De ec ion o empo al ends in a mosphe ic deposi ion o ino ganic
ni ogen and sulpha e o o es s in Eu ope
Pe e Waldne
a
,
*
, Aldo Ma che o
b
, Anne Thimonie
a
, Ma ia Schmi
a
, Michela Rogo a
b
,
Oli e G anke
c
, Volke Mues
d
, Ka in Hansen
e
, Gunilla Pihl Ka lsson
, Daniel 
Zlind a
g
,
Nicholas Cla ke
h
, A ne Ve s ae en
i
, Andis Lazdins
j
, Claus Schimming
k
,
Ca men Iacoban
l
, An i-Jussi Lind oos
m
, Elena Vanguelo a
n
, Sue Benham
n
,
Henning Meesenbu g
o
, Manuel Nicolas
p
, Anna Kowalska
q
, Vladisla Apuh in
,
Ulle Napa
s
, Zo a Lachmano 
a
, Fe dinand K is oe el
u
, Albe Bleeke
,
Mo en Inge sle
w
, La s Ves e dal
w
, Juan Molina
x
, Uwe Fische
y
, Wal e Seidling
y
,
Ma hieu Jona d
z
, Philip O'Dea
aa
, James Johnson
ab
,
ac
, Richa d Fische
ad
, Ma in Lo enz
ad
a
WSL, Swiss Fede al Ins i u e o Fo es , Snow and Landscape Resea ch, Zü che s asse 111, CH-8903 Bi mensdo , Swi ze land
b
ISE-CNR, Ins i u e o Ecosys em S udy, La go Tonolli 50, I-28922 Ve bania Pallanza, I aly
c
Digsyland, Ins i u ü Digi ale Sys emanalyse &Landscha sdiagnose, D-24975 Husby, Ge many
d
UHH, Uni e si y Hambu g, Cen e o Wood Sciences, Wo ld o es y, Leuschne s asse 91, D-21031 Hambu g, Ge many
e
IVL Swedish En i onmen al Resea ch Ins i u e, SE-100 31 S ockholm, Sweden
IVL Swedish En i onmen al Resea ch Ins i u e, SE-400 14 G€
o ebo g, Sweden
g
Slo enian Fo es y Ins i u e, SI-1000 Ljubljana, Slo enia
h
No wegian Fo es and Landscape Ins i u e, P.O. Box 115, N-1431 Ås, No way
i
Resea ch Ins i u e o Na u e and Fo es , Klinieks aa 25, B-1070 B ussels, Belgium
j
SILAVA, La ian S a e Fo es Resea ch Ins i u e, Riga S ee 111, LV-2169 Salaspils, La ia
k
Cen e o Ecology, Uni e si y o Kiel, Olshausens asse 40, D-24098 Kiel, Ge many
l
Expe imen S a ion o Sp uce Sil icul u e, Calea Buco inei 73, RO-5950 Campulung Moldo enesc, Sucea a, Romania
m
METLA, Finnish Fo es Resea ch Ins i u e, PL 18, FI-01301 Van aa, Finland
n
Fo es Resea ch, Alice Hol Lodge, W ecclesham, Fa nham, Su ey GU10 4LH, Uni ed Kingdom
o
NW-FVA, No dwes deu sche Fo s liche Ve suchsans al , G €
a zels asse 2, D-37079 G€
o ingen, Ge many
p
ONF, O fice Na ional des Fo ^
e s, D
epa emen Reche che e D
e eloppemen , B^
a imen B, Boule a d de Cons ance, F-77300 Fon ainebleau, F ance
q
FRI, Fo es Resea ch Ins i u e, Sekocin S a y, PL-05-090 Raszyn, Poland
Es onian En i onmen Agency, EE-33 Tallinn, Es onia
s
Uni e si y o Ta u, Vanemuise 46, EE-51013 Ta u, Es onia
FGMRI, Fo es y and Game Managemen Resea ch Ins i u e, CZ-156 04 P ague 5, Zb asla , Czech Republic
u
BFW, Fede al Resea ch Cen e o Fo es s, A-1131 Vienna, Aus ia
Ene gy Resea ch Cen e o he Ne he lands, 1755 ZG Pe en, Ne he lands
w
Depa men o Geosciences and Na u al Resou ce Managemen , Uni e si y o Copenhagen, Roligheds ej 23, DK-1958 F ede iksbe g C, Denma k
x
TECMENA, Cla a del Rey 22, ES-28000 Mad id, Spain
y
Thünen-Ins i u e o Fo es Ecosys ems, D-16225 Ebe swalde, Ge many
z
UCL-ELI, Uni e si 
e Ca holique de Lou ain, Ea h and Li e Ins i u e, C oix du Sud 2, B-1348 Lou ain-la-Neu e, Belgium
aa
Coill e Labo a o ies, Chu ch Road, New ownmoun kennedy, Wicklow, I eland
ab
T en Uni e si y, 1600 Wes Bank D i e Pe e bo ough, On a io K9J 7B8, Canada
ac
UCD School o Ag icul u e and Food Science, Uni e si y College Dublin, Belfield, Dublin 4, I eland
ad
Thünen Ins i u e o In e na ional Fo es y and Fo es Economics, Leuschne s asse 91, D-21031 Hambu g, Ge many
highligh s
Minimum de ec able end slopes depend on leng h o ime se ies.
Tempo al a iabili y o deposi ion was simila ac oss si es o many subs ances.
Despi e highe noise, mon hly da a we e be e han annual da a o end analysis.
Ni ogen and sulpha e deposi ion dec eased by 2% and 6% pe yea , espec i ely.
*Co esponding au ho .
E-mail add ess: pe e [email p o ec ed] (P. Waldne ).
Con en s lis s a ailable a ScienceDi ec
A mosphe ic En i onmen
jou nal homepage: www.else ie .com/loca e/a mosen
h p://dx.doi.o g/10.1016/j.a mosen .2014.06.054
1352-2310/©2014 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/3.0/).
A mosphe ic En i onmen 95 (2014) 363e374
a icle in o
A icle his o y:
Recei ed 13 Decembe 2013
Recei ed in e ised o m
17 June 2014
Accep ed 26 June 2014
A ailable online 26 June 2014
Keywo ds:
Th ough all
Bulk deposi ion
Ino ganic ni ogen deposi ion
Sulpha e deposi ion
Time end analyses
ICP Fo es s
abs ac
A mosphe ic deposi ion o o es s has been moni o ed wi hin he In e na ional Coope a i e P og amme
on Assessmen and Moni o ing o Ai Pollu ion E ec s on Fo es s (ICP Fo es s) wi h sampling and ana-
lyses o bulk p ecipi a ion and h ough all a se e al hund ed o es ed plo s o mo e han 15 yea s. The
cu en deposi ion o ino ganic ni ogen (ni a e and ammonium) and sulpha e is highes in cen al
Eu ope as well as in some sou he n egions. We compa ed linea eg ession and ManneKendall end
analysis echniques o en used o de ec empo al ends in a mosphe ic deposi ion. The choice o
me hod influenced he numbe o significan ends. De ec ion o ends was mo e powe ul using
mon hly da a compa ed o annual da a. The slope o a end needed o exceed a ce ain minimum in
o de o be de ec ed despi e he sho - e m a iabili y o deposi ion. This a iabili y could o a la ge
ex en be explained by me eo ological p ocesses, and he minimum slope o de ec able ends was hus
simila ac oss si es and many ions. The o e all dec easing ends o ino ganic ni ogen and sulpha e in
he decade o 2010 we e abou 2% and 6%, espec i ely. Time se ies o abou 10 and 6 yea s we e equi ed
o de ec significan ends in ino ganic ni ogen and sulpha e on a single plo . The s onges dec easing
ends we e obse ed in wes e n cen al Eu ope in egions wi h ela i ely high deposi ion fluxes,
whe eas s able o sligh ly inc easing deposi ion du ing he las 5 yea s was ound eas o he Alpine
egion as well as in no he n Eu ope. Pas educ ions in an h opogenic emissions o bo h acidi ying and
eu ophying compounds can be confi med due o he a ailabili y o long- e m da a se ies bu u he
educ ions a e equi ed o educe deposi ion o Eu opean o es s o le els below which significan
ha m ul e ec s do no occu acco ding o p esen knowledge.
©2014 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license
(h p://c ea i ecommons.o g/licenses/by/3.0/).
1. In oduc ion
Fo es ecosys ems ha e been exposed o inc eased a mosphe ic
deposi ion o sulphu (S) in he o m o sulpha e (SO
4
2
) and ino -
ganic ni ogen (N) since he 1950s, esul ing om an h opogenic
emissions o sulphu dioxide (SO
2
), ni ogen oxides (NO
x
) and
ammonia (NH
3
). Deposi ion o hese compounds is a majo d i e
o a ious changes in o es ecosys ems. I may al e nu ien
limi a ions and lead o inc eased o es g ow h and ca bon (C)
seques a ion (e.g. de V ies e al., 2008; Solbe g e al., 2009),
accele a e soil acidifica ion (e.g. Ul ich e al., 1980) and eu ophi-
ca ion e ec s (e.g. Abe e al.,1998) as well as mobilising aluminium
in soil solu ion o le els ha a e oxic o oo s (C onan e al., 1989).
Eu ophica ion e ec s include loss o nu ien s by leaching,
ele a ed ni a e (NO
3

) le els in pe cola ion and uno wa e (Dise
e al., 2009), nu ien imbalances in ees, and al e ed suscep ibili y
o pes s and diseases (Flückige and B aun, 1999).
De e mina ion o empo al ends o a mosphe ic deposi ion o
S and N compounds o o es s is he e o e o conside able in e es .
Measu es we e implemen ed o educe he emissions o S and N
compounds du ing he las h ee decades (Reis e al., 2012).
Deposi ion assessmen s in long ime se ies a e equi ed (i) o
moni o he success o hese measu es in educing deposi ion and
(ii) o in es iga e he impac o deposi ion on he long- e m s a-
bili y o o es and i s ecosys em se ices a selec ed in ensi ely
moni o ed si es (Paole i e al., 2010).
Fo his pu pose, empo al end analyses based on bulk p ecip-
i a ion and h ough all measu emen s pe o med unde he ame-
wo k o he In e na ional Coope a i e P og amme on Assessmen
and Moni o ing o Ai Pollu ion E ec s on Fo es s (ICP Fo es s) a e
egula ly ca ied ou a he in ensi ely moni o ed si es o he ICP
Fo es s Le el II ne wo k and published on pan-Eu opean le el (e.g.
Lo enz and G anke, 2009; G anke and Mues, 2010; Waldne e al.,
2012). Fu he end analyses o pa s o he ICP Fo es s deposi ion
da a and o he da a ha e been ca ied ou a he na ional, egional
and Eu opean le els using a ious me hods (Meesenbu ge al.,1995;
K aalen e al., 2002; H
uno 
a e al., 2004; Rogo a e al., 2006; Fage li
and Aas, 2008; Vanguelo a e al., 2010; G a Panna ie e al., 2011;
Oulehle e al., 2011; Pihl Ka lsson e al., 2011; S aelens e al., 2012;
Ve s ae en e al., 2012; Johnson e al., 2013; Ma che o e al., 2013).
Howe e , he commonly epo ed absolu e end slopes and
pe cen age o s a is ically significan ends a y and seem o be
pa ly con adic ing. This may be due o he a ia ion o me hods
used in hese s udies, e.g. di e en end analysis echniques, a -
ia ions in leng h and empo al esolu ion o ime se ies, spa ial
a ia ion o emission ime ends o o he ac o s influencing
deposi ion.
The main aims o his s udy we e o:
▪de e mine and explain he minimum de ec able end on a
single plo wi h deposi ion measu emen s ca ied ou acco ding
o he ICP Fo es s manual
▪in es iga e he influence o end analysis echnique, ime se ies
leng h and empo al esolu ion on he de ec ion o s a is ically
significan ends
▪assess bulk deposi ion (BD) and h ough all deposi ion (TF) o
SO
4
2
, ni a e (NO
3

) and ammonium (NH
4
þ
) and hei ends
ac oss Eu ope a ICP Fo es s si es
2. Me hods
2.1. Sampling and chemical analyses
Con inuous sampling o below canopy h ough all and open
field bulk deposi ion is ca ied ou on ICP Fo es s Le el II o es
moni o ing plo s and a nea by open field si es, espec i ely. The
me hods used in he a ious coun ies (F ance: Ul ich and Lanie ,
1993; No way: K aalen e al., 2002; Mo a e al., 2002; I aly:
Mosello e al., 2002; Swi ze land: Thimonie e al., 2005; Finland:
Lind oos e al., 2006; Denma k: Gunde sen e al., 2009; Czech Re-
public: Boh
a
co 
a e al., 2010; La ia: Lazdin¸ 
s e al., 2010; Uni ed
Kingdom: Vanguelo a e al., 2010; Swedish Th ough all Moni o ing
Ne wo k (SWETHRO): Pihl Ka lsson e al., 2011; Belgium:
Ve s ae en e al., 2012) ollow he ICP Fo es s manual (ea lie
e sions and ICP Fo es s, 2010).
In gene al, collec o s (3e20 eplica es) a e placed in he o es
based on a andom o fixed sys ema ic design in o de o co e he
spa ial a ia ion (Swi ze land: Thimonie , 1998; Uni ed Kingdom:
Hous on e al., 2002; Belgium: S aelens e al., 2006). Samples a e
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374364
collec ed a leas mon hly ( ypically o nigh ly o weekly), fil e ed,
and s o ed below 4

C be o e chemical analyses a e pe o med o
de e mine he concen a ions o SO
4
2
,NO
3

, and NH
4
þ
. The labo a-
o y esul s a e checked o in e nal consis ency based on he
conduc i i y, he ion balance, he concen a ion o o al N and he
sodium o chlo ide (Na/Cl) a io, and analyses a e epea ed i sus-
picious alues occu (Mosello e al., 2005, 2008; ICP Fo es s, 2010).
The quali y assu ance and con ol (QA/QC) p ocedu es u he
include he use o con ol cha s o in e nal e e ence ma e ial o
check long- e m compa abili y wi hin na ional labo a o ies, as well
as pa icipa ion in pe iodic labo a o y ing es s (e.g. Ma che o
e al., 2009) and field in e -compa isons (D aaije s e al., 2001;

Zlind a e al., 2011a) o check he in e na ional compa abili y.
Da a we e epo ed annually o he pan-Eu opean da a cen e,
checked o consis ency and s o ed in he p og amme da abase.
2.2. Da a p ocessing
Da a o he pe iod om 1999 o 2010 we e used in his analysis.
P ecipi a ion and h ough all da a sampled du ing mo e han 330
days pe yea , and wi h concen a ion alues o mo e han 300
days pe yea we e included. Sampling pe iods wi h mean p ecip-
i a ion below 0.1 mm day
1
we e coun ed e en i no chemical
analyses could be pe o med.
Da a om each sampling pe iod we e in e pola ed o egula
mon hly and annual da a by: (i) spli ing each sampling pe iod
o e lapping wo consecu i e mon hs by dis ibu ing p ecipi a ion
quan i y in p opo ion o he du a ion o he new sampling pe iods;
(ii) se ing deposi ion ¼0 o pe iods wi h missing concen a ions
and mean p ecipi a ion <0.1 mm day
1
; (iii) calcula ing TF and BD
(Q$c$10
2
,inkgha
1
) by mul iplica ion o he p ecipi a ion
quan i y (Q,inLm
2
), he concen a ions (c,inmgL
1
) and he
uni y con e sion ac o 10
2
; (i ) summing up o fluxes by mon h
and yea , espec i ely.
Mean annual fluxes o SO
4
2
eS and he ino ganic N species
NO
3

eN and NH
4
þ
eN o 2010 we e calcula ed o 286, 282, and
287 TF plo s and 266, 265, and 268 BD plo s, espec i ely.
2.3. T end analyses
We analysed he empo al ends o indi idual ime se ies o
se s o plo s wi h con inuous measu emen s om 2007 o 2010 (4
yea s), om 2005 o 2010 (6 yea s), om 2003 o 2010 (8 yea s),
om 2001 o 2010 (10 yea s) as well as o 1999 o 2010 (12 yea s).
We checked ha ime se ies we e no mally dis ibu ed and showed
a seasonal pa e n (see Annex).
T end analyses we e ca ied ou using (i) linea eg ession (LR),
(ii) ManneKendall (MK) es (Mann, 1945; Helsel and Hi sch, 2002)
using annual deposi ion fluxes, (iii) Seasonal ManneKendall (SMK)
(Hi sch e al., 1982; Hi sch and Slack, 1984), and (i ) Pa ial Man-
neKendall (PMK) es s (Libiselle and G im all, 2002) using
mon hly deposi ion da a. The PMK es includes es ing he influ-
ence o a co- a iable, and we chose p ecipi a ion quan i y o ha .
Linea eg ession and Kendall es s we e pe o med using he ‘ k ’
package (Ma che o, 2013) in he R so wa e (R De elopmen Co e
Team, 2009). Fo he Kendall es s (MK, SMK, PMK), end slopes b
(kg ha
1
y
2
) we e es ima ed ollowing Sen (1968).
Fo each ime se ies, we calcula ed a ela i e slope slope (y
1
),
as an es ima ed mean ela i e change pe yea , wi h
slope ¼b=meanðyÞ;(1)
whe e b(kg ha
1
y
2
) is he es ima o o he absolu e end
esul ing om he end analyses and mean (y) (kg ha
1
y
1
) he
mean alue o he ime se ies.
2.4. Tempo al a iabili y (backg ound signal)
The empo al a iabili y o he o iginal da a (CV0), da a a e
emo ing es ima ed empo al end (CV1), and da a a e emo ing
empo al end andseasonali y (CV2) we e de e mined o each ime
se ies and a e aged o each pa ame e (see equa ion (3) in Annex).
2.5. Minimum de ec able ends
Minimum de ec able ends slope
emp
min
we e de i ed empi ically
om he p- alues and he slope esul s o he indi idual end an-
alyses o each combina ion o pa ame e , ime se ies leng h and
end analysis echnique. The slope
emp
min
alue abo e which he ma-
jo i y o es s iden i y a end as significan , wi h p<0.05 (a signi -
icance le el 95%), was de e mined by fi ing a Gauss shaped unc ion
h ough he band o p- o slope alues o he es esul s (see Annex).
Secondly, minimum de ec able ends we e modelled based on
he empo al a iabili y o he o e all da ase wi h
slopemod
min ¼c72CV
nyea s
Tc i n
2
ffiffiffi
n
2
q(2)
whe e n
yea s
is he du a ion o ime se ies in yea s, nis he numbe
o obse a ions (n¼n
yea s
o annual and n¼12$n
yea s
o mon hly
da a), CV he coe ficien o a ia ion o he empo al a iabili y (see
Table 1) and T
c i
he es s a is ic o he T- es (e.g. T
c i
¼2.45 o n/
2¼6, T
c i
¼2.23 o n/2 ¼10, T
c i
¼1.98 o n/2 ¼100) and c
7
an
adjus men pa ame e (see Resul s and Annex).
3. Resul s and discussion
3.1. Cu en deposi ion
Clea egional a ia ion was obse ed in he deposi ions. High-
es SO
4
2
BD (no shown) and TF deposi ion was eco ded in o es
plo s in no he n cen al Eu ope and Poland eaching up o he
sou he n Bal ic and he cen al Hunga ian a ea, and in some
Medi e anean egions in Spain, F ance, sou he n I aly and G eece
(Fig. 1). Highes ino ganic N BD (no shown) and TF deposi ion was
eco ded in no he n cen al Eu ope, as o SO
4
2
, bu also in
sou he n Ge many and he Swiss Pla eau and u he o he wes , in
no he n F ance, he cen al UK and I eland. The egions bo de ing
he Alps in he sou h and some si es in Spain and in sou he n F ance
also showed ela i ely high N deposi ion.
Conside able pa s o he egionally highe ino ganic N and SO
4
2
deposi ion a e a ibu able o an h opogenic emission o NO
x
,SO
2
and NH
3
(Reis e al., 2012). O he con ibu ions a e o na u al o igin.
Table 1
Tempo al a iabili y o annual and mon hly deposi ion o NH
4
þ
eN, NO
3

eN, SO
4
2
eS
(kg ha
1
y
1
) and p ecipi a ion quan i y Q(L m
2
y
1
) ime se ies om plo s wi h
con inuous da a om 2001 o 2010 (10 yea s).
Flux Va iable Annual Mon hly Mon hly
CV1 CV1 CV2
BD NH
4
þ
eN 0.26 (±0.11) 0.75 (±0.30) 0.67 (±0.28)
NO
3

eN 0.18 (±0.07) 0.48 (±0.11) 0.44 (±0.11)
SO
4
2
eS 0.19 (±0.07) 0.48 (±0.08) 0.44 (±0.07)
Q0.18 (±0.05) 0.48 (±0.08) 0.49 (±0.07)
TF NH
4
þ
eN 0.30 (±0.20) 0.98 (±0.93) 0.86 (±0.87)
NO
3

eN 0.20 (±0.09) 0.63 (±0.31) 0.53 (±0.26)
SO
4
2
eS 0.16 (±0.07) 0.52 (±0.15) 0.46 (±0.13)
Q0.17 (±0.07) 0.52 (±0.15) 0.50 (±0.10)
CV1: coe ficien o a ia ion a e co ec ion o linea end, CV2: coe ficien o
a ia ion a e co ec ion o linea end and seasonali y.
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374 365
Fo example, pa s o he high SO
4
2
deposi ion along he coas
occu oge he wi h high Cl

deposi ion (e.g. a some No wegian
coas al si es), which is ypical o SO
4
2
o igina ing om sea sal
(G anke and Mues, 2010).
The measu emen s suppo he findings o modelling and
mapping app oaches (e.g. Posch e al., 2012) acco ding o which
a mosphe ic deposi ion o SO
4
2
and N compounds s ill exceeds
c i ical loads in pa s o Eu ope. C i ical loads apply o o al depo-
si ion (TD), i.e. he sum o we and d y deposi ion. In o es s, TD o N
is ypically a ac o o 1e2 highe han TF, due o up ake by plan
issue and h ough s oma a in he canopy (D aaije s and E isman,
1995). Fo SO
4
2
, TD is gene ally assumed o be equal o TF
(D aaije s and E isman, 1995). Fo N, he anges om 5 o 15 and
om 10 o 20 kg ha
1
y
1
ha e been p oposed as empi ical c i ical
loads o coni e ous and b oadlea ed deciduous woodland,
espec i ely (Bobbink and He elingh, 2011).
3.2. T end analyses and de i a ion o minimum de ec able ends
The slope es ima es esul ing om he end analysis echniques
LR, MK, SMK and PMK ag eed well. The ag eemen be ween end
echniques inc eased wi h leng h o he ime se ies, and wi h
inc easing slope (Fig. 2, le -hand side). The e was less ag eemen
be ween end analysis echniques in e ms o iden i ying a end as
being significan o no (Fig. 2, igh hand side).
The minimum de ec able end slope
emp
min
dec eased wi h
inc easing leng h o he ime se ies and was ypically smalle o
me hods applied o mon hly da a (SMK, PMK) compa ed o es s
applied o annual da a (MK, LR), as shown in Fig. 3 o SO
4
2
,NO
3

and NH
4
þ
in TF.
3.3. Tempo al a iabili y
The empo al a iabili y o deposi ion a ied li le om plo o
plo o om ion o ion (Table 1). The empo al a iabili y was on
a e age abou 20e60% highe o mon hly da a han o he annual
sums. The co ec ions o linea ends, and o seasonali y, educed
he empo al a ia ion on a e age by abou 5e10%.
The empo al a iabili y was qui e simila o all ions and no
much highe han ha o p ecipi a ion quan i y Q(L m
2
y
1
),
which migh be su p ising a fi s glance (Table 1). Ande sson e al.
(2006) used a chemis y anspo model (CTM) and es ima ed ha
he a e age Eu opean land-a ea in e -annual a iabili y o SO
4
2
and
ino ganic N deposi ion, due o me eo ological a iabili y, anged
om 11 o 14% o TD and o abou 20% o we deposi ion. K yza
e al. (2012) confi med ha me eo ology can lead o an in e -
annual a ia ion o 20% and s a ed ha p ecipi a ion quan i y is
gene ally he mo e impo an ac o , excep o egions such as he
UK, whe e he ci cula ion pa e n migh become mo e impo an .
The e o e, i is likely ha mos o he empo al a iabili y is
explained by he a iabili y o ai ci cula ion, i.e. he sou ce egion
and pollu ion le el o he ai masses, and he p ecipi a ion, i.e. he
sca enging o he gaseous and pa icula e compounds anspo ed
in he a mosphe e. Fo mos compounds, he empo al signals o
he emissions in a egion, e.g. om ossil uel bu ning, a e p obably
much smoo he han hose o deposi ion. Howe e , NH
4
þ
shows a
Fig. 1. Mean annual SO
4
2
eS (kg S ha
1
y
1
) and ino ganic ni ogen (NH
4
þ
eNþNO
3

eN) (kg N ha
1
y
1
) h ough all deposi ion in he yea 2010.
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374366
sligh ly highe o e all a iabili y han NO
3

and SO
4
2
which may be
caused by spa ially and empo ally mo e a iable emission sou ces.
The emissions om ag icul u al land in he o m o NH
3
a e a majo
sou ce o NH
4
þ
in p ecipi a ion and h ough all, and he emissions
a e hemsel es s ongly influenced by local wea he condi ions
(Wichink K ui e al., 2012).
The empo al a iabili y ound he e is likely o be alid o o he
subs ances anspo ed o e simila pa hways. I seems ha he
me hod o es ima e slope
min
p esen ed he e is gene ally applicable
o mos o he majo compounds in BD and TF, e en when using
jus he empo al a iabili y alues shown in Table 1.
3.4. Es ima ed minimum de ec able ends
The minimum de ec able end slope
emp
min
de e mined empi i-
cally om he end es esul s can o a la ge ex en be explained
Fig. 2. Rela i e slope ( slope) (le -hand side) and p- alue ( igh -hand side) o MK, SMK and PMK e sus LR o SO
4
2
eS TF deposi ion om 2001 o 2010 (n
yea s
¼10), and 2005 o
2010 (n
yea s
¼6).
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374 367

by he mean sho - e m empo al a iabili y. The slope
emp
min
(equa-
ion (5), Annex) co ela ed well o he slope
mod
min
(equa ion (2))
es ima ed om CV1 and CV2 alues in Table 3. The PMK es
showed he highes sca e . The co- a iable no conside ed in
equa ion (2) bu used in his PMK es may be a eason o his
highe sca e . Fo he pa ame e c
7
in equa ion (2)
ð slope
emp
min
= slope
mod
min
Þwe ound alues be ween 1 and 2.5 ha
ha e li le dependence on he end echnique applied, he
pa ame e o he ime se ies leng h (Table 3).
Mon hly da a in ol e mo e da a poin s han annual da a, which
seems o be a ou able o de ec ing ends despi e (i) he
unce ain y o mon hly da a in e pola ion and (ii) hei gene ally
highe empo al a iabili y.
3.5. Compa ison o minimum de ec able end o sou ces o
unce ain y o he measu emen s
This s udy sugges s ha he da a quali y objec i e o ‘de ec a
change o 30% in 10 yea s’which is defined in he ICP Fo es s
manual (ICP Fo es s, 2010) seems ealis ic. I has o be men ioned
ha his s udy only in es iga ed he unce ain y ela ed o he
s a is ical me hods. Howe e , unce ain ies ela ed o he s eps
p io o he end analyses (Thimonie , 1998; Hous on e al., 2002;
Bleeke e al., 2003; E isman e al., 2003; S aelens e al., 2006;
Ma che o e al., 2011; 
Zlind a e al., 2011b) we e on a e age
lowe in magni ude han he unce ain y esul ing om he em-
po al a iabili y o he deposi ion.
3.6. Deposi ion ends
The esul s on minimum de ec able ends a e eflec ed in he
s onge ag eemen o slopes be ween end echniques and he
smalle sca e ing o slope among plo s o longe ime se ies and
o PMK and SMK compa ed o LR and MK (Table 2).
The ends (Table 2) ag ee well wi h he findings o ea lie
s udies (Table 3), which is mo e ob ious when compa ing slope
alues. The low pe cen ages o significan ends ound in se e al
s udies a e o a la ge ex en due o he expec ed ends being low
compa ed o minimum de ec able end. No mean slope is gi en in
Table 3 because he slope alues o non-significan changes a e
o en omi ed in li e a u e.
Be ween he peak emission in he 1980s and he u n o he mil-
lennium as well as o he decade a ound he millennium, slope alues
o SO
4
2
we e ypically be ween 5% and 10% in cen al Eu ope and
be ween 12 and þ3% in no he n and wes e n Eu ope (Table 3). The
pe cen age o plo s wi h significan ends was especially high in
cen al Eu ope. In compa ison, he slope alueso significan and non-
significan changes o N deposi ion we e lowe , ypically be ween þ1
and 5%, and he pe cen age o plo s wi h significan ends was also
lowe , especially when he ime se ies we e sho .
Fo he 10 yea pe iod, ypical slope alues o N compounds
we e a ound 2% pe yea (Table 2). Hence, ypically abou 10 yea s
o da a we e equi ed o de ec such a end on a plo wi h s a is ical
significance wi h PMK (Fig. 4). Fo SO
4
2
wi h ypical slope alues o
4e6%, he co esponding equi emen was abou 6 yea s o da a.
The s onges dec easing ends du ing he 10 yea pe iod we e
ound in no he n cen al Eu ope om Belgium and he
Ne he lands o Ge many and o N compounds he egion o
s onges ends (0.2 o 0.15 kg ha
1
y
2
) ex ended u he o
Swi ze land, F ance, I aly, Czech Republic, Slo akia and Denma k.
Si es wi h non-significan changes in deposi ion we e dis ib-
u ed all o e Eu ope (Fig. 5). In he 6 yea pe iod, s able o sligh ly
inc easing SO
4
2
deposi ion was epo ed mainly o plo s in eas e n
cen al Eu ope and o N deposi ion o sou he n Ge many,
Swi ze land, Aus ia, I aly and he F anco-Belgian bo de egion as
well as in no he n Eu ope.
The gene ally dec easing ends o SO
4
2
and ino ganic N
deposi ion coincide and can be explained by he emission e-
duc ions achie ed be ween 1990 and 2001 (Table 1,Reis e al.,
2012). Fage li and Aas (2008) compa ed NO
3

and NH
4
þ
concen a-
ions in we p ecipi a ion modelled by he Eu opean Moni o ing
and E alua ion P og amme (EMEP) based on he emission in-
en o ies wi h measu emen s o he pe iod om 1980 o la e o
2003 a a ious si es in Eu ope. They s a ed ha mos o he e-
duc ions ook place in he yea s be ween 1985 and 1995.
Ve s ae en e al. (2012) poin ed ou ha he e ec o echnical
Fig. 3. Minimum de ec able ends de i ed om he p- alue o slope plo s o end
analyses wi h LR, and MK o annual, SMK and PMK o mon hly SO
4
2
,NO
3

and NH
4
þ
TF
deposi ion ime se ies wi h con inuous da a om 2007 o 2010 (4 yea s), om 2005 o
2010 (6 yea s), om 2003 o 2010 (8 yea s), om 2001 o 2010 (10 yea s), and om
1999 o 2010 (12 yea s).
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374368
measu es aken by indus y, a fic and ag icul u e in he 1980s and
1990s, which esul ed in a clea dec ease o SO
4
2
and NH
4
þ
, has
become ma ginal in ecen yea s, while inc easing a fic coun e -
ac s he e ec o s ic e emission no ms o ehicles.
O he easons o changes in deposi ion o o es a eas a e
changes in he ee s and s uc u e, such as he educ ion o he
numbe o ees due o ba k bee le a acks as epo ed o a Czech
o es (Boh
a
co 
a e al., 2010), o es age, o high le els o ni a e in
insec ass alling om he canopy as epo ed o si es in he UK
(Pi man e al., 2010).
Fo ino ganic N especially, he dec easing ends seem oo sligh
o a oid exceedance o he c i ical loads o acidifica ion and
eu ophica ion in di e en pa s o Eu opean o es s in he nea
u u e (Reis e al., 2012). Fu he educ ion o N emissions is needed
o p e en ai pollu ion e ec s on o es s.
4. Conclusions
The selec ion o he end analysis echnique had an e ec on
end de ec ion. The e was a s ong ag eemen be ween es ima ed
Table 2
Rela i e end ( slope in % y
1
) and s anda d de ia ion o slope and pe cen age o plo s wi h significan posi i e ends (þ), significan nega i e ends () o non-significan
(n.s.) changes o end analyses o NH
4
þ
,NO
3

,SO
4
2
bulk (BD) and h ough all (TF) deposi ion plo s wi h con inuous da a om 2005 o 2010 (6 yea s) and 2001 o 2010 (10
yea s).
Ion Pe iod Flux n slope n.s. þ slope n.s. þ
LR MK
SO
4
2
2001e2010 BD 78 4.1 (±3.5) 49 50 1 4.1 (±3.4) 42 56 1
TF 105 4.9 (±3.9) 65 35 0 4.9 (±3.2) 60 40 0
2005e2010 BD 143 5.2 (±6.1) 21 78 1 5.1 (±6.9) 9 91 0
TF 171 6.3 (±7.0) 33 67 0 6.6 (±6.1) 16 84 0
NO
3

2001e2010 BD 78 1.6 (±2.8) 22 77 1 1.7 (±2.6) 17 81 3
TF 105 1.5 (±3.6) 20 75 5 1.5 (±3.5) 20 78 2
2005e2010 BD 143 0.9 (±5.5) 6 93 1 0.9 (±5.7) 3 97 0
TF 171 3.1 (±6.3) 8 92 0 2.8 (±6.3) 6 94 0
NH
4
þ
2001e2010 BD 78 0.7 (±4.8) 15 82 3 0.9 (±4.2) 8 90 3
TF 105 1.6 (±4.8) 15 83 2 1.8 (±4.0) 14 84 2
2005e2010 BD 143 2.8 (±9.9) 4 93 3 2.6 (±8.6) 2 97 1
TF 171 4.9 (±9.5) 7 92 1 4.2 (±8.0) 2 97 1
SMK PMK
SO
4
2
2001e2010 BD 78 3.9 (±2.8) 79 19 1 3.9 (±2.8) 62 38 0
TF 105 4.5 (±2.6) 91 9 0 4.5 (±2.6) 71 29 0
2005e2010 BD 143 4.4 (±4.8) 60 38 1 4.4 (±4.8) 45 54 1
TF 171 5.5 (±5.0) 63 37 0 5.5 (±5.0) 46 54 0
NO
3

2001e2010 BD 78 1.4 (±2.1) 37 59 4 1.4 (±2.1) 24 73 3
TF 105 1.4 (±2.8) 35 61 4 1.4 (±2.8) 24 70 6
2005e2010 BD 143 0.1 (±4.9) 6 92 1 0.1 (±4.9) 6 90 3
TF 171 2.6 (±4.5) 23 76 1 2.6 (±4.5) 13 85 2
NH
4
þ
2001e2010 BD 78 0.9 (±2.4) 26 67 8 0.9 (±2.4) 22 72 6
TF 105 1.3 (±2.5) 31 65 4 1.3 (±2.5) 33 62 5
2005e2010 BD 143 1.8 (±5.3) 22 76 2 1.8 (±5.3) 15 82 3
TF 171 3.2 (±4.5) 30 69 1 3.2 (±4.5) 23 75 2
Legend: LR ¼Linea eg ession, MK ¼ManneKendall, SMK ¼Seasonal ManneKendall, PMK ¼Pa ial ManneKendall.
Table 3
Ranges o ela i e ends o S and N deposi ion (maxjmin slope in % y
1
) in Eu ope and pe cen age o plo s wi h significan ends ound by o he s udies.
Re e ence Region Pe iod Me h N * S N
Meesenbu g e al. (1995) NW-Ge many 1981e1994 LR 4/7 1 SO
4
2
:BD5j7 (100),
TF 5j9 (100)
NO
3

:BD0j3 (100), TF 1j5 (100)
Rogo a e al. (2006) Alps 1985e2002 SMK 7 SO
4
2
: BD (100) NO
3

: BD (29), NH
4
þ
: BD (86)
1990e2002 SMK 3/20 2 SO
4
2
:BD3j6 (95) NO
3

:BD1j2 (35), NH
4
þ
:BD2j5 (50)
S aelens e al. (2012) Flande s 2002e2010 K 9 SO
4
2
:TD2j14 (100) N: TD 0j6 (78)
Pihl Ka lsson e al. (2011) Sweden 1996e1999
2005e2008
a io, MK 14/52 3 SO
4
2
:BD8j4 (57),
TF 7j2 (90)
NO
3

: BD (0), NH
4
þ
: BD (11), N: BD (14)
K aalen e al. (2002) No way 1986e1997 SMK 13 SO
4
2
:BD3j12 (46),
TF 3j13 (54)
Vanguelo a e al. (2010) UK 1995e2006 SMK 10 SO
4
2
: BD (40), TF (80) NO
3

: BD (40), TF (30),
NH
4
þ
: BD (20),TF (0)
G a Panna ie e al. (2011) Swi ze -land 1994e2007 SMK, PMK 9 SO
4
2
:TF2j7 (100) N: TF (11)
Ma che o e al. (2013) I aly 1998e2010 SMK 9 SO
4
2
:BD5j12 (89) NO
3

:BD11j1 (77), NH
4
þ
:BD12j1 (66)
Ve s ae en e al. (2012) Flande s 1994e2010 SMK 5 SO
4
2
:TF5j6 (100) NO
3

:TF1j2 (60), NH
4
þ
:TF3j5 (100)
Johnson e al. (2013) I eland 1991e2010
(e2003)
SMK, PMK 2 4 SO
4
2
:BD3j4,
TF 4j12 (100)
NO
3

: BD (50), TF (0), NH
4
þ
: BD (0) TF (0)
Huno a e al. (2004) Czech
Republic (CZ)
1985e2000 mod mod 5 SO
4
2
: EM:6NO
3

:TD3, NH
4
þ
:TD3
1985e2000 mod mod 5 SO
4
2
:TD10 NO
3

:TD2, NH
4
þ
:TD3
Oulehle e al. (2011) Nace in, CZ 1995e1998
2004e2009
a io 1 SO
4
2
:BD7, TF 10 N: BD 1,TF 2
Fage li and Aas (2008) Eu ope 1980e2003 EMEP mod þsi es NO
3

and NH
4
þ
:WD1j3 (50)
Legend: Me h ¼ end analysis echnique: LR ¼Linea eg ession, SMK ¼Seasonal ManneKendall, K ¼Kendall, PMK ¼Pa ial ManneKendall, mod ¼Mapping model esul s,
a io ¼compa ison o means o he wo pe iods; n: numbe o si es; *: 1) n¼4 o BD and n¼7 o TF, 2) n¼3 o slope anges and n¼20 o pe cen age o plo s wi h
significan ends, 3) n¼52 o TF and n¼14 o BD, 4) gene ally 1991 o 2010, bu 1991e2003 o TF a one si e, 5) modelled; S, N: BD ¼bulk deposi ion, WD ¼we deposi ion,
TF ¼ h ough all, TD ¼ o al deposi ion, EM ¼emissions.
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374 369
end slopes om he di e en echniques, bu SMK and PMK es s
applied o mon hly da a ended o de ec smalle ends wi h s a-
is ical significance han LR o simple MK echniques applied o
annual da a and hese es s a e he e o e ecommended o end
analysis.
A consis en ela ionship be ween he slope and p- alue o he
end es s was ob ious o a gi en leng h o ime se ies. The choice
o he end analysis echnique, he in es iga ed fluxes and he
specific elemen o ion had less influence on he minimum
de ec able end slope slope
min
. I seems likely ha he minimum
de ec able end slope
min
can be de i ed om he mean empo al
a iabili y caused mainly by me eo ological phenomena.
Fo ime se ies wi h a leng h o 10 yea s, he slope
min
o ino -
ganic N compounds and SO
4
2
seemed o be a change o a ound
3e4% pe yea o es s applied o he mon hly da a in his s udy.
In mo e han hal o he si es a dec ease in SO
4
2
deposi ion was
s ong enough o be iden ified as s a is ically significan a he plo
le el in he pe iods 2001e2010 and 2005e2010. Fo deposi ion o
ino ganic N compounds, ela i e changes we e smalle and significan
dec easing ends we e only ound o abou a qua e o he plo s.
O e all, dec easing ends o SO
4
2
and ino ganic N compounds
o abou 6% and 2% pe yea espec i ely we e ypical o he 10
yea pe iod up o 2010. T end es ima es o indi idual si es howe e
anged om 15% o 7% pe yea . The s onges dec easing ends
we e ound o si es in wes e n cen al Eu ope in egions wi h
ela i ely high deposi ion fluxes whe eas s able o sligh ly
inc easing deposi ion du ing he las 5 yea s we e ound in and eas
o he Alpine egion as well as in no he n Eu ope.
Fo ino ganic N compounds, he ends in a mosphe ic deposi ion
(BD and TF) as a esul o emission educ ions in Eu ope a e unlikely
o be de ec ed wi h s a is ical significance in ime se ies sho e han
10 yea s. Fo SO
4
2
, ypical ends we e s onge , especially in he
1990s, and could be de ec ed e en in sho e ime se ies.
Fig. 4. Ra io o minimum de ec able end de i ed om end es s ( slopeemp
min ) o
minimum de ec able end de i ed om mean empo al a iabili y ( slopemod
min )(c
7
) o
end analyses wi h LR and MK o annual, SMK and PMK o mon hly SO
4
2
,NO
3

,NH
4
þ
TF
deposi ion ime se ies wi h con inuous da a om 2007 o 2010 (4 yea s), om 2005 o
2010 (6 yea s), om 2003 o 2010 (8 yea s), om 2001 o 2010 (10 yea s), and om
1999 o 2010 (12 yea s).
Fig. 5. T end o sulpha e sulphu (SO
4
2
eS) and ino ganic ni ogen (NO
3

eNþNH
4
þ
eN) TF deposi ion de e mined wi h PMK on plo s wi h con inuous measu emen s om 2005 o
2010. Non-significan posi i e and nega i e changes a e indica ed wi h ‘no change (þ)’and ‘no change ()’, espec i ely.
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374370
The deposi ion ends can o a la ge ex en be a ibu ed o he
educ ions o he emissions o ai pollu an s achie ed be ween
1990 and 2010. Despi e dec easing ends a nume ous plo s, o al
deposi ion o ino ganic N compounds and SO
4
2
o o es s s ill ex-
ceeds c i ical loads in pa s o Eu ope.
Con inued long- e m deposi ion moni o ing will be necessa y o
demons a e he e ec i eness o emission educ ion measu es and
o in es iga e obse ed e ec s on he ecosys ems caused by
deposi ion.
Acknowledgemen s
The applied me hods o de e mina ion o a mosphe ic deposi-
ion fluxes ha e been u he de eloped and ha monized by
nume ous scien is s in he Expe Panel on Deposi ion o ICP Fo es s
subsequen ly chai ed by G. L€
o blad, E. Ul ich, N. Cla ke and K.
Hansen. Applying hese me hods in ol ed nume ous echnicians o
he ins alla ion and main enance o abou 4000 sample s, collec ion
o oughly a million samples, and chemical analyses o app oxi-
ma ely 200,000 pooled samples, as well as on-going supe ision by
some 40þscien is s. Da a ansmission in ol ed na ional ocal
cen es and he da a cen e o ICP Fo es s ha was subsequen ly a
Al e a, Wageningen (FIMCI), a he Join Resea ch Cen e in Isp a
and a he P og amme Coo dina ion Cen e o ICP Fo es s in
Hambu g. Da a ansmission included sophis ica ed con o mi y and
plausibili y checks de eloped by he in ol ed da abase specialis s.
The compa abili y o he labo a o y analyses has been imp o ed by
he ac i i ies o he Wo king G oup on QA/QC in labo a o ies ini i-
a ed and suppo ed by R. Mosello, N. K€
onig, K. De ome, he la e John
De ome, A. Kowalska, A. Ma che o and o he s. Fo field in-
s alla ions, simila ac i i ies we e coo dina ed by G. D aaije s, A.
Bleeke , J. E isman, E. Ul ich, and D. 
Zlind a. Da a quali y objec i es
and o he me hodological imp o emen s we e he esul s o ac i -
i ies ini ia ed by he QA/QC commi ee o ICP Fo es s chai ed by M.
Fe e i. The a mosphe ic deposi ion measu emen s ypically
in ol ed access being g an ed by he land owne s, financial suppo
om he pa icipa ing coun ies and he EU, as well as suppo om
subo dina ed go e nmen al o ganisa ions such as communal o
o es se ices. The EU pa ially unded he deposi ion ne wo k
unde he Council Regula ion (EEC) 3528/86 on he ‘P o ec ion o
Fo es s agains A mosphe ic Pollu ion’and he Regula ion (EC) No
2152/2003 conce ning moni o ing o o es s and en i onmen al
in e ac ions in he communi y (Fo es Focus) and by he p ojec LIFE
07 ENV/D/000218 “Fu he De elopmen and Implemen a ion o an
EU-le el Fo es Moni o ing Sys em (Fu Mon)”.
The p esen ed e alua ion in ol ed na ional ep esen a i es
esponsible o he deposi ion measu emen s. Thanks o all who
con ibu ed.
Annex
Checking no mal dis ibu ion and seasonali y
The Shapi oeWilk es (R unc ion ‘shapi o. es ’, c. . Roys on,
1982) was applied o each da a se ies o check whe he deposi-
ion alues we e no mally dis ibu ed. To es o seasonali y, we
u he ca ied ou a linea eg ession (R unc ion ‘lm’) o he
mon hly da a (y) wi h a model o wo supe posed ha monic wa es
wi h wa eleng hs o one and hal a yea , espec i ely, i.e.
y¼aþb
yea s þc1sinð Þþc2cosð Þþc3sinð2 Þþc4cosð2 Þþε;
(3)
whe e y(kg ha
1
y
1
) is he deposi ion,
yea s
(yea s, as a con inuous
numbe ) he ime, ¼2
p
yea s
,ε he emainde and he in e cep a
(kg ha
1
y
1
), he slope b(kg ha
1
y
2
), and c
1
o c
4
(kg ha
1
y
1
)a e
pa ame e s. Seasonali y was assumed i a leas one o he seasonali y
e ms (c
1
o c
4
)wasiden ified as being significan (p- alue <0.05).
The seasonali y es confi med seasonali y o 85% o he ime
se ies. The emaining ime se ies o en had one o he seasonali y
e ms (c
1
ec
4
) almos eaching he p<0.05 h eshold o signifi-
cance (97% o p- alues <0.2). The e o e, seasonali y was assumed
and SMK and PMK we e applied o all ime se ies.
De e mining minimum de ec able end om indi idual end
esul s
The ela i e slope alues ( slope) we e plo ed agains he p-
alues (p) o each combina ion o end analysis echniques, flux, ion
and pe iod, o in es iga e pa e ns ha may be used o define a
minimum de ec able end o deposi ion da a. As shown in Fig. 6 o
he example o NO
3

TF se ies analysed wi h LR, we ound mos p-
alues o bewi hin a na owband wi h he shape o a Gaussian cu e
when plo ed agains he slope. This band was na owe o longe
ime se ies and wide o sho e ime se ies. In he 10 yea s ime
se ies o NO
3

in TF es ed wi h LR, mos plo s wi h absolu e alues o
slope abo e abou 5% pe yea ha e significan ends (p<0.05),
whils plo s wi h slope below 5% ha e ends ha a e no significan
(p>0.05) o many plo s. Hence, we can assign a minimum de ec -
able end slope
min
o abou 5% o he 10 yea s ime se ies.
Wi h a non-linea eg ession (R unc ion ‘nls’), we fi ed a
Gaussian shaped cu e o he poin s on he slope s. p- alue dia-
g am o end es esul s o bulk and h ough all deposi ion se ies
o he same a iable, he same leng h, and end analysis echnique.
The cu e was desc ibed by
p¼c5$e
1
2

slopem
s

2
;(4)
whe e c
5
¼0.8 is he ampli ude ha in con as o he no mal
dis ibu ion was fixed,
m
he slope alue o he peak and
s
a
measu e o he ho izon al ape u e o he Gaussian cu e, which
was used o de i e he minimum de ec able end.
We defined he minimum de ec able end slope
emp
min
as he
alue abo e which he majo i y o es s iden i y a end as signifi-
can , wi h p<0.05 (a significance le el 95%).
Fig. 6. Rela i e slope ( slope)andp- alueo linea eg ession (LR) end es o annual NO
3

h ough all deposi ion ime se iesg oups om 2007 o 2010 (4 yea s), om 2005 o 2010 (6
yea s), and om 2001 o 2010 (10 yea s) wi h a Gaussian shaped cu e fi ed o each g oup
using non-linea eg ession echniques. T end es s wi h p- alue <0.05 (black ho izon al
line) a e significan (a 95% significance le el). The in e sec ions o he cu es wi h he
ho izon al line (ci cles) we e used as empi ical alues o he minimum de ec able end
( slope
min
), i.e. he slope ange ou side which he majo i y o he ends a e significan .
P. Waldne e al. / A mosphe ic En i onmen 95 (2014) 363e374 371