Vol.:(0123456789)
1 3
Mine Wa e and he En i onmen (2022) 41:695–703
h ps://doi.o g/10.1007/s10230-022-00888-1
TECHNICAL ARTICLE
Na u al Radioac i i y andMe als inPi Lakes inSweden Analyzed
byP incipal Componen andClus e Analysis
RimonThomas1 · F anciscoPiñe o‑Ga cía1· E aFo ssell‑A onsson1,3· JuanMan e o1,2· Ma sIsaksson1
Recei ed: 1 Ma ch 2021 / Accep ed: 13 July 2022 / Published online: 1 Augus 2022
© The Au ho (s) 2022
Abs ac
The aim o his wo k was o de e mine which pa ame e s a e su icien o measu e in o de o desc ibe he wa e quali y o a
pi lake and o iden i y pa e ns in he da a among di e en kind o pi lakes. The da a consis ed o ambien dose equi alen
a e, elemen al and adionuclide concen a ion, pH, and speci ic conduc ance in su ace wa e and sedimen samples collec ed
om di e en ypes o mines. Da a we e es ed o no mali y and log-no mali y and used in p incipal componen analysis
(PCA) and hie a chical clus e analysis (HCA). The no mali y es s indica ed ha only 40K was no mally dis ibu ed, while
only he 234,238U iso opes we e log-no mally dis ibu ed. HCA pe o med on pa ame e s measu ed in su ace wa e p o ided
clus e s ha in mos cases sepa a ed he elemen s acco ding o hei chemical g oups. Howe e , when HCA was pe o med
on pi lakes, he clus e ing seemed o indica e ha su ace wa e migh no be he p e e ed sample o di e en ia e be ween
di e en ypes o pi lakes. PCA o su ace wa e da a esul ed in h ee componen s ha explained 72% o he a iance when
pH, SC, concen a ion o he elemen s Mg, K, Ca, Cu, Zn, S , Pb, ac i i y concen a ion o 234,238U and 210Po, and ambien
dose equi alen a e we e included. Fo su ace sedimen da a, he PCA esul ed in h ee componen s explaining 83% o he
a iance when he concen a ion o Na, Mg, Al, P, K, Ca, Rb, S , Y, Tl, ac i i y concen a ion o 234Th, 226Ra, 210Pb, 232Th
(se ies a e age), and 40K, and ambien dose equi alen a e we e included.
Keywo ds Mining· HCA· PCA· Radionuclides· U anium· Tho ium
In oduc ion
When a si e has been ound o con ain aluable o es, an
en i onmen al isk assessmen is equi ed be o e mining can
p oceed. One o he many aspec s ha mus be conside ed is
si e es o a ion, which o open cas mining usually in ol es
assu ing ha he esul ing pi lake will no con amina e he
su ounding en i onmen and g oundwa e (Geological Su -
ey o Sweden 2016). Many bio- and geochemical models
and ools o aqueous sys ems exis ha can aid in he isk
assessmen , such as PHREEQC (Pa khu s e al. 1980),
MAGIC (Cosby e al. 1985) and Bio-me (Bio-me 2013).
Some a e based on empi ical models, equilib ium p inci-
ples, and p ocess-o ien ed kine ics, bu a common ea u e
o mos o hem is ha hey a e only alid o a ce ain kind
o aqueous sys em, o a ce ain pH ange, o concen a ion
o dissol ed elemen s. The use o models is u he compli-
ca ed in lakes ha a e exposed o man-made in luences such
as liming and o lakes in Scandina ia whe e he su ace can
be co e ed by ice se e al mon hs pe yea .
Pi lakes di e in many ways om na u al lakes, which
ha e been s udied o much longe ime pe iods and whe e
expe ise, models, and ools a e eadily a ailable. One majo
di e ence wi h pi lakes is he unde lying bed ock, which
could con ain a highe concen a ion o o es o mine als
ha can signi ican ly al e he chemical p ope ies in he
lake, such as oxida ion o sul ide-bea ing mine als leading
o acidic wa e (Reddy e al. 1995; San o imia and López-
Pamo 2013; Vaughan and Lennie 1991). Fu he mo e, he
inal en i onmen al impac canno be seen un il decades
a e mine closu e.
* Rimon Thomas
imon. [email p o ec ed]
1 Depa men o Medical Radia ion Sciences, Ins i u e
o Clinical Sciences, Sahlg enska Academy a Uni e si y
o Go henbu g, 41345Go henbu g, Sweden
2 Depa men o Applied Physics II, ETSA, Uni e sidad de
Se illa, 41012Se ille, Spain
3 Depa men o Medical Physics andBiomedical Enginee ing,
Sahlg enska Uni e si y Hospi al, Region Väs a Gö aland,
41345Go henbu g, Sweden
696 Mine Wa e and he En i onmen (2022) 41:695–703
1 3
Rega ding na u ally occu ing adionuclides in pi lakes,
o me u anium mines ha e been s udied ex ensi ely (Fe a i
e al. 2015; IAEA echnical mee ing 2004; Rollings e al.
1999; S ømman e al. 2013). Howe e , he e a e ew such
s udies on non-u anium mine pi lakes. Mines and qua ies
con aining pi lakes can be ound in many places in Sweden
and hese we e he ocus in his s udy, along wi h hei na u-
ally occu ing adionuclide con en .
The da a used in his s udy consis s o pa ially p e i-
ously published da a o 51 pi lakes om 39 si es sp ead
ac oss Sweden, om Malmö in he sou h o Ki una in he
no h (Man e o e al. 2020; Thomas e al. 2022). Only a
ew had been p e iously s udied (Lu 2004) and mos we e
app oached as unknown sys ems wi hou any p io knowl-
edge wi h ega d o expec ed concen a ions. So, he aim
o his wo k was o de e mine which wa e pa ame e s a e
su icien o measu e in o de o desc ibe he wa e quali y
o a pi lake, and which a e edundan . Raw wa e and su -
ace sedimen s we e analyzed by spec ome ic echniques
o ob ain a la ge numbe o pa ame e s o build baseline
da a on hese pi lakes. Unpublished da a oge he wi h sup-
plemen a y da a gi en in p e ious publica ions (Man e o
e al. 2020; Thomas e al. 2022) we e analyzed by p incipal
componen analysis (PCA) and hie a chical clus e analysis
(HCA) o iden i y pa e ns and clus e s in he da a and clas-
si y Swedish pi lakes wi h simila dis ibu ions o wa e
quali y pa ame e s.
Ma e ial andMe hods
Among he pa ame e s ha we e measu ed in su ace wa e
and sedimen s in p e ious s udies (Man e o e al. 2020;
Thomas e al. 2022), hose included he e we e pH, speci ic
conduc ance (SC), concen a ions o Na, Mg, K, Ca, Mn,
Cu, Zn, S , Ba, Pb, and o he su ace wa e , ac i i y con-
cen a ions o 234,238U and 210Po. Fo su ace sedimen s, he
pa ame e s included we e he concen a ions o *Na, *Mg,
Al, Si, *P, S, *K, *Ca, *Ti, Mn, Fe, *Ni, Zn, As, *Rb, S ,
*Y, *Z , Ba, *Tl, Pb, and he ac i i y concen a ions o 40K,
234Th, 226Ra (214Pb, 214Bi), 210Pb and 232Th (* indica es da a
no p e iously published in Man e o e al. 2020 and Thomas
e al. 2022). The ac i i y concen a ion o 232Th is he se ies
a e age om he gamma emi e s 228Ac, 212Pb, 212Bi and
208Tl, which we e ound o be in secula equilib ium. An
a e age was no used o he 238U decay se ies, since a dis-
equilib ium was seen in he decay chain o he sedimen
samples. The disequilib ium was mos se e e o 238U o
210Po a ios (up o 0.36 ± 0.02). Fo some pa ame e s in a ew
lakes he concen a ion was below he de ec ion limi and
hese en ies we e eplaced by hal he de ec ion limi , which
has been ound o be a good subs i u e (Fa nham e al. 2002).
Fo ICP-MS de ec ion, limi s anged om 1µg/L–30ng/L,
depending on he elemen ; o alpha spec ome y, i was ≈
1mBq/kg o he wa e samples and 1Bq/kg o sedimen s.
Apa om he concen a ion o s able elemen s and adi-
onuclides in he su ace wa e and sedimen s, measu emen s
o ambien dose equi alen a e we e also included in he s a-
is ical analyses. This quan i y was measu ed wi h an ene gy
compensa ed GM- ube, SRV-2000 (RADOS, Finland) a a
heigh o 1m abo e g ound nea each pi lake (n = 50). Only
one alue was missing (si e 1) and i was subs i u ed by he
alue om a nea by sampled pi lake (si e 2).
The pi lakes we e di ided acco ding o hei loca ion
in Sweden, sou h and no h, ep esen ed by S o N in he
sample code, ollowing he bo de ou lined by limes no -
landicus (Ho n a e Ran zien 1948) and he Illies eco egion
14 (Illies 1978). The ype o mine o qua y was speci ied i
known, ollowed by a si e numbe and ending wi h a le e
A, B, o C, which indica es he p esence o se e al pi lakes
loca ed wi hin he same mining si e (a adius o 100m).
The s udied pi lakes in no he n Sweden (n = 21) included
16 Cu and Zn mines, wo Pb mines, one Au mine, and wo
si es whe e he ex ac ed o e was unknown. The pi lakes in
sou he n Sweden (n = 30) included one sil e and i e i on
mines, six limes one, i e ma ble, h ee eldspa , one g ani e,
and ou s one qua ies. The la e included di e en kinds o
ocks o a ious pu poses. A i e si es, he ype o ex ac ed
o e was unknown. 27 o hese 51 pi lakes we e sampled
o sedimen s: 16 in no he n Sweden and 11 in sou he n
Sweden (Fig.1).
The HCA and PCA s a is ical analyses we e pe o med
using IBM® SPSS S a is ics ( e sion 26) so wa e. HCA
is an unsupe ised s a is ic ool ha ca ego izes objec s
in o homogenous g oups by applying a simila i y c i e-
ion be ween objec s. In his wo k, he squa ed Euclidean
dis ance was used as he c i e ion. The analysis s a s by
conside ing all objec s (pi lakes) as indi idual clus e s and
calcula es he squa ed Euclidean dis ance be ween hem; he
minimum a iance me hod (Wa d 1963) was used o clus e
he pi lakes acco ding o hei simila i ies. A h eshold o
50% (linkage dis ance 12.5) was used o sepa a e clus e s.
The esul s o HCA we e plo ed as a escaled dis ance den-
d og am by he so wa e, whe e he dissimila i ies o he
objec s a e ep esen ed by he connec ed b anches.
PCA is a mul i a ia e s a is ical analysis ha can com-
p ess da a as well as iden i y endencies by educing ini-
ial pa ame e s o a ew unco ela ed componen s. These
componen s a e linea combina ions o he o iginal pa am-
e e s in such a way ha each componen explains as much
a iance as possible among he da a se (Mille and Mille
2005; Piñe o-Ga cía e al. 2012). The amoun o a iance
explained by each componen is quan i ied by an eigen-
alue and acco ding o Gu man–Kaise c i e ion (Gu man
1954; Kaise 2011), componen s wi h eigen alues abo e
1 should be e ained as hey explain su icien a iance
697Mine Wa e and he En i onmen (2022) 41:695–703
1 3
wi hou in o ma ion loss. This c i e ion was used o selec
componen s.
Also, a Va imax o hogonal o a ion was used o iden-
i y he pa ame e s in luencing each componen (Mille and
Mille 2005; Piñe o-Ga cía e al. 2012). The pa ame e s used
in PCA we e s anda dized in o no mal dis ibu ion (a e age
alue 0 and s anda d de ia ion 1) o a oid he i s compo-
nen being in luenced by he pa ame e s wi h highe a i-
ance. The pa ame e s a e lis ed in each o a ed componen
wi h loadings (weigh s) ha can be in e p e ed as co ela-
ions o each componen ; a high loading (> 0.6) would hus
indica e ha he speci ic pa ame e is impo an in explain-
ing he a iance o ha pa icula componen .
Resul s
Sample Cha ac e is ics
Table1 summa izes he desc ip i e s a is ics o he ac i i y
concen a ion o adionuclides in su ace wa e (n = 51) and
sedimen (n = 27). The dis ibu ion o ac i i y concen a-
ion o all adionuclides excep o 40K we e asymme ic
(skewness > 0), wi h la ge ails han o a no mal dis ibu-
ion (ku osis > 0). Compa ison be ween he pe cen iles and
he a e age ac i i y concen a ion indica es some pi lakes
wi h high ac i i y concen a ions in bo h su ace wa e and
sedimen s, e.g. whe e he a e age 234,238U, 234Th, 226Ra and
210Pb exceeded he hi d pe cen ile.
A es o no mali y was pe o med using bo h Kolmogo-
o –Smi no and Shapi o–Wilks es s, whe e bo h es s ga e
p- alues o 0 o ac i i y concen a ion o all adionuclides
excep o 40K (p = 0.055), indica ing ha 234,238U, 232,234Th,
226Ra, 210Pb and 210Po we e no no mally dis ibu ed (no mal
dis ibu ion i p > 0.05). Analysis o he log10 o he ac i i y
concen a ion alues ( es ing o log-no mali y dis ibu ion)
showed no o ela i ely weak log-no mal dis ibu ion o all
adionuclides excep 234U (p = 0.154). Simila esul s we e
seen o he s able elemen s analyzed by ICP-MS (in wa e )
and XRF (in sedimen s), bu Na, K, and S in sedimen s and
he pH in su ace wa e we e ound o be no mally dis ib-
u ed (da a no shown).
The a io o 234U o 238U ac i i y concen a ions is also
included in Table1 and e lec s o some ex en a se o chem-
ical and physical p ocesses ha causes and enhances his
disequilib ium. The la e can be caused by alpha pa icles
ejec ed om 238U, esul ing in 234Th and 234U ecoiling in o
he wa e (Gascoyne e al. 2002; Kigoshi 1971; Mahe e al.
2006). Also, high ca bona es concen a ions in he wa e is
known o inc ease he solubili y o U (Leh o and Hou 2010).
Thus, high concen a ions o ca bona es in limes one qua -
ies, o example, would p omo e he disequilib ium, since
he eadily a ailable and ejec ed 234U a om can be bound o
ca bona es and u he solubilized, in con as o 238U bound
o he la ice. Las ly, he disequilib ium may also e lec he
sou ce o he wa e inpu since he a io o 234U o 238U in
g oundwa e usually exceeds uni y (Suksi e al. 2006), and
has been epo ed o be as high as 10 (Cowa 1981). Tes
Fig. 1 O e iew o he loca ions o he mining si es included in his
s udy. Map da a ©2020 Google
698 Mine Wa e and he En i onmen (2022) 41:695–703
1 3
esul s indica e ha his a io is bo h no mally and log-no -
mally dis ibu ed.
Hie a chical Clus e Analysis
Figu e2 shows he dend og ams om HCA o he pa am-
e e s measu ed in su ace wa e (pH, SC, Mg, K, Ca, Cu, Zn,
S , Pb, 234,238U and 210Po) and sedimen s (Na, Mg, Al, Si, P,
S, K, Ca, Ti, Mn, Fe, Ni, Zn, Rb, S , Y, Z , Ba, Tl, Pb, 40K,
234Th, 226Ra, 210Pb and 232Th). Ambien dose equi alen a e
we e included in bo h analyses.
In he wa e , h ee clus e s can be seen e lec ing simila
chemical beha io o he elemen s. The i s clus e con ains
he adionuclides and ambien dose equi alen a e, he sec-
ond he ansi ional and pos - ansi ion me als, and he hi d
includes pH, SC, and he alkali and alkaline ea h me als.
In sedimen s, he clus e s o pa ame e s a e mo e di icul
o in e p e since hey do no necessa y e lec any chemical
beha io o p ocess; hey mos likely ep esen he chemical
composi ion o he sedimen s (bed ock) o hese pi lakes.
Fo example, he clus e wi h Ca and Mg could e lec he
p esence o he mine al dolomi e (CaMg(CO3)2 which can
be ound in limes one qua ies, Fe and S can occu as he
mine al py i e (FeS2), and he clus e o Y ( a e ea h ele-
men ) wi h he adionuclides is a known co-occu ence in
mine als (Ba akos e al. 2015; Jai e h e al. 2014).
Figu e3 shows he dend og am o 51 pi lakes whe e
he pa ame e s used a e hose p esen ed in Table2 (pH,
SC, Mg, K, Ca, Cu, Zn, S , Pb, 234,238U, 210Po and ambien
dose equi alen a e). A h eshold o 50% (dis ance 12.5)
was used o dis inguish be ween clus e s, esul ing in ou
clus e s. Clus e one con ains 38 pi lakes including e y di -
e en ypes o mines. The smalle clus e s a e mo e homog-
enous in e ms o he ype o mine included. Clus e wo
con ains Ag and Zn mines, clus e h ee a e mainly eldspa ,
g ani e, and s one qua ies, and clus e ou a e all Zn and
Cu mines.
The Ag and Zn mines had a ela i ely low pH (4.9 and
5.7) and Zn concen a ions o 5900 and 3900µg/L, which
we e he wo highes concen a ions measu ed in all o he
su ace wa e samples. The Zn concen a ion in he Ag mine
was likely due o he p esence o sul ide bea ing mine als
(Cox 1970), as galena was one o he o es iden i ied a he
si e, which was u he con i med since ha pi lake had
he highes measu ed concen a ion o Pb in su ace wa e
(690µg/L). Thus, i is likely ha he Pb, Zn, and Ag exis
as galena (PbS), sphale i e (ZnS), and a gen i e (Ag2S) a
his si e.
The pi lakes in clus e h ee di e om he es mainly
due o ela i ely high concen a ions o U. The clus e a e -
age o 238U was 490mBq/kg whe eas he a e age o all lakes
was 110mBq/kg (Table1). One s one qua y, in clus e one
a he han clus e h ee, had a su ace wa e 238U ac i i y
concen a ion o 1.4mBq/kg. O he wise, all eldspa , g an-
i e, and s one qua ies a e included in clus e h ee.
Clus e ou includes only h ee o me Cu and Zn
mines pi lakes. These lakes had a pH o 7–9 and he high-
es speci ic conduc ance, 1600–2000 µS/cm, whe eas he
a e age SC o all lakes we e 400 µS/cm. They also had
he highes su ace wa e concen a ions o Ca; he clus e s
a e aged 1000mg/L while he limes one qua ies (n = 6)
Table 1 Desc ip i e s a is ics o adionuclide concen a ion in su ace wa e (n = 51) and sedimen (n = 27) om p e iously published esul s
(Man e o e al. 2020; Thomas e al. 2022)
Ac i i y concen a ion o adionuclides in su ace wa e is gi en in mBq/kg, and o su ace sedimen s in Bq/kg. P alues om Kolmogo o –
Smi no (pKS) and Shapi o–Wilk (pSW) es s a e shown a he bo om o he no mal and log-no mal es s
Su ace wa e Su ace sedimen s
210Po 234U238U234U/238U234Th 226Ra 210Pb 232Th 40K
Range 0.8–94 < 0.5–1700 < 0.5–1200 0.3–2.3 6–1500 10–1300 14–1400 7–170 121–1400
Mean ± SD 8.5 ± 2 150 ± 45 110 ± 33 1.3 ± 0.1 120 ± 61 130 ± 59 140 ± 58 37 ± 7 680 ± 69
Median 3.4 16 12 1.3 24 35 47 24 740
SD 15 320 240 0.3 320 300 300 38 360
Skewness 4.3 ± 0.3 3.2 ± 0.3 3.0 ± 0.3 0.1 ± 0.3 3.7 ± 0.4 3.3 ± 0.4 3.4 ± 0.4 2.5 ± 0.4 − 0.35 ± 0.4
Ku osis 21 ± 0.7 12 ± 0.7 9.6 ± 0.7 1.3 ± 0.7 14 ± 0.9 11 ± 0.9 12 ± 0.9 6.7 ± 0.9 0.39 ± 0.9
Pe cen iles
25% 2 3 2 1.1 17 23 32 16 500
50% 3 16 12 1.3 24 35 47 24 740
75% 10 130 84 1.5 43 54 85 35 900
No mal es pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0.06
pSW = 0.02 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0.03
pSW = 0.06
Log-no mal es pKS = 0.01
pSW = 0.01 pKS = 0.15
pSW = 0.01 pKS = 0.05
pSW = 0.01 pKS = 0.2
pSW = 0.4 pKS = 0.01
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0 pKS = 0
pSW = 0
699Mine Wa e and he En i onmen (2022) 41:695–703
1 3
a e aged 200mg/L. The S concen a ions we e also he
highes ; his combina ion be ween high S and Ca concen-
a ions can be explained by he liming ac i i ies ha a e
conduc ed a hese si es o main ain good wa e quali y.
The sul ide-bea ing mine als would mos likely cause
acidic wa e and high me al concen a ions in he su ace
wa e i liming we e no done.
P incipal Componen Analysis (PCA)
Su ace Wa e
Table2 shows he esul s om PCA as well as he load-
ings o he o a ed componen s o he su ace wa e samples
(n = 51). Th ee componen s wi h eigen alues exceeding 1
explain 72% o he o al a iance. C1 explains 32% o he
2520151050
Y
Ra226
Pb210
Th234
Y
Th232
Dose a e
Rb
Al
K
K40
Na
Fe
As
S
Ni
Zn
Mn
Ti
P
S
Si
Z
Mg
Ca
Pb
Tl
Ba
b)
2520151050
Y
U238
U234
Po210
Dose a e
Zn
Pb
Cu
SC
Ca
S
Mg
K
pH
a)
Fig. 2 Resul s om hie a chical clus e analysis (HCA) o pa am-
e e s measu ed in a su ace wa e and in b sedimen s in pi lakes in
Sweden. The clus e s in a seem o sepa a e he di e en pa ame e s
acco ding o hei o igin ( adionuclides) and chemical g oups ( an-
si ional and pos - ansi ion me als, alkali and alkaline ea h me als).
While in b he clus e s appea o e lec he composi ion o he sedi-
men s
Fig. 3 Resul s om hie a chical clus e analysis o 51 su ace wa e
samples om pi lakes in Sweden including he ollowing pa am-
e e s: pH, SC, Mg, K, Ca, Cu, Zn, S , Pb, 234,238U, 210Po and dose
a e. The dend og am shows ou clus e s when applying a h eshold
o 50% o dis ance 12.5. Clus e one includes mos o he pi lakes,
while he smalle clus e s a e mo e homogenous in e ms o he ype
o mine included

700 Mine Wa e and he En i onmen (2022) 41:695–703
1 3
o al a iance wi h posi i e co ela ions wi h SC, Mg, K,
Ca, and S . C2 explains 21% o he o al a iance wi h posi-
i e co ela ions wi h 234,238U and 210Po and also wi h he
ambien dose equi alen a e. C3 explains 19% o he o al
a iance wi h a nega i e co ela ion wi h pH and a posi i e
co ela ion wi h Cu, Zn, and Pb. Thus, C1 is la gely in lu-
enced by he alkali and alkaline ea h me als, C2 p ima ily
by he na u ally occu ing adionuclides, and C3 by he pos -
and ansi ion me als. The ela ionship be ween he nega i e
loading o pH and he posi i e loadings o me al concen a-
ions in C3 is assumed o be due o he ac ha a low pH
inc eases me al solubili y. In Fig.4, he PCA sco es and
PCA loadings a e plo ed wi h espec o he h ee o a ed
componen s.
In he C1–C2 plo , mos o he pi lakes a e cen e ed
close o he o igin. Howe e , one g oup o pi lakes
ex ends o he igh , which co esponds o he eldspa ,
g ani e, and s one qua ies, which a e known o ha e
highe u anium and ho ium concen a ions (Mau ice
1980; Wilson and Åke blom 1980). These lakes a e all in
clus e h ee (Fig.3) and he s one qua y ha had a low
238U ac i i y concen a ion is seen o be g ouped oge he
wi h mos o he pi lakes, simila o clus e one (Fig.3).
The second pi lake g oup wi h a dis inc upwa d p ojec-
ion include he Cu and Zn mines (also seen in clus e 4 in
Fig.3), which align wi h he loadings o SC, Ca, and S .
The SC and Ca is because hese pi lakes a e being limed
o main ain a neu al pH. The highes sul u concen a ion
in su ace wa e was also ound among hese pi lakes, also
con ibu ing o he SC eading.
The esul s seen in he C2–C3 plo a e simila o hose
in C1–C2 o he pi lakes ela ed o C2 ( adionuclide load-
ings). The Cu and Zn mines a e, howe e , mo e closely
g ouped since he SC loading is now cen e ed a ound he
o igin oge he wi h ha he Cu and Zn mines no being
in luenced by nei he C2 no C3. One Zn mine (pi lake) is
posi ioned o he igh oge he wi h an Ag mine, and hese
a e he wo lakes ha had he highes Zn concen a ion in
su ace wa e . Since he loading o Zn in C3 is posi i e and
o he igh , so also a e hese wo lakes.
The ela i e posi ion o he pa ame e s in he loading
plo s in Fig.4 a e e y simila o he ou clus e esul s
ob ained by HCA. He e he adionuclides and ambien
dose equi alen a e a e seen o be g ouped oge he , SC
and alkali and alkaline ea h me als, he ansi ional and
pos - ansi ion me als, and as shown by C3, pH is seen
sepa a ing he es , o ming a ou h g oup.
Su ace Sedimen The esul s om PCA on he 27 su -
ace sedimen samples is shown in Table3, whe e h ee
componen s wi h eigen alues g ea e han 1 explain 83%
o he o al a iance. The plo o C1–C2 is ound in Fig.5.
C1 explains 51% o he a iance and is highly co ela ed
wi h he ac i i y concen a ion o adionuclides, ambien
dose equi alen a es, and concen a ions o Rb and Y. C1
is also nega i ely co ela ed wi h S concen a ion, which
seems o indica e ha he sedimen s in his da a se ha
a e posi i ely co ela ed wi h U and Th a e nega i ely co -
ela ed wi h S . Also, he posi i e loadings o Y and Rb
would imply ha highe concen a ion o hese elemen s
co ela es wi h highe ac i i y concen a ion o he na u-
ally occu ing adionuclides. The co ela ion be ween he
concen a ion o Rb and he ac i i y concen a ion o he
adionuclides is less appa en , excep a he eldspa mine.
This will be in es iga ed u he in ano he wo k. How-
e e , he co ela ion o he concen a ion o Y wi h he
ac i i y concen a ion o he adionuclides is mo e appa -
en (R2 = 0.97) and he i ed equa ion is:
This equa ion is alid o he ange o he 234Th ac i i y
concen a ions, om 6 o 1500Bq/kg and o concen a-
ions o Y om 17 o 63ppm. Simila i ing, hough wi h
lowe R2 alues, was achie ed using 226Ra o 210Pb ins ead
o 234Th.
C2 explains he emaining 23% o he o al a iance
and is nega i ely co ela ed wi h Mg and Ca and posi-
i ely co ela ed wi h he es o he pa ame e s. Since
C3 only explains an addi ional 9% o he o al a iance, i
(1)
234Th=0.57 ∙Y2−11.7 ∙Y−18
Table 2 Resul s om PCA a e a Va imax o a ion o 51 su ace
wa e samples om pi lakes in Sweden wi h h ee componen s
explaining 72% o he o al a iance
C1 is highly connec ed o alkali and alkaline ea h me als, C2 o na u-
ally occu ing adionuclides and ambien dose equi alen a e (dose
a e), and C3 o he pos - and ansi ion me als
Ro a ed componen s
C1 C2 C3
pH 0.15 − 0.31 − 0.76
SC 0.95 − 0.14 0.05
Mg 0.65 − 0.10 − 0.13
K 0.83 0.09 − 0.14
Ca 0.92 − 0.13 0.00
Cu − 0.16 0.03 0.66
Zn 0.12 − 0.16 0.93
S 0.94 − 0.09 − 0.05
Pb − 0.03 − 0.04 0.79
238U 0.05 0.95 0.04
234U 0.04 0.95 0.03
210Po − 0.17 0.76 0.01
Dose a e − 0.17 0.61 − 0.01
Va iance explained (%) 32 21 19
Cumula i e (%) 32 53 72
Eigen alue 4.2 2.7 2.4
701Mine Wa e and he En i onmen (2022) 41:695–703
1 3
was conside ed o mino impo ance, and he C1–C3 and
C2–C3 plo s a e no shown.
40K analyzed by gamma spec ome y and 39,41K analyzed
by XRF a e posi ioned closely in he PCA loading plo s. The
a io be ween 40K ac i i y concen a ion and K concen a-
ion analyzed by XRF gi es 252Bq/kg o 40K pe % o K.
Assuming a na u al abundance o 40K o 0.012% o na u al
K, his would co espond o ≈ 320Bq/kg 40K pe % o K,
26% highe han he alue quo ed he e. Howe e , his di e -
ence is in he same magni ude as he associa ed unce ain-
ies o he da a o bo h he XRF and gamma spec ome y
analysis.
Un o una ely, Cu was only de ec ed in a ew sedimen
samples (mainly Cu mines) and had o be excluded due o
oo many missing alues. This is one o he easons why he
Cu and Zn mines a e g ouped a ound he o igin since he e
is no pa ame e (o loading) ha ex ends hem a ma ked
di ec ion in he sco e plo . Since hese pi lakes con ain low
concen a ion o adionuclides, ocusing on o he elemen s
and excluding he adionuclides in he PCA would be e
explain he cha ac e is ics o hese lakes. Howe e , ha is
ou side o he scope o his s udy, which was ocused on he
na u ally occu ing adionuclides.
Conclusions
Resul s om he desc ip i e s a is ics showed ha mos na u-
ally occu ing adionuclides in su ace wa e and sedimen s
in pi lakes a e no no mally o log no mally dis ibu ed.
Fig. 4 Plo s o PCA sco e and loadings o 51 su ace wa e sam-
ples om pi lakes in Sweden. The di e en mines a e depic ed wi h
unique symbols and he smalle igu e o he op igh shows he plo -
ed loadings om each componen om Table2. a Shows he ela-
ion be ween C1 and C2, b be ween C2 and C3 and c be ween C1 and
C3. Two ends can be seen in a and b, i s he eldspa , g ani e and
s one qua ies a e aligned wi h he loadings o he adionuclides and
ambien dose equi alen a e, which is easonable since hey con ain
highe ac i i y concen a ion o adionuclides and consequen ly ea-
u es highe dose a es. The second end is seen in a whe e he Cu
and Zn mines a e mainly aligned wi h he loadings o speci ic con-
duc ance (SC), Ca and S
702 Mine Wa e and he En i onmen (2022) 41:695–703
1 3
Ca e should he e o e be aken when calcula ing s anda d
de ia ions o pe o ming o he s anda diza ion p ocedu es
ha assume a no mal dis ibu ion.
Hie a chical clus e analysis pe o med on pa ame e s
measu ed in su ace wa e p o ided clus e s ha ag eed wi h
assumed ela ions among elemen s in he chemical g oups.
By including only one o wo pa ame e s om each clus e ,
one could po en ially educe he numbe o pa ame e s ha
needs o be measu ed. This can educe analysis cos s while
main aining he pa ame e s needed o explain he di e en
chemical cha ac e is ics o a pi lake. The clus e ing o pi
lakes wi h espec o he ype o mines seems mo e plausible
o he smalle clus e s, which included simila mines and
qua ies. Howe e , he la ges clus e included many di -
e en kinds o mines and qua ies, indica ing ha su ace
wa e migh no be he bes sample o di e en ia e be ween
di e en ypes o pi lakes when using hie a chical clus e
analysis.
P incipal componen analysis wi h a imax o a ion
showed ha when including a la ge a ie y o pi lakes
(limes one qua ies o gold mines), he ollowing pa ame e s
(apa om he adionuclides and ambien dose equi alen
a e) explained mos o he a iance: pH, SC, and concen-
a ions o Mg, K, Ca, Cu, Zn, S , and Pb. Fo su ace sedi-
men s, he co esponding pa ame e s we e concen a ions o
Na, Mg, Al, P, K, Ca, Rb, S , Y, and Tl.
A co ela ion in he PCA be ween he concen a ion o Y
and he ac i i y concen a ion o 238U (234Th) in sedimen s
and ambien dose equi alen a e was ound. The e o e, we
ecommend including analysis o Y when s udying na u ally
occu ing adionuclides in he en i onmen o ollow his
ela ionship u he . Ambien dose equi alen a e was also
closely posi ioned wi h he adionuclides in he PCA load-
ing plo s o su ace sedimen , indica ing ha ambien dose
equi alen a e could be used as a subs i u e in PCA i adio-
nuclide concen a ions a e no measu ed. The subs i u ion is
expec ed o be less good o su ace wa e since he loadings
o ambien dose equi alen a e is u he away. Since ambi-
en dose equi alen a e is ela i ely simple o measu e, i
is ecommended o be included in gene al s udies on lakes.
Acknowledgemen s The au ho s acknowledge he suppo o and
aluable discussions wi h ou now-deceased colleague, P o esso Elis
Holm, who was pa o he ea ly s ages o his s udy. This esea ch
was suppo ed by esea ch g an s om he Swedish Radia ion Sa e y
Au ho i y (SSM) [G an no. SSM2014-3485].
Funding Open access unding p o ided by Uni e si y o Go henbu g.
Open Access This a icle is licensed unde a C ea i e Commons A i-
bu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a-
ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce,
p o ide a link o he C ea i e Commons licence, and indica e i changes
we e made. The images o o he hi d pa y ma e ial in his a icle a e
Table 3 Resul s om PCA a e a Va imax o a ion o 27 su ace
sedimen samples om pi lakes in Sweden
The h ee componen s explain 83% o he o al a iance, howe e , C1
alone explains 51% and his is due o se e al adionuclides included
in he analysis whe e hey also a e co ela ed o each o he
Ro a ed componen s
C1 C2 C3
Na − 0.06 0.39 0.80
Mg − 0.13 − 0.93 − 0.03
Al 0.40 0.42 0.73
P− 0.48 0.41 0.39
K 0.45 0.51 0.53
Ca − 0.15 − 0.90 − 0.13
Rb 0.89 0.24 0.19
S − 0.61 0.46 0.39
Y 0.95 0.17 0.12
Tl − 0.16 0.30 − 0.83
234Th 0.96 0.03 0.07
226Ra 0.97 0.03 0.07
210Pb 0.96 0.04 0.09
232Th (se ies a e age) 0.93 0.08 0.13
40K 0.61 0.38 0.35
Dose a e 0.94 0.08 0.05
% o a iance 51 23 9
Cumula i e (%) 51 74 83
Eigen alue 8.2 3.7 1.5
Fig. 5 Plo s o PCA sco e and loadings o 27 su ace sedimen sam-
ples om pi lakes in Sweden. The di e en mines a e depic ed wi h
unique symbols and he smalle igu e o he op le shows he plo -
ed loadings om each componen , om Table 3. Apa om he
clus e cen e ed a ound o igin, he pi lake o he a le aligns wi h
Ca and Mg, p obably since he da a is om a limes one qua y. Simi-
la ly, as seen in Fig.4 o su ace wa e , he lakes aligned wi h adio-
nuclides and ambien dose equi alen a e a e he eldspa /s one qua -
ies, he g ani e qua y is no included he e since no sedimen s we e
sampled om ha lake
703Mine Wa e and he En i onmen (2022) 41:695–703
1 3
included in he a icle's C ea i e Commons licence, unless indica ed
o he wise in a c edi line o he ma e ial. I ma e ial is no included in
he a icle's C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will
need o ob ain pe mission di ec ly om he copy igh holde . To iew a
copy o his licence, isi h p:// c ea i eco mmons. o g/ licen ses/ by/4. 0/.
Re e ences
Ba akos G, Mischo H, Gu zme J (2015) Ra e ea h unde g ound min-
ing e alua ions wi h espec o adioac i i y. In: De Lima B, Filho
L (eds) Ra e ea hs indus y, echnological, economic, and en i-
onmen al implica ions, ch 8. Else ie , Ams e dam, pp 121–138
Bio-me (2013) Bioa ailabili y ool. Eu opean Coppe Ins i u e, In e -
na ional Zinc Assoc, Nickel P oduce s En i onmen al Resea ch
Associa ion, bio-me .ne
Cosby BJ, Ho nbe ge GM, Galloway JN, W igh RF (1985) Modeling
he e ec s o acid deposi ion: assessmen o a lumped pa ame e
model o soil wa e and s eamwa e chemis y. Wa e Resou Res
21:51–63. h ps:// doi. o g/ 10. 1029/ WR021 i001p 00051
Cowa JB (1981) U anium iso opes and 226Ra con en in he deep
g oundwa e s o he T i-S a e egion, USA. J Hyd ol 54:185–193.
h ps:// doi. o g/ 10. 1016/ 0022- 1694(81) 90159-1
Cox DP (1970) Lead-zinc-sil e deposi s ela ed o he Whi e Moun ain
plu onic se ies in New Hampshi e and Maine. U.S. Geological
Su ey Bulle in 1312-D. h ps:// doi. o g/ 10. 3133/ b1312D
Fa nham IM, Singh AK, S e zenbach KJ, Johannesson KH (2002)
T ea men o nonde ec s in mul i a ia e analysis o g oundwa e
geochemis y da a. Chemome In ell Lab 60:265–281. h ps:// doi.
o g/ 10. 1016/ S0169- 7439(01) 00201-5
Fe a i C, Aze edo H, Wisniewski M, Rodghe S, Roque C, Nasci-
men o M (2015) An o e iew o an acidic u anium mine pi lake
(Caldas, B azil): composi ion o he zooplank on communi y and
limnochemical aspec s. Mine Wa e En i on 34:343–351. h ps://
doi. o g/ 10. 1007/ s10230- 015- 0333-9
Gascoyne M, Mille NH, Neyma k LA (2002) U anium-se ies disequi-
lib ium in u s om Yucca Moun ain, Ne ada, as e idence o
po e- luid low o e he las million yea s. Appl Geochem 17:781–
792. h ps:// doi. o g/ 10. 1016/ S0883- 2927(02) 00038-0
Geological Su ey o Sweden (2016) Vägledning ö p ö ning a g u -
e ksamhe . SGU- appo 2016:23 (in Swedish)
Gu man L (1954) Some necessa y condi ions o common- ac o
analysis. Psychome ika 19:149–161. h ps:// doi. o g/ 10. 1007/
BF022 89162
Ho n a Ran zien H (ed) (1948) S ensk Bo anisk Tidsk i . S enska
bo aniska ö eningen, S ockholm. 42(1) (in Swedish)
IAEA Technical Mee ing (2004) Annex. Pape s p esen ed a IAEA
echnical mee ing, Saska oon, Canada. h ps:// www- pub. iaea. o g/
MTCD/ publi ca io ns/ PDF/ P1406_ CD/ PDF/ P1406_ Annex. pd
Illies J (1978) Limno auna Eu opaea. Fische , S u ga
Jai e h S, Hoa son DM, Miezi is Y (2014) Geological se ing and
esou ces o he majo a e-ea h-elemen deposi s in Aus alia.
O e Geol Re 62:72–128. h ps:// doi. o g/ 10. 1016/j. o ege o e .
2014. 02. 008
Kaise H (2011) A no e on Gu man’s lowe bound o he numbe
o common ac o . B J S a Psychol 14:1–2. h ps:// doi. o g/ 10.
1111/j. 2044- 8317. 1961. b000 61.x
Kigoshi K (1971) Alpha- ecoil ho ium-234: dissolu ion in o wa e and
he u anium-234/u anium-238 disequilib ium in na u e. Science
173:47. h ps:// doi. o g/ 10. 1126/ scien ce. 173. 3991. 47
Leh o J, Hou X (2010) Chemis y and analysis o adionuclides. Labo-
a o y echniques and me hodology. Wiley-VCH, Weinheim
Lu M (2004) Pi lakes om sulphide o e mining, geochemical and lim-
nological cha ac e iza ion be o e ea men , a e liming and sew-
age sludge ea men s: cases s udies a Rä lidmy an and Udden,
Sweden. PhD hesis, comp ehensi e summa y, Luleå Uni o
Technology
Mahe K, DePaolo DJ, Ch is ensen JN (2006) U-S iso opic speedom-
e e : luid low and chemical wea he ing a es in aqui e s. Geo-
chim Cosmochim Ac 70:4417–4435. h ps:// doi. o g/ 10. 1016/j.
gca. 2006. 06. 1559
Man e o J, Thomas R, Holm E, Raa C, Vioque I, Ruiz-Cano as C,
Ga cia-Teno io R, Fo ssell-A onsson E, Isaksson M (2020) Pi
lakes om sou he n Sweden: na u al adioac i i y and elemen a y
cha ac e iza ion. Sci Rep-UK 10:13712. h ps:// doi. o g/ 10. 1038/
s41598- 020- 70521-0
Mau ice YT (1980) U anium in g ani es. In: P oc, wo kshop, Geologi-
cal Su ey o Canada Pape 81-23, ISBN 0 660 11141 1, ol 120.
h ps:// doi. o g/ 10. 1017/ S0016 75680 00255 77
Mille JN, Mille JC (2005) S a is ics and chemome ics o analy ical
chemis y, 5 h edn. Pea son/P en ice Hall, To on o
Pa khu s D, Tho s enson DC, Plumme LN (1980) PHREEQC, a
compu e p og am o geochemical calcula ions. U.S. Geological
Su ey, Den e
Piñe o-Ga cía F, Fe o-Ga cía MA, D ożdżak J, Ruiz-Samblás C
(2012) Explo a o y da a analysis in he s udy o 7Be p esen in
a mosphe ic ae osols. En i on Sci Pollu R 19:3317. h ps:// doi.
o g/ 10. 1007/ s11356- 012- 0849-5
Reddy K, Wang L, Gloss S (1995) Solubili y and mobili y o cop-
pe , zinc and lead in acidic en i onmen s. Plan Soil 171:53–58.
h ps:// doi. o g/ 10. 1007/ BF000 09564
Rollings P, Hend y MJ, Ke ich R (1999) Sequen ial il a ion o
su ace and g ound wa e s om he Rabbi Lake u anium mine,
no he n Saska chewan, Canada. Wa e Qual Res J Can 34:221–
247. h ps:// doi. o g/ 10. 2166/ wq j. 1999. 010
San o imia E, López-Pamo E (2013) The ole o su ace wa e and
mine g oundwa e in he chemical s a i ica ion o an acidic pi
lake (Ibe ian Py i e Bel , Spain). J Hyd ol 490:21–31. h ps:// doi.
o g/ 10. 1016/j. jhyd ol. 2013. 03. 018
S ømman G, Rosseland BO, Skippe ud L, Bu ki bae LM, U albeko
B, Heie LS, Salbu B (2013) U anium ac i i y a io in wa e and
ish om pi lakes in Ku day, Kazakhs an and Tabosha , Tajik-
is an. J En i on Radioac 123:71–81. h ps:// doi. o g/ 10. 1016/j.
jen ad. 2012. 05. 014
Suksi J, Rasilainen K, Ma cos N (2006) U iso opic ac iona ion—
a p ocess cha ac e ising g oundwa e sys ems. In: Me kel BJ,
Hasche-Be ge A (eds) U anium in he en i onmen : mining
impac and consequences. Sp inge , Be lin, pp 683–690. h ps://
doi. o g/ 10. 1007/3- 540- 28367-6_ 69
Thomas R, Man e o J, Ruiz-Cano as C, Holm E, Ga cía-Teno io R,
Fo ssell-A onsson E, Isaksson M (2022) Na u al adioac i i y and
elemen cha ac e iza ion in pi lakes in No he n Sweden. PLoS
ONE. h ps:// doi. o g/ 10. 1371/ jou n al. pone. 02660 02
Vaughan DJ, Lennie AR (1991) The i on sulphide mine als: hei chem-
is y and ole in na u e. Sci P og (1933-) 75:371–388
Wa d JH (1963) Hie a chical g ouping o op imize an objec i e unc-
ion. J Am S a Assoc 58:236–244. h ps:// doi. o g/ 10. 1080/ 01621
459. 1963. 10500 845
Wilson M, Åke blom G (1980) U anium en iched g ani es in Sweden.
SGU, Rappo e och meddelanden n 19. Uppsala, p 30