Air Quality Impacts on the Giant Panda Habitat in the Qinling Mountains: Chemical Characteristics and Sources of Elements in PM_2.5

Abstract

The wild giant panda habitat is inaccessible and far away from the main areas of human activity, so environmental pollutants entering the habitat are mainly the result of external migration and spread through the atmospheric advection and diffusion processes and particulate matter deposition. To research the variation, transmission route, chemical characteristics, and source of PM_2.5 in the habitat of wild giant pandas, we set up a PM_2.5 sampling point near the Shaanxi Foping National Nature Reserve (SFNNR), which is the area with the highest population density of wild giant pandas in the Qinling Mountains. The 12-month average concentration of PM_2.5 was 11.3 ± 7.9 μg/m³ from July 2021 to June 2022, and the 12-month average concentration did not exceed the limit value set in the standard. In the results of our analysis of element concentrations, As and Pb were much lower than the limit standard. Si, S, P, and Cl accounted for 99.60% of nonmetallic elements, while the proportion of the six metallic elements, Na, Mg, Al, K, Ca, and Fe, was 96.27%. According to the analysis results of enrichment factor (EF) and pollutant emission sources, there were four sources for the total elements in PM_2.5, which were mainly distributed in the areas around the reserve. These included dust, coal combustion, biomass burning, and traffic-related emissions, which contributed 55.10%, 24.78%, 11.91%, and 8.22% of the total element mass in PM_2.5, respectively. Additionally, Pb, Cu, Zn, As, Sc, Co, Ga, Mg, and, especially, Se were severely affected by human activities (coal burning, biomass burning, and traffic-related emissions). In the villages and towns around the wild giant panda habitat, the majority of energy for cooking and heating comes from coal and biomass burning, and older vehicles with high emissions are used more frequently. Therefore, to better protect the health of and reduce the impact of environmental pollution on wild giant pandas, we put forward relevant recommendations, including upgrading the energy structure of towns and villages near the habitat to increase the proportion of clean energy, such as photovoltaic power generation, natural gas, etc.; decreasing the combustion of coal and biomass; encouraging the upgrading of agricultural diesel machines and older vehicles used in these areas; and setting limits on vehicle emissions in areas surrounding the habitat.

1. Introduction

The giant panda (Ailuropoda melanoleuca), a national treasure and living fossil, is an extremely rare and wild endangered animal in China, and it is also the flagship species of Species Diversity Protection in the world [1,2]. The area of giant panda habitats in China is approximately 2.58 × 10⁶ ha, and they are distributed in the Qinling Mountains, Minshan Mountains, Daxiangling Mountains, Qionglai Mountains, Xiaoxiangling Mountains, and Liangshan Mountains. In 2013, the wild giant panda population in China numbered 1864, and their habitats were distributed in the Sichuan (1387, 74%), Shaanxi (345, 19%), and Gansu (132, 7%) provinces [3]. The giant panda habitat in the Qinling Mountains of the Shaanxi Province is distributed in the northernmost and easternmost region of the wild giant panda habitats in China (as shown in Figure S1) and is geographically isolated from the giant panda habitats in the other five mountains. Therefore, the giant pandas in Qinling can be evolutionarily differentiated from the giant pandas in the other five mountains. Due to the evolutionary processes and significant differences in their external biological characteristics and genetic characteristics, the giant panda population in Qinling is taxonomically identified as “Ailuropoda melanoleuca qinlingensis” [4]. Furthermore, the appearance of many brown giant pandas in the Qinling Mountains in recent years has made their particularity and biological and ecological value more obvious [5].

Human activities (mineral mining, metal smelting, agricultural activities, transportation, product manufacturing, etc.) have discharged significant amounts of toxic and harmful substances into the environment [6,7]. These environmental pollutants migrate and spread throughout the whole ecosystem through water flow, atmospheric movement, and other pathways, endangering the health of human beings, animals, and plants [8]. In recent years, the impact of environmental pollutants on the health of giant pandas has aroused wide concern [9]. According to past research, giant pandas in Qinling are threatened by heavy metal pollutants and persistent organic pollutants (POPs) to varying degrees [10,11]. Additionally, heavy metal pollutants (Cd, Pb, As, Hg, and other elements) are found in the soil, water, staple food (bamboo), and feces of giant pandas in their habitat in the Qinling Mountains, which are mainly the result of coal burning, garbage burning, and traffic pollution sources [12]. Meanwhile, high levels of pollutants have been detected in a giant panda’s body [13], which were related to abnormalities of alanine aminotransferase, lactate dehydrogenase, total bilirubin, etc., in the giant panda’s body [14,15]. The wild giant panda habitats are inaccessible and far away from the main areas of human activity, and administrators strictly manage their habitats. Therefore, there is almost no emission source of environmental pollutants in the wild giant panda’s habitat, and the pollutants entering the habitat come from external sources and migrate and spread through the processes of atmospheric motion and particulate matter deposition.

Fine particulate matter (PM_2.5, particulate matter with an aerodynamic diameter ≤2.5 μm) can, directly and indirectly, affect climates and ecosystems [16,17], e.g., by causing the air quality to deteriorate via a reduction in atmosphere visibility, etc. [18,19,20,21]. PM_2.5 environmental pollutants can also be transported over long distances by atmospheric movement, which can harm the health of humans and other organisms [22,23]. PM_2.5 can directly enter the bodies of giant pandas via breathing and eating (PM_2.5 adheres to the surface of bamboo leaves) and harm their health. Therefore, PM_2.5 is the most harmful kind of atmospheric particulate matter to giant pandas. In the cities around the wild giant panda habitat in the Qinling Mountains, the annual average concentrations of PM_2.5 are 169.3 ± 101.7 μg/m³, 135.5 ± 70.0 μg/m³, and 132.0 ± 78.5 μg/m³ in Xi’an, Weinan, and Baoji, respectively [24]. Additionally, the concentration of PM_2.5 in Hanzhong during autumn is 27.0 ± 14.6 μg/m³ [25]. However, there is no research on PM_2.5 in wild giant panda habitats. Due to the complex and variable chemical composition, variety of sources, and processes of PM_2.5 [26], it is important to research the variation, transmission route, and source of PM_2.5 in the habitat of wild giant pandas to evaluate the impact of environmental pollutants emitted by human activities on them, but it is quite challenging and pioneering.

Therefore, we carried out this study in order to understand the atmospheric environment of the wild giant panda habitat in the Qinling Mountains, to identify the pollutant emission sources in the atmosphere, and to provide reasonable and effective suggestions for reducing the impact of environmental pollution on wild giant pandas. A PM_2.5 sampling point was set up near SFNNR, which is the area with the highest population density of wild giant pandas in the Qinling Mountains. By collecting PM_2.5 samples from July 2021 to June 2022, the concentration variations in PM_2.5 and its elements were analyzed, and the pollutant emission sources and their contributions were identified. The results obtained with the present study (pollution sources identification) provided the scientific basis for policy recommendations to better protect the health of and reduce the impact of environmental pollution on wild giant pandas.

2. Materials and Methods

2.1. Study Area and Sampling Point

The SFNNR is located in the northwest of Foping County, Hanzhong City, Shaanxi Province, on the southern slope of the middle part of the Qinling Mountains, between 33°31′ N and 33°43′ N and 107°40′ E and 107°55′ E (Figure 1). SFNNR is at the core of the wild panda habitat in Qinling. Its west, northwest, north, northeast, and east are bordered by the Changqing National Nature Reserve, Huangbaiyuan National Nature Reserve, Laoxiancheng National Nature Reserve, Zhouzhi National Nature Reserve, and Guanyinshan National Nature Reserve, respectively. The altitude of the reserve increases from the southeast to the northwest. The Paotong Gully boasts the lowest altitude at 980 m, while Luban Peak stands at the highest altitude at 2904 m, resulting in a vertical drop of 1924 m. The SFNNR is extremely rich in terms of its biodiversity, with a wide variety of rare and endangered animals and plants. The reserve contains 452 species of terrestrial vertebrates from 32 orders and 101 families, including 73 mammalian species from 7 orders and 26 families and 316 bird species from 17 orders and 54 families [27,28]. Additionally, it contains 295 species of spore plants from 185 genera and 91 families and 1376 species of seed plants from 560 genera and 133 families, including 13 species of gymnosperms from 11 genera and 5 families and 1363 species of angiosperms from 549 genera and 128 families [29,30,31].

Our PM_2.5 sampling point (33°41′22.120″, 107°53′34.704″, 1759 m) was located at the Liangfengya Protection Station (as shown in Figure 1) in the northeast of the SFNNR. The sampling instrument was located on the top of the protection station, which consisted of only one floor. There was no signification electromagnetic interference in close proximity, and a dependable power supply and lightning protection equipment were present. Additionally, there were no obstructions from surrounding structures. Virgin forests cover dozens of square kilometers around the sampling point, and the land use forms near the sampling point are consistent. PM_2.5 samples were collected on 47 mm Teflon filters (Whatman Limited, Maidstone, UK) using a mini-volume sampler (Airmetrics, Springfield, OR, USA) that operated at a flow rate of 5 L min⁻¹. The sampling interval was one week, during which the sampler worked continuously. For mass determinations, the filters before and after sampling were equilibrated under a controlled temperature (20–23 °C) and relative humidity (35–45%) before the measurements were made. Additionally, the mass of the filter before and after sampling was determined gravimetrically using a Sartorius MC5 electronic microbalance with ±1 μg sensitivity (Sartorius, Göttingen, Germany). Collected samples were stored in a refrigerator at −4 °C before inorganic elemental analyses, and the field blanks were collected and analyzed to account for possible background effects.

2.2. Sample Analysis

The concentrations of elements (Na, Mg, Al, Si, P, S, Cl, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Br, Sr, Ba, and Pb) in the PM_2.5 were determined by energy-dispersive X-ray fluorescence (ED-XRF) spectrometry (Epsilon 5 ED-XRF, PANalytical B.V., Almelo, The Netherlands) [32]. The determination operation and quality control were in accordance with the ambient air determination of inorganic elements in ambient particle matter energy dispersive X-ray fluorescence spectroscopy (ED-XRF) method [33], and the analytical accuracy of the ED-XRF measurements was determined with the NIST Standard Reference Material 2783 (National Institute of Standards and Technology, Gaithersburg, MD, USA). The Method detection limit (MDL) of the energy-dispersive X-ray fluorescence spectrometry is shown in Table S1, where the concentration units are converted to μg/m³.

2.3. Data Analysis

2.3.1. EF of Elements in PM_2.5

The EF was used to represent the enrichment degree of elements in atmospheric particulate matter, and the equation is as follows:

$$EF=\frac{{\left(\raisebox{1ex}{$C_n$}\!\left/ \!\raisebox{-1ex}{$C_{ref}$}\right.\right)}_{sample}}{{\left(\raisebox{1ex}{$C_{Bn}$}\!\left/ \!\raisebox{-1ex}{$C_{Bref}$}\right.\right)}_{baseline}}$$

(1)

where $C_n$ and $C_{ref}$ are the concentration of elements and reference elements in PM_2.5, respectively, and $C_{Bn}$ and $C_{Bref}$ are the elements and reference elements’ background concentration values in the soil, respectively. The reference element was Fe in our research [34,35,36], and the background values of the elements in Shaanxi are shown in Table S2 according to the elemental soil background values in China [37]. If the EF is <1, the element mainly comes from a natural source. When the EF is >1, the elements are affected by different degrees of human activity, which is classified as primarily (10 < EF < 100) or entirely (EF > 100) from man-made sources of pollution [38]. The classification standards of the EF are slight enrichment (EF < 10), moderate enrichment (10 < EF < 100), and high enrichment (EF > 100) [39,40].

2.3.2. Receptor Model for Emission Sources

Positive matrix factorization (PMF) is a mathematical approach that is widely used in environmental science [21,41,42,43] to analyze the sources of atmospheric particulate matter by quantifying their contributions [44]. PMF decomposes a matrix of speciated sample data into two matrices, a factor contribution matrix ($g_{ik}$) and a factor profiles matrix ($f_{ki}$), and then minimizes the objective function, Q. The equations are as follows:

$$x_{ij}={\displaystyle \sum }_{k=1}^p{g}_{ik}{f}_{ki}+e_{ij}$$

(2)

$$Q={\displaystyle \sum }_{i=1}^n{\displaystyle \sum }_{j=1}^m{\left[\frac{x_{ij}-{\sum }_{k=1}^p{g}_{ik}{f}_{ki}}{u_{ij}}\right]}^2$$

(3)

where $x_{ij}$ is the concentration of the sample, $e_{ij}$ is the model residual, and $u_{ij}$ is the uncertainty. The PMF Model (PMF 5.0) from the US Environmental Protection Agency (EPA) determines the optimal number of factors by calculating the values of residuals (between −3 and +3) and using a small Q_true/Q_expect [45]. In our study, the PMF model was employed to analyze the source of the total elements (26 elements we determined) in PM_2.5.

2.3.3. Analysis Using MeteoInfo Software

The Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Model, developed by the National Oceanic and Atmospheric Administration of the USA and the Bureau of Meteorology of Australia, is a professional model widely used in environmental science research for analyzing the trajectory and simulating the diffusion of air masses and analyzing the transportation, diffusion, and settlement of pollutants in the atmosphere [46,47,48]. We used MeteoInfo software and its TrajStat package developed for meteorological data visualization and analysis to analyze the PM_2.5 backward trajectories, including the trajectory clustering, potential source contribution factors (PSCFs), and concentration weight trajectories (CWT) [49,50]. The meteorological data were obtained from the Global Data Assimilation System (GDAS) (ftp://arlftp.arlhq.noaa.gov/pub/archives/gdas1, accessed on 7 November 2022) database of the American National Centers for Environmental Prediction (NCEP) and had a spatial resolution of 1 degree by 1 degree.

In our research, the angle distance classification method in the TrajStat package of MeteoInfo was used to cluster the air mass trajectories (48 h, 500 m) in order to obtain the air transport characteristics in the SFNNR.

The PSCF value, which measures conditional probability, represents the contribution of a potential pollution source, and determines the spatial distribution by analyzing the trajectories of the air mass [51,52]. The equation is as follows:

$$PSC{F}_{ij}=\frac{m_{ij}}{n_{ij}}$$

(4)

where the $m_{ij}$ is the number of pollution trajectories corresponding to the PM_2.5 concentrations of the sampling point exceeding the limit value (15 μg/m³) in grid ij, and $n_{ij}$ is the total number of trajectories in grid ij.

The CWT calculates the weight concentration of the grid based on the weight concentration trajectory to reflect the pollution sources in the study area [53,54]. The equation is as follows:

$$C_{ij}=\frac{{\sum }_{l=1}^m{C}_l{\tau }_{ijl}}{{\sum }_{l=1}^m{\tau }_{ijl}}$$

(5)

where $C_{ij}$ is the average weight concentration (μg/m³) of grid ij; $l$ is the trajectory; $m$ is the number of trajectories; $C_l$ is the corresponding pollution concentration when trajectory $l$ reaches grid ij; and ${\tau }_{ijl}$ is the residence time of trajectory $l$ in grid ij, which is calculated based on the trajectory nodes falling within grid ij.

3. Results and Discussion

3.1. The Concentrations of PM_2.5 near the SFNNR

Figure 2A–C shows the weekly variation, monthly variation, and quarterly variation in PM_2.5 concentrations near the SFNNR during the sampling period. The monthly and quarterly average concentrations of PM_2.5 are shown in

Table 1. The monthly and quarterly average concentrations of PM_2.5 near Shaanxi Foping National Nature Reserve.

	Monthly Average Concentration (μg/m3)		Quarterly Average Concentration (μg/m3)
2021	Jul	6.39 ± 1.81	Q3	5.26 ± 1.90
	Aug	5.31 ± 2.65
	Sep	4.58 ± 1.98
	Oct	5.46 ± 2.07	Q4	10.29 ± 7.08
	Nov	8.61 ± 1.08
	Dec	13.28 ± 8.33
2022	Jan	17.88 ± 13.07	Q1	17.94 ± 10.24
	Feb	18.02 ± 5.86
	Mar	18.77 ± 9.42
	Apr	13.56 ± 3.44	Q2	11.50 ± 3.46
	May	11.71 ± 2.75
	Jun	8.30 ± 1.35

, and there were significant differences (p < 0.05) between the monthly and quarterly average concentrations of PM_2.5.

The 12-month average of PM_2.5 concentration near the SFNNR was 11.3 ± 7.9 μg/m³ from July 2021 to June 2022. From July 2021 to September 2021 and from March 2022 to June 2022, the concentration of PM_2.5 decreased each month. From September 2021 to March 2022, the concentration of PM_2.5 increased each month. It is well-known that climate factors, such as rainfall, snowfall, and wind, have a great impact on the concentration of PM_2.5 [55,56,57]. In the period from August to October 2021, there was a long-term rainfall record in the SFNNR; the rainfall was about 878.3 mm over these three months, and the concentration of PM_2.5 in this period was the lowest during our sampling period. Although the average concentrations were highest in the first quarter of 2022 in the SFNNR, the lowest value of PM_2.5 concentration occurred during the period from 10 to 16 January 2022 due to the continuous snowfall during this PM_2.5 sample collection period. The 12-month average concentration did not exceed the PM_2.5 limit value set out in the Ambient Air Quality Standards (15 μg/m³) [58] during the sampling period. The samples with PM_2.5 concentrations exceeding the average concentration were collected from December 2021 to May 2022. Meanwhile, the PM_2.5 concentrations were the lowest in the third quarter of 2021; the concentrations were highest in the first quarter of 2022.

According to the results of the variations in PM_2.5 concentration near the SFNNR, the concentration was different in different seasons, with the lowest concentrations being seen during the summer and the highest during the winter, which is consistent with other research results regarding PM_2.5 and other components ((NH₄)₂SO₄, NH₄NO₃, OM, EC, etc.) of the atmosphere in the surrounding area of Foping [41,59,60]. Meanwhile, the concentration of PM_2.5 near Foping was lower than in the surrounding area during different seasons. The annual average concentration of PM_2.5 was 169.3 ± 101.7 μg/m³, 135.5 ± 70.0 μg/m³, 132.0 ± 78.5 μg/m³, 120.3 ± 83.8 μg/m³ in Xi’an, Weinan, Baoji, and the northern Qinling Mountains, respectively [24]. The concentration of PM_2.5 in Hanzhong during autumn was 27.0 ± 14.6 μg/m³ [25]. The distances from the sampling point to the areas with frequent human activity in Baoji, Hanzhong, Xi ‘an, and Weinan are 97 km, 107 km, 118 km, and 176 km, respectively.

Figure 2. Concentrations (μg/m³) of PM_2.5 near the SFNNR from July 2021 to June 2022. (A) Weekly PM_2.5 concentrations. (B) Monthly PM_2.5 concentrations. (C) Quarterly PM_2.5 concentrations. The red dashed line indicates the ambient air quality standard of PM_2.5 concentrations in the nature reserve [58].

Figure 2. Concentrations (μg/m³) of PM_2.5 near the SFNNR from July 2021 to June 2022. (A) Weekly PM_2.5 concentrations. (B) Monthly PM_2.5 concentrations. (C) Quarterly PM_2.5 concentrations. The red dashed line indicates the ambient air quality standard of PM_2.5 concentrations in the nature reserve [58].

3.2. The Concentrations of Elements in PM_2.5 near the SFNNR

The concentration variations of 26 elements in PM_2.5 during the sampling period (July 2021 to June 2022) are shown in Figure 3 and Figure S2 (Na and Mg were semi-quantified).

The concentration variations of nonmetallic elements (Si, S, P, Cl, As, Br, and Se) are shown in Figure 3B,C,G,H,P,Q,X. The 12-month average concentrations were 0.52 ± 0.52 μg/m³ (Si), 0.59 ± 0.31 μg/m³ (S), 0.012 ± 0.004 μg/m³ (P), 0.049 ± 0.030 μg/m³ (Cl), 0.0014 ± 0.00078 μg/m³ (As), 0.0020 ± 0.0011 μg/m³ (Br), and 0.00074 ± 0.00042 μg/m³ (Se). Si, S, P, and Cl were the main elements (>1%) and accounted for 99.60% of the total nonmetallic elements in the PM_2.5, and the proportions of Br, As, and Se were 0.19%, 0.14%, and 0.07%, respectively. The highest concentrations of S, Si, and As all occurred in the same period (9 March 2022 to 16 March 2022), and those of Cl, Br, and Se also all occurred in the same period (10 January 2022 to 26 January 2022). Meanwhile, the lowest concentrations of S, Si, As, and Br all occurred in the same period from (27 December 2021 to 3 January 2022), and those of Cl and Se occurred in the period from 30 September 2021 to 13 October 2021 and 28 August 2021 to 8 September 2021, respectively. However, the variation in S was different from that of other nonmetallic elements, with the highest value, 0.019 μg/m³, occurring from 4 May 2022 to 11 May 2022, and the lowest value, 0.00001 μg/m³, occurring from 26 January 2022to 16 February 2022. Although the concentrations of S and Si elements were high among the nonmetallic elements, the variation trends in the period from the third quarter of 2021 to the second quarter of 2022 were different, with a decrease in S from 59.9% to 45.7% and an increase in Si from 32.9% to 50.3%. The concentration of As within the nature reserve was lower than the concentration limit value (0.006 μg/m³) of Ambient Air Quality Standards [58] over the entire sampling period.

The concentration variations of metallic elements (Al, K, Ca, Fe, Ti, Mn, Zn, Sc, Cr, Cu, Ga, Sr, Ba, Pb, V, Co, Ni, Na, and Mg) are shown in Figure 3 and Figure S2. The 12-month average concentrations were 0.31 ± 0.22 μg/m³ (Al), 0.22 ± 0.17 μg/m³ (K), 0.22 ± 0.31 μg/m³ (Ca), 0.20 ± 0.20 μg/m³ (Fe), 0.016 ± 0.017 μg/m³ (Ti), 0.012 ± 0.0058 μg/m³ (Mn), 0.011 ± 0.0062 μg/m³ (Zn), 0.0015 ± 0.0012 μg/m³ (Sc), 0.0021 ± 0.0009 μg/m³ (Cr), 0.0041 ± 0.00085 μg/m³ (Cu), 0.0020 ± 0.0019 μg/m³ (Ga), 0.0010 ± 0.00022 μg/m³ (Sr), 0.0081 ± 0.0029 μg/m³ (Ba), 0.0060 ± 0.0027 μg/m³ (Pb), 0.00081 ± 0.00041 μg/m³ (V), 0.00063 ± 0.00020 μg/m³ (Co), 0.00092 ± 0.00038 μg/m³ (Ni), 0.36 ± 0.10 μg/m³ (Na), and 0.50 ± 0.13 μg/m³ (Mg). A proportion of six metallic elements (Na, Mg, Al, K, Ca, Fe) accounted for 96.3% of the total (28.7% Mg, 20.9% Na, 15.7% Al, 11.8% K, 9.8% Ca, and 9.4% Fe), and the other 13 metallic elements made up the remaining 3.7%. The variations in metallic element concentrations and the mass proportions of total metallic elements were different and are shown below. The highest concentrations of all the elements occurred in the same period (9 March 2022 to 16 March 2022), and the lowest concentrations occurred in two periods (28 August 2021 to 13 October 2021 and 27 December 2021 to 3 January 2022) due to the long-term rainfall from August to October 2021 and the continuous snowfall in January 2022. In terms of the variation in mass proportion, the highest and lowest values of Na, Mg, V, Cr, Mn, Co, Ni, Cu, and Ga occurred from July to September 2021 and February to May 2022, respectively, and the opposite pattern was observed for K, Sc, and Sr. Meanwhile, the highest and lowest proportions of Al, Ca, and Ti occurred in March to June 2022 and January 2022, respectively, with the opposite pattern being observed for Zn, Ba, and Pb. The concentration of Pb within the nature reserve was much lower than its Ambient Air Quality Standards [58] limit value (0.5 μg/m³) over the entire sampling period.

3.3. The EF of Elements of PM_2.5 near the SFNNR

The analysis results of the elements’ EFs in PM_2.5 near the SFNNR are shown in Figure 4 and Figure 5. The correlation analysis results of the elements’ EFs are shown in Table S4.

The average value of the elements’ EFs in descending order were 2162.0 ± 1657.2 (Se), 78.6 ± 55.1 (Pb), 62.9 ± 59.5 (Cu), 42.9 ± 27.5 (Zn), 34.9 ± 23.9 (As), 31.3 ± 39.0 (Sc), 19.8 ± 22.2 (Co), 19.5 ± 21.4 (Ga), 14.9 ± 13.0 (Mg), 9.8 ± 7.9 (Cr), 9.1 ± 7.4 (Na), 8.8 ± 6.1 (Ni), 5.7 ± 3.7 (Mn), 4.6 ± 4.5 (Ba), 3.2 ± 2.3 (V), 2.3 ± 1.6 (K), 2.2 ± 1.0 (Sr), 1.3 ± 0.4 (Ca), 0.9 ± 0.4 (Al), and 0.6 ± 0.1 (Ti). According to the correlation analysis results, except for Ca, Ti, and Sr, which have low EF values and are least affected by human activities, the EF of the other elements showed a strong correlation. The highest EF values of those elements all occurred in January 2022, and the lowest EF values all occurred between March and May 2022. Meanwhile, the proportion of EFs greater than 100 (high enrichment standard) was 100% for Se, 27.3% for Pb, 13.6% for Cu, and 4.6% for Zn. The proportion greater than 10 (moderate enrichment standard) were 100% for Pb, 97.7% for Cu, 95.5% for Zn, 90.9% for As, 93.2% for Sc, 59.1% for Co, 61.4% for Ga, 54.5% for Mg, 40.9% for Cr, 31.8% for Na, 34.1% for Ni, 9.1% for Mn, 4.6% for Ba, and 2.3% for V.

Therefore, we should pay more attention to Pb, Cu, Zn, As, Sc, Co, Ga, Mg, and, especially, Se in PM_2.5 during our sampling period near the SFNNR. The emission sources of these elements are the main factors affecting the variations in concentration, and we discuss them below.

3.4. The Results of the Receptor Model for Emission Sources by PMF

The analysis results of source contribution from the PMF 5.0 model are shown in Figure 6 and Figure 7, which show the concentration and percentage of elements from the four sources of the total elements in PM_2.5 and the variations in these sources in the SFNNR during the sampling period. The Q/Qexp analysis results of different number factors are shown in Table S5, and the external validation (including a correlation analysis between different pollution sources and corresponding tracers in time series) is shown in Figure S4.

According to the PMF results, the total element mass in PM_2.5 was strongly correlated with the observed values (r² = 0.998, slope = 0.994), and the model-calculated concentrations of the elements exhibited good linearity and correlation with the measured values (r² = 0.668–0.996). Therefore, the above results show that the four sources account for much of the variability in the data. Through comparisons between the PMF profiles and reference profiles from previous research, there were four presumptive sources for the total elements in PM_2.5, including (i) dust, (ii) coal combustion, (iii) biomass burning, and (iv) traffic-related emissions.

The first source factor, which contributed 55.1% of the total element mass in PM_2.5 (Figure 7A) during the sampling period, was natural dust because it had a high loading of Al (66.0%), Si (80.1%), Ca (81.9%), Ti (84.6%), Fe (78.0%), and Sc (61.7%) [42,61,62,63]. Additionally, the variation of the natural dust factor contributing to the total elements in PM_2.5 was significant (Figure 7B), showing that the highest contribution was during spring because the wind is stronger in spring. The second source factor, which contributed 24.8% of the total element mass in PM_2.5 (Figure 7A) during the sampling period, was coal burning emissions because it had a high loading of S (56.2%), Zn (42.6%), As (40.0%), Se (51.3%), and Pb (31.5%). As is a significant tracer for coal combustion [64,65] and the main source of Se in the atmosphere is coal burning [66,67,68]. S, Zn, and Pb are the main elements released through coal burning [32,69,70]. Additionally, the third source factor, which contributed 11.9% of the total element mass in PM_2.5 (Figure 7A), was biomass burning as it has a high loading of Cl (48.3%) and K (40.6%). K was a tracer, and Cl was shown to be emitted by biomass burning in previous research [42,43,61,71]. Additionally, the contribution variations of coal burning emissions and biomass burning to the total elements in PM_2.5 made a greater contribution during the heating season. The fourth source factor, which contributed 8.2% of the total element mass in PM_2.5 (Figure 7A) during the sampling period, was traffic-related emissions because they have a high loading of Cu (54.1%), Ni (43.2%), Mn (38.4%), and V (34.5%). In transportation processes, such as gasoline combustion, tire brake pad wear, and the loss of lubricating oil, the elements Cu, Ni, Mn, and V are emitted via aerosol, according to previous research [69,72,73].

Combining the results of the analysis of the elements (concentrations and EF) in PM_2.5, the elements (Se, Cu, Zn, Pb, etc.) with high values of EF were mainly from pollutant sources, such as coal and biomass burning and traffic-related emissions. Additionally, the coal burning emission was the largest of the above three sources. The variations in coal and biomass burning contributions to the total elements in PM_2.5 were seasonal, and their contributions to the total elements in PM_2.5 were high during the heating season, as is shown in Figure 7, because they are the main sources of energy for heating and cooking in villages and towns around the reserve.

3.5. The Analysis Results of Cluster, PSCF, and CWT by MeteoInfo Software

Using MeteoInfo’s TrajStat package, the results of the trajectory clustering, PSCF, and CWT in the SFNNR during the sampling period are shown in Figure 8, Figure 9 and Figure 10.

The movement of air mass is the main transport route of PM_2.5 in the atmosphere [74,75]. According to our trajectory clustering results, shown in Figure 8, the clustering directions of the main air mass transport to the SFNNR were from the southeast (33.6% (July 2021), 47.3% (August 2021), 39.9% (September 2021), 36.2% (February 2022), 38.1% (June 2022)), northwest (40.9% (October 2021), 54.2% (November 2021), 39.4% (December 2021), 55.9% (March 2022), 49.9% (April 2022), 41.7% (May 2022)) and south (44.4% (January 2022)), respectively. Meanwhile, the long-distance air mass transport primarily came from the northwest areas of the reserve, including the Shaanxi Guanzhong region, Gansu Province, Ningxia Province, Qinghai Province, and Inner Mongolia. The air mass coming from the southeast direction of the reserve primarily came from the Shaanxi southern region, and a small portion comes from Hubei Province, Henan Province, and Chongqing City.

The PSCF and CWT results represented the possible impact and contribution of potential pollution sources to the SFNNR during the sampling period. Except for the period from December 2021 to May 2022, the concentrations of PM_2.5 near the SFNNR were all lower than the limit value. Thus, the PSCF results were only for the abovementioned period (December 2021 to May 2022). According to the results, the areas with a big impact on PM_2.5 were mainly distributed within Shaanxi Province and were rarely distributed in Gansu, Ningxia, Inner Mongolia, Shanxi, Henan, Hubei, Chongqing, and Sichuan provinces. According to the CWT results, the areas with high contributions of potential pollution sources during the sampling period were also mainly distributed within Shaanxi Province.

Combining the results of the source analysis with those of the PMF in the SFNNR, the concentration of PM_2.5 was the highest and the sources of coal combustion and biomass burning contributed the most (Figure 7) from December 2021 to March 2022. Meanwhile, the areas that made high contributions of potential pollution sources were mainly distributed in the areas around the reserve. Therefore, during this time, the pollutant emissions were large due to heating, and more attention needs to be paid to the impact of environmental pollutants on giant pandas at this time.

4. Conclusions

In our research, the concentration of PM_2.5 near the SFNNR, with 11.3 ± 7.9 μg/m³ being the 12-month average concentration from July 2021 to June 2022, was much lower than the concentrations in cities with more human activity around the reserve. Thus, the Qinling Mountains not only provide a suitable habitat for the wild giant panda but also reduce the harm done to them via environmental pollution. Si, S, P, and Cl accounted for 99.6% of nonmetallic elements, while a proportion of six elements (Na, Mg, Al, K, Ca, Fe) represented 96.3% of the total metallic elements. Additionally, we should pay more attention to the elements Pb, Cu, Zn, As, Sc, Co, Ga, Mg, and, especially, Se in PM_2.5 during the sampling period because these elements (EF > 10 or EF > 100) were more affected by human activity. According to the analysis results on the emission sources, the four sources (dust, coal combustion, biomass burning, and traffic-related emissions) contributed 55.1%, 24.8%, 11.9%, and 8.2% of the total element mass in PM_2.5, respectively. Additionally, when combining the analysis results of elements (concentrations and EF) with the results of emission sources for the total elements in PM_2.5, we can see that the elements (Se, Cu, Zn, Pb, etc.) with high EF values were mainly from the pollutant sources of coal, biomass burning, and traffic-related emissions. Because coal and biomass burning are the main sources of energy for heating and cooking in villages and towns around the reserve, they are important influencing factors for the total elements in PM_2.5 in the SFNNR. These two emission sources were, due to their large pollutant emissions, major contributors to the poor air quality during the heating season (December 2021 to March 2022), in particular, as illustrated in Figure 7.

The main energy for cooking and heating comes from coal and biomass burning, and older vehicles with high emissions are used more frequently in the villages and towns around the wild giant panda habitat. It is important to control the emission of environmental pollutants around the habitat to better protect the health of and reduce the impact of atmospheric pollution on wild giant pandas. Therefore, pollutant emissions should be managed, and relevant policies should be created and enforced. The energy structure of towns and villages near the habitat should be upgraded by the government as soon as possible. This may include increasing the use of clean energy, such as photovoltaic power generation, natural gas, etc., and decreasing the combustion of coal and biomass. Meanwhile, administrators need to set up preferential policies to encourage the upgrading of agricultural diesel machines and older vehicles used in the abovementioned areas and should set limits on vehicle emissions in areas surrounding the habitat as established by many first-tier Chinese cities.