Magnetic Biomonitoring of PM in a Semi-Arid Urban Park of North-Central Mexico Using Tillandsia recurvata as a Particulate Matter Biocollector

Abstract

This study assessed the spatial distribution and composition of airborne particulate matter within a 10 km long urban green corridor in Zacatecas, Mexico, using magnetic biomonitoring with Tillandsia recurvata and SEM-EDS particle characterization. A total of 44 samples were collected from distinct urban park contexts (e.g., commercial zones, malls, bus stops), revealing mass-specific magnetic susceptibility χ values ranging from −6.71 to 61.1 × 10⁻⁸ m³ kg⁻¹. Three compositional groups were identified based on a PCA performed using elemental concentrations from SEM-EDS and magnetic data, which are associated with traffic emissions and industrial inputs. SEM-EDS images confirmed abundant magnetite-like particles (1–8 μm) with hazardous metals including Pb (up to 5.6 wt.%), Ba (up to 67.6 wt.%), and Cr (up to 31.5 wt.%). Wind direction data indicated predominant SSW–NNE transport, correlating with hotspots in central and northeastern park areas. Overall, vegetated zones exhibited markedly lower magnetic loads (mean χ = 8.84 × 10⁻⁸ m³ kg⁻¹) than traffic-exposed sites (mean χ = 17.27 × 10⁻⁸ m³ kg⁻¹), representing an approximate 50% reduction in magnetic particle accumulation, which highlights the effective role of continuous vegetation cover as a functional green barrier that attenuates the lateral transport and deposition of airborne particulate matter within the park. This research highlights the applicability of combined magnetic and microscopic techniques for evaluating the dynamics of airborne pollution in urban parks and supports their use for identifying both pollution hotspots and mitigation zones, reinforcing the role of urban green spaces as biofunctional filters in cities facing vehicular air pollution.

1. Introduction

Air pollutants, particularly atmospheric particulate matter, are a global concern due to their harmful effects on human health, as they can cause serious diseases [1,2,3]. Heavy metals are among the most toxic components of atmospheric dust [4]. In Latin America, emissions from mobile sources are a major contributor to air pollution. These emissions frequently contain potentially toxic elements (PTEs) such as Pb, Cd, Cu, As, Ni, Zn, and Cr [3]. In recent years, studies examining the environmental and public health impacts of total suspended particles (TSPs) and particulate matter (PM) have largely focused on downtown areas of major cities or metropolitan zones with intensive industrial activity and heavy vehicular traffic [5,6,7,8]. However, this particulate matter is also present in urban parks—key spaces for recreation and essential urban ecosystem services. These green areas expose vulnerable populations to air pollutants derived from nearby traffic. Urban parks are often located along major city roads to improve accessibility and serve a larger number of people. Consequently, the heavy traffic surrounding these parks increases user exposure to vehicle-related air pollutants, including PM, which raises the risk of adverse health effects [9] or even short-duration events with high rates of PM emission into the atmosphere, such as recreational activities involving fireworks or forest fires [10].

Urban parks are typically chosen for activities aimed at promoting health, leisure, and physical exercise, with users often expecting cleaner air. Nevertheless, air quality within a park can vary considerably depending on multiple factors, such as the amount and type of vegetation, nearby pollution sources, wind patterns, and climatic conditions. Trees within these parks can mitigate air pollution through various mechanisms, including the direct interception of airborne particles on their leaf surfaces, thereby improving ambient air quality. Vegetation is widely recognized as a natural filter for pollutants and plays a role in reducing PM levels [11]. However, the effectiveness of vegetation in capturing atmospheric dust varies significantly by plant type and species [12]. Numerous studies have demonstrated the capacity of vegetation to act as a physical barrier, altering pollutant dispersion patterns and contributing to localized air purification effects [13]. One species that has received attention for its PM accumulation capabilities is Tillandsia recurvata. This epiphytic bromeliad has been identified as a powerful bioaccumulator of atmospheric contaminants [14,15,16]. Due to its specialized trichomes, T. recurvata absorbs a significant portion of its nutrients and moisture directly from the atmosphere, making it particularly sensitive to airborne pollutants and thus a suitable bioindicator for assessing air quality in urban environments [17].

However, evaluating air quality in urban parks presents challenges due to the heterogeneity of conditions across the landscape. High-resolution spatial assessments require simultaneous measurements at multiple locations. In Latin American cities, such efforts are often limited by equipment costs and the risk of vandalism. Traditional high-precision particle monitors are expensive and scarce, with many cities possessing only one to five monitoring units, as is the case in our study area. Consequently, alternative low-cost approaches such as magnetic monitoring have gained traction for urban air pollution studies [18]. Magnetic monitoring is particularly effective because it selectively targets iron-rich particles, such as magnetite, which are a known toxic component of particulate matter derived from combustion and friction-related processes. These magnetically responsive particles have been widely associated with adverse neurological and cardiovascular effects, making magnetic properties a relevant proxy for health-oriented air quality assessments [19,20,21]. Moreover, magnetic monitoring provides an efficient and cost-effective method for assessing air quality with high spatial resolution and has shown strong correlations with chemical analyses in previous studies.

In this study, we conduct magnetic monitoring of air quality in an urban park using native Tillandsia recurvata as biomonitors. We also present a quantitative analysis of tree species within the park and report the results of magnetic characterization of the particles accumulated on the biomonitors after one year of exposure. Additionally, we provide spatial distribution and concentration maps of airborne particle pollutants across three distinct sections of the urban park in our study area. Therefore, this study aims to (1) map the spatial distribution of particulate matter within a Zacatecas urban park using the magnetic properties of T. recurvata; (2) characterize the chemical and morphological composition of the accumulated particles; and (3) evaluate the role of vegetation as a natural barrier for the interception and attenuation of airborne pollutants in urban green spaces.

2. Materials and Methods

2.1. Study Area

The study site is an urban park located in the metropolitan area of Zacatecas–Guadalupe, in the state of Zacatecas, Mexico (Figure 1). This park consists of three main sections intersected by two of the busiest roadways in the region: Adolfo López Mateos Blvd and José López Portillo Blvd, both forming part of Federal Highway 45 (also known as the Pan-American Highway). This road corridor carries a high volume of vehicular traffic, handling approximately 105,000 vehicles per day according to official data from [22]. Although no public traffic records are available for the secondary roadway that borders the commercial zone on the opposite side of the park, the continuous exposure to the heavy vehicular load of Federal Highway 45 represents a major mobile emission source influencing all three park sections. The city of Zacatecas lies in north-central Mexico and exhibits a semi-arid temperate climate (BSh in the Köppen classification), characterized by dry winters and moderate summer rainfall. The average annual temperature is 17 °C, with average highs reaching up to 30 °C in May and lows near 3 °C in January. Annual precipitation is approximately 510 mm, concentrated between June and September [23].

Water supply in this region relies heavily on the exploitation of aquifers, as surface water sources are scarce and insufficient to meet urban demand. Zacatecas suffers from severe aquifer overexploitation, particularly in the Guadalupe–Bañuelos and Calera systems, both classified as critical by the National Water Commission [24]. The state’s meteorological monitoring infrastructure includes 38 automatic weather stations operated by SMN-CONAGUA, equipped with sensors to measure temperature, relative humidity, precipitation, wind speed and direction, solar radiation, and soil moisture. These stations generate data every 15 min. The nearest stations to the study site are Guadalupe (Academic Unit of Biology) and Zacatecas (Academic Unit of Agronomy), which provide key environmental data for atmospheric biomonitoring strategies. These hydrological limitations contribute to persistently dry surface conditions throughout most of the year, which favor the resuspension of mineral dust and metal-bearing particles. This environmental context is relevant because dry conditions enhance particulate deposition on Tillandsia recurvata, directly influencing the magnetic load measured in this study.

In the Zacatecas–Guadalupe urban sprawl, two main traffic flows move in directions opposite to Federal Highway 45. The first is driven by commercial and service activities toward the city center; the second corresponds to bureaucratic and academic commuting, mainly toward the western sector. Both converge along Adolfo López Mateos Blvd, causing traffic saturation in the area where the park under study is located [25]. Although Zacatecas experiences slower urban growth compared to other major Mexican cities such as Monterrey and Santiago de Querétaro, it has recorded a population increase of 8.27% over the past decade [23]. The main sources of air pollution in the region include vehicular traffic, brick kilns, forest fires, and mining tailings—typical of a city with a strong mining heritage and ongoing urban development.

2.2. Sampling

One individual of Tillandsia recurvata was collected from each of the 44 sites distributed across the three sections of the urban park (Figure 1). Several of these locations were subject to higher pollutant loads and exhibited variations in physical structure and tree density. For the sampling design, the park’s total length of 10 km was initially divided into 50 segments, corresponding to the predetermined number of sampling sites for detailed analysis. Consequently, a sampling interval of 200 m was established. However, T. recurvata was not found at eight locations, resulting in a final total of 42 sites, along with two additional points (sites 22 and 23), for a complete set of 44 sampling points. Within each 200-m segment, a sampling point was selected based on the availability of Tillandsia individuals.

The two additional samples (sites 22 and 23) were collected near the local football stadium, an adjacent urban area frequently used for recreational and athletic activities. These sites were included to enable a comparative analysis between the main park and neighboring urban zones with similar pedestrian density.

The dominant tree species in the park is Schinus molle (Peruvian pepper tree), from which all Tillandsia recurvata individuals were collected. These S. molle trees are mature specimens, typically 15–20 m tall, and have been established in the park for several decades. Based on municipal planting records and long-term local observation, they are estimated to be at least 30–40 years old and already exhibited a fully developed canopy structure by the early 1990s. Their stable growth habit and extensive planting—favored by their drought tolerance—provide consistent and suitable hosting conditions for T. recurvata across all sampled sites.

T. recurvata samples were collected at heights above 1.5 m to minimize the influence of resuspended urban soil particles. To avoid cross-contamination between sites, plastic scrapers and disposable gloves were used throughout the sampling process. Since this species consists of leaves with varying ages, individuals of similar estimated age were selected by choosing specimens approximately 10–12 cm in diameter, in order to reduce variability in exposure times. Because Tillandsia recurvata individuals were not transplanted but naturally established, the exposure time cannot be defined as a fixed period. Instead, this study focuses on the comparison of spatial differences in accumulated particulate matter under similar exposure conditions. By selecting individuals of comparable size, we minimize variability related to plant development and assume a broadly similar long-term exposure. Therefore, the magnetic signal is interpreted as an integrated record of historical particulate deposition rather than as an absolute accumulation rate.

Sampling was conducted in late October 2023, during what was considered an atypical climatic year for Mexico and various other regions worldwide, marked by unusually dry conditions. Although the rainy season typically spans from June to September, virtually no precipitation was recorded during that period. Instead, rainfall events began in October, coinciding with what is generally classified as the dry season. It is important to note that T. recurvata does not self-clean completely during rainfall events [26]. Several studies have demonstrated that Tillandsia retains atmospheric particles even during rainfall events, as precipitation does not act as a cleansing mechanism but rather represents an additional deposition pathway. According to [27], the plant’s specialized trichomes mechanically trap mineral and organic particles on the leaf surface. When dry, the trichomes remain elevated, forming a shield that captures airborne material; upon wetting, this shield adheres closely to the leaf surface, yet this movement does not remove previously accumulated particles. Consequently, Tillandsia continues to accumulate contaminants under both dry and wet conditions, reinforcing its suitability as a reliable atmospheric biomonitor regardless of precipitation. Therefore, a single sampling campaign was conducted one week after the last precipitation event, following the methodological protocol established by [28]. This post-precipitation interval allows sufficient time for renewed atmospheric deposition while avoiding the immediate effects of rainfall redistribution, ensuring that the collected material represents stabilized accumulated particles rather than transient wet deposition. All material was placed in paper bags and stored at room temperature in the laboratory until subsequent magnetic and complementary analyses were carried out.

2.3. Magnetic Measurements

Biological specimens were enclosed in non-magnetic plastic capsules (8 cm³) and subsequently weighed, with masses reaching up to 3.2 g, prior to magnetic characterization. All magnetic analyses were performed at the Laboratory of Paleomagnetism and Rock Magnetism, located at the Geosciences Center of the National Autonomous University of Mexico (UNAM).

Complementary analyses were carried out on selected specimens (100–220 mg) using a custom-designed horizontal magnetic balance to generate thermomagnetic curves. A total of six specimens were analyzed by thermomagnetic methods, selected to represent the full range of magnetic susceptibility values and environmental settings observed across the park. An inducing magnetic field of 0.5 T was applied, and both heating and cooling procedures were performed in air at a controlled ramp rate of 30 °C·min⁻¹. Samples were heated up to approximately 700 °C and then cooled to room temperature. The relative induced magnetization (M/MRT) was recorded as a function of temperature (T), and the resulting M/MRT–T curves are reported due to their significance in identifying thermal magnetic transitions

Low-field magnetic susceptibility was measured in both volumetric and mass-specific forms using a KLY-3 Kappabridge susceptometer (AGICO, Brno, Czech Republic).

Finally, anhysteretic remanent magnetization (ARM) was imparted using a custom-built alternating field (AF) demagnetizer. During the process, a DC bias field of 90 μT was superimposed on an alternating magnetic field with a peak amplitude of 100 mT. The induced ARM was measured using a JR-5 induction magnetometer (AGICO, Brno, Czech Republic), from which the anhysteretic susceptibility (χARM) was calculated. The relationship between χARM and low-field susceptibility (χ) was also explored using King’s plots (χARM vs. χ) to assess variations in magnetic grain size.

2.4. Microscopy and Elemental Studies

Samples of Tillandsia individuals were examined by scanning electron microscopy (SEM) using a Philips XL30 microscope (Philips, Eindhoven, The Netherlands) at the Institute of Geosciences, UNAM. Prior to imaging, all samples were sputter-coated with a thin layer of graphite to improve electrical conductivity and enhance imaging contrast, particularly for metallic particles.

High-resolution micrographs were obtained to document surface morphology and particulate deposition. Special attention was given to the trichomes, where particulate matter was frequently observed embedded within their internal cavities. Energy-dispersive X-ray spectroscopy (EDS) was conducted using an integrated EDAX DX4 detector (EDAX, Mahwah, NJ, USA), with a detection threshold of approximately 0.5 wt%. This system enabled elemental characterization of both surface-bound and trichome-associated particles.In total, six samples (IDs 17, 19, 27, 37, 39, and 43) were selected for SEM-EDS analysis. These samples were intentionally chosen to represent the main environmental conditions across the park, including vegetated vs. non-vegetated areas, traffic-exposed zones, wall-bounded microenvironments, and localized internal sources such as barbecue sites. This subset captures the principal contrasts needed for compositional assessment.

2.5. Statistical Analysis Methods

To identify underlying patterns and reduce data dimensionality, a Principal Component Analysis (PCA) was performed using a combined dataset of magnetic and chemical variables. The magnetic variables included χ and ARM. The chemical variables, obtained through energy-dispersive X-ray spectroscopy (EDX), included the following elements: Fe, Ba, Si, Cr, Al, Ca, S, Br, Ti, K, Cu, Mn, Pb, Mg, In, V, P, Nd, Ce, La, and W. All variables were z-score-normalized prior to analysis to ensure comparability across different units and scales. Sample identification numbers were used only for labeling data points in the PCA plots. For the PCA, only the six SEM-EDS samples were included, corresponding to the subset for which complete magnetic and chemical data were available. The sample IDs were used exclusively for labeling points in the PCA plots and, of course, not used as computational variables.

The PCA was executed in the R environment (version 4.3.1) using the FactoMineR and factoextra packages for both computation and visualization of the principal components. In parallel, a hierarchical cluster analysis (HCA) was conducted using Ward’s linkage method and Euclidean distance as a dissimilarity metric, allowing for the identification of natural groupings among the samples. The resulting dendrograms were visualized using the ggdendro package, an R extension that converts hierarchical clustering objects into data frames compatible with the ggplot2 system to enable clear dendrogram representation. Additionally, Pearson correlation coefficients were calculated between magnetic and chemical variables to explore linear associations. Correlation matrices were visualized using the corrplot package, a graphical tool in R designed specifically for plotting color-coded correlation structures. All graphical outputs were generated in (RStudio Team, Boston, MA, USA).

3. Results and Discussion

3.1. Magnetic Properties

Magnetic parameters obtained from Tillandsia samples (see Supplementary Data) exhibited substantial variability, reflecting the diversity of pollution sources and intensities across the urban park. These parameters are critical for interpreting airborne particulate matter concentration, mineralogy, and grain size—key factors tightly linked to anthropogenic activity. Mass-specific magnetic susceptibility ranged from 0.87 to 40.00 × 10⁻⁸ m³·kg⁻¹, with a mean value of 13.18 × 10⁻⁸ m³·kg⁻¹ and a standard deviation of 10.07 × 10⁻⁸ m³·kg⁻¹. Background magnetic susceptibility levels were interpreted using previously published values for Tillandsia recurvata obtained in Mexico and Colombia. Prior biomonitoring studies conducted under low-exposure or control conditions have shown that unexposed Tillandsia tissues consistently exhibit negative mass-specific magnetic susceptibility values, reflecting the intrinsic diamagnetic behavior of the plant. These baseline values typically range between approximately −1 × 10⁻⁸ and −6 × 10⁻⁸ m³·kg⁻¹, a range reported both in Mexican field studies [15] and in Colombian transplant experiments [26]. The close agreement between these independent studies provides a robust reference for defining the conditionally background magnetic susceptibility signal for T. recurvata in this work.

Elevated χ values suggest an enrichment in ferrimagnetic minerals, most likely magnetite, commonly linked to vehicular emissions and urban dust sources. In contrast, near-zero or negative values suggest a dominance of diamagnetic or weakly magnetic materials, such as biological tissue. Anhysteretic remanent magnetization, a proxy for low-coercivity, fine ferrimagnetic grains, ranged from 1.06 to 9.38 × 10⁻³ A·m²·kg⁻¹ (mean: 3.04 × 10⁻³ A·m²·kg⁻¹). The wide range reflects spatial heterogeneity in the intensity of magnetic particle deposition, likely influenced by proximity to traffic routes, topographical configuration, and vegetative cover.

Figure 2 (King’s plot) shows the logarithmic relationship between χ and χARM. Most samples fall between the 0.1 μm and 1 μm isolines, corresponding to fine, submicron particles in the respirable range. These grains are of concern due to their ability to penetrate the alveolar region, cross epithelial barriers, and transport toxic metals—such as Pb, Cd, and As—into systemic circulation. Their large surface area promotes reactive oxygen species generation, leading to oxidative stress, inflammation, and neurotoxicity, particularly in children and vulnerable populations [29,30].

Some samples approach or are beyond the 0.1 μm isoline, indicating the presence of ultrafine particles (UFPs), which are often produced by high-temperature combustion. UFPs can bypass physiological barriers and have been linked to neurodevelopmental deficits, cognitive decline, and systemic inflammation. On the opposite end, a few samples approach or surpass the 5 μm line, suggesting the presence of coarse dust from soil, mechanical abrasion, or resuspension. These may deposit in the upper airways but still pose risks through metal adsorption and respiratory irritation. The coexistence of particles ranging from <0.1 to >10 μm within a recreational park is concerning, especially considering its impact during physical exercise and children’s play. Increased respiratory activity during exercise heightens the inhalation of fine and ultrafine particles, exacerbating potential health risks. Children are particularly susceptible due to higher ventilation rates per unit body mass and developmental sensitivity.

Thermomagnetic analysis (Figure 3) revealed two distinct magnetic behaviors. Group (a) displayed abrupt drops in magnetization near 580 °C, typical of magnetite. These samples were predominantly located in high-traffic and densely used sectors of the park (e.g., sites 17–20 and 34–39), consistent with elevated χ values shown in the magnetic susceptibility map (Figure 6). This pattern suggests substantial anthropogenic influence. In contrast, group (b) showed progressive demagnetization between 675–700 °C, characteristic of hematite. Samples 29 and 43 belonged to this group and exhibited paramagnetic behavior at high temperatures, likely due to oxidation processes. Field inspection confirmed their proximity to rusted metal barbecue structures, which may act as localized sources of oxidized iron particles.

Principal component analysis as shown later in the multivariate analysissupported this interpretation. Sample 43 separated clearly along PC1, driven by high levels of Si, Al, and Ca—elements associated with mineral dust and thermal degradation. The agreement between magnetic and chemical data underscores the mineralogical heterogeneity of particles across the park. Overall, the spatial segregation of magnetic phases—magnetite near roads, hematite near rusted infrastructure—highlights the importance of considering both mineralogy and toxicological potential in urban air quality assessments. These findings reinforce the relevance of magnetic biomonitoring in identifying localized pollution sources in spaces frequented by vulnerable populations.

3.2. SEM Observations and Elemental Content

The analysis performed via SEM-EDS on the composite image (Figure 4) reveals a wide diversity of suspended magnetic particles collected in the Zacatecas urban park. The particles exhibit varied morphologies and compositions, indicative of multiple anthropogenic sources. Quantitative SEM-EDS measurements (Supplementary Table S1) revealed marked variability in the concentrations of toxic metals across the study area, with Pb reaching up to 5.6 wt.%, Ba up to 8.9 wt.%, and Cr up to 3.2 wt.%. These values, although spatially heterogeneous, are fully consistent with the particle morphologies observed in Figure 4 and provide important context for interpreting source contributions.

In previous research, ref. [15] demonstrated that Tillandsia recurvata is capable of retaining a wide spectrum of particle sizes, not only the fine and ultrafine fractions (<10 µm) typically associated with atmospheric pollution. SEM observations from that study revealed the presence of coarse, angular mineral particles, with several examples exceeding 10 µm and reaching ~20 µm, naturally trapped between the foliar trichomes. These findings confirm that the particle-size distribution collected by the plant does not depend solely on a selective retention mechanism imposed by trichome dimensions, but rather reflects the particle-size availability in the surrounding atmosphere. Therefore, the predominance of particles in the 0.2–5 µm range found in our study area corresponds to the actual environmental particle-size spectrum, not to a limitation of the biocollector itself.

Barium (Ba)-bearing particles were frequently identified, mainly as angular or semi-spherical grains ranging between 1 and 3 µm in diameter (Figure 4a,g). The most likely origin of these Ba-containing particles is vehicular activity, specifically brake pad wear and diesel fuel additives. The presence of barite (BaSO₄) with characteristic cleavage has been previously documented in urban centers such as Italy [31], Querétaro, Mexico [32], and Warsaw, Poland [33]. Due to their small size, these particles significantly contribute to the PM_2.5 fraction, representing a health risk upon inhalation.

Iron-rich (Fe) particles dominate the SEM-EDS field and display both spheroidal and irregular morphologies. Spherical Fe oxides (e.g., magnetite), as observed in Figure 4f,j, are typical of high-temperature processes such as combustion and welding [34,35]. Irregular particles associated with Cr (Figure 4k) suggest mechanical abrasion and the degradation of vehicular components. These morphologies have also been reported in cities such as Singapore [36] and Argentina [37].

Notably, titanium (Ti)-bearing particles were detected in several samples (e.g., Figure 4b,j), frequently associated with Fe and Al. The presence of Ti—commonly originating from paints, pigments (e.g., TiO₂), or vehicular abrasion—raises concerns due to its capacity to form ultrafine particles with photocatalytic properties. These particles may promote the generation of reactive oxygen species (ROS) upon exposure to sunlight, exacerbating oxidative stress and cellular damage in the respiratory system.

In some cases, the Fe-rich particles displayed a high oxygen content and an irregular platy morphology, which is compatible with hematite (Fe₂O₃), particularly in Figure 4d,l. Hematite is typically formed through oxidative processes and has been reported in environments with intense vehicular emissions and atmospheric aging. Although less magnetically responsive than magnetite, hematite contributes to the particulate matter burden and may be a marker of long-range transport or secondary atmospheric processes. Its presence in a park environment suggests infiltration from nearby traffic corridors or Local pollution events.

Lead (Pb)-bearing particles, as seen in Figure 4e, appear as irregular bright grains, possibly composed of PbO or PbS. In the context of a long-established urban park that traverses much of the city, the presence of Pb-bearing particles is most plausibly linked to a combination of legacy contamination and ongoing resuspension processes rather than to a single contemporary source. Previous studies conducted in Zacatecas and Guadalupe have documented elevated Pb concentrations in urban soils, attributed to the city’s long mining history, traffic-related deposition, and the persistence of historical pollutants in surface materials [38]. In a heavily used recreational corridor such as the studied park, these contaminated soils can be readily remobilized by wind action, pedestrian activity, and vehicular turbulence from surrounding roads. Additional local contributions may arise from the weathering of old park infrastructure and Pb-based paints, as well as from recreational activities such as barbecue grilling, where biomass and charcoal combustion can release trace amounts of Pb. Together, these factors indicate that the detected Pb reflects an integrated urban background shaped by historical deposition, park use, and ongoing resuspension processes. These particles are commonly associated with fossil fuel combustion, deteriorating paint, and wear on metal structures. Their high toxicity, particularly due to their small size, raises concerns considering that the park is a recreational zone frequently visited by children. Even trace exposure to lead has been linked to irreversible neurodevelopmental effects, especially in young populations.

Aluminum (Al)-rich particles (Figure 4b,c,h,l) were also observed, likely originating from construction dust or soil resuspension. A critical finding is the small size of most particles (<4 µm), placing them within the respirable fraction (PM_2.5 and PM₁). These fractions are associated with severe health effects, including respiratory inflammation, cardiovascular diseases, and neurological impacts [39,40].

The particle sizes observed in SEM-EDS are consistent with the results from the King’s Plot (Figure 2), where most data points are concentrated in the region corresponding to particles ranging from 0.2 to 5 µm, typical of ultrafine particles. More than 60% of the particles fall within this range, which is particularly alarming in a recreational urban environment. Particles of these dimensions exhibit high pulmonary penetration capability and can reach the bloodstream, increasing the risk of chronic diseases.

In summary, the presence of ultrafine and metal-bearing particles—including Ti, Pb, Ba, and Fe oxides—in a public recreation space highlights the environmental injustice and potential health risks faced by park users. The detection of hematite and titanium compounds further supports the hypothesis of multi-source pollution. Given that children frequently visit the park and are particularly susceptible to airborne pollutants, these findings underscore the urgent need for pollution mitigation strategies and enhanced urban air quality monitoring.

Overall, the quantified ranges of Pb, Ba, and Cr provide essential support for the interpretation of pollution sources in the study area. The presence of particles containing up to 5–9 wt.% of these elements indicates that even in an urban park setting, the deposition environment is shaped by a mixture of vehicular emissions, dust resuspension, and sporadic combustion activity.

Although the presence of fine and ultrafine metal-bearing particles is of clear environmental and health relevance, it is important to note that toxic effects depend on exceeding element-specific concentration thresholds established by air-quality and health regulations. Because magnetic techniques do not quantify absolute elemental concentrations, the present study cannot yet determine whether the detected deposition levels represent a direct toxicological risk. However, the occurrence of respirable, Fe-rich particles is consistent with previous studies showing that magnetite and other redox-active particles can accumulate in neural and cardiopulmonary tissues and contribute to adverse health outcomes [41]. These findings indicate that follow-up chemical analyses are necessary to establish potential toxicity, ideally by focusing on the magnetically anomalous sites identified here. This targeted approach substantially reduces analytical cost and effort, as only a small subset of representative samples would require detailed elemental quantification.

3.3. Statistical Analysis of the Results

To identify underlying compositional patterns among particulate matter samples collected in the Zacatecas urban park, a PCA was performed using all available quantitative variables, including elemental concentrations obtained via SEM-EDS and magnetic parameters (Supplementary Table S1). Figure 5a presents the PCA projection, where each point corresponds to an individual sample plotted along the first two principal components (PC1 and PC2), which together capture the dominant directions of variance in the multielement dataset. Samples are color-coded according to their assigned cluster and labelled with their identification numbers, allowing direct interpretation of group separation. Three distinct clusters are evident: Cluster 1 (blue; Samples 17, 19, and 30) forms a compact group in the negative PC1 and low PC2 region, Cluster 2 (orange; Samples 27 and 37) is positioned toward higher PC2 values, and Cluster 3 (green; Sample 43) appears isolated at the extreme positive end of PC1. This multivariate structure is consistent with contrasting geochemical signatures among the groups.

Figure 5b complements the PCA by illustrating radar (spider) plots that highlight the relative abundance of the major detected elements. (e.g., Fe, Si, Al, Ca, Ti, S, Ba, K, Mg, and In). The polygonal patterns highlight compositional differences that drive the multivariate separation observed in Figure 5a. Cluster 1 exhibits high relative Fe content and similar radar profiles among its samples, whereas Cluster 2 shows elevated contributions of elements such as Ti, P, In, and Ba. Cluster 3, represented by Sample 43, displays a distinctly different profile dominated by Si and Al. Together, Figure 5a,b provide a coherent visualization of both the statistical grouping and the specific elemental contributions responsible for the observed segregation.

The concentration patterns derived from SEM-EDS also help explain the multivariate structure illustrated in Figure 5a,b. Samples with the highest Fe contents (up to 76.4 wt.%) consistently exhibited detectable Pb, Ba, and Cr, suggesting that their magnetic and geochemical signatures are governed by the combined influence of traffic emissions, mechanical abrasion, and localized combustion processes occurring within the park. In samples where Ba reached concentrations approaching 9 wt.%, the compositional profiles were dominated by particle morphologies characteristic of vehicular wear, whereas samples with elevated Cr showed a stronger association with metallic fragments embedded in resuspended dust. These chemical contrasts underpin the separation of clusters, demonstrating that differences in elemental abundances exert a primary control on the observed PCA and radar-plot groupings.

Cluster 1, composed of Samples 17, 19, and 30, was grouped closely within the PCA space (Figure 5a) and exhibited high relative concentrations of Fe, as shown in the radar diagram (Figure 5b), indicating a strong magnetic signal and a dominant influence from vehicular activity and combustion-related sources. These samples displayed abrupt magnetization drops around 580 °C in their thermomagnetic curves (Figure 3), consistent with the presence of magnetite. The sampling points for this cluster are located adjacent to heavily trafficked roads and dense commercial corridors, where little or no vegetation is available to buffer pollutant dispersion. Additionally, recent observations of welding and metallic cutting during building maintenance in nearby commercial structures support the hypothesis that Fe-rich aerosols were emitted and deposited at these sites.

Similar findings were reported in urban parks in Barcelona, where ref. [42] observed that magnetite-rich particles originating from traffic emissions dominated the magnetic signal in green areas close to major roads. Likewise, ref. [43] demonstrated the high concentration of ultrafine magnetite particles in roadside tree leaves, linking them to vehicular emissions and oxidative stress markers. In Mexico, ref. [15] also documented high χ and IRM values in Tillandsia recurvata near highways, directly associated with traffic intensity. These parallels reinforce the interpretation that the elevated magnetic parameters in Cluster 1 are primarily traffic-related. From a health perspective, magnetite nanoparticles are particularly concerning due to their submicron size and redox-active surface. Their inhalation facilitates translocation to the brain via the olfactory nerve and bloodstream, promoting neuroinflammation, oxidative stress, and the aggregation of amyloid plaques [41]. Given that children and elderly individuals widely use urban parks—two especially vulnerable groups—the presence of such particles demands urgent mitigation efforts.

Cluster 2, comprising Samples 27 and 37, presented lower Fe content and greater compositional heterogeneity. The presence of In should be interpreted with caution, as SEM-EDS provides semi-quantitative information and trace-level signals may result from spectral overlap or complex particle assemblages; however, its detection does not affect the overall interpretation of this cluster, which is primarily controlled by elevated Ti, Ba, and mixed anthropogenic inputs. The radar plots (Figure 5b) reveal elevated concentrations of Ti, P, In, and Ba. These sites, also situated in commercial areas, may reflect the influence of ongoing construction, handling of composite materials, and diffuse emissions from nearby businesses, rather than direct industrial sources. Although no heavy industry is located near the park, the presence of small-scale workshops and building activity could explain the diversity of elemental signatures observed. Similar results were obtained in studies from Massachusetts, United States [44], where commercial zones showed elevated Ba and Ti levels in deposited dust due to vehicular wear and civil works. The moderate magnetic signals of this group may indicate the presence of particles that are less enriched in ferrimagnetic phases but are still potentially hazardous upon inhalation.

Cluster 3, represented solely by Sample 43, was markedly distinct in both compositional and spatial terms. The elemental profile was dominated by Al and Si, indicating a mineral origin consistent with urban soil resuspension or clay-derived dust. Its thermomagnetic curve (Figure 3) confirmed the presence of hematite, which, along with that of Sample 29, likely resulted from surface oxidation of iron-containing silicates rather than anthropogenic combustion. Notably, Sample 43 was collected in a green zone physically isolated from traffic by tall walls and vegetation, confirming the protective role of urban design. This observation aligns with findings from the Triangle Area Barriers Study (TABS), where ref. [45] demonstrated that tree cover and physical barriers significantly reduced magnetic loadings in soil and vegetation near highways.

3.4. Magnetic Monitoring

The spatial distribution map of mass-specific magnetic susceptibility, obtained using Tillandsia recurvata as a biomonitor, reveals a highly heterogeneous deposition of magnetic particles across the urban park (Figure 6). Well-defined maxima are observed in sectors adjacent to primary and secondary roads with heavy traffic, as well as in urban microenvironments characterized by slow-moving vehicles and high-turnover parking areas, particularly within the commercial surroundings of the park. These zones exhibit the highest χ values (>50 × 10⁻⁸ m³·kg⁻¹) and correspond to areas with prolonged vehicle presence, frequent stop-and-go conditions, and congestion. In contrast, internal park sectors with higher tree density and continuous vegetation cover display consistently low to moderate magnetic susceptibility values (<7 and 7–26 × 10⁻⁸ m³·kg⁻¹, respectively), clearly evidencing the role of vegetation as an effective physical and functional green barrier for the interception and retention of airborne particles. The Kernel Density Estimation (KDE) interpolation highlights accumulation corridors aligned with the road network and intensively used traffic areas, but also reveals distinctive attenuation gradients toward sectors protected by green barriers.

This fact indicates the simultaneous role of parks that act as a receptor surface for urban-derived pollutants and as an active environmental filter. These spatial patterns are fully consistent with SEM-EDS results, showing that sites associated with magnetic hotspots are dominated by abundant Fe-rich particles with spherical and irregular morphologies typical of combustion processes and vehicular wear [32], accompanied by potentially toxic elements such as Pb, Ba, Cr, and Ti, whereas zones with lower magnetic loading are characterized by a predominance of silicate particles and more oxidized iron phases. King plots further confirm that samples from high-χ areas cluster within the fine to ultrafine grain size domain (SD–PSD), with dominant particle sizes below 1 µm, explaining the magnetic enhancement and indicating that the mapped magnetic signal is primarily controlled by a fine ferrimagnetic fraction dominated by magnetite. By contrast, sites exhibiting low to moderate susceptibility, generally associated with sectors shielded by continuous vegetation barriers or physically isolated from traffic and localized recreational facilities, such as barbecue sites, where wood-based fuel combustion during cooking activities contributes to localized PM loading, show contributions from hematite-magnetite mixtures and Si–Al-rich materials, whose weak magnetic response does not govern the overall spatial pattern. Taken together, the convergence of magnetic mapping, microchemical characterization, and granulometric information demonstrates that air particle quality within the park is shaped not only by external mobile emission sources and localized internal activities, but also effectively modulated by vegetation barriers, reinforcing their key role as green filters and confirming the utility of magnetic biomonitoring with Tillandsia recurvata for identifying both high particle-load microenvironments and mitigation zones within urban recreational spaces.

Wind direction and intensity data for the same period (October) are shown in Figure 7a, revealing dominant flows from the south (S to SSW) and the north–northeast (NNE), with wind speeds categorized between 0.5 and 5.0 m·s⁻¹. These directions are consistent with the orientation of primary emission sources surrounding the park. The overlapping of wind corridors and areas of high χ further suggests that prevailing wind patterns actively shape advective transport and subsequent surface deposition of magnetically susceptible particles. Notably, the sampling campaign was conducted during the transition to the cold–dry season in Zacatecas, a period characterized by lower ambient temperatures, with mean daily values ranging from 10 to 15 °C, according to long-term climatological records from [46]. These meteorological conditions favor the development of thermal inversion layers and a shallow planetary boundary layer, which reduces turbulent mixing and limits the vertical dispersion of suspended particles. In addition, the park is embedded within a predominantly built-up urban setting surrounded by major roadways and dense infrastructure, consistent with a compact Local Climate Zone (LCZ), which promotes reduced airflow, pollutant trapping, and enhanced near-surface particle accumulation. As a result, particles remain concentrated near the surface, leading to enhanced deposition rates on biological and artificial collectors—particularly in topographic depressions and wind-sheltered sectors of the park.

While previous studies have primarily investigated pollutant dispersion along transects perpendicular to roads or based on linear distances from emission sources, often under simplified wind assumptions, these approaches may underestimate spatial variability in environments with complex use patterns. In contrast, urban parks require a surface-based assessment, with multiple sampling points distributed across a two-dimensional matrix. This methodology enables the capture of cumulative deposition effects, providing a more realistic representation of the pollutant load experienced by visitors upon entering and moving within the park.

The elevated values of χ observed in recreational sectors align with those reported in other urban environments with high traffic density. For instance, ref. [47] reported χ values between 30 and 60 ×10⁻⁸ m³·kg⁻¹ near arterial roads in southern China, while ref. [48] documented similar levels in parks adjacent to traffic corridors in Italy. Likewise, ref. [49] observed magnetically enriched particles in urban green spaces in Antwerp, Belgium, with SIRM values ranging from 19 to 40 × 10⁻⁸ m³·kg⁻¹, depending on the distance from roadways and vegetation coverage.

A limitation of this component of the study is that the PCA and SEM-EDS interpretations are based on six samples (n = 6). Although these samples were intentionally selected to represent the dominant environmental contrasts of the park, the reduced sample size may restrict the generalization of multivariate patterns. This limitation has been acknowledged and will guide future work involving expanded SEM-EDS datasets.

We note that Sample 29 exhibits a thermomagnetic curve that shows a hematite contribution (Figure 3b), while in the PCA, it clusters with Samples 17 and 19, which show magnetite-related characteristics. This is not contradictory because the thermomagnetic analysis reflects magnetic mineralogy and thermal stability, whereas the PCA is driven by multielement chemical composition and the magnitude of χ. Hematite has intrinsically low magnetic susceptibility—consistent with the low χ values of Sample 29—yet its Fe abundance detected by SEM-EDS will dominate the magnetic signal, if small amounts of magnetite are present. In addition, atmospheric oxidation of fine magnetite particles can generate hematite while preserving Fe content, providing a plausible mechanism for the observed behavior. For these reasons, a thermomagnetic curve showing the presence of hematite may coexist with a PCA position governed by chemical similarity rather than magnetic phase. This clarification has been incorporated to avoid misinterpretation of the results.

4. Conclusions

This study provides compelling evidence that Tillandsia recurvata is a reliable bioindicator for detecting spatial variability in airborne particulate pollution within urban parks. The combined use of magnetic monitoring and multivariate analyses enabled the identification of distinct pollution hotspots, characterized by elevated magnetic parameters—particularly χ and ARM—concentrated in areas adjacent to high-traffic corridors. These parameters correlated with elevated concentrations of Fe, Ba, Ti, and Pb, suggesting multiple anthropogenic sources, including vehicular abrasion, combustion, and surface material degradation.

King’s plots and SEM-EDS confirmed the predominance of (sub)micron magnetite-rich particles, many of which fall within the PM_2.5 and PM₁ ranges. The presence of ultrafine and redox-active particles is of particular concern, given the vulnerability of children and older adults who frequently use these recreational areas. Thermomagnetic and morphological evidence further revealed the coexistence of magnetite and hematite particles, associated respectively with vehicular emissions and localized corrosion processes.

Magnetic susceptibility mapping demonstrated that conventional linear transect sampling may underestimate exposure variability in open urban spaces. By implementing a surface-based grid with multiple sampling points, this study captured fine-scale deposition gradients shaped by wind transport, site morphology, and vegetation buffers. This approach provides a more accurate representation of actual exposure conditions for park users.

Magnetic measurements also revealed the consistent presence of respirable metal-bearing particles in the study area. Although magnetic techniques do not quantify elemental concentrations directly, each metal requires specific threshold concentrations to be considered toxic; therefore, this study cannot yet determine whether the detected levels exceed hazardous limits. Nevertheless, the identification of magnetically anomalous sites indicates where targeted chemical analyses should be performed in future work. This strategy substantially reduces analytical costs by allowing only a few representative samples to undergo detailed chemical characterization rather than all samples collected.

Ultimately, the findings reinforce the need for non-invasive, cost-effective tools for urban air-quality assessment, particularly in under-monitored regions. Magnetic biomonitoring with T. recurvata emerges as a scalable, sensitive, and economically efficient approach for identifying pollution gradients and informing decision-making processes aimed at improving environmental equity and public health in urban green spaces.

Based on the spatial patterns identified in this study, targeted and ecologically informed mitigation measures can be proposed, with particular emphasis on reinforcing green barriers. In the most polluted sectors of the park—especially the central and southeastern areas adjacent to major traffic corridors—the effectiveness of vegetation as a particle filter can be enhanced by reinforcing existing tree barriers rather than introducing isolated or shaded understory species. In areas already dominated by mature Schinus molle (Peruvian pepper tree), reinforcement with additional individuals of the same species is recommended to increase canopy density and continuity, thereby improving particle interception and reducing lateral transport. In contrast, in park sections where tree cover is absent or discontinuous, the establishment of sun-tolerant, endemic vegetation adapted to semi-arid conditions should be prioritized to create functional green barriers. Importantly, planting new species beneath large, established trees without adequate light availability is unlikely to provide additional mitigation benefits, as shaded vegetation has limited growth and particle-capture capacity. Therefore, effective green infrastructure design in urban parks should focus on reinforced, continuous, and light-exposed vegetation barriers aligned with dominant pollution pathways, complemented where possible by traffic-calming measures along adjacent roads.