Different characteristics of microbial diversity and special functional microbes in rainwater and topsoil before and after 2019 new coronavirus epidemic in Inner Mongolia Grassland

Abstract

Grassland ecosystems are vital terrestrial ecosystems. As areas sensitive to climate change, they are critical for assessing the effects of global climate change. In China, grasslands account for over 40% of the land area. There is currently limited information on microbial diversity evolution in different grassland areas, particularly microorganisms with ice nucleation activity (INA) and their potential resources with potential influence to regulate regional precipitation and climate. We used Illumina MiSeq to sequence the 16S rRNA V3–V4 hypervariable region and performed a simple droplet freezing experiment to determine the variation in the grassland microbial community species composition and community structure. Rainwater and topsoil samples from the Hulunbuir Grassland in Inner Mongolia collected over three years were characterized. The dominant bacterial genus in the rainwater was Massilia, and the dominant fungus was Cladosporium. Additionally, the dominant bacteria in the soil were Sphingomonas, and the dominant fungus was Gibberella. There were differences in the microbial communities before and after the coronavirus disease epidemic. Pathogenic microorganisms exhibited inconsistent responses to environmental changes. The low relative abundance of known high-INA microorganisms and the higher freezing temperature indicated that unknown high-efficiency biological ice nucleating particles may be present. We found significant differences in species diversity and richness between the rainwater and soil populations in grassland areas by analyzing the sample community structures. Our research results revealed the species composition and structure of the microbiota in grassland ecosystems in China, indicating that environmental media and human activities may affect the microbiota in the grassland area and indicating underlying microorganisms with high INA.

Graphical abstract

1. Introduction

Environmental microbiota exhibit unique relationships with their environmental media and local ecological conditions. Many factors, such as moisture, temperature, and environmental pollutants, can affect the community structure (Cao et al., 2014; Du et al., 2018c; Sun et al., 2018; Azua-Bustos et al., 2020). Advances in genomics have renewed these relationships, indicating the adaptation of microbiota to the living environment (Schulze-Makuch et al., 2018; Chen et al., 2021).

Environmental microbiota play a vital role in biogeochemical cycles, such as the water, carbon, and nitrogen cycles. Soil microorganisms play a critical function as decomposers in the ecosystem, promoting the conversion of soil nutrients (Griffiths et al., 2003). Therefore, changes in the microbial community affect the nutrient cycling process and function. Microorganisms can act as biological ice nuclei to facilitate water cycle processes (Morris et al., 2008). Biological particles in the Earth's atmosphere are unique types of ice nucleating particles (INPs) because they can promote ice crystal formation in clouds at warm temperatures (Huang et al., 2021). Biological INPs (e.g., pollen, bacteria, fungal spores, and plankton), most of which contain ice nucleation proteins, can induce freezing at warmer temperatures (mostly ≥−15 °C) (Murray et al., 2012). Above this temperature, the only materials known to nucleate ice are biological. S. Zhang et al. (2020) denoted INPs with freezing temperatures of ≥−10 °C efficient INPs. Ice crystals play critical roles in modulating cloud microphysical properties and chemical compositions in the troposphere through precipitation formation processes, cloud radiative forcing, cloud electrification, etc. Thus, they indirectly influence the global hydrological cycle and climate change (DeMott et al., 2010; Mulmenstadt et al., 2015). Microbes well-known to exhibit ice nucleating activity include four bacterial genera (Pseudomonas, Lysinibacillus, Erwinia, and Xanthomonas) (Morris et al., 2004; Failor et al., 2017; Amato et al., 2015; Santl-Temkiv et al., 2015), and two fungal genera (Fusarium and Mortierella) (Froehlich-Nowoisky et al., 2015; Froehlich-Nowoisky and Poeschl, 2013; Pouleur et al., 1992). Morris et al. (2008) analyzed the bacterial population composition in precipitation samples from different regions and explained the relationship between P. syringae and the water cycle, and then proposed a hypothetical model of the life history of P. syringae driven by the environmental water cycle process. Delort et al. (2010) reviewed the analysis results of microbes in major cloud water and rainwater samples. The concentration of total bacteria is 10³–10⁵ cell mL⁻¹; the proportion of cultivable bacteria was low, almost less than 1%. Ahern et al. (2007) showed that over 60 fluorescent Pseudomonas strains were isolated from Hebridean cloud and rain samples. In general, there are few investigations on microbial populations and community structure in atmospheric precipitation, and further research is needed.

Society has been greatly affected by the coronavirus disease (COVID-19), which continues to affect the public. Bioaerosols have been reported as potential transmission routes for COVID-19 (Zhang et al., 2021). Due to the movement of people between cities and countries, COVID-19 infection has spread globally. Efforts have been made at the local, regional, and national levels to reduce the movement of people and quarantine the infected people to stop the spread of COVID-19. Traffic, markets, and small industries were temporarily shut down, which caused environmental changes. Venter et al. (2020) analyzed the concentration of pollutants during the COVID-19 epidemic in 34 global regions based on satellite remote sensing and on-ground air quality monitoring data. The results demonstrated that after considering the impact of meteorological conditions, the control measures during the epidemic reduced ρ(NOx), ρ(PM_2.5), and ρ(O₃) by 60%, 31%, and 4%, respectively. Chen et al. (2020) reported that during the epidemic period in China, traffic source pollutant emissions decreased significantly, residential heating and industrial emissions remained stable or slightly decreased, and air quality in most areas improved significantly in Wuhan and throughout the country. Research on most domestic regions in China, such as the Yangtze River Delta (L. Li et al., 2020; Lu et al., 2021) and the eastern region (R. Zhang et al., 2020), illustrates that pollutant emissions were significantly reduced during the national epidemic control period. The ambient air ρ(SO₂), ρ(PM_2.5), ρ(PM₁₀), ρ(NO₂), and ρ(CO) also significantly decreased, while ρ(O₃) rose in some areas (Lu et al., 2021; L. Li et al., 2020; Le et al., 2020). However, few studies have compared the microbiome before and after the COVID-19 epidemic. Considering the significant role of microorganisms in the environment, this area needs further exploration.

In our previous study (Du et al., 2017) on INPs in rainwater in the HunlunBuir grassland ecosystem, samples were collected from May–August for the years 2011–2013. The median freezing temperature (T₅₀) ranged from −7.7 °C to −10.3 °C. Therefore, we postulated that certain unknown efficient INPs are likely to exist in the regional precipitation. We further investigated the abundance, distribution, and sources of the efficient INPs in an additional study (S. Zhang et al., 2020), wherein we focused on efficient INPs in rainwater and soil. We found that biological INPs dominated the efficient INP population. The distinct distribution of specified INPs and known ice nucleation activity (INA) genera between rainwater and soils indicated that efficient INPs in rainwater may originate from remote sources instead of local ones.

This study aims to solve the following scientific questions: 1) What are the main bacteria and fungi in different environmental media in the grassland ecosystem and is there a difference? 2) Further study the scope of effective INP in the environment, determine the distribution of well-known biological INP and explore which species may be potentially undiscovered biological INP. 3) Research pathogenic microorganisms in the environment. Especially, whether the aerosol composition of environmental microorganisms be affected by human activities? To answer these questions, the droplet freezing experiment was used to determine the freezing temperature of rainwater samples, also, high-throughput genetic sequencing was used to characterize the microbiota of rainwater and soil. The results are expected to provide important references for exploring potential biological efficient ice nuclei in rainwater and their possible sources, the impact of human activities on outdoor air microorganisms, biological pollution, and biosafety related issues.

2. Materials and methods

2.1. Sampling site

All samples (Table 1 ) were collected from the Hulunbuir Grassland Ecosystem Research Station of the Chinese Academy of Agricultural Sciences at Xiertala Farm. The farm is located in the center of the Hulunbuir meadow steppe (49°19′ N, 120°03′ E, altitude: 628 m) in Inner Mongolia in the eastern part of the Eurasian steppe (total area: 88,000 km²). The sampling site is a grass-covered area with an ecosystem that features chestnut soil. The region is characterized by a semi-arid climate with an annual precipitation of 400 mm. There is a large variation in precipitation, which mostly occurs from June to August. Our collection campaigns were conducted in June, July and August in summer at temperatures of approximately 20 °C during the daytime and 12 °C at nighttime.

Table 1.

Sampling information and geographic coordinates for rainwater and topsoil.

Sample type	Sample ID	Sampling time (China Standard Time UT + 8:00) and duration of the rain	Rainfall intensity	Location
Rainwater	rainwater201808	3rd Aug 2018 (2 h)	Heavy rain (25–49.9 mm/24 h)	49°19′N 120°03′E
	rainwater201906	26th Jun 2019 (6 h)	Moderate rain (10–24.9 mm/24 h)
	rainwater201907	15th Jul 2019 (6 h)	Drizzle (<10 mm/24 h)
	rainwater201908	20th Aug 2019 (3 h)	Heavy rain (25–49.9 mm/24 h)
	rainwater202007	30th Jul 2020 (3 h)	Drizzle (<10 mm/24 h)
	rainwater202008	2nd Aug 2020 (<1 h)	Heavy rain(hail) (25–49.9 mm/24 h)
Topsoil	soil201808	2nd Aug 2018
	soil201906	25th Jun 2019
	soil201907	14th Jul 2019
	soil201908	19th Aug 2019
	soil202007	29th Jul 2020
	soil202008	1st Aug 2020

2.2. Sample collection and preparation

Rainwater samples were collected in several sterilized disposable plastic bags with an open area of 1 m² to ensure adequate rainwater volume collection. The plastic bags were held in barrels to form cylinders. The sampling equipment was placed in an open area at a suitable height (1.5–2 m) to avoid splashing from plants and the ground. The equipment was washed once with rainwater and then used to collect the rainwater. The volume of each sample was at least 4 L. Each sample was immediately stored in a refrigerator at 4 °C prior to processing. Part of the crude rainwater (volume of 0.5 L) was filtered through a sterile polytetrafluoroethylene filter with a pore diameter of 0.22 μm (Millipore, USA). It was used for DNA extraction and microbial community composition analysis. The crude rainwater samples were then transferred to sterile plastic containers. We took 10 mL of the crude rainwater sample and the filtrate and heated them to 100 °C with a water bath for 10 min to inactivate the proteinaceous INPs by disrupting the structure of membrane-bound proteins (Hill et al., 2016). Subsequently, the samples were subjected to an immersion freezing test. A thin layer of topsoil (3 cm in depth) was collected from 20 collection spots scattered in an S-shape. The soil samples were collected one day before the precipitation event. The soils were ground and passed through sieves with pore sizes of 0.150 mm and 0.045 mm for pretreatment. Then, 250 mg of ground soil was weighed for DNA extraction. All of the materials used for sampling were rinsed with ultrapure Milli–Q water and then sterilized via autoclaving to ensure that no contaminants remained.

2.3. Drop-freezing assay

Immersion freezing tests were performed using a modified device based on the Vali method (Vali, 1971). Immersion freezing, which is considered the predominant mode in the atmosphere, was tested to determine the ice nucleation temperature (Hande and Hoose, 2017). Forty-seven 10-μL droplets were equally distributed on a sterile plate with a cooling rate of 2 °C min⁻¹ and for five repetitions for a total of 235 droplets tested in each sample. The initial temperature was 0 °C, which then declined at a rate of 2 °C min⁻¹ until all the droplets froze. The modified device automatically detected the frozen drop signals and processed the experimental data (Yang and Feng, 2007). A study by Du et al. (2017) describes the parameters of the instrument.

2.4. Ice nucleation data analysis

In this manuscript, T₁₀ refers to the temperature at which 10% of the droplets froze, and T₅₀ refers to the temperature at which 50% of the droplets froze. The cumulative INP concentration at each temperature was calculated using the following equation (Vali, 1971):

$$\mathrm{K}\left(\mathrm{T}\right)=\left[ln{\mathrm{N}}_0-ln\left(\mathrm{N}\left(\mathrm{T}\right)\right)\right]/\mathrm{V}.$$

where K(T) is the concentration of INPs at temperature T, N₀ is the number of droplets tested, N(T) is the number of unfrozen droplets at a given temperature T and V is the volume of the droplet.

2.5. DNA extraction and PCR amplification

Microbial DNA was extracted using a PowerSoil DNA isolation kit (MoBio Laboratories, Carlsbad, CA, USA) according to the manufacturer protocol. The tools used in the experiments were clean and sterile. The extracted DNA was quantified using a Q-bit nucleic acid protein analyzer (Thermo Fisher Scientific). The extracted DNA samples were stored at −80 °C until further analysis. The V3–V4 region of 16S rRNA was amplified using the bacterial universal PCR primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The internal transcribed spacer (ITS)1 region of the fungal rRNA gene was amplified using the primers 1737F (5′-GGAAGTAAAAGTCGTAACAAGG-3′) and 2043R (GCTGCGTTCTGCATCGATGC). PCR amplification was performed in a 25-μL reaction mixture containing 0.5 μL of dNTP, 1 μL of primer, 2.5 μL of PCR buffer, 0.125 units of Taq DNA polymerase, and 2 μL of DNA template. PCR was performed as follows: 94 °C for 5 min, 30 cycles of 94 °C for 5 s, 50 °C for 30 s (for bacteria) or 56 °C for 30 s (for fungi), 72 °C for 45 s, 72 °C for 5 min, and then the temperature was maintained at 4 °C (Du et al., 2018b; Du et al., 2018c). The PCR products were sequenced using the Illumina MiSeq platform (Illumina, San Diego, CA, USA) by the Majorbio Company (Shanghai, China).

2.6. Sequence analysis

The gene sequences were processed and analyzed using the open-source software package Mothur. Sequences less than 400 bp, greater than 470 bp, containing homopolymer stretches of over 8 bp, or containing ambiguous bases were removed. PCR chimeras were filtered using the Chimera.uchime command in Mothur (http://www.mothur.org/wiki/Chimera.uchime). The 16S rRNA gene was processed and analyzed as described by (Kozich et al., 2013). Sequences of the ITS rRNA gene were aligned using MAFFT version 7 (https://mafft.cbrc.jp/alignment/software). After quality control, the relative abundances were calculated using a series of processing protocols documented by (Du et al., 2018a). The sequences were clustered into operational taxonomic units (OTUs) by setting a distance of 0.03. The bacterial sequences were assigned to phylotypes using the Bayesian approach and the Ribosomal Database Project 16S rRNA gene training database with a confidence threshold of 70%. The method and database employed in the taxonomic classification of the fungal sequences were the K-nearest neighbor algorithm and the UNITE ITS database, respectively (Abarenkov et al., 2010; Wang et al., 2007). Raw sequencing data were deposited in the National Center for Biotechnology Information Sequence Read Archive by accession number PRJNA749903. IBM SPSS Statistics 26 software was used for all data analyses, and drawings were created using Origin 2019b.

3. Results and discussion

3.1. Variation in microbic community diversity in two environmental media during sampling period

By sequencing the DNA of rainwater and soil samples collected during the summer from August 2018 to August 2020, approximately 515,040 sequences of the bacterial 16S rRNA gene and 840,349 sequences of the fungal ITS gene were obtained after excluding non-qualified gene sequences and chimeric sequences (Table 2 ). Normalization was performed by randomly selecting 26,886 bacterial sequences and 44,197 fungal sequences from each sample to compare the species richness and community diversity objectively. A 97% similarity cutoff was adopted to delineate the OTUs, and 2561 and 2905 OTUs were identified for the bacteria and fungi, respectively.

Table 2.

Comparison of diversity estimators of the bacterial and fungal communities in rainwater and topsoil from 2018 to 2020.

	Sequencesa	OTUsb	ACE	Chao 1	Shannon	Coverage
Bacteria
rainwater201808	70,049	193	269	277	2.649	0.998
rainwater201906	35,897	161	221	216	2.666	0.998
rainwater201907	41,928	165	231	254	2.236	0.998
rainwater201908	33,632	95	131	126	2.177	0.999
rainwater202007	38,480	411	560	534	3.384	0.995
rainwater202008	40,488	829	973	972	5.279	0.994
soil201808	44,626	1436	1699	1699	5.648	0.988
soil201906	26,886	1326	1586	1592	5.381	0.988
soil201907	35,423	1448	1720	1749	5.705	0.987
soil201908	62,879	1152	1441	1427	5.085	0.988
soil202007	49,768	1490	1758	1794	5.925	0.987
soil202008	34,984	1523	1804	1840	5.879	0.987
Fungi
rainwater201808	53,824	612	735	724	3.665	0.997
rainwater201906	44,197	340	403	400	1.809	0.998
rainwater201907	95,639	282	376	349	1.810	0.998
rainwater201908	52,272	679	786	780	3.520	0.997
rainwater202007	66,002	510	650	663	2.706	0.997
rainwater202008	117,957	520	725	693	3.116	0.996
soil201808	95,821	733	831	824	4.122	0.997
soil201906	50,031	1081	1248	1216	4.191	0.995
soil201907	54,749	544	704	689	2.725	0.996
soil201908	67,408	532	601	583	3.177	0.998
soil202007	73,689	687	749	756	4.190	0.998
soil202008	68,760	881	1050	1046	4.257	0.995

^aReads after quality controls and chimera removal. For each samples, 26,886 of 16S rRNA gene sequences and 44,197 of ITS gene sequences were randomly selected for calculating the species richness indexes (ACE and Chao 1) and diversity index (Shannon).

^bThe operational taxonomic units (OTU) were defined with 3% dissimilarity.

The Abundance-based Coverage Estimator and Chao1 indices in the alpha diversity index reflect the abundance of the microbial community. The higher the value, the higher the abundance. The Shannon index characterizes the diversity of microbial communities. The higher the value of the Shannon index, the greater the number of species and uniformity. The Shannon index diversity dilution curve gradually flattens, indicating that the sequence is sufficiently representative, and the diversity data obtained is credible (Figs. S1 and S2). The statistical test found that in the three years, the species richness and community diversity of bacterial communities and the community diversity of fungal communities exhibited significant differences between the rainwater and soil samples, while the species richness of the fungal communities did not exhibit such differences (Table 3 ).

Table 3.

Analysis of different bacterial and fungal communities in different environmental media.

	Alpha-diversity index	Analysis of variance	Mann-Whitney test
Bacteria	Chao 1	P = 0.000⁎	–
	Ace	P = 0.000⁎	–
	Shannon	–	P = 0.006⁎
Fungi	Chao 1	P = 0.067	–
	Ace	P = 0.066	–
	Shannon	P = 0.041⁎	–

Both non-biological and biological particles can accumulate by scrubbing, including in-cloud and below-cloud scavenging, as precipitation droplets fall. Therefore, some primary biological aerosol particles (PBAPs) in the atmosphere can ultimately be deposited in rainwater through wet deposition (Lu et al., 2016). Rainfall can promote the release of bioaerosol particles from the soil into the atmosphere. When raindrops hit the soil surface, soil bacteria-containing aerosols are formed. One drop of rain can release 0.01% of the soil bacteria into the air. Joung et al. (2017) comprehensively considered factors such as the aerosolization rate, average surface density of bacteria, global land area, and precipitation type. They estimated that the total number of bacteria spread in raindrops worldwide each year is (1.2–85.0) × 10²². The microbial community in the soil is characterized by extremely rich species diversity (Veresoglou and Rillig, 2014). The soil is rich in organic matter and is a natural medium suitable for the growth and reproduction of microorganisms. Studies have demonstrated that in forest soil, the number of prokaryotic soil cells (non-nucleated organisms, including bacteria and archaea) is approximately 4 × 10⁷ cells·g⁻¹, while in other soils (including desert and agricultural soils), the number of prokaryotic cells is approximately 2 × 10⁹ cells·g⁻¹ (Whitman et al., 1998). Many soil microorganisms can be aerosolized and released into the atmosphere by the wind. Cao et al. (2014) illustrated that the relative abundance of Geodermat obscurus in winter air in Beijing is high, and the bacteria often exist in dry soil environments. Mu et al. (2020) found that leaf surface is the main local source of airborne bacteria, and the correlation between leaf surface samples and air samples is greater than the correlation between surface soil and air samples. The atmospheric environment rarely contains the nutritional conditions required for the survival of microorganisms and is often only a temporary residence for microorganisms. Therefore, the microbial content and community composition of the air largely depend on the source of the bioaerosols. This may lead to differences in the alpha diversity index of bacteria in various environmental media.

Fungi can form a dormant spore structure, which can help in its survival for an extended period when the external environment is unsuitable for growth. The spore structures of the fungi can also spread in the air. Qi et al. (2020) analyzed the source distribution of airborne fungi in Xi'an from the summer of 2018 to the spring of 2019 and found that 63.5% of airborne fungi originated from unknown sources. Long-distance transportation may also be a critical source of biological aerosols. In addition, airborne microorganisms primarily originate from the surface of leaves throughout the year, and their contribution accounts for 26.8% of the total fungi. In contrast, the contribution of soil is small, accounting for only about 6% of the fungi. Fungi are typically easily dispersed, can colonize many substrata, and can tolerate diverse environmental conditions (Santiago et al., 2018). Bioaerosols come from many different sources on the earth's surface (X. Li et al., 2020). The complexity of the source of fungi and the small proportion of fungi from the soil may lead to no significant difference in microbial species richness indices between rainwater and soil.

In the Bray–Curtis similarity matrix, if different samples in the coordinate system are closer, the community similarity is higher. The matrix results for this study are displayed in Fig. 1 . In the two different environmental media, the samples exhibit clusters, indicating that the microbial community composition in the same environmental media was similar. There are some differences between the rainwater samples from 2018 to 2019 and those from 2020. Therefore, the bacterial and fungal community composition of Hulunbuir Grassland exhibits distribution characteristics according to environmental media.

Fig. 1.

Principal coordinates analysis (PCoA) of bacterial (a) and fungal (b) communities of different environmental media at the genus level in Hulunbuir, Inner Mongolia.

In summary, during the sampling period, the bacterial community in the Hulunbuir Grassland of Inner Mongolia exhibited significant differences in species abundance, community composition, and community diversity among different environmental media. Additionally, the differences between fungal communities in different environmental media were less than those between bacterial communities.

3.2. Difference of community composition in rainwater and topsoil over three years

According to taxonomic analysis, 29 phyla and 635 genera of bacteria and 14 phyla and 690 genera of fungi were identified. Of these, 416 and 576 bacterial genera and 496 and 534 fungal genera were identified in the rainwater and soil samples, respectively. The number of species in the rainwater was less than that in the soil. Based on the results displayed in Fig. S3, 34 genera of bacteria can simultaneously exist in the three-year rainwater samples, accounting for 8.17% of the total bacteria and 44 genera of fungi, accounting for 8.87% of the total fungi. This result demonstrates that few species can exist in the local atmosphere for an extended period and that the species composition of bacteria and fungi in rainwater has changed significantly, which may be because the atmospheric environment is unsuitable for the survival of microorganisms. There are 276 genera of bacteria that can persist in the soil samples for three years, accounting for 47.92% of the total bacteria, and 98 genera of fungi, accounting for 18.35% of the total fungi (Fig. S4). This result demonstrates that a large proportion of bacteria can exist in the local soil environment for a long time, and the proportion of fungi is small. There are 22 genera of bacteria that simultaneously existed in the three-year rainwater and soil samples, accounting for 3.46% of the total bacteria, and 31 genera of fungi, accounting for 4.49% of the total fungi. We define these microorganisms that can exist in different environmental media for a long time as “core species.” Thus, we inferred that microorganisms in the atmosphere washed by rainwater may partly originate from the local soil. The 22 core species of bacteria accounted for 5.29% and 3.82% of the bacteria in the rainwater and soil samples, respectively. The 31 core species of fungi accounted for 6.25% and 5.81% of the fungi in the rainwater and soil samples, respectively (Fig. 2 ). These species may originate from local sources and participate in local hydrological cycles. Among them, dominant species with a relative abundance greater than 1% were Massilia (14.11%), Sphingomonas (3.11%), Arthrobacter (2.82%), Janthinobacterium (2.69%), Pseudomonas (2.23%), Noviherbaspirillum (1.78%), and Methylobacterium (1.21%), which belong to the Proteobacteria and Actinobacteria phyla, respectively. Existing studies have shown that Proteobacteria and Actinobacteria are widely present in the soil, water, and atmosphere, and they can survive in extreme environments and endure harsh environments such as low temperatures, dryness, and strong ultraviolet rays (Wei et al., 2020; Maki et al., 2015). Therefore, it is speculated that species with sufficient resistance can exist in various environmental media in the Hulunbuir Grassland for a long period.

Fig. 2.

A Venn diagram of all the 12 samples (rainwater and soils) for bacteria (a) and fungi (b) at the genus level.

The dominant fungal species with a relative abundance greater than 1% were Gibberella (15.49%), Cladosporium (7.51%), Fusarium (4.69%), Alternaria (4.35%), Ophiobolopsis (2.08%), Preussia (1.92%), Didymella (1.81%), Microdochium (1.63%), Knufia (1.34%), Epicoccum (1.09%), which belong to the Ascomycota phylum. Previous studies have demonstrated that Ascomycetes are the most dominant fungal phylum. We detected Fusarium in the core species, which is a known high-INA genus (Pouleur et al., 1992; Frohlich-Nowoisky and Poschl, 2013).

As illustrated in Fig. 3 , by calculating and comparing the relative abundance of the top 20 species with abundance levels in different samples, Massilia (26.83%) was identified as the genus with the highest relative abundance of bacteria in the rainwater samples for three years; however, its relative abundance in the soil was only 1.38%. Sphingomonas (5.23%) was the genus with the highest relative abundance in the soil, while its relative abundance in rainwater was only 0.99%. Compared with the rainwater samples, the relative abundance of species was more evenly distributed in the soil samples. Cladosporium (12.54%) was the most abundant genus among the rainwater samples with a three-year timespan for fungi, while in the soil, its relative abundance was only 2.47%. Gibberella (30.41%) was the genus with the highest relative abundance in the three-year soil samples; however, its relative abundance was only 0.57% in the rainwater samples.

Fig. 3.

Phylogenetic classification of the bacterial communities (a) at the genus level and fungal communities (b) at the genus level.

We found that in both the rainwater and soil samples, the dominant bacteria changed significantly over time. Plants spread spores, pollen, and debris into the atmosphere. The discharged plant particles and microorganisms in the atmosphere return to the soil through sedimentation (dry sedimentation) and/or precipitation (wet sedimentation) and form part of the global material cycle. This process includes the biological precipitation cycle, in which microorganisms related to plants and soil are transported to the height of the clouds as aerosols and trigger precipitation through INA (Dong et al., 2019; Woo and Yamamoto, 2020). PBAPs play a vital role in atmospheric chemistry and physics. During rainfall, the entrapment of suspended particulates in the atmosphere leads to a particle removal process. This process leads rainwater to contain both ice nuclei particulate matter that freezes in high-altitude clouds and atmospheric particulate matter carried by atmospheric washing that clears under and near the clouds, regardless of whether the particle is an interstitial aerosol or not. Therefore, the biological INPs in cloud water or PBAPs in the atmosphere may eventually be deposited into the rain. As the water falls, both non-biological and biological particles may accumulate through scouring. Therefore, microorganisms in the soil and rainwater samples have a certain similarity. Precipitation is an effective way to remove airborne particles, including bacteria and fungi. The microorganisms in rainwater may originate from clouds and particles in the air. According to previous studies conducted worldwide, the total number of bacteria in rainwater is approximately 10³–10⁵ cells mL⁻¹ (Herlihy et al., 1987; Sattler et al., 2001; Amato et al., 2005). Microorganisms can also be transported at high altitudes via clouds, which increases the complexity of the sources of microorganisms in rainwater. DasSarma et al. (2020) stated that from the top of the troposphere (approximately 10 km) to a height of 50 km is the stratosphere. In the thin and dry air of the stratosphere, the temperature can drop to −60 °F (approximately −51 °C). The stratosphere contains a group of small but tenacious microorganisms, such as Pseudomonas syringae. In addition to their potential impact on the weather system, these high-altitude residents may also spread allergens and diseases. Thus, we must investigate the transmission process of microorganisms in the atmosphere in more detail and study their survival mechanisms.

In summary, from the microbial community perspective, few species in the rainwater can exist stably for a long time. In contrast, most of the species in the soil can exist in the local grassland environment for an extended period. Some core species can exist stably in both rainwater and soil for a long time. We speculate that in the grassland environment, some microorganisms, such as Massilia and Gibberella, can adapt to the local environment and participate in the hydrological cycle, which may impact the local climate. The diversity of microbial communities in different environmental media exhibited significant differences over the three years. Subsequent analysis of the sample composition demonstrated that only a small number of species can exist in different environmental media for an extended period, which explains the significant differences in the composition of microbial communities between the various environmental media. The varying abundance of the above microorganisms in different environmental media further demonstrates significant differences in the compositions of the microbial communities of different environmental media.

3.3. Difference of efficient biological INPs during the different sample years

By analyzing three years of sequences, we investigated genera known to have high-efficiency INA. We detected four known genera of high-INA microorganisms, including Pseudomonas, Lysinibacillus, Mortierella and Fusarium. In contrast, Erwinia and Xanthomonas were not identified in any of the samples. These two high-INA genera were also not identified in our previous research (S. Zhang et al., 2020). This result is presented in Table 4 . Lysinibacillus has rarely been detected in rainwater, and it only accounted for 0.01% of the rainwater-202008 sample. Its relative abundance in the soil sample was only 0.02%. The relative abundance of the genus Pseudomonas was 3.91%, 12.33%, and 9.13% regarding the bacterial genera in the rainwater-201906, rainwater-201908, and rainwater-202007 samples, respectively. The maximum relative abundance among the other rainwater and soil samples was only 0.62%. The abundance distribution of Mortierella in the rainwater was similar to that of Lysinibacillus, with only a 0.09% abundance in the rainwater-201907 sample and a 0.02% abundance in the rainwater-202008 sample. Mortierella accounted for 1.34%, 3.12%, and 3.00% of the fungal genera in the soil-201808, soil-202007, and soil-202008 samples, respectively. The maximum relative abundance among the other soil samples was 0.64%. By investigating the known INA fungi, we found that the genus Fusarium accounted for 9.29%, 7.18%, 17.73%, 3.87% and 10.11% of the fungal genera in the soil-201906, soil-201907, soil-201908, soil-202007, and soil-202008 samples, respectively. The maximum relative abundance among the other soil and rainwater samples was 1.80%. Overall, the abundance distribution of the above microorganisms exhibited differences between the rainwater and soil samples.

Table 4.

The relative abundance of certain known efficient INA bacterial and fungal genera in the rainwater and topsoil samples.

Sample ID	Relative abundance
	Bacterial genera		Fungal genera
	Lysinibacillus	Pseudomonas	Mortierella	Fusarium
rainwater201808	–	0.62%	0.00%	0.42%
rainwater201906	–	3.91%	–	0.01%
rainwater201907	–	0.34%	0.09%	1.14%
rainwater201908	–	12.33%	–	1.42%
rainwater202007	–	9.13%	0.00%	1.76%
rainwater202008	0.01%	0.12%	0.02%	0.15%
soil201808	0.02%	0.00%	1.34%	1.80%
soil201906	0.00%	0.01%	0.64%	9.29%
soil201907	0.15%	0.00%	0.06%	7.18%
soil201908	0.02%	0.15%	0.14%	17.73%
soil202007	0.15%	0.06%	3.12%	3.87%
soil202008	0.02%	0.06%	3.00%	10.11%

– The genus was not detected in this sample.

Biological INPs (e.g., pollen, bacteria, fungal spores, and plankton), most of which contain ice nucleation proteins, can trigger freezing at warmer temperatures (predominately ≥−15 °C) (Murray et al., 2012). Above this temperature, the only materials known to nucleate ice are biological. S. Zhang et al. (2020) denoted INPs with freezing temperatures ≥−10 °C efficient INPs. For the rainwater samples from 2018 to 2020, two samples of crude rainwater (August 2018 and June 2019) were active at approximately −3.8 °C, while the other samples induced freezing at −4.0 °C to −6.4 °C. T₅₀ values of crude rainwater were −9.8 °C (August 2018), −9.2 °C (June 2019), −10.4 °C (July 2019), −10.4 °C (August 2019), −6.8 °C (July 2020), and −6.6 °C (August 2020) (Table S1). All T₅₀ values were ≥−15 °C, and the T₁₀ values (−7.2 °C, −6.4 °C, −7.8 °C, −8.2 °C, −6 °C, and −6 °C, respectively) of the crude rainwater were ≥−10 °C. On average, more than 228 INP cm⁻³ were active at −10 °C (Fig. 4 ). We hypothesize that these findings are due to the widespread distribution of known INA bacterial and fungal species in the rainwater (Christner et al., 2008a; Christner et al., 2008b).

Fig. 4.

Cumulative INP spectra of ultrapure water, a Pseudomonas syringae suspension and rainwater samples collected from a single precipitation event in the Hulun Buir grassland in the summers of 2018, 2019 and 2020. The symbol * means that the sample received heat treatment.

Protein has been proposed as an effective biological INP, and INA can easily be reduced by heat treatment. The size of the INP directly affects the nucleation ability of ice (Zobrist et al., 2007; Pummer et al., 2015). Filtration should eliminate INA bacteria or the INA of intact cells. The freezing temperature of heated or filtered rainwater was shown in Tables S2 and S3, respectively. For the 2020 rainwater samples, the T₅₀ of the filtered rainwater with particles less than 220 nm after heating were −10.2 °C and −9.8 °C (Table S4), and the freezing temperatures of the droplets were primarily between −10 °C and −6 °C. The average cumulative ice nucleus concentration at −10 °C was 71 IN/mL. These INs are not sensitive to heat and are not known efficient IN. Moreover, the heat-resistant Lysinibacillus was rarely detected in the rainwater samples, and it only accounted for 0.01% of the bacteria in the rainwater-202008 sample. Therefore, we can infer that there is probably a biological or non-biological heat-resistant INP <220 nm in the 2020 rainwater samples. This result can help us understand the possible ice nucleus formation mechanism of sub-micron or nano-scale particles, expand our understanding of high-efficiency INPs in the natural environment, and contribute to the meteorological field by proposing candidate materials for efficient IN. These results are consistent with previous research results (Du et al., 2017), where rainwater samples were collected at the same site in August 2011, 2012, and 2013.

Furthermore, the disparity in the concentrations of efficient INPs between the two samples may have resulted from variations in meteorological conditions and rainfall intensity. Rainfall for the rainwater-202007 sample occurred as a drizzle (<10 mm, 24 h⁻¹) and lasted nearly 3 h. However, the rainfall for the rainwater-202008 sample occurred as heavy rainfall in the form of hail (25–49.9 mm, 24 h⁻¹) and lasted less than 1 h. Because the relationship between INPs and the distribution, occurrence, and intensity of precipitation was not explicitly due to the lack of quantitative data (Stopelli et al., 2016), more research must be conducted to reveal a correlation.

The rainwater samples from 2018 to 2019 exhibited similar ice nucleus concentration ranges after different treatments. Compared with crude rainwater, heat treatment reduced the INP concentration by 66–100% at −6 °C, 24–100% at −8 °C, and 3–98% at −10 °C. The filtration treatment reduced the IN concentration by 0–100% at −6 °C, 24–93% at −8 °C, and 40–96% at −10 °C. After the filtering and heating treatments (Table S4), the initial freezing temperature remained above −10 °C, T₅₀ was −13.7 ± 1.7 °C, and the freezing temperatures of the droplets were primarily between −15 °C and −10 °C, which further verified the existence of heat-resistant INPs.

The low relative abundance of INA genera contrasted with the INA of biological INPs in the rainwater. This result indicated that there may be unknown efficient biological INPs likely derived from other unknown INA microbes. Therefore, there are unknown biological or non-biological INPs that are heat-resistant and efficient in precipitation in this area.

We found that Massilia exhibited the highest relative abundance of the bacteria genera in the three-year rainwater samples by assessing the species composition and community structure. It is also the genus with the highest relative abundance of core species existing in different environmental media for long periods. In our previous research (S. Zhang et al., 2020; Du et al., 2017), Massilia was also the genus with the highest relative abundance among the rainwater bacteria. In another study (Lu et al., 2016), Massilia was the second-most abundant genus among rainwater bacteria. Massilia is a gram-negative bacterium, which is consistent with known high-efficiency IN bacteria. The genus used to be separated in ice and water environments (Shen et al., 2015; Gallego et al., 2006). Fig. 3 illustrates how Massilia can survive in different environmental media, that it is a core species, and that it is the genus with the highest relative abundance in the rainwater samples. Based on its abundant presence in rainwater, we suspect that Massilia may affect the formation of rainwater, and it may be a genus of bacteria with high-efficiency INA. However, we do not currently understand the corresponding mechanism, which warrants a scope for further investigation.

Cladosporium and Gibberella were the top two genera in relative abundance among the core species. Through indoor simulation experiments, Iannone et al. (2011) found that Cladosporium, the most abundant fungus in the atmosphere, does not exhibit effective INA for its spores, and its freezing temperature ranges from −25 °C to −35 °C. There are few studies on the INA of Gibberella. There is currently a lack of knowledge of the INA of potential bacterial species, and our experimental conditions cannot be used for isolation and culture determination. Therefore, this may be a direction for future research.

To explore the role of Pseudomonas in precipitation and topsoil over a period of three years, the 16S rRNA gene sequence of the typical Pseudomonas OTU and the verified INAs of Pseudomonas P. syringae, P. fluorescens, and P. meridian were used to construct phylogenetic trees for Pseudomonas fluorescens and Pseudomonas meridian (Fig. 5 ). The Pseudomonas OTU with the highest sequence richness was OTU26, followed by OTU23, OTU19, and OTU892. OTU26 was similar to the Pseudomonas fluorescens strain H40 16S ribosomal RNA gene (partial sequence, EU862079.2). Phylogenetic studies based on 16S rRNA genes illustrated that OTU23, OTU675, OTU19, OTU892, OTU860, and OTU1852 appeared in new branches. These results indicated that some representative OTU sequences may not exhibit INA. However, the distribution sequence of OTU26 was much greater than that of the other representative OTUs, and it appears to have the ability to catalyze ice formation. Therefore, additional sequences related to OTU26 were involved in precipitation, and the relative abundance of OTU26 in the soil was only 0.15% at the maximum. As the relative abundance of sequences belonging to OTU26 increased, the T₅₀ and T₁₀ values also significantly increased. The increased abundance of Pseudomonas may help increase INA during precipitation. Unfortunately, due to a lack of in-situ microbial culture experiments on rainwater and soil samples and the limitations of high-throughput sequencing technologies, we were unable to identify potentially effective strains. According to the phylogenetic tree analysis, we can infer that not every strain of bacterial genera with potentially effective INA exhibit INA. Among Pseudomonas, Erwinia, and Xanthomonas, only 20 species and variants of the bacteria were INA bacteria. Comparing the relative abundance of well-known INA bacterial sequences for particulate matter in the air can help clarify the source of effective INPs in rainwater. Based on the bacterial abundance distribution in the rainwater and soil samples, the effective biological INPs are likely depleted during precipitation, including Pseudomonas, and they may be transported to the troposphere from remote areas.

Fig. 5.

Neighbor-joining phylogenetic tree constructed on the basis of 16S rRNA gene sequences with similarities to those of the known INA bacterial strains of genus Pseudomonas.

3.4. Evolution of pathogenic microorganism community composition before and after COVID-19 epidemic

The principal coordinates analysis results (Fig. 1) demonstrate that the microbial communities of the rainwater samples from 2018 to 2019 are clustered, indicating that the microbial community structure in these two years is similar. However, the 2020 rainwater samples are farther from those of the previous two years and cannot be clustered, indicating that there are significant differences in the composition of the 2020 sample microbial communities. Therefore, the microbial communities in the rainwater have changed. The soil samples for the three years exhibited apparent clusters, indicating that the soil microbial community composition in the area was highly stable. This also demonstrates the accuracy of our experiments and reliability of the results. Both bacteria and fungi displayed the same trends.

We combined the rainwater samples from 2018 and 2019 into a “before” sample group, and the 2020 samples constituted the “after” sample group. A total of 181,506 bacterial sequences were obtained from the before samples (rainwater samples from 2018 and 2019). The most abundant bacterial genera were Massilia (39.08%), Comamonas (11.48%), Duganella (10.65%) and Janthinobacterium (7.85%). The 2020 sample group contained 78,968 bacterial sequences. The four most abundant bacterial genera were Prosthecobacter (17.47%), Pedobacter (8.78%), Pseudomonas (4.63%) and Novosphingobium (3.12%). Sphingomonas (5.23%), Arthrobacter (5.06%), Microvirga (3.26%) and Rubrobacter (2.58%) were the four most abundant genera in the soil group.

A total of 245,932 and 183,959 fungal sequences were identified in the before and after sample groups, respectively. Sporobolomyces (16.14%), Bullera (16.02%), Cystofilobasidium (15.33%) and Ophiobolopsis (6.08%) were the four most abundant genera in the before sample group. The most abundant fungal in the after sample group were Cladosporium (31.11%), Alternaria (1.32%) and Schizothecium (1.17%), which differed from the before sample group. Gibberella (30.41%), Fusarium (8.79%), Alternaria (6.14%) and Preussia (3.78%) were the four most abundant genera in the soil group.

The Venn diagram in Fig. S5 illustrates that the number and proportion of overlapping parts of the genera are significantly different between the sample groups. After the epidemic, the number of bacterial genera increased compared with that before, and the proportion of overlapping parts increased from 26.3% to 52.6%. This result indicates that the microbial community structure changes with the degree of human activity. Human activities may introduce many foreign microorganisms to the local environment and alter the structure of the local microbial community. For fungi, the number of genera declined, and the proportion of overlapping parts almost remained unchanged. The epidemic limited the flow of people and resulted in a cleaner air environment. Through the STAMP analysis (Welch's t-test) (Fig. 6 ), the differences of the dominant microorganisms at phylum and family levels in the rainwater samples before and after the epidemic were shown. It was found that at phylum level for bacterial, Proteobacteria, Bacteroidota, Cyanobacteria; and Oxalobacteraceae and Caulobacteraceae at the family level had significant differences (P < 0.05). The fungus Basidiomycota at the phylum level and Sporidiobolaceae at the family level were significantly different (P < 0.05). We speculate that certain microbial groups were affected by human activity. However, the relative abundance of some microorganisms did not change significantly before and after the epidemic, indicating that different microorganisms may respond differently to changes in the external environment.

Fig. 6.

Comparison of microbial communities before and after the COVID-19 pandemic. Welch's t-tests were performed between rainwater samples before and after the COVID-19 pandemic at phylum ((a) bacteria and (c) fungi) and family ((b) bacteria and (d) fungi) levels. Dominant phyla and families are shown.

We found that the rainwater microbial community in 2018 and 2019 was significantly different from that in 2020. Since December 2019, COVID-19 has spread rapidly worldwide. To control the spread of the virus, the Chinese government has implemented a nationwide blockade policy, large-scale industrial companies have suspended operations, and the national government has advised people to remain at home (Tian et al., 2020; Wang et al., 2020). These actions have led to a sharp reduction in human activities and a subsequent reduction in the emission of major air pollutants, improving the air quality. Studies have found that during the epidemic, the primary pollutants PM_2.5, PM₁₀, SO₂, NO_X, and CO all demonstrated a significant decline during the COVID-19 epidemic (Berman and Ebisu, 2020; Chauhan and Singh, 2020; Collivignarelli et al., 2020; L. Li et al., 2020; Wang et al., 2020). Humans continuously produce and release large amounts of bioaerosols. Additionally, many interior and exterior parts of the human body carry many types of bacteria (Costello et al., 2009). For example, the skin and digestive pathways of the human body carry approximately (1.0–100.0) × 10¹² microorganisms (Luckey, 1972). Human activities will affect the microbiota in the air, which in turn affect the local human health and ecosystem (X. Li et al., 2020). We can infer that the changes in the environment and the reduction in human activities may cause shifts in the microbial community.

Our sampling site was located in the Hulunbuir Grassland in Inner Mongolia, a remote area and a tourist attraction in China. Air quality during the sampling period was excellent. Rainwater fully washed the atmosphere, and we used the microorganisms in the rainwater to represent the atmosphere. At present, extensive research has focused on atmospheric environmental research under severe pollution conditions (Cao et al., 2014; Li et al., 2015; Xu et al., 2017). However, although long-term exposure to these conditions poses a risk to human health, little consideration has been given to the impact of the atmospheric composition on human health under suitable air conditions (Air Quality Index <100). For example, Shi et al. (2016) has demonstrated that short-term and long-term exposure to PM_2.5 is associated with all-cause mortality (total mortality), even if the exposure level is less than 10 μg/m³. Zhang et al. (2019) investigated the microbial activity under non-severely polluted conditions in the suburbs of Beijing and found that the air contained the pathogenic bacterium Streptococcus, which can cause wound infections and pinkeye, with a relative abundance of 0.23%.

Three pathogenic bacteria were identified in the rainwater samples in 2018, 2019, and 2020, including Rickettsia (0.005%), Acinetobacter lupus (0.015%), and pathogenic Escherichia coli (0.010%). Two pathogens were identified in the soil samples: Bacillus anthracis (0.023%) and Acinetobacter rouxii (0.002%). In 2018, 2019, and 2020, pathogenic bacteria in rainwater samples accounted for 0.019%, 0.002%, and 0.078% of all bacteria, respectively. The proportions of pathogenic bacteria in the soil samples were 0.022%, 0.024%, and 0.028%, respectively. Five pathogenic fungi were identified in rainwater samples in 2018, 2019, and 2020, including Alternaria (2.559%), Cephalosporium (0.006%), Mucor (0.008%), Botrytis (0.001%), and Trichoderma (0.002%). Six pathogenic fungi were identified in the soil samples: Alternaria (6.142%), Cephalosporium (0.003%), Mucor (0.021%), Botrytis (0.026%), and Trichoderma (0.090%) and Trichothecium (0.002%). In 2018, 2019, and 2020, pathogenic fungi in the rainwater samples accounted for 6.276%, 2.171%, and 1.329% of all fungi, respectively. The proportions of pathogenic fungi in the soil samples were 6.403%, 5.551%, and 7.327% in 2018, 2019, and 2020, respectively. In addition, other pathogens have been found in different studies. For example, (Du et al., 2018c) found four pathogens (Streptococcus, Prevotella, Erysipelas, and Rickettsia) and five pathogens (Trichothecium, Stachybotrys, Alternaria, Trichoderma, and Arthrinium) in the suburbs of Beijing. Cao et al. (2014) used metagenomic sequencing technology to study the composition of atmospheric microorganisms under severely polluted weather conditions in Beijing and found three human pathogens, including Streptococcus pneumoniae, Aspergillus fumigatus, and Human Adenovirus C. Hurtado et al. (2014) used culture methods to isolate Escherichia coli, Staphylococcus aureus, Pseudomonas aeruginosa and Enterococcus faecalis in Mexico.

Acinetobacter lwoffii, a key opportunistic pathogen, typically causes sepsis and gastroenteritis (Ku et al., 2000; Regalado et al., 2009). Bacillus anthracis is highly pathogenic and can cause anthrax, a zoonotic acute infectious disease. People can be infected by contact with herbivores and their pollutants, and severe infections can cause anthrax meningitis and even death (Beyer and Turnbull, 2009; Hicks et al., 2011). Rickettsia may cause epidemic typhus, spotted fever, and Brill-Zinsser disease (Kelly et al., 2002; Perlman et al., 2006). Pathogenic Escherichia coli is a gram-negative bacterium that causes disease outbreaks by polluting drinking water, food, and recreational waters. It can cause severe diarrhea and sepsis and can be life-threatening in severe cases (Kaper et al., 2004). Pathogenic fungi, such as Trichothecene, Botrytis, and Trichoderma can produce metabolites, such as trichothecenes, that can inhibit immune regulation and protein synthesis (Sudakin, 2003). Alternaria can cause skin tinea, onychomycosis, jaw osteomyelitis and other diseases (Woudenberg et al., 2015; Fraeyman et al., 2017). Certain mycotoxins produced by Alternaria alternata are critical carcinogens; the spores of this fungus are spread through air.

We compared the relative abundance of pathogenic bacteria before and after the SARS-CoV-2 outbreak. The outbreak led to the COVID-19 pandemic, during which the relative abundance of pathogenic bacteria and pathogenic fungi in the soil remained stable. However, in rainwater, we found that among the detected pathogenic microorganisms, Rickettsia, Acinetobacter reuteri, pathogenic Escherichia coli, Mucor, and Botrytis spp. showed an increase in relative abundance, while the relative abundances of Alternaria, Cephalosporium, and Trichoderma exhibited a downward trend. Overall, the COVID-19 epidemic has led to a substantial reduction in human activities and pollutant discharge, improving environmental conditions. However, various pathogens may respond differently to environmental changes.

The outbreak of COVID-19 reminds us that we must quickly find and establish a national environmental and biological monitoring and defense system suitable for national and regional conditions to improve biosafety. This can help us solve biological pollution in the environment and control and prevent the spread of diseases. The results of this study provide an important reference for achieving these goals and a theoretical basis for a comprehensive understanding of the potential risks of bioaerosols to human health. In addition, we must also monitor microorganisms under excellent and good weather conditions.

4. Conclusions

In this study, we identified shared and unique bacterial taxa present in the environment from two different environmental media. Human activities may significantly affect species richness and microbial community diversity. The known high-INA genus is low in abundance, and we suspect that the high-abundance species (Massilia and Gibberella) present in various local environmental media may be unknown high-efficiency biological ice nuclei. However, due to the limitation of experimental conditions, the strains were not isolated and cultured, and the INA of the potential microbial genus was not measured. Further research is still required to determine whether a specific genus exhibits a valid INA to verify this postulation. The relative abundance of different pathogenic bacteria demonstrated different trends before and after the COVID-19 epidemic. Our study provides insights into the composition and structure of the environmental microbiota in grasslands, enhances our understanding of the effect of environmental media on the microbiota species diversity, and highlights the importance of biological pollution and safety.

CRediT authorship contribution statement

Yongtao Zhang: Writing – original draft, Formal analysis, Investigation, Validation, Data curation, Writing – review & editing. Rui Du: Conceptualization, Methodology, Writing – review & editing, Supervision, Validation, Investigation, Resources. Hanlin Chen: Investigation, Resources. Pengrui Du: Investigation, Software. Sujian Zhang: Investigation, Resources. Weishan Ren: Resources.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Editor: Pingqing Fu

Appendix A. Supplementary data

Supplementary material