The industrial section of the ancient city of historical Pundranagar, Bangladesh, and the farmlands used for long-term banana farming were chosen as the study sites (Figure 1). We selected some virgin fields in the study region to compare the consequences of longstanding banana cultivation on the physicochemical properties of topsoil. The selected site of the current study is one of the earliest urban archaeological locations in Bangladesh, known as Paundravardhanapura or Pundranagara, in the region of Pundravardhana (Hossain, 2006). Pundranagara, a significant city of the Maurya Empire, was occupied until the eighth century. The economy of Bangladesh in the study area comprises various activities, such as agrarian activities, manufacturing events like convenience garments, poultry feed production, food packaging, and agrochemical industries (Zakir et al., 2013). During sampling, we observed that farmers have been using untreated wastes from various industries and chemical fertilizers/pesticides in the banana fields for 10 years or more. A number of surface soil samples (51) were taken at a 0–20 cm depth, and 56 different stem, root, leaf, and fruit tissue samples were collected during January–June 2019. Triplicate plant specimens of the banana (Musa spp.) species were randomly selected from the banana fields. The plant tissues and soil samples (approximately 500 g each) were placed in polyethylene bags, carefully labeled, and brought to the lab, where they were dried in the sun for a few days. To achieve the constant weight, samples were then ground, homogenized, and dried for 72 h at 65°C in the oven. Before the chemical analysis, the prepared soil and plant specimens were preserved hermetically in clean Ziploc bags in a refrigerator (4°C).

Trace elements in soil and plant specimens were analyzed in the Laboratory of Plant Nutrition and Fertilizer, using inductively coupled plasma mass spectrometry (ICP-MS, 7800, Agilent Technologies). Approximately 2–4 mg of plant tissues and 3.0 g of dry soil samples were digested by adding 4.0 mL of 65% HNO₃, 1.0 mL concentrated HCl, and 2 mL of 30% H₂O₂ (Wako Chemical Co., Japan), using a microwave digesting apparatus with a Teflon tube (Kühnlenz et al., 2016). A standard solution XSTC-13 (Spex CertiPrep^® USA) was employed (R ² > 0.999) to create the calibration plot. Indium, yttrium, beryllium, telium, cobalt, and titanium (1.0 mg/L) were used as internal standards. Ultrapure-grade 5% (v/v) HNO₃ was used for sample preparation and dilution. A similar process was followed to verify the batch-specific accuracy throughout the experimental performance; triple measurement protocol blanks were run and verified using Japanese-certified reference material (NMIJ CRM 7303-a) and Polish-certified reference material (INCT-CF-3) of corn flour (Supplementary Table S1).

The PLI, which is calculated as the following CF multiplication of the trace elements, is known as the nth root: P L I = C F 1 × C F 2 × C F 3 × … … … … × C F n 1 / n , (5)

Based on the value of PLI, Sadhu et al. (2012) classified three grades of soil due to trace elements contamination in soil (Supplementary Table S2).

For contamination assessment, the geoaccumulation index (I _geo) has been calculated as follows: I g e o = log 2 C n / 1.5 B n , (6)

where C_n is the determined abundance of trace element “n” in the soil sample and B_n is the baseline value of element “n” in the background soil sample. A factor of 1.5 is applied to reduce any fluctuations in the background values caused by lithogenic influences. Müller (1981) interpreted seven classes of I _geo for trace elements in soil (Supplementary Table S2).

The magnitude of trace elements in any contaminated environment can be assessed by the enrichment factor (EF) (Franco-Uria et al., 2009). For the present study, the Fe value from each field site was used for trace element normalization and calculated as follows: E F = C n / C element   normalized   by   Fe sample in target / C n / C element   normalized   by   Fe background sample . (7)

The ratio of the trace elements in the examined sample (C_n) to the element in the background is standardized by normalizing the amount of the element in the soil and background to the amount of Fe (mg/kg) in the soil and background (Islam et al., 2022a). Birch and Olmos (2008) classified seven CF groups presented in Supplementary Table S2.

The Nemerow pollution index (PN) is utilized to evaluate the comprehensive status of soil due to trace element contamination (Islam et al., 2022a). Natural changes of trace elements can be seen in the soil environment at any particular location; PN differentiates from other pollution assessment indices in that it considers both the maximum and average abundances of trace elements under consideration. It is calculated as follows: P N = C F m e a n 2 + C F m a x 2 / 2 , (8)

where CF indicates the mean contamination factor value of trace elements in the soil under investigation and CFmax is the highest CF for a specific element in a specific soil sample (Chen, 2010).

The mean effects range median quotient (MERM-Q) of trace elements can be calculated by the following equation: M E R M − Q = C n / E R M , (9)

where C_n is the concentration of the metal in the soil sample, and the ERM values for Cr, Ni, Cu, As, Cd, Pb, Fe, Mn, and Zn are 370, 52, 270, 70, 10, 218, 15000, 550, and 410, respectively (Islam et al., 2022b).

The potential ecological risk index (RI) was used to evaluate the degree of trace element contamination of the examined soil that can be used for long-term banana cultivation, and the calculation procedure is as follows: E r i = T r i   x   C F , (10) R I = ∑ E r i , (11)

where T r i represents the toxic-response factor for a particular element and RI represents the total risk factors for trace elements. The T r i values for Cr, Ni, Cu, As, Cd, Pb, Fe, Mn, and Zn are 2, 6, 5, 10, 30, 5, 20, and 1, respectively (Guo et al., 2010). Five toxicity levels and four RI categories were established by Håkanson (1980) for the potential ecological risk factor E r i ) of a single element (Supplementary Table S2).

The data were evaluated using the statistical package SPSS 20.0 (SPSS, USA). Nine trace elements in the soil of banana farms were chosen, and their means, standard deviations, variances, relative standard deviations, kurtosis, and skewness were computed. Principal component analysis (PCA) was used to identify potential pathways or origins of the trace elements by measuring the variable’s average value. The isolated component’s eigenvalue and loading value were greater than 1 and 0.5, respectively. Statistical analysis, such as analysis of variance (ANOVA), was carried out to check the variability of metal concentration in soil and banana tissues. We applied bivariate correlations analysis with a Tukey post hoc test to verify significant differences in metal concentrations in soil and different tissues of banana plants among sampling sites. Microsoft Excel 2010 was also used for other calculations.

Table 1 and Supplementary Figure S1 show the descriptive statistics of Cr, Cu, Ni, As, Pb, Cd, Fe, Mn, and Zn in the topsoil of a banana field used for long-term banana farming. There was a perceptible change in the amounts of the examined trace elements in the soil samples. Among the measured elements, the highest mean value was observed for Fe (15,390 mg/kg), followed by Zn (49.37 mg/kg). The average values of trace elements in soil samples followed the descending order of Fe > Zn > Mn > Cu > Ni > Pb > Cr > As > Cd. As highlighted in Table 1, an elevated standard deviation was observed for the evaluated trace elements in surface soil samples, which exhibited relatively substantial spatial variability. This variation might be due to the application of various metal-contaminated agrochemicals, such as diazinon, trichlorfon, carbofuran, chlorpyrifos, carbaryl, 1,3-dichloropropene, thiabendazole, atrazine, and glufosinate ammonium (Henriques et al., 1997; Dong et al., 2019; Rosa et al., 2020). There was a finding that trace element extents in banana fields were far higher than in the virgin field (Supplementary Figure S1), indicating that the application of various insecticides or pesticides and excavation in the banana fields could be the causes of the increased contaminant content in banana cultivated soil.

The average extents of Cu, Cd, As, and Pb in the topsoil of banana fields in the present study were higher than the average shale values (ASV) (Turekian and Wedepohl, 1961), the BV (background value) in Bangladesh (Kashem and Singh, 1999), and the standard of Dutch soil quality (VROM, 2000), indicating their contamination in the soil and the possibility of potential ecological risk to the banana field soils. Islam et al. (2021b) conducted a study in the soils of Bangladesh, and they stated that toxic-element-contaminated waste from burning or metal processing sites of an industrial area and imbalanced use of pesticides in agricultural fields are responsible for the elevated levels of As, Cu, Pb, and Cd. This study highlighted similar reasons for elevated trace elements in the soil of banana fields around the industrial area in Bangladesh. In the current study, surprisingly an anomalous concentration of Cd was observed, which was 17.3 times higher than the average shale standard of Cd in soil (Table 1). On the earth’s surface, Cd is the most eco-poisonous component; its elevated level in farming soil is a notable concern of environmental scientists due to its adverse effect on the soil ecosystems, especially on soil flora and fauna (Ekere et al., 2020).

The coefficient of variation (CV) of As, Zn, Cu, Cd, and Pb is distinctly greater than that of other examined elements, which demonstrated that the studied trace elements, especially As, Zn, Cu, Cd, and Pb, are distributed unevenly in the surface soil of banana fields and are affected by the human activities of banana cultivation (Zhuang et al., 2021). The value of kurtosis indicates that As in the examined soil had a typical distribution and was likely due to a variety of banana cultivation activities and industries in the study area. Either a positive or negative attribute of skewness is possible where the high values of skewness for the examined trace elements confirm the present situation for banana cultivation in soil over a long period of time. The skewness values of As, Fe, Zn, and Mn were 0.88, 0.84, 0.75, and 0.66, respectively, indicating a positive skewness toward the higher concentrations in soil in the present study (Table 1). In this work, the skewness of the examined metalloids in surface soil was positive, which revealed their asymmetric distribution (Sadhu et al., 2012; Abbasi, 2019).

The quantities of Ni, Cr, Cu, Cd, As, Pb, Fe, Zn, and Mn (mg/kg dw) in the four main tissues (leaves, stems, fruits, and roots) of the banana plant are presented in Table 2. Varied abundances of trace elements were observed in the soil, which could be reflected in the considerable variations of trace elements in the examined four banana plant tissues. The average level of trace elements in the tissues of banana plants followed the descending order of roots > leaves > stems > fruits. Among the plant tissues, roots accumulate more trace elements than other parts, which may be due to their direct contact with soil. The second highest abundance of trace elements was found in banana leaves, which can be explained by the fact that the large surface area of leaves is regarded as the first recipient of contaminants from industrial emission (Liu et al., 2009; Rosa et al., 2020). An exploration of weed species plants was led by Zhao and Duo (2015), who stated that leaves accumulate higher levels of trace elements than other parts. Interestingly, Cu was detected in substantial amounts in plant leaves, possibly as a result of the chloroplasts’ photosynthetic activity, plant respiration, and assistance in the plant metabolism of carbohydrates and proteins. Because Cu regulates enzymatic activity for lignin formation, and chloroplasts are the typical component of leaves, substantial Cu buildup is anticipated in plant leaves (Islam et al., 2021b).

Mean abundances of Cr (mg/kg) were found to be 5.98 in leaves, 5.03 in stems, 2.30 in fruits, and 8.54 in roots (Table 2). Islam et al. (2015a) reported that Cr content in banana fruits in Bangladesh was 1.2 mg/kg (range 0.06–3.6 mg/kg), and Shaheen et al. (2016) also observed 0.32 mg/kg Cr in banana, which was in line with the present study. The mean concentrations of Ni, Cu, As, Cd, Pb, Fe, Mn, and Zn in leaves, stems, fruits, and roots are presented in Table 2. Significant differences were observed for trace elements in different tissues of plants, which indicated that plant tissues had different accumulation potentials for trace elements. In contrast to soil and water, food is the predominant supplier of possibly dangerous substances to humans, including Cr, Pb, Cd, and As (Ma et al., 2018; Kuerban et al., 2020). However, mean extents of Cr, As, Cd, Pb, Mn, and Zn in banana plants were higher than the permissible limits proposed by FAO/WHO (2011) (Table 2), suggesting the contamination of different banana plant tissues by those elements and potentially imposing risks to consumers. The enhanced levels of Zn, Mn, Cr, As, Pb, and Cd in banana tissues might be due to the effects of applying various wastes from the waste piles around the industrial area and imbalanced application of fertilizers, pesticides, and other agrochemicals to the banana fields (Lin et al., 2010; Ahmed et al., 2016; Rosa et al., 2020). The mean extents of trace elements in various tissues of the banana plant followed the descending order of Fe > Zn > Mn > Cu > Ni > Cr > As > Pb > Cd.

Trace elements are absorbed from the soil by the banana plant roots and are dispersed to other parts of the plant. The calculated values of the BCF are shown in Figure 2. The BCF values of As, Cr, Cd, Fe, and Pb in roots were higher than those in stems, fruits, and leaves (Figures 2A, B), indicating that roots have more accumulation potential of these elements. This was in line with earlier research by Bonanno (2013), who reported that essential micronutrients can be easily ingested by the lower parts of plants like the roots. Although the BCF value for trace elements in fruits was relatively lower than the other three tissues, fruits could have a direct adverse health impact on the consumers who frequently consume bananas from the study sites. Considering all the examined tissues, the mean BCF values of trace elements followed in the descending order of Zn > As > Mn > Cr > Ni > Cu > Cd > Pb > Fe. Among the studied elements, the most concerning toxic elements are Cd and Pb, which showed a relatively lower BCF value, possibly due to their stable form in the soil of the cultivated field (Khairiah et al., 2009; Islam et al., 2021a) or the situation of their oxide form in the topsoil (Juen et al., 2014; Islam et al., 2022b). The BCF values of Zn in leaves, stems, fruits, and roots were 1.03, 1.16, 0.80, and 1.04, respectively, which were slightly higher than 1.0, suggesting that the bioavailability of Zn is high, and plants can easily accumulate this element from soil and disperse it to the aboveground tissues. Previous studies by Subramanian et al. (2022) and Islam M. D. et al. (2020) stated that variations in the plants’ element adsorption capacity, species of plant, the anchoring behavior of nutrients by root systems, and the prevailing ecological conditions of soil account for the variation of trace element accumulation by banana plants.

The contamination of soil by trace elements was assessed by various chemometric approaches such as the geoaccumulation index (I _geo), the contamination factor (CF), the enrichment factor (EF), the pollution load index (PLI), the degree of contamination (C_d), the modified degree of contamination (mC_d), the mean effects range median quotient (MERM-Q), the potential ecological risk index (PERI), and the Nemerow pollution index (P_N). The data are presented in Figure 3 and Figure 4 and Supplementary Figures S2, S3. The I _geo values are presented in Figure 3. Considering the virgin and banana fields, the I _geo values are in the descending order of Cu > Cd > Pb > Ni > Cr > As > Zn > Mn > Fe (Figure 3). According to Müller (1981), I _geo values indicated that the studied soil was moderately to heavily contaminated with Cu, Cd, Pb, and Ni, whereas it was uncontaminated to moderately contaminated with Cr, As, and Zn (Figure 3). Based on I _geo values, the soil of virgin land was not contaminated by trace elements, unlike the soil of long-term banana cultivation, confirming that long-term banana cultivation practices are responsible for elevated trace elements in banana field soil.

EF is a tool that can categorize the natural or anthropogenic origin of trace elements (Proshad et al., 2019; Asomba et al., 2023). Trace elements were somewhat enriched in the soils of the virgin field, whereas severe to very severe modification was observed for trace elements except Fe and Mn in the banana field (Supplementary Figure S2). Altogether, the EF of the studied elements were in the descending order of Cu > Pb > Cd > Ni > As > Cr > Zn > Mn > Fe (Supplementary Figure S2). The EF value of studied trace elements in the banana field was larger than 1.5, indicating man-made activity. Industries are responsible for the high enrichment of trace elements in the surface soil of the examined positions (Islam M. S. et al., 2020; Ustaoğlu and Islam, 2020). Shirani et al. (2020) stated that anthropogenic activity accounted for the higher enrichment of Cd, Pb, Cr, and Ni in the aquatic setting of Iran, which is consistent with this study.

According to the classification system of Håkanson (1980), the examined trace elements in the soil of virgin fields demonstrated little significant pollution whereas, in banana fields, soil presented very high contamination for most of the trace elements, except for Fe and Mn, which showed moderate contamination (Supplementary Figure S3).

The PLI acts as an effective technique for pollution evaluation. A PLI of zero indicates the suitability of soil for crop production, a PLI value of 1 indicates background levels of trace elements, and PLI values greater than 1 designate the soil as contaminated by trace elements (Proshad et al., 2021). The mean values of PLI were observed as 1.49 and 8.12 for virgin and banana fields, respectively (Figure 4A), which indicated that the soils of the current study are polluted by trace elements (PLI >1). The PLI of the banana field was significantly higher than that of the virgin field, indicating the effect of long-term banana cultivation on trace element accumulation in soil. The PLI value for the virgin field was greater than one, confirming the operation of metal processing industries and open waste-burning activities in the studied area (Islam et al., 2015b).

Similar spatial patterns to PLI were observed for the degree of contamination (C_d) and modified degree of contamination (mC_d), where their values in banana field samples were significantly higher than those of the virgin field samples (Figures 4B, C). The values of C_d and mC_d for the virgin field samples demonstrated a low to moderate degree of trace element contamination. However, the banana field samples presented a high degree of contamination (C_d > 32 and mC_d value as 8 ≤ mC_d < 16) (Figures 4B, C), which could be due to the impact of agricultural operations and banana plantation cultivation methods (Lin et al., 2010; Kumar et al., 2021). The obtained values of MERM-Q for studied trace elements in surface soil showed 21% toxicity (Figure 4D). Håkanson (1980) described a potential ecological risk factor (E_r ⁱ) for evaluating soil ecological hazards based on trace element contamination. The value of potential ecological risks ( E r i ) for specific elements in soil samples decreased in the order of Cd > Cu > Pb > As > Ni > Fe > Cr > Zn > Mn. The PERI was observed as being high risk for banana field soil and low risk for the virgin field (Figure 4E), which indicated that banana cultivation might create a probable biological danger compared to virgin field. High latent environmental jeopardy observed in the banana field was mainly contributed by Cd, Cu, Pb, Ni, and As, which might originate from industrial effluents used to add carbon-based substances to the banana field and from the agrochemicals applied for banana cultivation (Islam et al., 2015b; Kabir et al., 2020). Rodrguez Martin et al. (2013) concluded that the use of phosphate fertilizers in crop fields and untreated industrial waste created high ecological risks in agricultural soil after the release of Cd, Cu, Pb, Ni, and As. The Nemerow pollution index (PN) value also showed high soil pollution in the banana fields of the study area in Bangladesh (Figure 4F). Therefore, special attention should be paid to managing the banana field soil using sustainable agricultural practices to reduce the accumulation of trace elements, especially Cd, Cu, As, Pb, and Ni, and protect public health in Bangladesh.

In this study, the sources of trace elements in surface soil and banana plants were identified using PCA (Figure 5). The PCA was utilized based on the profusions of trace elements in soil and plant trials with the varimax rotation technique. Three principal components accounted for the variability of trace elements in both soil and banana plants and cumulatively explained 58.58% and 60.30% in soil and plant samples, respectively (Figure 5). The first principal component (PC1) contributed 22.18% and 23.25% of the total variance in soil and plant samples, respectively, and loading with Cr, Zn, Pb, and As in soil and Zn, Cd, and Pb in plant samples (Figure 5). The sources of Cr, Zn, Pb, As, and Cd in the soil and plant samples from the study area may arise from paint flakes, leakage of oil and lubricants, waste from old scrap batteries, metal workshops, and industrial exhaust emissions (Ekere et al., 2017; Zglobicki et al., 2018; Roy et al., 2019). The second component explained 21.48% of the total variance in soil and presented positive loading for Ni, Mn, Cu, and Cd, whereas 23.021% of the total variance in plant showed positive loading with Cr, Cu, Ni, As, and Mn (Figure 5). These elements in soil and plant samples could arise from multiple origins, viz., waste burning, industrial emissions, excess use of fertilizers and pesticides in banana fields, or lithogenic sources (Pandey et al., 2012; Islam et al., 2021a). The PC3 accounted for 14.91% and 14.03% of the total variance in soil and plant samples, respectively, and presented positive loading for Fe (Figure 5). A study by Iqbal and Shah (2011) mentioned that Fe is a crucial component of the earth’s topsoil and crust and is mainly attributed to some general causes. From the PCA, we have observed diverse distributions of identical types of trace elements in soil and plant samples that may be caused by the element’s source emission characteristics as well as the variation of its accumulation in the soil and plants. The ANOVA results showed that there were significant differences in metal concentrations in soil and different banana plant tissues, whereas no significant difference was observed for Zn concentration in soil and different tissues of banana plants (Table 3). Multivariate analysis revealed that anthropogenic sources, rather than geogenic, are the leading factors which agrees with the previous demonstrations of Bhuiyan et al. (2021).

Data are shown in Table 4 for the exposure of adults and children to the trace elements in banana fruits. The mean values of EDI from banana fruits followed the descending order of Fe > Zn > Mn > Ni > Cu > Cr > As > Pb > Cd. Compared to the maximum acceptable daily ingestion of trace elements (JECFA 2011), EDI levels for the studied elements were found to be lower than the MTDI (Table 4), indicating that the studied elements had a minor effect on human health via banana consumption in the considered area. The non-carcinogenic risk (THQ) for eight elements (except Pb) and TR of Cr and As of adult and child consumption of banana fruit are shown in Table 4. Except for children, the THQs for the studied elements were lower than the threshold value (THQ <1), indicating no probable non-carcinogenic concerns associated with banana consumption. However, the total THQ of multiple elements from banana fruits was higher than the allowable levels for adults and children (THQ >1), indicating a potential risk to human health.

The target cancer risk for Cr and As was determined to be 1.3 × 10⁻³ and 7.9 × 10⁻⁴, respectively (Table 4). These values were somewhat larger than the permitted value (10^–4), indicating a cancer risk to adult consumers (USEPA, 2010, USEPA 2015). The current study considered 50% of inorganic As in banana fruit for carcinogenic risk calculation (Saha and Zaman, 2013). In the current study, only the edible part of the banana was considered for health risk estimation. Other food items, such as cereals, vegetables, other fruits, and non-piscine protein source foodstuffs (meat, milk, and egg), from the research area should be considered in order to gain the real scenario of health benefits from dietary intake of bio-accessible forms of trace elements (Praveena and Omar 2017). This is a limitation of this study.

Findings from the present study also supported the idea that the manufacturing of deleterious substances, especially those that generate trace elements, should be curtailed by adopting tough measures such as the imposition of higher taxes on those products that cause food insecurity or have a negative impact on public health. Proper education on soil nutrients in banana plant systems could assist residents in comprehending some likely risks from trace elements. Therefore, result-oriented approaches should be adopted by the concerned authorities for better management of trace-element-contributing sources in the banana field soil in Bangladesh. Presently, no efficient management approach has been developed for addressing trace elements pollution in Bangladesh. However, the present research revealed different untapped research themes such as economic and ecological aspects of soil–banana plant systems and ecotoxicological profiles, trace elements in agro-food, their hazards, and potential remedies through intervention options and proper regulatory measures (Figure 6). The available literature on soil–plant management systems should be incorporated into these untapped research themes to address the present challenges of assessment and identification of trace elements in the banana field soils. Strict enforcement of any existing policy concerning the use of agrochemicals on agrarian soils will restrict trace element pollution in the soil–banana plant systems of Bangladesh. More so, the transfer of trace elements along the food chain and their toxicity in other agroecosystems should be studied. Additionally, eco-friendly and reliable technology should be incorporated with modern global approaches for the mitigation of potentially toxic elements. Thus, a careful evaluation of existing research gaps and global policy synchronization should be undertaken by authorized bodies to ensure proper management of trace element pollution in the soils of banana fields and other agroforestry systems in Bangladesh.