Two lichens differing in element concentrations have similar spatial patterns of element concentrations responding to road traffic and soil input

Two epiphytic lichens (Xanthoria alfredii, XAa; X. ulophyllodes, XAu) and soil were sampled at three sites with varied distances to a road in a semiarid sandland in Inner Mongolia, China and analyzed for concentrations of 42 elements to assess the contribution of soil input and road traffic to lichen element burdens, and to compare element concentration differences between the two lichens. The study showed that multielement patterns, Fe:Ti and rare earth element ratios were similar between the lichen and soil samples. Enrichment factors (EFs) showed that ten elements (Ca, Cd, Co, Cu, K, P, Pb, S, Sb, and Zn) were enriched in the lichens relative to the local soil. Concentrations of most elements were higher in XAu than in XAa regardless of sites, and increased with proximity to the road regardless of lichen species. These results suggested that lichen element compositions were highly affected by soil input and road traffic. The narrow-lobed sorediate species were more efficient in particulate entrapment than the broad-lobed nonsorediate species. XAa and XAu are good bioaccumulators for road pollution in desert and have similar spatial patterns of element concentrations for most elements as response to road traffic emissions and soil input.

REE patterns. Figure 3a shows the chondrite-normalized REE distribution patterns of the lichen and soil samples, upper continental crust (UCC), post-Archean Australian shale (PAAS) and argillaceous rocks in the eastern part of China (ECA). Despite great REE concentration differences, patterns of these samples are roughly similar to one another.       and XAu (n = 21) and among sites (S1: 100-200 m from the road; S2: 400-500 m; S3: 900-1,000 m). Statistic: "s" and "ns" denote the significant effects and nonsignificant effects at α = 0.05, respectively. Different capitalized letters denote the significant differences in concentration between XAa and XAu. Different small letters denote the significant differences in concentration among sites. Elements in bold are enriched elements with a mean EF of > 3.0 (Fig. 2b). XAa-Xanthoria alfredii, XAu-X. ulophyllodes.
Scientific Reports | (2020) 10:19001 | https://doi.org/10.1038/s41598-020-76099-x www.nature.com/scientificreports/ Cs, Nb, Sr, and Tl are separated from the other elements at a correlation similarity of 0.80 (Fig. 4a). There is a significant interaction effect of the lichen species and sites on concentrations of these metals (Fig. 4b). The concentrations of these metals tend to decrease with distance to the road in XAu but are nearly identical among the sites in XAa; XAu is higher than XAa at site S1, but XAu and XAa are not significantly different at sites S2 and S3 (Fig. 4b).
The remaining 36 elements form a cluster at a correlation similarity matrix of > 0.80 (Fig. 4a). There is no significant interaction effect of the lichen species and sites on the concentrations of these metals (Fig. 4b). Concentrations of these elements are higher in XAu than in XAa (XAu:XAa = 1.12-1.66) and are higher in site S1 than in S2 and/or S3, with the exception of 3 elements (K, P, and S), of which the concentrations are not significantly different between the sites (Fig. 4b).

Discussion
Deposition degree. OSES can be considered a fairly contaminated place when comparing the lichen data with those of epiphytic lichens in other studies. The concentrations of most elements in the lichen samples are higher than or similar to those in epiphytic lichens from the desertified sites or sites near roads (Supplementary  Table S1), such as similar ecosystems in Xilinhot, Inner Mongolia, China 17,18 . This finding is also the case when the data in this study are compared with the data from epiphytic lichens near roads in Turkey 9,29 , India 30 , and France 31 (Supplementary Table S1). However, our data of most elements are lower than or at the lower range of 26 elements in Flavopunctelia soredica transplanted along the two busy roads in a highly polluted area of Hebei, China 7 (Supplementary Table S1).
Soil contribution. Thirty-two elements (Al, Ba, Ce, Cs, Dy, Er, Eu, Fe, Gd, Ho, La, Lu, Mg, Mn, Na, Nb, Nd, Ni, Pr, Rb, Sc, Sm, Sr, Tb, Th, Ti, Tl, Tm, U, V, Y, and Yb) in XAa and XAu are highly affected by soil input. These elements show similar multielement patterns between the lichen and soil samples (Fig. 2a) and have EFs of < 3.0 (Fig. 2b). An EF of < 3.0 suggests crustal input 19,32 . In this ecosystem, the vegetation is sparse, and the soil is vulnerable to wind erosion. Most of these elements, such as Al, Fe, Rb, Sc, Ti and lanthanoids are attributed to windblown dust input in similar ecosystems of Inner Mongolia 17,18 . Fe:Ti ratios are similar among XAa (5.90 ± 0.24), XAu (5.66 ± 0.13), and soil (5.90 ± 0.51; Table 1) in all sites, suggesting the trapping of local soil particulates in lichen thalli 26,27 . In a similar ecosystem of Inner Mongolia, the similar Fe:Ti ratios between the epiphytic foliose lichens (Phaeophyscia hirtuosa and XAa; 12.30-13.12) and the local soil samples (12.27) are attributed to an entrapment of windblown soil particulates in lichen thalli 18 .
The high soil contribution is also supported by the REE patterns (Fig. 3). The lichen and soil samples, UCC, PAAS and ECA have roughly similar REE distribution patterns (Fig. 3a) (Fig. 3b), indicating that the REE composition in the lichen samples is highly related to local soil. This conclusion agrees with the results of other studies: the REE accumulation in mosses and lichens is attributed to soil dust deposition 20,21,28,33,34 .
The soil contribution to lichen element burdens might be marked by redeposition of local soil contaminants from human activities such as coal mining and transport. The road traffic effect is evident in data of the 22 elements (the 14 lanthanoids, Al, Na, Ni, Sc, Th, U, V, and Y), which are closely correlated (Fig. 4a) and have the highest concentrations at the site close to the road (S1) regardless of lichen species (Fig. 4b). This spatial pattern is also the case for the other 4 elements (Sr, Nb, Cs, and Tl) observed in XAu (Fig. 4b). Other lichen biomonitoring studies conducted close to roads also have found similar distance-dependent concentration patterns attributed to the enhancement of the deposition/redeposition of soil dust by traffic 7,12,18,30 . Enriched elements and road traffic effects. The results of EFs (Fig. 2b) show that ten elements (Ca, Cd, Co, Cu, K, P, Pb, S, Sb, and Zn) are enriched in XAa and XAu relative to the local soil. An EF of > 3.0 is an indicator of anthropogenic and/or nonlocal sources or bioregulation of these elements in lichen thallus 19,32 .
Five enriched metals (Cd, Cu, Pb, Sb, and Zn) are likely to have come from traffic emissions. These metals are typical traffic-related pollutants emitted through fossil fuel combustion, fuel additives, tire and brake pad abrasion, corrosion, and lubricating oils 4,6 . These metals have the highest concentrations at S1 and lowest concentrations at S3, regardless of lichen species (Fig. 4b). In Negev deserts, the amount of Pb in lichens has been found higher at one site close to a road than at other sites 14 . The higher concentrations of these metals in lichens close to roads or at sites with high traffic levels have also been reported in other studies 7-10,30,35-37 .
The concentrations of 4 enriched elements (Co, K, P, and S) did not undergo any changes with distance from the road regardless of lichen species (Fig. 4b). One explanation for the spatial pattern of S may be the impact of the coal emissions. Sulfur is rich in coals and is an important contaminant during coal combustion in China 6 . Sulfur emissions from several coal mines around the study site may represent a significant source of S in lichens and surface soils. Bioregulation of these essential nutrients in lichen thallus may also be responsible for this pattern. In fact, the trends of nutrients are often different from or even inverse to those of pollutants in lichens. For example, concentrations of traffic-related heavy metals increase with proximity to the road, while some nutrients (K, P and Mn) show a reverse trend due to nutrient leakage as a result of road pollution 8 . The metals (Cu, Pb, and Zn) in Xanthoparmelia scabrosa decrease from urban to rural areas, whereas three nutrients (K, P, and S) show an inverse pattern 38 .
Calcium appears to come from traffic-related dust redeposition superimposed on local soil deposition. The spatial pattern for Ca is similar to typical soil-derived metals such as Ti and Sc (Fig. 4a, b). This metal is seldom released by vehicle emissions. The enrichment of Ca in XAa and XAu (Fig. 2b)  Lichen species differences. The research results show a species-and element-specific accumulation of elements in lichens. The narrow-lobed sorediate lichen XAu (Fig. 1e) has a higher accumulation capability for 40 elements (all elements barring Rb and Co; Fig. 4b) than the broad-lobed nonsorediate lichen XAa (Fig. 1f). These results are in accordance with the other studies suggesting that the presence of soredia and narrower lobes allows a higher surface/volume ratio to enhance the capability of the entrapment and retention of atmospheric particles [1][2][3][4]40 . The degree of concentration difference between XAu and XAa is highest for the 14 lanthanoids (XAu:XAa: 1.32-1.66) and lowest for the 5 elements (Co, Cs, K, P, and Rb; XAu:XAa: 1.06 ~ 1.12; Table 1, Fig. 4b). Other studies also reported that different lichens accumulate different elements to different extent 2 .
Despite the species-and element-specific contrasts in element concentrations, XAa and XAu share similar multielement patterns (Fig. 2a), EFs (Fig. 2b) and REE patterns (Fig. 3), and show similar concentration trends with the variation of distance from the road for most elements (Fig. 4b). These results are consistent with those of other studies reporting that the element concentration differences among lichen species mainly manifest different accumulation rates, but the spatial/temporal trends of individual elements remain similar 7,8,21,26 .

Conclusions
The element compositions in XAa and XAu are highly affected by road traffic and local soil. Five metals (Cd, Cu, Pb, Sb, and Zn) accumulated in lichens can be traced to traffic emissions. Local soil input has great influence on the concentrations of 33 elements (Al, Ba, Ca, Ce, Cs, Dy, Er, Eu, Fe, Gd, Ho, La, Lu, Mg, Mn, Na, Nb, Nd, Ni, Pr, Rb, Sc, Sm, Sr, Tb, Th, Ti, Tl, Tm, U, V, Y, and Yb) in lichen thalli and their content reaches highest in the places close to the roads due to the redeposition of road dust. Concentrations of 4 nutrients (Co, K, P, and S) in XAa and XAu show little changes with proximity to the road, possibly due to the interaction between lichen physiology and air pollution. Concentrations of the most elements are higher in XAu than those in XAa. The two lichens can serve as bioaccumulators to monitor atmospheric element deposition near roads in deserts and yield similar spatial patterns of element concentrations in most cases.

Methods
Investigation area. Ordos Sandland Ecological Station (N 39°29′, E 110°11′; OSES, Institute of Botany, Chinese Academy of Sciences) is located at Mu Us Sandland, southeastern Ordos Plateau, Inner Mongolia, China (Fig. 1a, b). This area has a semiarid monsoon climate with a mean annual evaporation of 2093 mm. The mean annual precipitation is 350-380 mm, largely (60-80%) in the form of rainfall during June to August. The elevation is approximately 1290 m a.s.l. The soil is sandy loam and aeolian sandy soil. The region has been severely desertified due to overgrazing, mining and other anthropogenic activities and is one of the most important sources of sand dust storms in China. In 2013, the landscape was characterized by semifixed and moving sand dunes with patches of cultivated trees (mainly Poplus spp.), psammophytic shrubs and herbs.
The station lies at a rural site approximately 35 km from the nearest city (Fig. 1b). However, the station is surrounded with several coal mines and mine tailings and is adjacent to industrial roads for coal transportation (Fig. 1c). The nearest coal mine is 3 km away and its operation commenced in Dec 2009. About a dozen workers and students stayed at the station mainly from late May to early October. There were some private paths with few, if any, vehicles (Fig. 1d).
Sample collection. XAa, XAu and soil were sampled during 8-10 August 2013. To investigate the road traffic effects on lichen element burdens, three sites of 100 × 800-1000 m each were selected at an increasing distance from the road: S1 (100-200 m from the road), S2 (400-500 m), and S3 (900-1000 m), with the longest side parallel to the industrial road. The area within 100 m of the road was not included because there were very few trees and epiphytic lichens (Fig. 1d).
In each site, 6-8 homogeneous plots [i.e., the plots had Poplar trees with uniform density, similar stem diameter (15-20 cm) and abundant lichen individuals], each with an area of 5-8 × 5-8 m, were selected for each of the two lichens. Each plot was represented by a single composite sample made up of 15-25 thalli (6-10 g dw) randomly collected from all aspects of 3-5 Poplar trees at a height of approximately 1.0-2.0 m from the ground by using a knife. An influence of inter-individual differences in size, age, or microclimatic factors on lichen element concentrations is nonnegligible 31 . Thus the large composite samples has been frequently adopted in the biomonitoring studies to reduce the effects of sample/habitat heterogeneity 28,41,42 . Due to the complex vegetation conditions and the high dependency of plot selection on the availability of trees and lichens, the experiment is an unbalanced design with unequal number of samples for XAa and XAu in each site. For most cases, XAa and XAu were collected from different plots. A total of 41 composite samples were collected, with 20 for XAa and 21 for XAu (Table 2).
Three samples of approximately 100 g of shallow (5 cm deep) neighboring soil, each composed of five subsamples, were also randomly collected in each site. All samples were placed in plastic bags to prevent contamination and were taken to the laboratory for later identification and analysis.
Sample preparation and measurement. Apothecia of XAa were removed manually. All samples were carefully cleaned under a low-powered stereomicroscope, dried in oven to a constant weight at 60 °C for 72 h, ground and homogenized in a grinding mill equipped with tungsten carbide jars (Retsch MM400; Retsch GmbH, Haan, Germany). Aliquots of 200-300 mg of each homogenized sample were mineralized in a mixture of HNO 3  where A and B are the elements in question, the subscript "NC" indicates that the samples are normalized to the chondrite values 44 , and the subscripts "sample" and "chondrite" indicate which medium the concentration refers to.
The average values of the upper continental crust (UCC) 45 , post-Archean Australian shale (PAAS) 44 and argillaceous rocks in the eastern part of China (ECA) 46 are used for comparison in the study of REE distribution and fractionation.
The enrichment factor (EF) is calculated according to Eq. (2): where X is the element in question, Al is the reference crustal element, and the subscripts "lichen" or "soil" indicate which medium the concentration refers to.

Statistical analyses.
Concentrations of each element are tested for normality using Shapiro-Wilk's test and for homogeneity of variance using Levene's test (α = 0.05). For each of the three sites, an independent samples t test is conducted to check whether the element concentration and Fe:Ti ratio (log10-transformed) in the soil are significantly different between sites and significantly different from those in XAa and in XAu (α = 0.05). The raw concentrations of the lichen combined dataset and the soil samples are z-score standardized [(x-mean)/SD] respectively for subsequent analyses. A cluster analysis is conducted with the unweighted pairgroup method plus arithmetic means (UPGMA) linkage method based on the correlation distance as a measure of similarity. A two-way analysis of variance (ANOVA) is performed to test the main and interactive effects of the lichen species (fixed factor of two levels, either XAa or XAu) and sites (fixed factor of three levels: either S1, S2, or S3) on each element (α = 0.05). A Tukey's honestly significant difference (HSD) test is conducted for post hoc comparisons. Harmonic means are used in this analysis to correct the variations in sample size. A simple effect analysis is conducted in the case of significant interactive effects. All statistical analyses are performed using PAST 3.26 software (Ø. Hammer, April 2019). Plots are generated using PAST 3.26 software and Inkscape 0.92 software (Free Software Foundation Inc., USA).