![Cartographic representation of Pseudevernia furfuracea transplant sites, corresponding classes of the accumulation/loss scale [27] (here, "previous scale") and the bioaccumulation scale (Table 5; here, "new scale"), with indication of EC ratios (Table 1) for the former, percentile thresholds (%ile), and EU ratios for the latter. Transplant sites are identified by alphanumeric codes (as also reported in Table S6).](https://www.researchgate.net/profile/Juri-Nascimbene/publication/331739051/figure/fig2/AS:736349736935428@1552570866310/Cartographic-representation-of-Pseudevernia-furfuracea-transplant-sites-corresponding_Q320.jpg)
Content uploaded by Juri Nascimbene
Author content
All content in this area was uploaded by Juri Nascimbene on Mar 15, 2019
Content may be subject to copyright.
Available via license: CC BY
Content may be subject to copyright.
atmosphere
Article
New Interpretative Scales for Lichen Bioaccumulation
Data: The Italian Proposal
Elva Cecconi 1, †, Lorenzo Fortuna 1 ,† , Renato Benesperi 2, Elisabetta Bianchi 2,
Giorgio Brunialti 3, Tania Contardo 4, Luca Di Nuzzo 2, Luisa Frati 3, Fabrizio Monaci 4,
Silvana Munzi 5, Juri Nascimbene 6, Luca Paoli 7, Sonia Ravera 8, Andrea Vannini 4,
Paolo Giordani 9, Stefano Loppi 4and Mauro Tretiach 1, *
1Department of Life Sciences, University of Trieste, Via L. Giorgieri 10, 34127 Trieste, Italy;
elva.cecconi@phd.units.it (E.C.); lfortuna@units.it (L.F.)
2Department of Biology, University of Firenze, Via La Pira 4, 50121 Firenze, Italy;
renato.benesperi@unifi.it (R.B.); e.bianchi@unifi.it (E.B.); luca.dinuzzo@stud.unifi.it (L.D.N.)
3TerraData environmetrics, Spin Off of the University of Siena, Via L. Bardelloni 19,
58025 Monterotondo Marittimo (GR), Italy; brunialti@terradata.it (G.B.); frati@terradata.it (L.F.)
4Department of Life Sciences, University of Siena, via Mattioli 4, 53100 Siena, Italy;
tania.contardo2@unisi.it (T.C.); fabrizio.monaci@unisi.it (F.M.); andrea.vannini@unisi.it (A.V.);
stefano.loppi@unisi.it (S.L.)
5Centre for Ecology, Evolution and Environmental Changes (cE3c), Faculdade de Ciências,
Universidade de Lisboa, Campo Grande, 1749-016 Lisbon, Portugal; ssmunzi@fc.ul.pt
6Department of Biological, Geological and Environmental Sciences, University of Bologna, Via Irnerio 42,
40126 Bologna, Italy; juri.nascimbene@unibo.it
7Department of Biology, University of Pisa, via L. Ghini, 13, 56126 Pisa, Italy; luca.paoli@unipi.it
8Department of Biosciences and Territory (DiBT), University of Molise, C.da Fonte Lappone,
86090 Pesche (Isernia), Italy; sonia.ravera@unimol.it
9DIFAR, University of Genova, Viale Cembrano 4, 16148 Genova, Italy; giordani@difar.unige.it
*Correspondence: tretiach@units.it; Tel.: +39-040-5583885
† These authors contributed equally to this work.
Received: 26 February 2019; Accepted: 8 March 2019; Published: 13 March 2019
Abstract:
The interpretation of lichen bioaccumulation data is of paramount importance in
environmental forensics and decision-making processes. By implementing basic ideas underlying
previous interpretative scales, new dimensionless, species-independent “bioaccumulation scales” for
native and transplanted lichens are proposed. Methodologically consistent element concentration
datasets were populated with data from biomonitoring studies relying on native and transplanted
lichens. The scale for native lichens was built up by analyzing the distribution of ratios between
element concentration data and species-specific background concentration references (Bratios), herein
provided for Flavoparmelia caperata and Xanthoria parietina (foliose lichens). The scale for transplants
was built up by analyzing the distribution of ratios between element concentration in exposed and
unexposed samples (EU ratio) of Evernia prunastri and Pseudevernia furfuracea (fruticose lichens). Both
scales consist of five percentile-based classes; namely, “Absence of”, “Low”, “Moderate”, “High”,
and “Severe” bioaccumulation. A comparative analysis of extant interpretative tools showed that
previous ones for native lichens suffered from the obsolescence of source data, whereas the previous
expert-assessed scale for transplants failed in describing noticeable element concentration variations.
The new scales, based on the concept that pollution can be quantified by dimensionless ratios between
experimental and benchmark values, overcome most critical points affecting the previous scales.
Keywords:
biomonitoring; native lichens; lichen transplants; air pollution; trace elements;
background levels; Flavoparmelia caperata;Xanthoria parietina;Evernia prunastri;Pseudevernia furfuracea
Atmosphere 2019,10, 136; doi:10.3390/atmos10030136 www.mdpi.com/journal/atmosphere
Atmosphere 2019,10, 136 2 of 19
1. Introduction
Air quality standards are fundamental references in environmental policy. These standards
are mostly set up from data on atmospheric pollutant concentrations obtained by continuous
measurements within fully- or semi-automatic apparatus [
1
]. Nevertheless, the need for evaluating
biological effects of low concentrated air pollutants over large geographical areas has made
biomonitoring of paramount importance to provide and integrate information on pollutant
depositions [
2
,
3
]. A fortiori, this is true for trace element pollution. In this respect, the extensive
use of lichens as effective bioaccumulators has provided valuable data through the years [
4
]. A crucial
point with potential outcomes for decision making and environmental forensics is the appropriate
interpretation of biomonitoring results [
5
]. This issue has been extensively addressed in the context
of human biomonitoring and chemical risk assessment (e.g., [
6
,
7
]), but it has also been faced in the
field of biomonitoring by mosses and lichens. Indeed, several efforts have been made to achieve high
quality standards for the moss bag technique, hence to improve cross-study comparison and guarantee
proper result interpretation [
8
,
9
]. Even for lichens, efforts have been made to enhance methodological
standardization and result interpretation. In particular, in this case, interpretative tools were purposely
developed for bioaccumulation data from both native and transplanted lichens.
In biomonitoring techniques based on native lichens (see Table 1for a glossary), measured
element concentrations are usually compared with background element concentration values (BECs;
Table 1) and, when these are not available, results are expressed as deviations from the minimum
values revealed in the study area, considered as an “internal baseline” [
10
]. BECs can either be
obtained by analyzing literature data (review-based BECs, [
2
,
11
]) or assessed by direct large-scale
field campaigns (field-assessed BECs [
12
–
14
]). Both approaches should follow robust methodological
guidelines and, in fact, the discussion on their proper assessment in lichen matrices is of current
interest [
14
,
15
]. Another approach to the interpretation of biomonitoring data from native lichens
is the use of interpretative scales [
16
] based on thresholds identifying classes of increasing element
concentrations, and obtained by the meta-analysis of a large set of bioaccumulation data for epiphytic
lichens at the national level [
17
]. The so-called “naturality/alteration scales”, extensively applied
until now (e.g., [
18
–
23
]), were originally proposed by Nimis and Bargagli and consist of seven
classes of element concentrations built up on hundreds of data points collected in Italy between
the 1980s and the 1990s [
16
]. Source data referred to 17 elements with a minimum of 100 records each,
obtained from at least three biomonitoring surveys carried out in areas characterized by different
pollution levels and geomorphology [
16
]. The simple core idea behind these interpretative scales is
undoubtedly powerful. However, as recognized by the same authors, some important issues remained
unsolved. In particular, naturality/alteration scales are multi-specific, meaning that the source data
referred to manifold lichen species, hence posing an important problem related to the acknowledged
species-specificity of lichen bioaccumulation [
17
,
24
–
26
]. Moreover, in the intentions of authors, source
data should have been reported in an accessible database, so as to allow the inclusion of new records
together with important methodological information (e.g., lichen species, geographical location, sample
mineralization technique). Unfortunately, this did not occur and the same source data on which the
scales were built up remained unpublished.
For biomonitoring techniques based on lichen transplants (Table 1), the interpretation is generally
based on the comparison of the elemental concentrations measured in samples exposed in the target
study area for 0.5–6(–12) months and those measured in unexposed samples (Table 1) immediately after
collection of the bulk material in a proximate-natural site. Even in this case, an interpretative scale is
available [
27
]. The core idea behind this scale, originally proposed by Frati et al. [
27
], was substantially
different from that of naturality/alteration scales, because the data are expressed as a ratio, the so-called
exposed-to-control (EC) ratio (Table 1), calculated by dividing the element concentration values of
exposed samples (eventually the mean values, if more samples are exposed at the same site) by those
of unexposed samples. The resulting “accumulation/loss scale” consisted of five classes built up on
Atmosphere 2019,10, 136 3 of 19
arbitrary cutoffs (that is, progressive
±
25% deviations from the unitary EC value [
27
]) on the basis of
considerations derived from lichen bioindication studies [27,28].
The aim of this work was to develop new scales, using a methodologically consistent pipeline
and revised terminology, based on the meta-analysis of biomonitoring data. In particular, we
conceptualized two dimensionless scales for native and transplanted lichens, based on (i) the ratio
between element concentration data and species-specific review-based BECs (herein contextually
provided), and (ii) the ratio between element concentration values measured in exposed and unexposed
lichen samples, respectively. Both scales, along with the previously available ones, were also applied
to real case studies in order to assess their relative performance. The new scales are valid for Italy,
but being based on a robust conceptual framework, they may easily be implemented in other countries.
2. Data and Methodology
2.1. Data Collection
A literature search was undertaken between April and May 2018, in order to compile a list of
eligible biomonitoring studies targeting lichens as bioaccumulators of trace elements, with the main
aim of populating two datasets, including (i) bioaccumulation data from biomonitoring studies relying
on native lichens (herein, dataset N: i.e., Native) and (ii) bioaccumulation data from biomonitoring
studies relying on lichen transplants (herein, dataset T: i.e., Transplants). References were included
when based on (a) native lichens, referring to thalli grown at different environmental conditions,
from proximate-natural to variously human-impacted ones; and (b) lichen transplants, referring to
lichen material purposely collected in areas unaffected by significant levels of airborne pollutants and
afterwards exposed in polluted areas for relatively short periods. Studies were then reviewed and
excluded when meeting at least one of the following conditions: (i) the study was carried out outside
the Italian territory; (ii) data were pooled for different lichen species; or (iii) transplanted samples were
exposed for more than four months (non-routine exposure time spans).
Element concentration values were recorded just as reported in the papers, with the exception
of values below the limit of instrumental detection (LOD), which were recorded as LOD values.
All values were expressed in
µ
g
×
g
−1
dry weight (DW). In addition to the element concentration data,
methodological information concerning the acid sample digestion was also recorded, because such a
procedure is known to affect the results of elemental analytical determination [
15
,
29
,
30
]. Moreover,
relevant information concerning the lichen species, the administrative region of study areas, and the
year of data collection or publication (when the former was missing) was recorded as well. Data
from lichen transplants were labelled according to their type: (i) element concentration data from
unexposed samples, and (ii) element concentration data from exposed samples. The exposure time
span of transplants was also recorded, using the week as base unit.
2.2. Data Processing
The dataset Nwas initially subjected to a methodological screening. In order to enhance data
homogeneity, element concentration data obtained with partial acid digestion of lichen samples
(i.e., without hydrofluoric acid, HF) were discarded to enhance methodological uniformity and because
this mineralization approach, although largely used and safer for operator health, may determine
unsatisfactory recoveries for typical tracers of soil contamination [
15
,
31
]. Moreover, data before 2008
were also removed (temporal data filtering) to increase methodological comparability. Indeed, older
biomonitoring studies were often deficient in methodological details concerning sample processing
procedures, with special reference to sample cleaning (i.e., washing vs. manual cleaning) [
14
] and
selection of suitable parts of thalli (i.e., peripheral parts vs. whole thalli) [
32
], all procedures known
to bias the lichen elemental concentration [
17
]. Also, when these studies reported such information,
a substantial methodological heterogeneity was highlighted. By contrast, in the most recent literature,
sample washing and the use of whole thalli were abandoned in favor of a manual debris cleaning
Atmosphere 2019,10, 136 4 of 19
and the use of peripheral portions of thalli. In addition, all the elements with data deriving from
lichen samples collected from less than three administrative Italian regions were excluded, along with
those characterized by less than 40 records. Afterwards, the dataset was carefully screened for the
occurrence of duplicated records, typing errors, or inconsistent units of measurements. Descriptive
statistics were finally calculated for different elements and lichen species; these included the number of
records, mean, median, range, quartiles, as well as skewness and kurtosis of the element concentration
data distribution.
The construction of the dataset Nenabled easy assessment of review-based, methodologically
uniform background element concentration values (BECs) for frequently used species. Indeed,
following the rank-based approach used for the assessment of quality levels of soils and sediments [
33
],
the 25th percentile of species- and element-specific bioaccumulation data distributions was selected as
a background benchmark, and each value below this threshold was regarded as a result of unpolluted
conditions. Descriptive statistics, that is, mean, standard deviation, median, and median absolute
deviation [
34
], were provided for the sub-dataset consisting of element concentration values below
the BEC threshold (BEC dataset). Median values of the BEC dataset were also tested for inter-specific
significant differences using Mann–Whitney’s U test for independent samples.
After having identified species-specific BECs as the 25th percentile of dataset N, each element
concentration value in the same dataset was divided by the corresponding BEC value to obtain a
new dataset of the same size that included dimensionless values, namely the ratios between element
concentration and background values. This simple procedure was inspired by common practices
in soil geochemistry. Indeed, many authors [
35
–
37
] have suggested that the calculation of ratios of
element concentrations observed in topsoil to those in the subsoil may provide a reliable indication of
contamination [
38
]. Moreover, this approach is methodologically similar to that used for the expression
of results in transplant-based studies (see infra). Such a unitless entity, the Bratio (i.e., Bioaccumulation
ratio; Table 1), indicates absence of bioaccumulation with respect to the national background when
it is lower than or equal to 1, whereas it indicates bioaccumulation occurrence when it exceeds 1.
Bratios, organized in a single column vector (i.e., Bratio dataset; Data S1), were then tested for
significant inter-specific differences (Methods S1; Table S1). Finally, skewness and kurtosis of the B
ratio distribution were calculated. After appraisal of Bratio distributional shape, the 25th, 75th, 90th,
and 95th percentiles were used as interval thresholds to define a five-class interpretative scale.
The dataset Twas also subjected to a preliminary methodological screening; data obtained with
partial acid digestion of lichen samples were discarded. However, because the dataset was far smaller
than the dataset N, no temporal data filtering was performed, and only those elements characterized
by less than 25 records were removed.
Each element concentration value referring to exposed samples in the dataset Twas divided by
the corresponding unexposed mean value, so as to obtain a new dataset that included dimensionless
values, namely the ratios between element concentration values of exposed and unexposed lichen
samples. This further unitless entity, the EU ratio (i.e., exposed-to-unexposed ratio; Table 1), previously
proposed by Frati et al. [
27
] as EC ratio, and herein terminologically revised, indicates absence of
bioaccumulation with respect to a local unaltered situation when it is lower than or equal to 1, whereas
it indicates bioaccumulation occurrence when it exceeds 1. EU ratios, organized in a single column
vector, were then tested for significant inter-specific differences, as done for the Bratios of native
lichens (Methods S1; Table S1). EU ratio data were analyzed to assess the most frequent exposure time
spans (expressed in weeks). Data obtained from the analysis of samples exposed for commensurable
time spans were then uniformly labelled (e.g., 8 and 9 week-transplants; Section 3.2). Subsequently,
data were split into three sub-datasets, homogeneous for transplant exposure time span (i.e., EU
ratio sub-datasets; Data S2–S4). Each sub-dataset was further screened for upper outliers according
to the Tukey method (i.e., values higher than the third quartile of the distribution plus three times
the interquartile range), which makes no distributional assumptions and is applicable to skewed or
non-bell-shaped data distributions [
39
]. After outlier removal, each sub-dataset was further tested
Atmosphere 2019,10, 136 5 of 19
for inter-specific differences (Methods S1; Table S1). Finally, skewness and kurtosis of EU ratio
distributions were calculated. After appraisal of EU ratio distributional shape, the 25th, 75th, 90th,
and 95th percentiles were calculated and corrected to account for the overall uncertainty associated
with small-sized datasets. In particular, EU ratio values were corrected by subtracting from them a
percentage corresponding to the semi-range of the 95% confidence interval of EU ratio data divided by
the mean [
40
]. The corrected-percentiles were then used as interval thresholds to define a five-class
interpretative scale, with the exception of the boundary between Class 1 and Class 2, which was
aprioristically established at the unitary value because this represents the discernibility threshold
between the occurrence of bioaccumulation (EU ratio > 1) and its absence (EU ratio ≤1).
All data analyses and graphics were performed with the software packages Statistica v. 10 (StatSoft
Inc., Tulsa, OK, US) and Microsoft Excel (Microsoft Office Professional Plus 2010), with statistical
significance tested at α= 0.05 in all cases. Figures were edited with CorelDraw X7.
Table 1. Glossary of main terms and concepts.
Glossary
Native lichens Lichens grown in a target study area.
Background element
concentration values (BECs)
Species-specific element concentration values measured in lichen samples
reflecting proximate-natural, unaltered conditions.
Bioaccumulation ratio
(B ratio)
The dimensionless ratio between species-specific element concentration values
measured in native samples and the corresponding background values.
Lichen transplants
Lichens collected in a proximate-natural site and afterwards transplanted for a
certain exposure time span to a target study area.
Exposed samples
In the context of lichen transplants, samples transplanted to a target study area,
exposed to pollutant depositions for a certain exposure time span, and then
subjected to the determination of elemental concentration.
Unexposed samples
In the context of lichen transplants, samples collected in a proximate-natural
site and subjected to the determination of elemental concentration. These
samples are used as a benchmark to assess the magnitude of lichen
bioaccumulation after transplantation.
Exposed-to-unexposed ratio
(EU ratio)
The dimensionless ratio between species-specific element concentration values
measured in exposed samples and the corresponding element concentration
values measured in unexposed samples.
Exposed-to-control ratio
(EC ratio) Previous name of the EU ratio [27], here terminologically revised.
2.3. Working Examples: Case Studies from NE Italy
The results obtained with the new interpretative scales and with the previous ones for native [
16
]
and transplanted lichens [
27
] were compared using two case studies from NE Italy, respectively
obtained by analyzing (a) native samples of Flavoparmelia caperata and Xanthoria parietina and (b)
transplanted samples of Pseudevernia furfuracea.
Samples of F. caperata and X. parietina were collected in 2014 at 40 sampling sites around a
coal-fired thermoelectric power plant in the municipality of Monfalcone (NE Italy) [
41
,
42
]. Sampling
sites were distributed according to a systematic design (regular grid, 2
×
2 km) and, when possible,
lichen thalli were collected on the same host tree species. Element concentration (expressed in
µ
g g
−1
dry weight (DW)) was measured by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) after
total sample mineralization [
41
]. The severity of pollutant depositions was expressed according to (i)
naturality/alteration scales [
16
] (Table S3) and (ii) the bioaccumulation scale for native lichens proposed
here. The results reported were limited to As, Cd, and Cr—three elements of environmental and
health concern [
43
,
44
] that were randomly extracted from the set of elements included in the dataset N.
The random selection was meant to avoid a potential bias due to expert-based selection of elements,
Atmosphere 2019,10, 136 6 of 19
possibly leading to a “desired outcome”. For this reason, the random extraction prevailed on other
criteria, such as the accuracy achieved in the determination of the elemental content. In this respect,
As, Cd, and Cr were characterized by recovery percentages of 109%, 70%, and 98%, respectively [41].
For the transplant case study, we referred to ancillary data of a biomonitoring study aimed at
evaluating the contamination of mercury around a waste incinerator located in the northern Friulian
plane (NE Italy) [
45
]. Samples of P. furfuracea were collected in 2008 in a remote area of the eastern Alps
and transplanted for 12 weeks to 30 sites distributed along three linear transects centered on the waste
incinerator, mostly characterized by agricultural land use [
45
]. Element concentration (expressed in
µ
g g
−1
DW) was measured by ICP-MS after total sample mineralization [
45
]. EU ratios were calculated
and used to assess the severity of pollutant depositions according to (i) the accumulation/loss scale [
27
]
and (ii) the bioaccumulation scale for lichen transplants proposed here. Even in this case, the results
reported were limited to As, Cd, and Cr, for which recovery percentages were 99%, 94%, and 101%,
respectively [45].
Cartographic representations showing sampling and transplant sites and the outcome of the
application of different interpretative scales were provided. Cartographic elaborations were performed
with QGIS 2.18.27 ‘Las Palmas’.
3. Results and Discussion
3.1. Native Lichens
3.1.1. Source Data and Species-Specific BECs
The dataset Nincluded 32,187 bioaccumulation data points from native lichen samples. Element
concentration data referred to 42 elements measured in samples of five lichen species collected in 18
administrative Italian regions. Species included Flavoparmelia caperata,Parmelia sulcata, and Xanthoria
parietina (foliose lichens), as well as Evernia prunastri and Pseudevernia furfuracea (fruticose lichens). After
the methodological and temporal data filtering, the dataset Nincluded 3773 records for 11 elements of
environmental concern analyzed in the context of 11 studies (either published or not; in the latter case,
methodologically consistent data were provided by the authors; Data S1). Data referred to samples of
the lichen species F. caperata and X. parietina (Table 2; Figure S1) collected in five Italian regions (Friuli
Venezia Giulia, Lazio, Liguria, Molise, and Toscana).
Table 2.
Descriptive statistics of element concentration values included in dataset Nfor the lichen
species Flavoparmelia caperata and Xanthoria parietina. Statistics refer to the data counts (n), mean and
median values (Mean, Med), minima and maxima (Range), interquartile range (IQR), skewness (S),
and kurtosis (K). Mean and median values, minima and maxima, as well as interquartile ranges are
expressed in µg g−1dry weight (DW) (n.a., data not available).
Element
Flavoparmelia caperata Xanthoria parietina
n Mean Med Range IQR S K n Mean Med Range IQR S K
Al
244
551 348 110–4224 252–526
3.24 11.63
68 656 605 150–3408 371–722 3.58
16.28
As 367
0.34 0.25 0.06–1.90
0.18–0.40 2.63
9.08 79 0.35 0.28 0.06–2.31
0.15–0.40
3.41
14.59
Cd 298
0.30 0.25 0.06–1.69
0.18–0.37 2.62 13.68
80 0.15 0.09 0.04–1.46
0.07–0.15
4.73
27.97
Cr
321
2.44 1.84
0.35–24.94 1.17–2.66 4.80 33.16
77 2.39 1.91
0.69–10.52 1.61–2.73
2.82
10.87
Cu 321
8.58 7.38
2.50–78.29 6.23–9.34 6.99 67.44
98 5.83 5.48
3.20–19.27 4.45–6.31
3.17
13.29
Hg 182
0.09 0.08 0.01–0.43
0.06–0.11 1.91
8.97 77 0.06 0.05 0.01–0.63
0.04–0.07
5.55
37.18
Ni
296
3.14 2.67
0.32–19.01 1.27–4.03 2.61
9.42 51 3.39 2.66 0.82–11.2
1.55–4.68
1.45 2.68
Pb
321
6.0 4.0 0.8–114.2
2.40–6.30 7.24 69.56
98 2.37 1.64
0.36–15.40 1.00–2.67
3.03
11.78
Ti
184
41.8 26.4 0.3–309.0
19.4–40.7 2.80
9.21 42 59.8 48.9
15.2–262.0 37.2–60.0
3.00
10.57
V
150
1.71 0.94
0.34–13.22 0.75–1.60 2.87 10.67
n.a. n.a. n.a. n.a. n.a. n.a.
Zn 321
47.3 44.0
17.7–330.8 35.3–53.0 6.17 65.08
98 30.0 25.6
13.0–168.0 21.2–34.4
4.88
33.51
The foliose lichen species F. caperata and X. parietina are the most used species in biomonitoring
based on native lichens across Italy [
4
]. Indeed, such species are widespread, with very abundant
Atmosphere 2019,10, 136 7 of 19
populations from the submediterranean to the submontane belt [
46
], providing adequate sampling
density for a reliable assessment of pollutant deposition patterns [
10
]. Overall, F. caperata and
X. parietina accounted for 79.6% and 20.6%, respectively, of data. All elements, except for V, included
data from both lichen species (Table 2). The methodological data filtering resulted in a substantial
reduction of the dataset (
−
88.3%). However, the final data sizes, separately reported for each element
and species, were comparable to those reported by Nimis and Bargagli [
16
] in their multi-specific
interpretative scales (Table S3).
When inter-specific differences were tested on median values of the BEC dataset (Section 2.2),
significant differences were highlighted for 9 out of 10 elements (p< 0.05), with the exception of Hg,
characterized by very low values in both species (Table 3). F. caperata exhibited higher concentrations
of As, Cd, Cu, Pb, and Zn, whereas X. parietina had more Al, Cr, Ni, and Ti.
Table 3.
Review-based BECs (
µ
g g
−1
DW) for the epiphytic lichen species Flavoparmelia caperata and
Xanthoria parietina in Italy. Descriptive statistics refer to the data counts (n), mean and associated
standard deviation (Mean
±
SD), and median and median absolute deviation (Med
±
MAD) for
11 (F. caperata) and 10 elements (X. parietina) included in the BEC dataset (Section 2.2). Results of
statistical testing (Mann–Whitney U test for independent samples) for differences between median
element concentration in the two species are also reported. Significant p-values are highlighted in italic
(n.a., data not available).
Element
Flavoparmelia caperata Xanthoria parietina Mann–Whitney
U Test
BEC BEC Dataset BEC BEC Dataset
n Mean ±SD Med ±MAD n Mean ±SD Med ±MAD U Z p-Value
Al 253 61 201 ±37 209 ±26 372 17 295 ±59 300 ±34 87.5 −5.215 1.8 ×10−7
As
0.18 91 0.14 ±0.03 0.15 ±0.02 0.15 19 0.11 ±0.02 0.11 ±0.01
329.0
4.238 2.3 ×10−5
Cd
0.18 68 0.14 ±0.03 0.13 ±0.02 0.07 19 0.05 ±0.01 0.05 ±0.01 13.0 6.505 7.8 ×10−11
Cr 1.17 80 0.85 ±0.24 0.90 ±0.20 1.61 19 1.20 ±0.28 1.10 ±0.20
293.0
−4.149 3.3 ×10−5
Cu
6.2 80 5.2 ±0.9 5.5 ±0.5 4.5 25 4.1 ±0.3 4.1 ±0.2
221.0
5.859 4.7 ×10−9
Hg 0.057
45 0.031 ±0.021 0.038 ±0.019 0.035 18 0.019 ±0.009 0.021 ±0.008
328.0
1.308 0.191
Ni 1.27 73 0.91 ±0.20 0.93 ±0.17 1.64 13 1.28 ±0.22 1.33 ±0.16
108.5
−4.408 1.0 ×10−5
Pb 2.37 80 1.71 ±0.45 1.82 ±0.48 1.00 24 0.67 ±0.21 0.70 ±0.21 21.0 7.243 4.4 ×10−13
Ti 19.5 46 12.8 ±5.8 14.8 ±4.3 37.3 11 29.3 ±7.3 31.6 ±4.2 22.5 −4.652 3.3 ×10−6
V 0.75 37 0.61 ±0.11 0.62 ±0.07 n.a.
n.a.
n.a. n.a. n.a. n.a. n.a.
Zn
35.3 80 29.6 ±4.3 30.1 ±3.2 21.3 25 17.9 ±2.6 19.0 ±1.9 30.0 7.296 3.0 ×10−13
Our findings were in agreement with those of Nimis et al. [
17
]. Indeed, these authors highlighted
higher Cd and Zn in F. caperata than in X. parietina and an opposite pattern for Al and Fe [
17
]. Limited
to Cd, Zn, and Al (Fe was excluded from our analyses), such a pattern fully matched our results,
both when inter-specific differences were statistically tested in the BEC dataset (cf. the outcome of
non-parametric statistical testing carried out for BECs in Table 3) and in the entire dataset N(Table 2).
The review-based BECs for F. caperata and X. parietina were generally comparable in terms of
order of magnitude to those previously published for other lichens. Nevertheless, our BEC values were
often lower than review-based BECs reported for pooled foliose lichen species [
2
] and for Hypogymnia
physodes [
11
], interestingly with the only exception of Al and Ti for X. parietina (cf. Table 3and Table S2).
BECs for F. caperata and X. parietina were also compared to field-assessed BECs for the fruticose lichen
Pseudevernia furfuracea based on total acid sample digestion [
15
]. Even in this case, the reference values
for the two foliose species were either lower than or comparable with the lowest BECs reported by the
authors (cf. Table 3and Table S2).
Such data comparisons highlighted an overall pattern of comparability between the magnitude
of different species-specific sets of BECs (with few exceptions, e.g., Hg). With respect to P. furfuracea,
the lower BECs of F. caperata and X. parietina may reflect both different approaches (review-based vs.
field-based BEC assessment) and lichen morphology [
47
]. By contrast, the higher review-based
BECs reported by Bargagli are plausibly the result of aged source data, which likely included
methodologically inconsistent records and bioaccumulation data from improperly defined background
contexts. In this light, the assessment of review-based BECs for biological matrices should be regarded
Atmosphere 2019,10, 136 8 of 19
as an accurate and dynamic process, providing for the collection of methodologically uniform data for
single species and involving periodical adjustments aimed at including the most recent data.
3.1.2. Bioaccumulation Scale for Native Lichens
When bioaccumulation data in the dataset Nwere divided by the corresponding BECs (Bratio
dataset; Data S1), inter-specific differences became negligible. Indeed, the non-parametric testing
(Methods S1) did not highlight significant Bratio differences between the two species (Mann–Whitney
U test, p> 0.05; Table S1; Figure S2). Therefore, the simple operation of dividing element concentration
data by matched BECs had useful effects for interpretative purposes. As element concentration data
and BECs are element- and species-specific, such specificity resulted flattened in the Bratio, allowing
to develop a unique scale based on a high samples size (n = 3773).
The distribution of Bratios (Figure 1) was unimodal, right-skewed, and strongly leptokurtic
(skewness > 0, kurtosis > 3), as previously highlighted for the distribution of bioaccumulation data in
epiphytic lichens, either pooled or not [17].
Atmosphere 2019, 10, x FOR PEER REVIEW 8 of 20
the lower BECs of F. caperata and X. parietina may reflect both different approaches (review-based vs.
field-based BEC assessment) and lichen morphology [47]. By contrast, the higher review-based BECs
reported by Bargagli are plausibly the result of aged source data, which likely included
methodologically inconsistent records and bioaccumulation data from improperly defined
background contexts. In this light, the assessment of review-based BECs for biological matrices
should be regarded as an accurate and dynamic process, providing for the collection of
methodologically uniform data for single species and involving periodical adjustments aimed at
including the most recent data.
3.1.2. Bioaccumulation Scale for Native Lichens
When bioaccumulation data in the dataset N were divided by the corresponding BECs (B ratio
dataset; Data S1), inter-specific differences became negligible. Indeed, the non-parametric testing
(Methods S1) did not highlight significant B ratio differences between the two species (Mann–
Whitney U test, p > 0.05; Table S1; Figure S2). Therefore, the simple operation of dividing element
concentration data by matched BECs had useful effects for interpretative purposes. As element
concentration data and BECs are element- and species-specific, such specificity resulted flattened in
the B ratio, allowing to develop a unique scale based on a high samples size (n = 3773).
The distribution of B ratios (Figure 1) was unimodal, right-skewed, and strongly leptokurtic
(skewness > 0, kurtosis > 3), as previously highlighted for the distribution of bioaccumulation data in
epiphytic lichens, either pooled or not [17].
Figure 1. B ratio data distribution with indication of data counts per interval, skewness and kurtosis,
and percentile values corresponding to the thresholds defining five bioaccumulation classes (the B
ratio axis ends at the first ‘zero’ count).
The percentile thresholds for the elaboration of the new interpretative scale were redefined with
respect to those of Nimis and Bargagli [16] after the appraisal of B ratio distributional shape. The 50th
percentile was discarded because the corresponding B ratio value (B = 1.4) was too close to the BEC
threshold of 1.0. The selection of the 90th percentile as upper/lower threshold of Class 3/4 (Figure 1,
Table 4) was based on toxicological considerations [33,48]. In particular, the 90th percentile of
concentration data was recently proposed as “environmentally relevant” [48], thus well suited as
Figure 1.
Bratio data distribution with indication of data counts per interval, skewness and kurtosis,
and percentile values corresponding to the thresholds defining five bioaccumulation classes (the Bratio
axis ends at the first ‘zero’ count).
The percentile thresholds for the elaboration of the new interpretative scale were redefined
with respect to those of Nimis and Bargagli [
16
] after the appraisal of Bratio distributional shape.
The 50th percentile was discarded because the corresponding Bratio value (B= 1.4) was too close to
the BEC threshold of 1.0. The selection of the 90th percentile as upper/lower threshold of Class 3/4
(Figure 1, Table 4) was based on toxicological considerations [
33
,
48
]. In particular, the 90th percentile of
concentration data was recently proposed as “environmentally relevant” [
48
], thus well suited as cutoff
between the occurrence of “Moderate” and “High” bioaccumulation. Finally, the 95th percentile was
chosen, instead of 98th, as the upper/lower threshold of Class 4/5 based on a precautionary approach.
On these grounds, the ranges of the Bratio classes were characterized by similar amplitudes (Table 4).
The interpretative bioaccumulation scale was definitely improved with respect to previous
multi-specific naturality/alteration scales. Indeed, besides being based on the most recent and
methodologically consistent data, the Bratio scale is also readily understandable and provides easy
Atmosphere 2019,10, 136 9 of 19
implementation. As with species-specific BECs, the Bratio-based interpretative scale will need to be
updated with the most recent data. We estimate that this might occur approximately every ten years.
Table 4.
Bratio, percentile-based, five-class interpretative scale for bioaccumulation data from native
lichens. Class codes, description and abbreviations, percentile thresholds, corresponding Bratio values,
RGB and HTML color codes associated to bioaccumulation classes are reported.
Bioaccumulation Class Percentile
Thresholds BRatio Color Code
ID Description (Abbreviation) RGB HTML
1 Absence of bioaccumulation (A) ≤25th ≤1.0 0 0 255 #0000FF
2 Low bioaccumulation (L) (25th, 75th] (1.0, 2.1] 0 128 0 #008000
3 Moderate bioaccumulation (M) (75th, 90th] (2.1, 3.4] 255 243 15 #FFF30F
4 High bioaccumulation (H) (90th, 95th] (3.4, 4.9] 255 0 0 #FF0000
5 Severe bioaccumulation (S) >95th >4.9 128 0 64 #800040
The terminological shift from the previous “naturality/alteration” (Table S3) to the more cautious
“bioaccumulation level” (Table 4) may apparently pose some issues. Indeed, the latter form suggests a
mere assessment of the magnitude of bioaccumulation levels in lichens, whereas the former stresses
the link between lichen bioaccumulation and pollution, expressly indicating the use of scales to “assess
environmental alteration in terms of a deviation from natural backgrounds” [
10
] (i.e., “naturality”).
Yet, despite the inspiring terminology, previous scales did not rely on any operational definition of
quantitative threshold for “naturality” (e.g., proper background reference), instead being based on
a circular definition of “alteration” with respect to “naturality” (and vice versa) [
28
]. By contrast,
a statistically-based element concentration benchmark (i.e., review-based BECs) is inherent to the
Bratio, thus the new scale is actually able to assess whether or not deviations from a national unaltered
reference occurred. In this light, the terminological shift was contextually driven by (i) the need to
underline the novelty of the Bratio scale and (ii) a harmonization intent with the new scale provided
for lichen transplants (see infra).
3.2. Lichen Transplants
3.2.1. Source Data
Before the data cleaning (Section 2.1), the overall EU ratio dataset included 820 bioaccumulation
data from lichen transplant studies published over the last 25 years. Element concentration data referred
to 15 elements (Al, As, Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Ni, Pb, V, and Zn) analyzed in the context of
18 studies. Data referred to samples of two fruticose species, Evernia prunastri and Pseudevernia furfuracea,
collected in 10 Italian administrative regions (Calabria, Campania, Emilia Romagna, Friuli Venezia Giulia,
Lazio, Liguria, Lombardia, Piemonte, Toscana, Veneto). Fruticose species are usually preferred over
foliose species for lichen transplants [
4
] because they ensure greater biomass per lichen thallus, as well
as easier cleaning and installation, thus contextually reducing processing time and enhancing sample
homogeneity [
49
]. Overall, E. prunastri and P. furfuracea accounted for 18.3% and 81.7%, respectively, of
data. All elements, except for Mg, included data from both lichen species.
The transplant exposure time span varied across studies: 21%, 8%, 14%, 20%, 1%, and 36% of data
relied on 4, 6, 8, 9, 11, and 12 week transplants, respectively. Data relying on comparable exposure
periods (i.e., 6, 8, and 9 weeks, as well as 11 and 12 weeks) were labelled as 8-week transplant and
12-week transplant, respectively, in order to obtain three numerically balanced sub-datasets for equally
spaced exposure periods. Outliers identification led to the removal of 5, 17, and 11 EU values from
the three sub-datasets (4, 8, and 12 weeks, respectively). Eventually, the 4-week EU ratio sub-dataset
accounted for 21% of data (n = 169 records), the 8-week sub-dataset for 42% (n = 330), and the 12-week
sub-dataset for 37% (n = 288) (Data S2–S4).
Atmosphere 2019,10, 136 10 of 19
3.2.2. Bioaccumulation Scale for Lichen Transplants
Even in the case of bioaccumulation data from lichen transplants, inter-specific differences were
negligible when addressed by non-parametric testing on EU ratios. Indeed, the output of statistical
testing (Methods S1) did not highlight significant EU ratio differences between the two species
(Mann–Whitney U test, p> 0.05; Table S1; Figure S3), thus the same considerations spelt out for Bratios
apply. EU ratio distributions were right-skewed and slightly platykurtic (skewness > 0, kurtosis < 3;
Figure 2). The positive skewness was consistent (although lower) with that of the Bratio distribution.
Atmosphere 2019, 10, x FOR PEER REVIEW 11 of 20
Figure 2. EU ratio data distribution with indication of data counts per interval, skewness and kurtosis,
and percentile values corresponding to the thresholds defining five bioaccumulation classes. Data are
separately reported for different transplant exposure time spans (from the top to the bottom: 4, 8, and
12 weeks).
Certainly, the marked differences between the distributions of B and EU ratios were because of
different (i) sample sizes (Section 3.1.2 and 3.2.1), (ii) benchmark values in the ratio denominators
(i.e., BECs vs. elemental concentration values of unexposed samples), and (iii) duration of lichen
exposure to pollutants and bioaccumulation mechanisms. Concerning the latter point, transplanted
lichens are exposed for few weeks to new, and often harsh, environmental conditions, thus rapidly
accumulating mostly through passive mechanisms [50,51]. By contrast, the bioaccumulation in
lifespan-exposed native lichens is the result of a long-term interplay of both passive phenomena and
Figure 2.
EU ratio data distribution with indication of data counts per interval, skewness and kurtosis, and
percentile values corresponding to the thresholds defining five bioaccumulation classes. Data are separately
reported for different transplant exposure time spans (from the top to the bottom: 4, 8, and 12 weeks).
Atmosphere 2019,10, 136 11 of 19
Certainly, the marked differences between the distributions of Band EU ratios were because of
different (i) sample sizes (Sections 3.1.2 and 3.2.1), (ii) benchmark values in the ratio denominators
(i.e., BECs vs. elemental concentration values of unexposed samples), and (iii) duration of lichen
exposure to pollutants and bioaccumulation mechanisms. Concerning the latter point, transplanted
lichens are exposed for few weeks to new, and often harsh, environmental conditions, thus rapidly
accumulating mostly through passive mechanisms [
50
,
51
]. By contrast, the bioaccumulation in
lifespan-exposed native lichens is the result of a long-term interplay of both passive phenomena
and slower active intracellular uptakes characterized by element-specific kinetics [
52
]. Such interplay
eventually results in the achievement of a dynamic equilibrium with the surrounding environment [
53
]
and likely in higher elemental concentration levels in the case of important pollutant loads. Indeed, EU
ratios corresponding to 90th and 95th percentiles of the distributions were lower than the corresponding
Bratio values (cf. Tables 4and 5).
In transplants, the importance of the exposure time span [
54
,
55
] emerged in the EU ratio data
in the case of high pollutant loads. Indeed, an increasing trend of EU ratio values corresponding
to 90th and 95th percentiles was observed when moving from 4 to 12 week exposure (Figure 2,
Table 5). Utmost differences were highlighted for the EU ratio corresponding to the 95th percentile
between the 4 week exposure and 8 or 12 week exposures (2.8 vs. 3.5 and 3.7; Table 5). Interestingly,
no trend of increasing values from 4 to 12 weeks was highlighted for EU ratio values corresponding
to 25th and 75th percentiles, confirming that the exposure time span mostly affects bioaccumulation
results in the case of high levels of airborne pollutant depositions (EU ratio above 90th percentile,
i.e., environmentally relevant bioaccumulation [
48
]). On this basis, three different series of values have
been reported, to be alternatively used according to the selected exposure time span. Nevertheless,
it should also be pointed out that in most biomonitoring literature targeting mosses, short exposure
times (i.e., 3–4 weeks) are discouraged because unclear “accumulation signals” would lead to the
construction of derived datasets of limited reliability [
9
]. Such a methodological issue has been dealt
more rarely for lichen transplants, but it is generally agreed that lichens should be exposed for at least
6–8 weeks, based on the following considerations: detectable accumulated concentrations, replicability,
and exposure time spans within the limits of practical considerations [55].
Table 5.
EU ratio, percentile-based, five-class interpretative scale for bioaccumulation data from lichen
transplants. Class codes, description and abbreviations, percentile thresholds, corresponding EU
ratio values for different exposure time spans, and color codes (RGB and HTML) associated with
bioaccumulation classes are reported.
Bioaccumulation Class Percentile
Thresholds
EU Ratio Color Code
ID Description (Abbreviation) 4 Weeks 8 Weeks 12 Weeks RGB HTML
1 Absence of bioaccumulation (A) ≤25th * ≤1.0 ≤1.0 ≤1.0 0 0 255
#0000FF
2 Low bioaccumulation (L) (25th, 75th] (1.0, 1.8] (1.0, 1.9] (1.0, 1.8] 0 128 0
#008000
3 Moderate bioaccumulation
(M)
(75th, 90th] (1.8, 2.5] (1.9, 2.7] (1.8, 3.1] 255 243 15
#FFF30F
4 High bioaccumulation (H) (90th, 95th] (2.5, 2.8] (2.7, 3.5] (3.1, 3.7] 255 0 0
#FF0000
5 Severe bioaccumulation (S) >95th >2.8 >3.5 >3.7 128 0 64
#800040
* The EU ratio values corresponding to 25th percentile threshold (upper/lower threshold of Class 1/2) are actually
equal to 1.1, 1.0 and 0.9 for exposure time spans of 4, 8 and 12 weeks, respectively (see text for explanation).
In the bioaccumulation scale proposed for lichen transplants, the upper/lower EU ratio threshold
of Class 1/2 was aprioristically established at EU = 1 (Section 2.3). However, it is worth noting that the
corrected EU ratio values corresponding to the 25th percentile are actually very close to such a value
(ranging from 0.9 to 1.1, see footnote in Table 5; Data S2–S4).
Even in this case, we decided to abandon the old class description based on the concept of “loss”,
because an element concentration decrease may either reflect actual “pristine” ambient air conditions at
the transplant sites or a “washing effect” caused by rainfall in the presence of non-negligible pollutant
emissions [
56
]. Finally, it must be pointed out that, given the relatively limited amount of source data,
Atmosphere 2019,10, 136 12 of 19
this new bioaccumulation scale for lichen transplants has to be regarded as preliminary and should be
used with caution, pending the inclusion of new available bioaccumulation data.
3.3. Comparison between Previous and New Interpretative Scales
The naturality/alteration scales applied in the last twenty years in Italy [
16
] and the brand-new
bioaccumulation scale for native lichens (Table 4) were both applied to the same case study. In this case,
a direct comparison between classes attributed to sampling sites resulting from different interpretative
scales would be pointless because of the substantial differences in scale conceptualization; however, a
comparative analysis of outcomes permitted some interesting considerations.
Overall, the application of previous scales provided a rather optimistic description of the study
area. According to the seven-class scale, the vast majority of sampling sites were characterized by
“very high” and “high naturality”. In particular, 96.5%, 89.6%, and 86.2% of sites belonged to such
classes for As, Cd, and Cr concentration in Flavoparmelia caperata, as well as 88.9% (As and Cd) and
77.7% (Cr) for Xanthoria parietina (Figure 3, Table S5). “Low alteration” characterized only two sites for
Cr (G6 and D7 for F. caperata and X. parietina, respectively; Figure 3).
When the new bioaccumulation scale was applied, the majority of sampling sites were consistently
characterized by “Low bioaccumulation”, with the exception of As in X. parietina, instead characterized
by a majority of sites belonging to Class 3 (“Moderate bioaccumulation”). In particular, 69.0%, 79.3%,
and 82.8% belonged to Classes 1 and 2 for As, Cd, and Cr concentration, respectively, in F. caperata,
as well as 33.3% (As), 66.7% (Cd), and 88.9% (Cr) for X. parietina. However, by applying this scale,
some cases of “High” and “Severe bioaccumulation” were also highlighted (Figure 3, Table S5), thus
determining a more conservative interpretation.
The study area is characterized by high anthropogenic pressure and the presence of a coal-fired
thermoelectric power plant, shipbuilding industries, and other small industrial activities [
41
]. Previous
investigations demonstrated that, overall, the elemental concentrations in lichens grown in the study
area were not impressively high; however, a certain contamination of As and Cr occurred. These
elements are acknowledged tracers of coal combustion [
57
,
58
]; therefore, the enrichment observed in
thalli collected at specific sites was ascribed to the power plant emissions [
41
], although these were
compliant with threshold limits [
59
]. In particular, the evidence that Cr concentration in lichen samples
was related to the plant emissions was confirmed by the results of an air particulate matter sampling
carried out during both operational and non-operational state of the plant [42].
The pattern revealed by the new scale correlates well with the deposition plume highlighted
by traditional modelling approaches, particularly for As. Indeed, the deposition plume starts from
the power plant (E6) and develops over the east–west axis following the prevailing wind direction
blowing from the east (as modulated by the local orography) [
41
,
42
]. By contrast, the application of
previous scales failed to represent the actual variations in element depositions affecting the whole area,
especially for As. Regarding Cd, it should be preliminary stated that the rather low recovery (70%,
Section 2.3) could introduce a certain bias in element content results and cartographic output. Having
said that, previous and new scales identified a consistent pattern for sites located in the proximity of
shipbuilding activities (i.e., C6-7, D6-7, E6, F6, G6), but again, previous scales provided a certainly
more optimistic scenario (cf. sites D7 and G6; Figure 3).
The reasons for the general worse performance of the naturality/alteration scales have to be
sought in the source dataset. Indeed, this included rather old studies often reporting high element
concentration values, which consequently affected data distributions and resulted in a general
underestimation of pollutant depositions (cf. median values of Table 2with values corresponding
to 50th percentiles in Table S3), thus explaining the misleading outcome obtained for As. This is
further evidence that interpretative scales obtained through a meta-analytical approach may quickly
become obsolete as a result of rapidly changing scenarios, for example, variations in pollutant emissions
determined by a plethora of anthropogenic and non-anthropogenic causes (i.e., abatement or increasing
Atmosphere 2019,10, 136 13 of 19
traffic-related pollution, introduction of environmental protection measures, long-range atmospheric
transport, and so on [60–62]).
Atmosphere 2019, 10, x FOR PEER REVIEW 14 of 20
Figure 3. Cartographic representation of sampling sites and corresponding classes of the
naturality/alteration scale [16] (here, “previous scales”) and bioaccumulation scale (Table 4; here,
“new scale”), with indication of percentile thresholds (%ile), corresponding element concentration
values (left), and B ratios (right). Sampling sites are identified by alphanumeric codes (as also reported
in Table S5) followed by the letter F or X for Flavoparmelia caperata and Xanthoria parietina, respectively.
Figure 3.
Cartographic representation of sampling sites and corresponding classes of the
naturality/alteration scale [
16
] (here, “previous scales”) and bioaccumulation scale (Table 4; here,
“new scale”), with indication of percentile thresholds (%ile), corresponding element concentration
values (left), and Bratios (right). Sampling sites are identified by alphanumeric codes (as also reported
in Table S5) followed by the letter F or X for Flavoparmelia caperata and Xanthoria parietina, respectively.
Atmosphere 2019,10, 136 14 of 19
The accumulation/loss scale [
27
] and the brand-new bioaccumulation scale for lichen transplants
(Table 5) were both applied to the same case study. The concentration of As, Cd, and Cr significantly
increased in the samples of Pseudevernia furfuracea after 12-week exposure, although enrichment levels
were not indicative of strong contamination [45].
According to the previous scale [
27
], transplant sites were characterized by “normal” accumulation
(Table S6) in 66.7%, 46.7%, and 40% of cases for As, Cd, and Cr, respectively. Instead, “accumulation” or
“severe accumulation” occurred in 33.3% (As), 53.3% (Cd), and 60.0% (Cr) of cases. When the new scale
was applied, the great majority of sites were characterized by “Absence of bioaccumulation” or “Low
bioaccumulation”; in particular, 96.7% (As), 86.7% (Cd), and 90.0% (Cr). “Severe bioaccumulation”
was highlighted in samples exposed in a single site (3.3%) limited to As, whereas “Moderate
bioaccumulation” characterized 13.3% (Cd) and 10.0% (Cr) of sites.
The main limitations of the accumulation/loss scale are evident. Indeed, its use determines
(i) a heavy flattening of element concentration variations concerning enrichments exceeding 75%
(which are uncompromisingly identified as “severe accumulation”), and (ii) an exacerbation of slighter
variations (i.e., enrichments between 24% and 76%, which are considered to range between “normal”
and “severe” accumulation”). A case in point is represented by the highest values revealed for As in
the study area: indeed, the two highest exposed values were 0.37
µ
g g
−1
(EU = 1.50) and 2.34
µ
g g
−1
(EU = 9.35), measured in samples exposed in D5 and B2, respectively (Figure 4; Table S6). Using the
accumulation/loss scale, the difference between an increase of 50% (D5) and an increase of 835% (B2)
with respect to the unexposed levels is poorly reflected by a single class step (from “accumulation” to
“severe accumulation”; Table S4). By contrast, such a large difference is far better reflected by the three
class steps of the bioaccumulation scale (from “Low” to “Severe” bioaccumulation; Table 5).
Another issue inherent to the use of the previous scale concerns the precision achieved in
determining mean element concentration values of unexposed samples (i.e., the closeness of
agreement among the set of element concentration results [
63
]). A proper assessment of such
a reference value may indeed be a non-trivial task. Operators usually average element concentration
values measured in a certain number of samples taken from thalli randomly selected from bulked
lichen material. Obviously, this should be based on adequate sample size, which in turn should be
established on the basis of a preliminary characterization of the elemental concentration variability
of the target lichen matrix in the background site. However, the mean value of unexposed samples
is often assessed by analyzing too few samples (frequently n = 3), and this could have potential
interpretative consequences when using a scale based on classes of limited width such as the
accumulation/loss scale [
27
]. As a matter of fact, when element concentration values are few and
highly dispersed (e.g., coefficient of variation > 1), and especially in case of rather low enrichments,
the ascription of the EU value to a bioaccumulation class may result in a pointless procedure, as these
conditions would not guarantee repeatability.
Atmosphere 2019,10, 136 15 of 19
Atmosphere 2019, 10, x FOR PEER REVIEW 16 of 20
Figure 4. Cartographic representation of Pseudevernia furfuracea transplant sites, corresponding classes
of the accumulation/loss scale [27] (here, “previous scale”) and the bioaccumulation scale (Table 5;
here, “new scale”), with indication of EC ratios (Table 1) for the former, percentile thresholds (%ile),
and EU ratios for the latter. Transplant sites are identified by alphanumeric codes (as also reported in
Table S6).
Figure 4.
Cartographic representation of Pseudevernia furfuracea transplant sites, corresponding classes
of the accumulation/loss scale [
27
] (here, “previous scale”) and the bioaccumulation scale (Table 5;
here, “new scale”), with indication of EC ratios (Table 1) for the former, percentile thresholds (%ile),
and EU ratios for the latter. Transplant sites are identified by alphanumeric codes (as also reported in
Table S6).
4. Conclusions
In biomonitoring, interpretative scales are fundamental to the assessment of the magnitude
of pollution phenomena. Until now, scales based on very different assumptions have been
Atmosphere 2019,10, 136 16 of 19
developed: the so-called “naturality/alteration scales”, for biomonitoring with native lichens; and
the “accumulation/loss scale”, for transplant-based applications. Despite their popular use in
Italy and abroad, both scales were never critically reappraised, notwithstanding some evident
methodological flaws.
By recovering some core ideas from previous scales, we developed new interpretative scales
based on the meta-analysis of methodologically consistent bioaccumulation data from the most
recent Italian literature. The distributions of the ratios between element concentration data and
species-specific background (Bratio, native lichens) or element concentration of unexposed samples
(EU ratio, transplants) were analyzed. On this basis, two easily enforceable, percentile-based,
five-class “Bioaccumulation scales” were set up. A critical revision of scale-associated terminology
was also proposed. For both native lichens and transplants, the five classes refer to (1) “Absence
of bioaccumulation” (A), (2) “Low bioaccumulation” (L), (3) “Moderate bioaccumulation” (M),
(4) “High bioaccumulation” (H), and (5) “Severe bioaccumulation” (S), with Band EU ratio thresholds
corresponding to the 25th, 75th, 90th, and 95th percentiles of their distributions.
The comparative application of previous and new scales to two case studies suggested a better
and more consistent performance of the latter. Moreover, it also demonstrated that scales developed on
the basis of real biomonitoring data may become obsolete owing to changing scenarios, thereby leading
to the need for periodical updating with the inclusion of new available data to the source datasets.
Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4433/10/3/136/s1:
Supplementary material (including Methods S1, Tables S1–S6, and Figures S1–S3) and Supplementary Data
(Data S1–S4).
Author Contributions:
Conceptualization, M.T., E.C., L.F. (Lorenzo Fortuna), S.L., P.G., G.B., and L.F. (Luisa
Frati); methodology, E.C., L.F. (Lorenzo Fortuna), M.T., and S.L.; software, E.C. and L.F. (Lorenzo Fortuna); formal
analysis, E.C. and L.F. (Lorenzo Fortuna); investigation, E.C., L.F. (Lorenzo Fortuna), M.T., R.B., E.B., G.B., T.C.,
L.F. (Luisa Frati), L.D.N., P.G., S.L., F.M., S.M., J.N., L.P., S.R., and A.V.; resources, M.T., P.G., S.L., L.P., S.R.,
L.F. (Lorenzo Fortuna), G.B., L.F. (Luisa Frati), F.M., and A.V.; data curation, E.C. and L.F. (Lorenzo Fortuna);
writing—original draft preparation, E.C. and M.T.; writing—review and editing, E.C., M.T., P.G., S.L., G.B., L.F.
(Luisa Frati), R.B., S.M., F.M., L.P., J.N., L.F. (Lorenzo Fortuna), E.B., S.R., and A.V.; visualization, E.C. and M.T.;
supervision, M.T. and P.G.
Funding: This research received no external funding.
Acknowledgments:
Thanks are due to Elena Pittao (University of Trieste) for useful suggestions and critical
remarks provided during data collection and conceptualization, and to Lucy Sheppard for language revision.
This work was conceived and developed by the Working Group on Biomonitoring of the Italian Lichenological
Society (SLI).
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Bargagli, R.; Nimis, P.L.; Monaci, F. Lichen biomonitoring of trace element deposition in urban, industrial
and reference areas of Italy. J. Trace Elem. Med. Biol. 1997,11, 173–175. [CrossRef]
2.
Bargagli, R. Plants as biomonitors of atmospheric pollutants. In Trace Elements in Terrestrial Plants.
An Ecophysiological Approach to Biomonitoring and Biorecovering; Springer: Berlin, Germany, 1998; pp. 79–103.
3.
Ferretti, M.; Erhardt, W. Key issues in designing biomonitoring programmes. In Monitoring with
Lichens—Monitoring Lichens; Nimis, P.L., Scheidegger, C., Wolseley, P.A., Eds.; Springer: Dordrecht,
The Netherlands, 2002; pp. 111–139.
4.
Brunialti, G.; Frati, L. Bioaccumulation with lichens: The Italian experience. Int. J. Environ. Stud.
2014
,71,
15–26. [CrossRef]
5.
Brunialti, G.; Giordani, P.; Ferretti, M. Discriminating between the good and the bad: Quality assurance is
central in biomonitoring studies. In Environmental Monitoring; Wiersma, B., Ed.; CRC Press LLC: Boca Raton,
FL, USA, 2004; pp. 443–464.
6.
Hays, S.M.; Becker, R.A.; Leung, H.W.; Aylward, L.L.; Pyatt, D.W. Biomonitoring equivalents: A screening
approach for interpreting biomonitoring results from a public health risk perspective. Regul. Toxicol. Pharmacol.
2007,47, 96–109. [CrossRef]
Atmosphere 2019,10, 136 17 of 19
7.
Clewell, H.J.; Tan, Y.M.; Campbell, J.L.; Andersen, M.E. Quantitative interpretation of human biomonitoring
data. Toxicol. Appl. Pharmacol. 2008,231, 122–133. [CrossRef] [PubMed]
8.
Giordano, S.; Adamo, P.; Spagnuolo, V.; Tretiach, M.; Bargagli, R. Accumulation of airborne trace elements in
mosses, lichens and synthetic materials exposed at urban monitoring stations: Towards a harmonisation of
the moss-bag technique. Chemosphere 2013,90, 292–299. [CrossRef] [PubMed]
9.
Capozzi, F.; Giordano, S.; Aboal, J.R.; Adamo, P.; Bargagli, R.; Boquete, T.; Di Palma, A.; Real, C.; Reski, R.;
Spagnuolo, V.; et al. Best options for the exposure of traditional and innovative moss bags: A systematic
evaluation in three European countries. Environ. Pollut. 2016,214, 362–373. [CrossRef] [PubMed]
10.
Bargagli, R.; Nimis, P.L. Guidelines for the use of epiphytic lichens as biomonitors of atmospheric
deposition of trace elements. In Monitoring with Lichens—Monitoring Lichens; Nimis, P.L., Scheidegger, C.,
Wolseley, P.A., Eds.; Springer: Dordrecht, The Netherlands, 2002; pp. 295–299.
11.
Bennett, J.P. Statistical baseline values for chemical elements in the lichen Hypogymnia physodes.
In Environmental Pollution and Plant Responses; Agrawal, S.B., Agrawal, M., Eds.; CRC Press: Boca Raton, FL,
USA, 1999; pp. 343–353.
12.
Bargagli, R.; Sanchez-Hernandez, J.C.; Monaci, F. Baseline concentrations of elements in the Antarctic
macrolichen Umbilicaria decussata.Chemosphere 1999,38, 475–487. [CrossRef]
13.
Monaci, F.; Fantozzi, F.; Figueroa, R.; Parra, O.; Bargagli, R. Baseline element composition of foliose and
fruticose lichens along the steep climatic gradient of SW Patagonia (Aisén Region, Chile). J. Environ. Monit.
2012,14, 2309–2316. [CrossRef]
14.
Cecconi, E.; Incerti, G.; Capozzi, F.; Adamo, P.; Bargagli, R.; Benesperi, R.; Candotto Carniel, F.;
Favero-Longo, S.E.; Giordano, S.; Puntillo, D.; et al. Background element content of the lichen Pseudevernia
furfuracea: A supra-national state of art implemented by novel field data from Italy. Sci. Total Environ.
2018
,
622, 282–292. [CrossRef] [PubMed]
15.
Cecconi, E.; Incerti, G.; Capozzi, F.; Adamo, P.; Bargagli, R.; Benesperi, R.; Candotto Carniel, F.;
Favero-Longo, S.E.; Giordano, S.; Puntillo, D.; et al. Background element content in the lichen Pseudevernia
furfuracea: A comparative analysis of digestion methods. Environ. Monit. Assess. 2019, under review.
16.
Nimis, P.L.; Bargagli, R. Linee-guida per l’utilizzo di licheni epifiti come bioaccumulatori di metalli in
traccia. In Atti del Workshop “Biomonitoraggio della QualitàDell’aria sul Territorio Nazionale”; ANPA Serie Atti;
Piccini, C., Salvati, S., Eds.; Roma: Amsterdam, The Netherlands, 1999; pp. 279–289.
17.
Nimis, P.L.; Andreussi, S.; Pittao, E. The performance of two lichen species as bioaccumulators of trace
metals. Sci. Total Environ. 2001,275, 43–51. [CrossRef]
18.
Brunialti, G.; Frati, L. Biomonitoring of nine elements by the lichen Xanthoria parietina in Adriatic Italy:
A retrospective study over a 7-year time span. Sci. Total Environ. 2007,387, 289–300. [CrossRef] [PubMed]
19.
Demiray, A.D.; Yolcubal, I.; Akyol, N.H.; Çobano˘glu, G. Biomonitoring of airborne metals using the Lichen
Xanthoria parietina in Kocaeli Province, Turkey. Ecol. Ind. 2012,18, 632–643. [CrossRef]
20.
Pirintsos, S.; Bariotakis, M.; Kalogrias, V.; Katsogianni, S.; Brüggemann, R. Hasse diagram technique
can further improve the interpretation of results in multielemental large-scale biomonitoring studies of
atmospheric metal pollution. In Multi-Indicator Systems and Modelling in Partial Order; Brüggemann, R.,
Carlsen, L., Wittmann, J., Eds.; Springer: New York, NY, USA, 2014; pp. 237–251.
21.
Cocozza, C.; Ravera, S.; Cherubini, P.; Lombardi, F.; Marchetti, M.; Tognetti, R. Integrated biomonitoring
of airborne pollutants over space and time using tree rings, bark, leaves and epiphytic lichens. Urban For.
Urban Green. 2016,17, 177–191. [CrossRef]
22.
Paoli, L.; Winkler, A.; Guttová, A.; Sagnotti, L.; Grassi, A.; Lackoviˇcová, A.; Senko, D.; Loppi, S. Magnetic
properties and element concentrations in lichens exposed to airborne pollutants released during cement
production. Environ. Sci. Pollut. Res. 2017,24, 12063–12080. [CrossRef] [PubMed]
23.
Lucadamo, L.; Corapi, A.; Gallo, L. Evaluation of glyphosate drift and anthropogenic atmospheric trace
elements contamination by means of lichen transplants in a southern Italian agricultural district. Air Qual.
Atmos. Health 2018,11, 325–339. [CrossRef]
24.
Tretiach, M.; Baruffo, L. Contenuto di elementi in traccia in talli di Parmelia borreri eXanthoria parietina raccolti
sullo stesso forofita. Not. Soc. Lich. Ital. 2001,14, 70.
25.
Minganti, V.; Capelli, R.; Drava, G.; De Pellegrini, R.; Brunialti, G.; Giordani, P.; Modenesi, P. Biomonitoring
of trace metals by different species of lichens (Parmelia) in north-west Italy. J. Atmos. Chem.
2003
,45, 219–229.
[CrossRef]
Atmosphere 2019,10, 136 18 of 19
26.
Bergamaschi, L.; Rizzio, E.; Giaveri, G.; Loppi, S.; Gallorini, M. Comparison between the accumulation
capacity of four lichen species transplanted to a urban site. Environ. Pollut. 2007,148, 468–476. [CrossRef]
27.
Frati, L.; Brunialti, G.; Loppi, S. Problems related to lichen transplants to monitor trace element deposition in
repeated surveys: A case study from central Italy. J. Atmos. Chem. 2005,52, 221–230. [CrossRef]
28.
Loppi, S.; Giordani, P.; Brunialti, G.; Isocrono, D.; Piervittori, R. A new scale for the interpretation of lichen
biodiversity values in the Thyrrenian side of Italy. Bibl. Lichenol. 2002,82, 235–242.
29.
Baffi, C.; Bettinelli, M.; Beone, G.M.; Spezia, S. Comparison of different analytical procedures in the
determination of trace elements in lichens. Chemosphere 2002,48, 299–306. [CrossRef]
30.
Yafa, C.; Farmer, J.G. A comparative study of acid-extractable and total digestion methods for the
determination of inorganic elements in peat material by inductively coupled plasma-optical emission
spectrometry. Anal. Chim. Acta 2006,557, 296–303. [CrossRef]
31.
Tam, N.F.Y.; Yao, M.W.Y. Three digestion methods to determine concentrations of Cu, Zn, Cd, Ni, Pb, Cr,
Mn, and Fe in mangrove sediments from Sai Keng, Chek Keng, and Sha Tau Kok, Hong Kong. Bull. Environ.
Contam. Toxicol. 1999,62, 708–716. [CrossRef]
32.
Fortuna, L.; Tretiach, M. Effects of site-specific climatic conditions on the radial growth of the lichen
biomonitor Xanthoria parietina.Environ. Sci. Pollut. Res. 2018,25, 34017–34026. [CrossRef] [PubMed]
33.
Gaudet, C.L.; Bright, D.; Adare, K.; Potter, K. A rank-based approach to deriving Canadian soil and sediment:
Quality guidelines. In Species Sensitivity Distributions in Ecotoxicology, 1st ed.; Posthuma, L., Suter, G.W., II,
Traas, T.P., Eds.; CRC Press: Boca Raton, FL, USA, 2001; pp. 255–273.
34.
Reimann, C.; Garrett, R.G. Geochemical background—Concept and reality. Sci. Total Environ.
2005
,350,
12–27. [CrossRef] [PubMed]
35.
Facchinelli, A.; Sacchi, E.; Mallen, L. Multivariate statistical and GIS-based approach to identify heavy metal
sources in soils. Environ. Pollut. 2001,114, 313–324. [CrossRef]
36.
Massas, I.; Ehliotis, C.; Gerontidis, S.; Sarris, E. Elevated heavy metal concentrations in top soils of an Aegean
island town (Greece): Total and available forms, origin and distribution. Environ. Monit. Assess.
2009
,151,
105–116. [CrossRef]
37.
Yang, C.L.; Wu, Z.F.; Zhang, H.H.; Guo, R.P.; Wu, Y.Q. Risk assessment and distribution of soil Pb in
Guandong, China. Environ. Monit. Assess. 2009,159, 381–391. [CrossRef]
38.
Reimann, C.; de Caritat, P. Establishing geochemical background variation and threshold values for 59
elements in Australian surface soil. Sci. Total Environ. 2017,578, 633–648. [CrossRef]
39.
Hoaglin, D.C.; Iglewicz, B.; Tukey, J.W. Performance of some resistant rules for outlier labeling. J. Am.
Stat. Assoc. 1986,81, 991–999. [CrossRef]
40.
Cicchitelli, G.; Herzel, A.; Montanari, G.E. Il Campionamento Statistico, 2nd ed.; Il Mulino: Bologna, Italy, 1992.
41.
ARPA FVG. Available online: http://www.arpa.fvg.it/cms/hp/news/Studio-licheni-centrale-a2a.html
(accessed on 19 February 2019).
42.
Fortuna, L. Development of Biomonitoring Techniques of Persistent Airborne Pollutants Using Native
Lichens. Ph.D. Thesis, University of Trieste, Trieste, Italy, 27 March 2018.
43.
EU Directive 2008/50/EC. Available online: https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=
CELEX:32008L0050&from=en (accessed on 19 February 2019).
44.
IARC. 2018. Available online: https://monographs.iarc.fr/wp-content/uploads/2018/06/mono100C- 9.pdf
(accessed on 19 February 2019).
45.
Tretiach, M.; Candotto Carniel, F.C.; Loppi, S.; Carniel, A.; Bortolussi, A.; Mazzilis, D.; Del Bianco, C.
Lichen transplants as a suitable tool to identify mercury pollution from waste incinerators: A case study
from NE Italy. Environ. Monit. Assess. 2011,175, 589–600. [CrossRef] [PubMed]
46. Nimis, P.L. The Lichens of Italy. A Second Annotated Catalogue; EUT: Trieste, Italy, 2016.
47.
Incerti, G.; Cecconi, E.; Capozzi, F.; Adamo, P.; Bargagli, R.; Benesperi, R.; Candotto Carniel, F.; Cristofolini, F.;
Giordano, S.; Puntillo, D.; et al. Infraspecific variability in baseline element composition of the epiphytic
lichen Pseudevernia furfuracea in remote areas: Implications for biomonitoring of air pollution. Environ. Sci.
Pollut. Res. 2017,24, 8004–8016. [CrossRef] [PubMed]
48.
Weltje, L.; Sumpter, J.P. What makes a concentration environmentally relevant? Critique and a proposal.
Environ. Sci. Technol. 2017,51, 11520–11521. [CrossRef] [PubMed]
49.
Wolterbeek, H.T.; Bode, P. Strategies in sampling and sample handling in the context of large-scale plant
biomonitoring surveys of trace element air pollution. Sci. Total Environ. 1995,176, 33–43. [CrossRef]
Atmosphere 2019,10, 136 19 of 19
50.
Nieboer, E.; Richardson, D.H.S.; Tomassini, F.D. Mineral uptake and release by lichens: An overview.
Bryologist 1978,81, 226–246. [CrossRef]
51.
Nash, T.H., III. Nutrients, elemental accumulation, mineral cycling. In Lichen Biology, 2nd ed.; Nash, T.H., III,
Ed.; Cambridge University Press: Cambridge, UK, 2008; pp. 236–253.
52.
Brown, D.H.; Beckett, R.P. Uptake and effect of cations on lichen metabolism. Lichenologist
1984
,16, 173–188.
[CrossRef]
53.
Paoli, L.; Vannini, A.; Faˇckovcová, Z.; Guarnieri, M.; Ba ˇckor, M.; Loppi, S. One year of transplant: Is it
enough for lichens to reflect the new atmospheric conditions? Ecol. Ind. 2018,88, 495–502. [CrossRef]
54.
Mikhailova, I. Transplanted lichens for bioaccumulation studies. In Monitoring with Lichens—Monitoring
Lichens; Nimis, P.L., Scheidegger, C., Wolseley, P.A., Eds.; Springer: Dordrecht, The Netherlands, 2002;
pp. 301–304.
55.
Gailey, F.A.Y.; Lloyd, O.L. Methodological investigations into low technology monitoring of atmospheric
metal pollution: Part 2—The effects of length of exposure on metal concentrations. Environ. Pollut. B
1986
,
12, 61–74. [CrossRef]
56.
Bargagli, R.; Mikhailova, I. Accumulation of inorganic contaminants. In Monitoring with Lichens—Monitoring
Lichens; Nimis, P.L., Scheidegger, C., Wolseley, P.A., Eds.; Springer: Dordrecht, The Netherlands, 2002;
pp. 65–84.
57.
Zhang, J.; Zhao, Y.; Ding, F.; Zeng, H.; Zheng, C. Preliminary study of trace element emissions and control
during coal combustion. Front. Energ. Power Eng. China 2007,1, 273–279. [CrossRef]
58.
Wang, C.; Liu, H.; Zhang, Y.; Zou, C.; Anthony, E.J. Review of arsenic behavior during coal combustion:
Volatilization, transformation, emission and removal technologies. Prog. Energy Comb. Sci.
2018
,68, 1–28.
[CrossRef]
59.
ARIANET. Available online: http://www.a2aenergiefuture.eu/gruppo/export/sites/default/ene/
impianti/documenti_impianti/ARIANET-R2014.23-MONFALCONE.pdf (accessed on 19 February 2019).
60.
Wilson, N.; Horrocks, J. Lessons from the removal of lead from gasoline for controlling other environmental
pollutants: A case study from New Zealand. Environ. Health 2008,7, 1. [CrossRef] [PubMed]
61.
Madsen, C.; Gehring, U.; Håberg, S.E.; Nafstad, P.; Meliefste, K.; Nystad, W.; Lødrup Carlsen, K.C.;
Brunekreef, B. Comparison of land-use regression models for predicting spatial NOx contrasts over a
three year period in Oslo, Norway. Atmos. Environ. 2011,45, 3576–3583. [CrossRef]
62.
Kollanus, V.; Tiittanen, P.; Niemi, J.V.; Lanki, T. Effects of long-range transported air pollution from vegetation
fires on daily mortality and hospital admissions in the Helsinki metropolitan area, Finland. Environ. Res.
2016,151, 351–358. [CrossRef] [PubMed]
63.
Gotelli, N.J.; Ellison, A.M. A Primer of Ecological Statistics, 2nd ed.; Sinauer Associates: Sunderland, MA,
USA, 2012.
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
