Introduction

A constant increase in human population has led to the increase in anthropogenic pollution of the environment. Anthropogenic sources of pollution can be different: industry, traffic, urban development, intense agriculture, etc. All of these sources have a negative influence on the environment. Emitted dust, smoke, industrial and exhaust gases all heavily burden the atmosphere, soil, water and living organisms (Migula et al. 1991; Diels et al. 2002). One of the results of these processes is increasing metal concentrations in soil, water, air and other elements of the environment (Diels et al. 2002; Free et al. 1983; Wang et al. 2001).

Thermal power plants (TPP) have a long history of negative environmental impact due to their emission of particulate matter, inorganic and organic pollutants. During coal combustion, a significant amount of various metals, some of which are very toxic (such as Pb, Cd, Hg), is emitted to the environment through different pathways. Numerous investigations and publications are already available, mainly elaborating monitoring results based on air and ash samples which were directly collected in the vicinity of the TPPs (Davison et al. 1974; Goodarzi 2006; Hower et al. 1996; Meij 1994, Meij and te Winkel 2007; Pavageau et al. 2002; Pudasainee et al. 2010).

Urban areas are burdened with pollution which comes from various sources. Traffic is one of the major pollution sources in city centers (Silva et al. 2014). Burning of coal, oil and natural gas for purposes of heating is another source of urban pollution (Junninen et al. 2009).

As the population on the planet grows, so grows the need for food. To produce more food, farmers started using more fertilizers, herbicides, pesticides, fungicides and insecticides. Many of them leave metals as pollutants in the soil, air and water of agricultural areas (Mortvedt 1996).

Constantly growing pollution requires more sophisticated and efficient monitoring methods, such that can cover bigger areas, give faster results and cost less. Consequently, interest in use of bioindicators that own all these characteristics, for the detection and evaluation of environmental contaminants, has increased in the past few decades. Several different animal and plant species have been suggested as suitable for monitoring environmental pollution (Akguc et al. 2010; García-Hernández et al. 2015; Kaonga et al. 2010; Ugulu et al. 2012;). Since 1970, honeybees (Apis mellifera) have been increasingly used to monitor toxic metal environmental pollution (Bromenshenk et al. 1985; Crane 1984; Leita et al. 1996; Porrini et al. 2002).

Metals generally do not cause bee mortality, but they can accumulate in their body. Thus, the contamination in the form of metals accumulating in the honeybee’s body may reflect the contamination of plants, soil, air and water in the areas that are covered by foraging activities of bees (Sadeghi et al. 2013). Suspended particulate matter (PM), originated from TPP fly ash, enrich the atmosphere and contaminate the ground surface with toxic metals present in ash. PM and toxic metals contained in it can directly deposit on the hairy body of the bees or can reach the insect by nectar, pollen, honeydew, or through the water they drink during foraging.

Honeybees have been proposed as suitable bioindicators of metal pollution by many researchers (Crane 1984; Leita et al. 1996; Perugini et al. 2011; Porrini et al. 2003; Raes et al. 1992; Sadeghi et al. 2013; Tong et al. 1975; Wallwork-Barber et al. 1982). Bees, as foragers, fly over a 7-km2 area and usually complete 12–15 foraging trips per day (Perugini et al. 2011). As suggested by Leita et al. (1996), a network of apiaries located within polluted areas can supply data for the constant monitoring of toxic metal emissions from either point or diffuse sources. Most of these results prove that bees are good bioindicators, but not many results demonstrate the capability of bees as bioindicators to significantly distinguish different concentrations of metals in different foraging regions.

All organisms have requirements for minerals which are essential for their growth and development, and honeybees are no different. Even newly emerged bees have same mineral content in their bodies (Black 2006). The mineral content in young bees can be related to the diet they were exposed to. But since they did not fly out of the hive, this content does not represent the full impact of the environment. Diet exposure is only one of the ways bees are exposed to the environment contamination. Some of the other ways bees accumulate contaminants are directly on their body, by water they drink, or by pollen and nectar they process. That is the reason why forager bees are selected for the analysis.

The aim of this study was to determine levels of Al, Ba, Cd, Co, Cr, Cu, Fe, Li, Mn, Na, Ni, Pb, Sr and Zn in bodies of honeybees in different time of year and at sampling sites characterized by different environmental impact. In this study, we determined impact of the environment on the contents of the analyzed elements in bodies of honeybees.

Materials and methods

Sampling sites and sampling

The research materials were samples of honeybees obtained from six stationary apiaries—three tightly bound to TPP Kostolac, two from the urban area of Belgrade (Capital of Serbia) and one from mainly agricultural–woodland area near the village of Mesić (eastern border of Serbia with Romania).

Kostolac (TPP)

Kostolac is some 60 km to the East from Belgrade. Three apiaries (site number: 1, 2 and 3) are located around two thermal power plants located in municipality Kostolac, Braničevo district, Serbia (Fig. 1). The two power plants are roughly 3.5 km apart. This means that the bees from all apiaries are affected by both power plants. Sampling site 1 was 2 km to the south-southwest from thermal power plant Kostolac A, located in a small town of Kostolac, and 4.5 km southwest from thermal power plant Kostolac B. The sampling site is surrounded by mostly agricultural land. Sampling site 2 is in the village Drmno located 1.5 km to the south of power plant Kostolac B, and 4 km east from thermal power plant Kostolac A. The village is surrounded by an open pit coal mine. Sampling site 3 is located 1.7 km to the north-northeast of thermal power plant Kostolac A and 2.5 km to the northwest of thermal power plant Kostolac B. Ash disposal site for the two power plants is 500 m to the north-northwest of sampling site 3. The area covered with these three apiaries is about 40 km2. Prevalent wind direction for TPP region during the sampling period is given in Fig. 2. Most frequent wind direction was southeast, and the second most frequent wind was north.

Fig. 1
figure 1

Map of sampling sites

Full size image
Fig. 2
figure 2

Frequencies of prevalent wind directions during the sampling period for TPP and Belgrade (BG) urban region

Full size image

Belgrade (BG)

Two sampling sites (4 and 5) are located in the urban area of Belgrade (Fig. 1). Both of those sites are located in the center of the city. Sampling site 4 is located adjacent to a highway, and sampling site 5 is located in Zemun, northern part of the city. The distance between the sampling sites is 7 km. Most frequent wind directions during the sampling period were west and northwest (Fig. 2).

Mesić (MS)

Sampling site (6), located near the small village Mesić, municipality Vršac, South Banat District, Serbia, contained one single apiary. The sampling locality is in a rural region containing mainly agricultural–woodland areas. To the north it is bordering a protected area, “Landscape of outstanding features ‘Vršačke Planine.’” To the west vineyards cover an area of around 600 ha. From the other sides, it is surrounded by agricultural land.

Sampling was performed twice during 2014. The first sampling was done at the beginning of July and second at the beginning of September. Samples were taken from two hives at each of the sampling sites, except at Mesić, where samples were taken from three hives. Between 5 and 10 g of sample (50–100 bees) was collected from the outer frame of the hive that was occupied with bees but without brood. Samples were directly transferred into sterile plastic containers. The collected bees were killed in the laboratory by instant freezing and kept at −21 ± 3 °C until analyzed.

Sample preparation

Frozen samples of bees were first dried in open air for 24 h to remove any residual ice. They were measured and dried in the oven at 60 °C until constant mass. A test portion of dried sample between 0.5 and 1.0 grams was taken and digested according to the US EPA SW-846 Method 3052, under high pressure in closed Teflon vessels, with 7 ml of concentrated HNO3 and 2 ml of concentrated H2O2. Mineralization was performed in a closed microwave digestion system (ETHOS 1, Advanced Microwave Digestion System, Milestone, Italy) by heating the samples up to 200 °C (15 min), followed by another 15 min at the same temperature. Each sample was cooled, transferred to a 25-mL volumetric flask and diluted to 25 mL with deionized water.

Instrumental analysis

Quantitative determination of metals was performed by inductively coupled plasma–optical emission spectrometry, ICP-OES (iCAP 6500Duo, Thermo Scientific, Cambridge, UK). Multi-element standard solutions were used (Multi-Element Plasma Standard Solution 4, Specpure®Alfa Aesar, Karlsruhe, Germany) for the determination of elements of interest. Quality control was based on the analysis of blanks (prepared following the whole sample preparation procedure), duplicates and analysis of the standard solutions (Rakić et al. 2013).

Statistical analysis

The results from analyses were statistically processed. Mean concentrations and standard deviations of elements concentrations for the three regions, over the entire study period, were calculated. The assessment of the data normality was done by Kolmogorov–Smirnov test. For some elements, data were not normally distributed; therefore, they were log-transformed and normality was reassessed. The results were processed by two-way analysis of variance (ANOVA), followed by Tukey’s HSD (Honestly Significant Difference) multiple comparisons test. Differences in concentrations between dates or locations were considered significant if p values were ≤0.05. Differences in concentration between sampling sites within TPP region (sampling sites 1, 2 and 3) and urban region (sampling sites 4 and 5) were also assessed.

Results and discussion

In the first step, concentrations of metals found in honeybee bodies were analyzed disregarding the sampling sites from which the bees were harvested. The range and the average values of analyzed metal concentrations are given in Table 1.

Table 1 Range and mean concentrations of metals in adult honeybees [mg kg−1 dry matter (d.m.)]
Full size table

Most of analyzed metals are above detection limits. This is good indication that honeybees are excellent bioindicators for metal pollution of the environment.

A comparison of the results for individual sites, presented in Table 2 and Fig. 3, enables us to compare the levels of contamination and to determine whether statistically significant differences are present between metal concentrations determined in bees collected from agricultural–woodland (MS—Mesić) region, urban (BG—Belgrade) region and thermal power plant region (TPP—Kostolac). Bees from the same apiary may fly in different directions and sample different sites showing different levels of contamination (Formicki et al. 2013). This can explain relatively big standard deviations for the analyzed elements in this study.

Table 2 Minimum, maximum and mean metal concentrations in the bodies of honeybees (mg kg−1 d.m.) and standard deviation in samples from the three regions for the entire study period
Full size table
Fig. 3
figure 3

Mean concentration comparison of all analyzed elements for three regions: BG—urban region, MS—agricultural–woodland region, TPP—thermal power plant region (shown in log scale). Different uppercase letters indicate per location statistically significant differences between metal concentrations over the entire study period

Full size image

Na has the highest concentration of all the elements discussed in this study. An average Na concentration was 426 mg kg−1. It is followed by Fe, Zn, Al, Mn, Cu, Sr, Ba, Ni, Pb, Cr, Cd, Co and Li. (Table 1).

Statistically significant differences between analyzed regions are represented by different uppercase letters in Fig. 3. Significant differences were found for Al, Ba, Cr, Cu, Fe, Li and Ni between at least two of analyzed regions. Cd, Co, Mn, Na, Pb, Sr and Zn showed no significant differences between analyzed regions. For Al, significant differences were found between the TPP region and agricultural–woodland region, as well as between TPP region and urban region. Mean concentration of Al was 113 mg kg−1 d.m. in TPP region, 24.04 mg kg−1 d.m. in urban region, and 29 mg kg−1 d.m. in agricultural–woodland region. Ba shows significant differences between agricultural–woodland region and TPP region, with mean values of 2.77 and 1.74 mg kg−1 d.m., respectively. Cr showed significant differences between TPP region and other two regions agricultural–woodland region and urban region. The mean concentration of Cr in TPP region was 0.36 mg kg−1 d.m. In urban and agricultural–woodland region, Cr had the same mean concentrations of 0.15 mg kg−1 d.m. Cu showed significant differences between all three regions, TPP, urban and agricultural–woodland region, with mean concentrations of 17.1, 25.7 and 21.8 mg kg−1 d.m., respectively. For Fe, there were significant differences between the TPP region and agricultural–woodland region, and TPP and urban region. TPP region had the highest mean concentration of Fe which was 247 mg kg−1 d.m., followed by urban region with 169 mg kg−1 d.m. and agricultural–woodland region 139 mg kg−1 d.m. Li showed significant difference between TPP region and urban region, with mean concentrations of 0.092 and 0.014 mg kg−1 d.m., respectively. Statistically significant differences for Ni were found between agricultural–woodland region and TPP region, as well as between agricultural–woodland region and urban region. Mean Ni concentration in agricultural–woodland region was 1.76 mg kg−1 d.m., in urban region 0.71 mg kg−1 d.m. and in TPP region 0.86 mg kg−1 d.m. As mentioned above for Cd, Co, Mn, Na, Pb, Sr and Zn, no statistically significant differences were found. Explanation can be found in very similar environmental exposure of bees to these elements in the three regions examined in this study. Moreover, high standard deviations are present for some of the analyzed elements. Although high standard deviations were expected because of the nature of the bioindicators used, it should be mentioned that it can influence statistical analysis.

Statistically significant differences between samples collected at different dates in the study were found for Al, Ba, Co, Cr, Cu and Fe, while no differences were detected for Cd, Li, Mn, Na, Ni, Pb, Sr and Zn (Table 3). For all of these elements, concentrations were higher in July than in September. This can be explained by higher bee activity during July. More plants are flowering and bees fly out more frequently to collect pollen and nectar, which makes them more exposed to environmental pollution. In September, fewer plants are in flower, and bees do not fly out as frequent, so they are less exposed to pollution. Also a reduction in the amounts of nectar and pollen coming into the hive causes reduced brood rearing and diminishing population. The proportion of old bees in the colony decreases (Kauffeld 1980). Along with the decrease in the old bee population, their percentage in the analyzed sample also decreases, and that can also lead to lower concentration of analyzed elements in September compared to July.

Table 3 Mean concentrations of metal in bodies of honeybees (mg kg−1 d.m.) per location in July and September
Full size table

TPP region

TPP region had significantly higher concentrations of Al, Cr and Fe compared to both other regions, and higher Li concentrations compared to urban region. At the same time, it had significantly lower concentrations of Ba and Ni compared to agricultural–woodland region, and Cu when compared to both other regions. The fly ash produced by TPP Kostolac has the highest concentrations of Al and Fe, compared to other analyzed elements by Popović et al. (2011). This study showed the same results (Table 2). The concentrations of Al, Cr and Fe leached from the TPP Kostolac fly ash are positively correlated with each other (Popović et al. 2011). These elements are also positively correlated in the bees from this region analyzed in this study (Table 4). The ratios between leached Al, Cr and Fe found in Popović’s data were not the same as the ratios in this study, but they show a similar pattern, with Fe being the most abundant among them, followed by Al and with very small amounts of Cr compared to the first two. This shows that higher concentrations of Al, Cr and Fe in this region came possibly from fly ash produced by power plants Kostolac A and B.

Table 4 Correlation coefficient (r) for Al, Cr and Fe found in bodies of honeybees from TPP region
Full size table

Higher concentrations of Al, Cr and Fe in TPP region can influence living organisms. Since Al is not a proven nutrient for living organism, its higher concentrations could be toxic. Al toxicity to plants is largely dependent on soil pH. Soil with low pH contains more dissolved Al ions. For this reason, certain crops cannot be grown on acid soils. However, there are some plants that are known as “aluminum accumulators” and may contain as much as 37 % aluminum in ash (Foy et al. 1967). Inhalation of Al containing dust can cause lung problems in animals and humans (Burge et al. 2000). Nervous system of animals and humans is also sensitive to aluminum toxicity and may cause Alzheimer’s disease (Iregren et al. 2001). High concentrations of Cr can also be toxic to plants. It can effect plants growth and development including growth of roots, stems and leaves. Cr also has negative effects on plant physiological processes such as photosynthesis, water relations and mineral nutrition (Shanker et al. 2005). For mammals, Cr is an essential element that is required for normal sugar and fat metabolism. Cr toxicity is characterized by changes in the lungs, kidneys and liver. The lungs are affected with hyperemia, erosion and an inflammatory change in the respiratory system mucosa, or even nose septum perforation (Lee et al. 2002). It can also cause small cell cancer of the lung tissue (Park and Stayner 2006). Fe is an essential nutrient needed to produce chlorophyll in plants (Hänsch and Mendel 2009), while in animals it can be found in hemoglobin, myoglobin, cytochromes, iron-containing enzymes and in iron storage proteins, ferritin and hemosiderin. Although Fe is an essential element for living organisms, it can be toxic at higher concentrations. Fe toxicity to plants is mainly consequence of free radical production that irreversibly impairs cellular structure and damages membranes, DNA and proteins (de Dorlodot et al. 2005). Fe toxicity can also cause reduction in plant photosynthesis, yield and increase oxidative stress (Sinha et al. 1997). In humans and animals, Fe toxicity is not very frequent, but when it happens it can cause liver and heart disease, neurodegenerative disorder, hormonal abnormalities and immune system abnormalities (Berg et al. 2001; Wilkinson 1996; Walker and Walker 2000).

Within TPP region, there were three sampling sites: 1, 2 and 3. Cd, Mn and Zn showed statistically significant differences among sampling sites in TPP region. Cd had higher concentrations at sampling site 3 compared to sampling site 2. Mn had significantly higher concentrations at sampling site 1 when compared to both of the other sampling sites in this region, and Zn had higher concentrations at sampling site 1 when compared to sampling site 2. Since Mn and Zn had higher concentrations at sampling site 1 which is not located in the direction of prevalent winds, it cannot be concluded that wind direction had a determining impact on the distribution of the analyzed elements within this region.

Significant differences between two sampling periods for TPP region were found for Al, Ba, Co, Cr and Fe (Table 3). All of these elements had significantly higher values in July compared to September.

Urban region (BG)

The urban region (BG) had significantly higher concentration of Cu compared to TPP and agricultural–woodland region. At the same time, it had significantly lower concentrations of Al, Cr, Fe and Li compared to TPP region, and lower Ni concentration compared to agricultural–woodland region (Fig. 3). Presence of Cu in urban areas is usually explained with anthropogenic and industrial activities. Significant source of Cu is vehicle brake pads. Recent study by Whiley (2011) revealed that the average copper content in brake pads, based on a broad spectrum of manufacturers and vehicle models analysis, varied between 4 and 12 %. Constant use of brakes in urban traffic is actually the major source of Cu in urban environments (Christoforidis and Stamatis 2009; Ho and Tai 1988; Li et al. 2001).

Cu is an essential element for living organism (Hänsch and Mendel 2009). In plants, animals and humans, it plays an important role in many enzymes. However, excessive absorption of Cu can be toxic to plants, affecting growth, root elongation and mitosis (Fargasova 2001; Ebbs and Kochian 1997). Chronic Cu poisoning is rear in humans and is mainly associated with liver damage. Acute Cu poisoning can cause abdominal pain, cramps, nausea, diarrhea and vomiting (Fraga 2005).

Urban region contained two sampling sites: 4 and 5. Almost uniform distribution of elements was shown within this region. Significant difference was noticed only for Li which had higher concentrations at sampling site 5 which is located in the direction of prevalent wind. However, only one element is not enough to conclude that wind direction has influence on the distribution of analyzed elements within this region.

There were no significant differences in concentrations of elements analyzed between sampling dates in this region.

Agricultural–woodland region (MS)

The agricultural–woodland region (MS) had significantly higher concentrations of Ni when compared to TPP and urban region. It had higher concentrations of Ba and Cu when compared to TPP region, while at the same time significantly lower concentrations of Al, Cr and Fe in relation to TPP region and Cu compared to urban region. It has been found that Ni concentrations in this study are highest in agricultural–woodland region. Ni can be found in some commercial phosphate fertilizers, as well as in animal manure that is the main organic fertilizer (Cordell et al. 2011; Gonnelli and Renella 2013; Mortvedt 1996; Smolders and Mertens 2013; Westfall et al. 2005). The use of barium fluorosilicate and carbonate as insecticides contributes to the presence of Ba in agricultural soils (Meister 2004). The vineyards in the vicinity to sampling site 6 located in agricultural–woodland region are treated with Bordeaux mixture, which is used as a fungicide. The main component of this mixture is copper sulfate. Constant use of fertilizers, fungicides and insecticides in the area of Mesić can easily explain why the concentrations of Ba, Cu and Ni in this region are high.

Ni is a part of metalloenzyme urease that can be found in plants, animals and humans; therefore, it is considered to be an essential element (Hänsch and Mendel 2009). In higher concentrations, it can be toxic to plants causing chlorosis and necrosis, disorder of cell membrane functions and changes in water balance (Rahman et al. 2005; Gonnelli et al. 2001). Since higher Ni concentrations are found in prevalently agricultural region and the origin of Ni could be related to the application of the fertilizers, farmers should pay attention to Ni contents in fertilizers to avoid elevated concentrations of Ni in crops. The toxicity of Ni in animals and humans could be manifested by its interference with normal physiological roles of Mg, Zn, Ca and Mn (Alcon et al. 1991). Ni can also cause increase in immunoglobulins and serum proteins concentrations (Bencko et al. 1986).

In agricultural–woodland region, no significant differences were shown between sampling dates.

All of the bee colonies that were used for sampling survived the season of bee activity. This means that the concentrations of elements detected in this study, although high, did not affect the viability of the bee colonies from which the samples were taken.

Comparison with literature data

Measured concentrations of Ba, Cd, Co, Cr, Cu, Mn, Pb and Zn in this study had values (Table 1) that fall in the range of values reported in previous studies (Bromenshenk et al. 1985; Conti and Botré 2001; Fakhimzadeh and Lodenius 2000; Porrini et al. 2002; Roman 2005, 2010; Sadeghi et al. 2013; van der Steen et al. 2012). Concentrations of Al, Fe, Na and Ni found in this study are higher than values reported in earlier studies. The Al concentrations in this study [6.8–327 mg kg−1 d.m. (Table 1)] are significantly higher than the ones reported by van der Steen et al. (2012) who has found them to be in the range 4.6–15.52 mg kg−1 d.m. Fe concentration in earlier studies was below 117.12 mg kg−1 d.m. (Zhelyazkova et al. 2010), but this study shows that Fe concentrations are almost always above those levels and vary from 101 to 421 mg kg−1 d.m. A. Sadeghi et al. (2013) reported Na concentrations that varied from 12.65 to 16.18 mg kg−1 d.m. In this study, Na concentrations were reported to be much higher (216–687 mg kg−1 d.m.). Ni concentrations in this study (0.25–2.16 mg kg−1 d.m.) are higher than the ones reported in previous studies where Ni concentrations varied from 0.12 to 0.50 mg kg−1 d.m. (Roman 2005; Porrini et al. 2002).

Conclusion

This study showed that concentrations of some of the analyzed elements (Al, Ba, Cr, Cu, Fe, Li and Ni) were significantly different among considered TPP, urban and agricultural–woodland regions, while the concentrations of rest of the analyzed elements were not significantly different. Similar concentrations of some elements indicate that the environmental exposure of honeybees to these metals is very similar in the three regions. Also most of elements did not show temporal differences suggesting that their concentrations do not depend on the time of sampling.

For Al, Ba, Cr, Cu, Fe, Li and Ni, significant differences were observed between at least two of the analyzed regions. The TPP region had the highest concentrations of Al, Cr, Fe and Li, which are the constituents of fly ash produced by TPP. The agricultural–woodland region had significantly higher concentrations of Ni, Ba and Cu, which were attributed to the use of artificial fertilizers, fungicides and insecticides. The urban region had highest concentration of Cu. Elevated Cu concentrations in urban areas are related to intensive traffic, because Cu is one of constituents of brake pads, which emit considerable amount of this metal via aerosol particles into the atmosphere.

Significant differences between the two sampling dates were found in only one region, TPP region. Al, Ba, Co, Cr and Fe in TPP region had significantly higher concentrations in July, which can be contributed to higher bee activity during this period.

The results in this study show that each region has specific pollution associated with anthropogenic activities that take place in it.

It can be concluded that honeybees are excellent bioindicators that can detect trace elements present in the environment. They are capable of distinguishing different concentrations of metals in different foraging regions.