H in the active root zones.
Secondly, the correlation curves tend to flatten off after about 40
cumulative days, suggesting that additional information after that
time did not add anything to improve the interpretation.
Thirdly, there was considerable variation in correlation between
branches in the period up to about 10 days. This implied that short-
term variability in rainfall and evaporation had an imprecise effect
on transpirate 3H shift. Experimental studies with sunflower and
tobacco plants showed that equilibrium between the 3H in the
solution to which their roots were exposed and the stems and
petioles of mature leaves was reached within half a day (Raney and
Vaadia, 1965). Citrus seedlings, being slightly larger plants at 1e2
years old, required up to 40 h to achieve equilibrium between a 3H
enriched nutrient solution bathing the roots and their leaves
(Mantell et al., 1979). It follows that the larger trees sampled in this
study will take somewhat longer times for such transfer to take
place. All the literature sited inSection3.1 (above) also indicated that
therewas a delay inmass transport ofwaterwithin the tree in that it
took fromdays toweeks betweenuptake from roots and subsequent
transpiration. In addition, there is the time required for the infil-
tratingwater to reach andmix in the root zonebefore anyuptake can
occur. Hence, it is not unexpected that recent, short-term, rainfall
patterns will not be consistently reflected in the transpirate.
Lastly from Fig. 5, there was a period from approximately two to
three weeks where most of the correlations were more stronglyTable 1
Tritium activity concentrations in tree transpirate collected over or adjacent to the
trenches at LFBG, or in nearby transects.
Sample set Average  S.E. n Range Genera
(Bq L
1) (Bq L
1) sampled
Standard 116  11 100 0e667 See text
North East 2  1 13 0e10 Eucalyptus &
Transect Acacia
West 12  3 11 4e29 Eucalyptus &
Transect Syncarpia
South East 23  7 7 0e52 Eucalyptus &
Transect Acacianegative. This implied that a lag period in this range was optimal
and, hence, the data in Fig. 4 were reasonable estimates for this site.
3.3. Spatial variability in transpirate tritium activity concentrations
3.3.1. Differences between the standard set and nearby transects
Average transpirate 3H activity concentrations for the standard
set of trees and for the nearby transects indicated in Fig. 1b are
presented in Table 1.
It is quite clear that significantly higher average activity
concentrationswere registered over the trenches than in the nearby
transects. The transect to the north and east was at background
levels. There was no evidence of any near-surface seepage plume in
that direction. The transect to the west was also low but slightly
higher on average than the northern transect. The LFBG is bordered
by a disused landfill site in that direction which exhibited leachate
3H levels ranging from 240 to 415 Bq L
1 in late 2008 (Hughes et al.,
2010). It is possible than the minimal transpired 3H levels detected
could have been the result of seepage from the landfill into the LFBG
instead of seepage from the trenches in this case. The transect to the
south east follows a natural drainage line (Fig.1b)where elevated 3H
levels havebeen found in surface runoff (Hughes et al., 2010). Hence,
the slightly higher values again in that case may well have been
associated with surface expression of a near-surface seepage plume
or runoff from the LFBG trenches. However, the values were still
much lower than those detected in the standard set and hence the
seepage, if it existed, was negligible or mixed with lower activity
waters. It should be noted that in none of the transectswas there any
signi 3ficant difference in transpirate H activity concentration
between the different species that were sampled.
At the time of sampling, there was another potential nearby
source of 3H, theHIFAR research reactor approximately 1.6 km to the
south (Fig.1a), that may have contributed some activity by localised
atmospheric transport, particularly to the south east transect. This is
particularly so as it is reported that atmospheric 3H uptake by plants
tends to be themost commonexposure pathway (Boyer et al., 2009).
Whilst 3H in rainfall at the Lucas Heights meteorological station
averaged14Bq L
1 over the period 2004e2008 (Hughes et al., 2010),
during the period of transpirate sampling the rainfall weighted
average 3H was 2.5 Bq L
1 following the final shutdown of HIFAR in
the first quarter of 2007. To assess the atmospheric contribution to
transpirate, a transect fromHIFAR to thenorth through the LFBGwas
sampled twice. The results are shown in Fig. 6.
The local effect of HIFAR was clearly evident with transpirate
levels well in excess of those observed in rainfall; however its100.0 
50.0 
0.0 
-50.0 
Northing 
Fig. 6. Tritium activity concentrations in transpirate samples collected on two days
along a transect from HIFAR through the LBFG and further north. The standard tree set
is indicated by open symbols.
B
HI  F A R
LF  B G 
TR E  NC H E  S 
HEATHCOTE
RO A  D 
HOL S  WO R T  HY  
J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558 557influence declined to background before the LFBG was reached.
Background levels were re-established to the north of the trenches
before the Heathcote Road samples. Therefore the standard set of
samples clearly showed the effect of 3H sourced from the trenches.
In addition, samples near the trenches showed an increase above
background which implied some movement, albeit small, into the
wider environment at LFBG from either the LFBG trenches or the
adjacent Harringtons Quarry landfill. This was reinforced by recent
data acquired as a consequence of drilling a series of shallow core
holes with depths of 3e5m surrounding and between the trenches,
including one transect of 6 core holes perpendicular to the trenches
at distances of w1 to w15 m. The core extraction work indicated
a shallow, perched, saturated zone within 2e3.5 m of the ground
surface in the area of the trenches and corresponding approxi-
mately with the base of the burial trenches. Samples were gathered
from seepage waters that collected in the core holes over the 24-h
period immediately after drilling in Aug 2009. The 3H concentra-
tions in the transect waters were up tow2.7  104 Bq L
1 closest to
the trenches. However, they declined quickly to <5  103 Bq L
1
within 15 m.
3.3.2. Groundwater interactions
Four sets of measurements of groundwater 3Hmade in 1975 and
1976 by Isaacs and Mears (1977) showed that the highest levels
(>2  105 Bq L
1) occurred in samples taken from shallow bore-
holes located between the two sets of trenches, and this was
confirmed during the Aug 2009 core hole sampling. Tritiated water
has beenmeasured in groundwater samples from LFBG twice yearly
since 1975 as part of the routine monitoring program and an
analysis of the results has been reported by Hughes et al. (2010).
The monitoring indicated that 3H in groundwater has been slowly
dispersing from the trenched area, particularly towards the north
and south, broadly following the surface gradient. The current 3H
maxima are from 3 to 4  104 Bq L
1 as indicated by the Aug 2009
sampling of core holes.
The potential for capillary rise of 3H contaminated groundwater
in the clay soils of the LFBGwould be in the order of 10m (Holtz and
Kovacs, 1981, Tables 6e1) allowing 3H to be transported vertically
up the soil profile during dry periods. This vertical transport
process will mediate any potential tree interaction with the gro-
undwater plume.500
450
400
350
300
250
200
150
100
50
0
0 1000 2000 3000 4000 5000 6000 7000
Interpolated groundwater tritium (Bq/L)
Average Individual samples
Trendline excl Sy 2 high value All data trendline
Fig. 7. Transpirate vs groundwater 3H at LFBG. Groundwater values were estimated
from interpolation of AugeNov 2009 core hole and groundwater samples. The rela-
tionship between transpirate (TR) and groundwater (GW) for all data is H3TR ¼ 0.017
H3GW þ 29.0 (r2 ¼ 0.24). When the highest Sy 2 data point is removed the correlation
improves and the intercept becomes closer to zero: H3 ¼ 0.021 H3TR GW þ 4.9
(r2 ¼ 0.51).
Measured transpirate tritium (Bq/L)
Sy 2
Eu 1
Ac 1
Ba 1
Le 1
Ac 2
Sy 1Given that the locations of the available groundwater boreholes
do not correspond with those of the tree samples, groundwater 3H
values from the Aug 2009 core hole samples and selected OcteNov
2009 borehole samples were interpolated to estimate the
groundwater 3H concentration at the seven standard transpirate
sampling locations. The interpolated groundwater values are
plotted against the transpirate values in Fig. 7. Soil water samples
collected using lysimeters to a depth of w30 cm at three locations
at the LFBG were reported to have 3H levels ranging from 28 to
253 Bq L
1 (Hughes et al., 2010), similar to those found in tran-
spirate in this study; however their locations were not directly
comparable with the vegetation sampled in this study and were
intentionally located in surface seepage zones.
The 3H activity concentrations in the transpirate (and lysimeter)
samples were substantially less than those measured in waters
collected from boreholes across the LFBG site. Average transpirate
concentrations from individual trees and sampling dates ranged
from 0 to 91% of the estimated groundwater concentration, with an
overall average of 4.5% and median of 2%. However, a number of
factors were likely to influence the relationship between tran-
spirate and groundwater, such as recent rainfall, depth to ground-
water, rooting depth and the presence of preferential flow
pathways. Similarly, there were simplifications made in interpola-
tion of the groundwater at the transpirate sampling points
including assumed linear variation in activity (unlikely to be true),
ignoring differences in bore depth and bore construction, variation
in standing water levels and poorly defined 3H distribution in the
trenches. Despite that, an overall positive correlation between
groundwater and transpirate 3H can be seen in Fig. 7 with an r2
value of 0.24. Syncarpia 2, however, exhibited a much higher
average transpirate to groundwater ratio, which may indicate that
the groundwater concentrationwas underestimated at this location
and/or that preferential flow pathways or root locations close to the
trenches provided the tree roots access to more concentrated
groundwater. When the highest data point from Sy 2 is excluded
the r2 increases to 0.51 and the intercept comes nearer to zero.
Nonetheless, the observation that all of the average transpirate
values were less than any interpolated groundwater value points to
the fact that the soil water being accessed by the plants is a mixed
source, likely to comprise rainwater diluted seepage. This is
consistent with the observation, above, that the transpirate 3H
concentrations tended to respond inversely to rainfall intensity.
4. Conclusions
Transpired water from trees growing over, or adjacent to, the
low-level nuclear waste disposal trenches at the LFBG contained
activity concentrations of 3H that were significantly higher (up to
a factor of about 70 times) than local background levels. There was
a correlation between proximity to likely seepage water plumes
and 3H activity concentration in transpired water.
The levels in transpirate were substantially lower than grou-
ndwater concentrations measured across the site (by a median
factor of 50). The observed shifts in tree transpirate 3H concentra-
tion between sampling periods, together with local meteorological
observations, indicate that the soil water within the active root
zones comprises a mixture of seepage and rainfall infiltration. The
degree of mixing was variable given that the soil water activity
concentrations were heterogeneous at a scale equivalent to the
effective rooting volume of the trees. In addition, water taken up by
roots was not well mixed within the trees.
Whilst the nearby HIFAR research reactor was a source of
atmospheric 3H and local fallout over the period of sampling, its
influence reduced to background within the distance to the LFBG.
However, the low-level waste at the LFBG had a slight, but
558 J.R. Twining et al. / Journal of Environmental Radioactivity 102 (2011) 551e558measurable, effect on transpirate 3H activity concentrations beyond
the disposal trenches. Those effects declined rapidly with distance
to be approaching, or at, background levels within 10s of metres
from the trenches. It is possible that some 3H contribution may be
coming from an alternative low-level source, the adjacent landfill
to the west of the trenches, but if so, it is minimal.
Acknowledgements
This work was carried out collaboratively under the supportive
Project leadership of Dr Timothy Payne. A number of technical and
analytical staff also provided valuable assistance. These particularly
include Nicola Creighton, Michelle Vine, Emmy Hoffmann, Barbora
Neklapilova, Robert Chisari, Carla Howe and Kerry Wilsher. The
authors appreciated the cooperation of the Waste Operations
Section at ANSTO in permitting access to the site and other oper-
ational staff (AFP and Site Maintenance) for their consideration of
our field activities.
References
Amano, H., Garten, C.T., 1991. Uptake of tritium by plants from atmosphere and soil.
Environ. Int. 17, 23e29.
Amano, H., Garten, C.T., 1992. Seasonal trends in environmental tritium concen-
trations in a small forest adjacent to a radioactive waste storage area. Fusion
Technol. 21, 700e705.
Boyer, C., Vichot, L., Fromm,M., Losset, Y., Tatin-Froux, F., Guetat, P., Badot, P.M., 2009.
Tritium in plants: a review of current knowledge. Environ. Exp. Bot. 67, 34e51.
Canadell, J., Jackson, R.B., Ehleringer, J.R., Mooney, H.A., Sala, O.E., Schulze, E.D., 1996.
Maximum rooting depth of vegetation types at the global scale. Oecologia 108,
583e595.
Clark, G.H., 2003. Updated Analysis of the Lucas Heights Climatology 1991e2003.
ANSTO/E754 available from: http://apo.ansto.gov.au/dspace/handle/10238/255.
Eichrom Technologies Incorporated, 1996. OTW02 Analytical Procedures: Tritium in
Water. Eichrom Technologies Inc., Darien IL.
Evenden, W.G., Sheppard, S.C., Killey, R.W.D., 1998. Carbon-14 and tritium in plants
of a wetland containing contaminated groundwater. Appl. Geochem. 13, 17e21.
Guswa, A.J., 2008. The influence of climate on root depth: a carbon cost-benefit
analysis. Water Resour. Res. 44, W02427. doi:10.1029/2007WR006384.Hoffmann, E.L., Loosz, T., Ferris, J.M., 2008. Environmental and Effluent Monitoring
at ANSTO Sites, 2006e2007 ANSTO Report E-762.
Holtz,R.D.,Kovacs,W.D.,1981.AnIntroductiontoGeotechnicalEngineering.Prentice-Hall Inc,NJ.
Hughes, C.E., Cendon, D.I., Harrison, J.J., Hankin, S., Johansen, M.P., Payne, T.E.,
Vine, M., Collins, R.N., Hoffmann, E.L., Loosz, T., 2010. Movement of a tritium
plume in shallow groundwater at a legacy low-level radioactive waste disposal
site in eastern Australia. J. Environ. Radioact.. doi:10.1016/j.jenvrad.2010.05.009.
IAEA, 2003. Modelling the environmental transport of tritium in the vicinity of long
term atmospheric and sub-surface sources. Report of the Tritium Working
Group of the Biosphere Modelling and Assessment (BIOMASS) Programme,
Theme 3. IAEA, Vienna. http://www-pub.iaea.org/MTCD/publications/PDF/
Biomass3_web.pdf.
ICRP 108, 2008. Environmental Protection: The Concept and Use of Reference
Animals and Plants. International Commission on Radiological Protection
Publication 108. Approved by the Commission in October 2008.
Isaacs, S., Mears, K., December 1977. A Study of the Burial Ground Used for
Radioactive Waste at the Little Forest Area Near Lucas Heights, New South
Wales Australian Atomic Energy Commission Report AAEC/E427.
Jordan, C.F., Kline, J.R., 1977. Transpiration of trees in a tropical rainforest. J. Appl.
Ecol. 14 (3), 853e860. http://www.jstor.org/stable/2402816.
Kline, J.R., Reed, K.L., Waring, R.H., 1976. Field measurement of transpiration in
Douglas fir. J. Appl. Ecol. 13 (1), 273e283.
Mantell, A., Monselise, S.P., Goldschmidt, E.E., 1979. Movement of tritiated water
through young citrus plants. J. Exp. Bot. 30 (114), 155e164.
Mathew, G., Vasu, K., Vamadevan, V.K., Wahid, P.A., 1988. Measurement of tran-
spiration rate in coconut palm with tritiated water: tritium profile in coconut
crown. J. Nucl. Agric. Biol. 17, 110e112.
McColl, J.G., 1969. Soileplant relations in a Eucalyptus forest on the south coast of
New South Wales. Ecology 50 (3), 354e362.
Raney, F., Vaadia, Y., 1965. Movement and distribution of HTO in tissue water and
vapour transpired by shoots of Helianthus and Nicotiana. Plant Physiol. 40,
383e388.
Specht, R.L., Rayson, P., 1957. Dark Island Heath (Ninety-Mile Plain, south Australia).
III. The root systems. Aust. J. Bot. 5, 103e114.
Stringer, J.W., Kelisz, P.J., Volpe, J.A., 1989. Deep tritiated water uptake and pre-dawn
xylem water potentials as indicators of vertical rooting extent in a Quercus-
Carya forest. Can. J. For. Res. 19, 627e631.
Twining, J.R., 1994. Measurement of Tritiated Water Transpiration from Tree
Leaves Following Root Injection. Proc SPERA Biennial Workshop. ANU,
Canberra.
Twining, J., Harrison, J., Vine, M., Creighton, N., Neklapilova, B., Hoffmann, E., 2009.
Analytical Method Development for Tritium in Tree Transpirate from the Little
Forest Burial Ground. Institute for Environmental Research. ANSTO, NMESP/
TN1.p. 19.
Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics with S, Fourth Ed.
Springer ScienceþBusiness Media, ISBN 978-0-387-95457-8.