Microbial Community and Fermentation Quality of Alfalfa Silage Stored in Farm Bunker Silos in Inner Mongolia, China

Abstract

Alfalfa is conserved in silo-type bunkers in the cold and humid regions of Inner Mongolia, China. Its quality is essential to ensure a healthy and sustainable dairy production. However, the impact of environmental factors on the microbiota and fermentation products of alfalfa silage remains unclear. The present research examined changes in the microbiota and fermentation products and their association with environmental parameters in 72 samples collected from 12 farms located at 4 different latitudes and longitudes across four regions. The samples were labeled with distinct codes, A, B, and C, from the cold–rainy region, D, E, and F, from the warm–rainy region, G, H, and I from the cold–dry region, and J, K, and L from the warm–dry region. The lactic acid levels ranged from 14.25 to 24.27 g/kg of DM across all samples. The pH and concentrations of NH₃-N and butyric acid in samples A, B, and H were higher (p < 0.01) than in the other samples. Samples D and E had higher acetic acid concentrations and 1, 2-propanediol content (p < 0.01). The fresh material was dominated by Pantoea and Pseudomonas, whereas Lactobacillus was the most dominant genus in all silages, except for the B silage. The A, B, and H silages contained more Clostridium but less Lactobacillus than the other silages. The lactic acid levels were strongly associated with Lactobacillus plantarum, Weissella paramesenteroides, Lactobacillus acetotolerans, Pedobacter borvungensis, and Lactobacillus brevis (p < 0.01). In contrast, the pH and the NH₃-N and butyric acid concentrations were strongly associated (p < 0.01) with the presence of Clostridium estertheticum. A correlation analysis revealed that precipitation, temperature, longitude, and latitude were the most critical factors influencing epiphytic microbes in the fresh material. After silage fermentation, low-temperature conditions significantly affected the fermentation products and microbial community composition. In conclusion, the microbial community of silages is distinctive in cold and humid regions, and climatic parameters ultimately affect the microbiota and fermentation products. Furthermore, the findings of this study demonstrate that Illumina MiSeq sequencing combined with environmental factor assessment might provide new information about the microbiota composition and fermentation quality of silages, facilitating the achievement of high-quality silage.

1. Introduction

The Inner Mongolia Autonomous Region plays a crucial role as a milk producer and is the foremost grower of alfalfa (Medicago sativa L.) in China. In 2020, this region had 1.43 million dairy cattle, producing 6.11 million tons of fresh milk. Alfalfa silage is the primary legume utilized in dairy farming, and there is an increased focus on producing high-quality silage due to technological and economic developments. Ensiling is a traditional anaerobic fermentation-based technique for preserving forage grasses or crops that use lactic acid bacteria (LAB) to accomplish fermentation [1]. Compared to other forages, legume silage preservation is more difficult because the buffering capacity is high and the water-soluble carbohydrate (WSC) content is low during ensiling [2]. However, the degree of successful fermentation can be assessed by determining the pH, the composition of organic acids, and the level of ammonia nitrogen (NH₃-N). In addition, understanding the microbiota linked to aerobic stability and fermentation is crucial for managing bunker silos. Moreover, a review of previous studies revealed that the climatic conditions affect the four phases (initial aerobic, fermentation, stable storage, and feed-out phases) of crop ensiling, particularly in cold and pluvial regions [3], such as Inner Mongolia in North China.

These climatic conditions promote the growth of fungal and bacterial pathogens that cause yellow spots, anthracnose, and rust in alfalfa, as well as the growth of mycotoxin-producing pathogens, such as Aspergillus, Byssochlamys, Penicillium, and Fusarium [4]. Besides affecting feed crop yield and disease incidence, the short and cold growing seasons and high humidity affect the silage’s nutritional value and fermentation. Moreover, rain during harvest can cause Clostridial fermentation in the silo, resulting in poorly preserved silage. Clostridial fermentation may lead to the anaerobic deterioration of silage by butyric-producing Clostridia, such as Clostridium butyricum, Clostridium beijerinckii, and Clostridium tyrobutyricum, posing potential risks to milk quality and safety, as well as to animal health [5]. Furthermore, ensiling at low temperatures delays and stops the production of organic acids in cold regions because the activity of LAB in silage is low when the temperature is low, compared to those of undesirable microbes, such as mold and yeast, that cause silage spoilage in the presence of air [6]. Previous studies primarily focused on corn, oat, and total mixed ration silages, with experiments examining how dry matter concentration or temperature influences silage fermentation quality, microflora, and aerobic stability [7,8]. However, to date, no research has been carried out on how environmental factors influence the fermentation production of alfalfa silage stored in farm bunker silos in the Inner Mongolia Plateau, China, by influencing the bacterial microflora.

Silage production is a fermentation process that relies entirely on microorganisms. The microbial community of silage has a significant impact on silage quality and animal performance. However, the microbiota composition of alfalfa silage stored in farm bunker silos is not well understood, even though it could be crucial for the future production and utilization of bunker silos. Previous studies primarily focused on changes in fermentation production and microbial communities in laboratory silos during the ensiling process, which fail to reflect the changes in bunker silo production. In our previous study, which utilized denaturing gradient gel electrophoresis (DGGE) and culture-based isolation techniques, little information was obtained regarding the microflora in alfalfa silage fermentation and aerobic stability tests [9]. Next-generation sequencing (NGS) has recently been used to analyze the microflora of silage in this environment. However, these studies only focused on how heterofermentative and homofermentative LAB influence the microbial community during ensiling [10,11], and silage ecology in different environmental factors and regions have not been characterized.

Alfalfa is cultivated in four distinct regions of Inner Mongolia that experience varying climates, namely, cold–rainy, warm–rainy, cold–dry, and warm–dry regions. Therefore, we hypothesized that there would be differences in the microbial communities in fresh material and silage from different geographical locations. This study examined the impact of environmental factors (precipitation, temperature, humidity, and location) on the fermentation quality and microflora of alfalfa silage produced in farm bunker silos.

2. Materials and Methods

2.1. Site Description and Sample Collection

Thirty-six fresh material and thirty-six silage samples were collected from the bunker silo of 12 farms in Inner Mongolia, China. The samples were obtained from four different regions in the Inner Mongolia Autonomous region, which were the cold–rainy region (48°18′–49°14′ N, 119°21′–122°20′ E), the warm–rainy region (44°19′–45°27′ N, 118°24′–119°28′ E), the cold–dry region (42°30′–43°16′ N, 114°43′–116°31′ E), and the warm–dry region (40°15′–41°23′ N, 109°32′–110°37′ E). Figure 1 illustrates the three representative dairy farms selected from each region. Specifically, bunker silos A, B, and C were located in the cold–rainy region, while bunker silos D, E, and F were situated in the warm–rainy region. Similarly, bunker silos G, H, and I were constructed in the cold–dry region, and bunker silos J, K, and L were located from the warm–dry region. Silage was collected from 20 October 2021 to 10 November 2021, from 12 dairy farms. The fermented samples were stored for 60 d before sampling. At each dairy farm, silage (500 g) was sampled at a depth of 20 cm from the bunker face using clean disposable gloves. The samples were transported in ice boxes and frozen-preserved at −80 °C. All the samples examined in this study were prepared without inoculants. The samples were evaluated for their microbial community, fermentation production, and chemical composition.

2.2. Chemical Composition, Fermentation Products, and Microbial Composition Analysis

The DM contents of the fresh materials and silages were dried in a 48 h oven-drying process at 65 °C, followed by pulverization of each sample using a 1 mm-sieve Wiley mill (ZM200, Retsch GmbH, Beijing, China). Crude protein (CP) was tested using the standard Association of Official Analytical Chemists [12] procedures. A previously reported method [13] was used to evaluate acid detergent fiber (ADF) and neutral detergent fiber (NDF), while a previously published protocol was used to identify water-soluble carbohydrates [9]. The NH₃-N levels were estimated using the method described by Broderick and Kang [14].

Aqueous extracts of silage samples were prepared for pH, lactic acid, alcohol, and short-chain fatty acid determination by homogenizing 20 g of fresh sample with 180 mL of sterilized water inside a blender for 1 min and then filtering through a 0.22 µm membrane. The pH of the extracts was measured using an SX-620 glass electrode pH meter (Sanxin, Shanghai, China). The fermented products were measured using ion-exclusion polymeric HPLC with an RI detector [9].

A clean bench was used for serial dilutions (10⁻¹ to 10⁻⁸). Enterobacteria were quantified using violet red bile agar (CM0107B; Oxoid, UK), whereas LAB were quantified using Rogosa, Man, and Sharpe agar (CM0361B, Oxoid, UK). Potato dextrose agar (CM0139B, Oxoid, UK) plates spread at pH 3.5 (kept using sterile lactic acid) were used to quantify molds and yeasts.

2.3. Bacterial Community Analysis

The refrigerated samples (10 g) were blended in sterile PBS (40 mL; pH 7.4) for 2 h at 120 rpm using an electronic oscillator and then filtered using double-gauze masks. The supernatant was discarded after a 10 min centrifugation of the filtrate at 13,000× g and 4 °C, and the pellet was stored on dry ice. Bacterial sequencing, which consisted of DNA extraction and PCR amplification, was performed using Majorbio Bio-Pharm Technology (Shanghai, China). The data were processed using Illumina MiSeq sequencing and final sequencing. Operational taxonomic unit (OTU) clustering was performed using Uperse software ver. 7.1 at a 97% similarity cut-off [15]. Following UCHIME-aided identification and elimination of chimeric sequences, a taxonomic assessment of each 16S rRNA gene sequence was performed using the RDP Classifier ver. 2.2 against Silva ver. 138 at a 0.7 confidence level [16]. The bacterial sequencing data were examined using the Majorbio Cloud, an open and accessible platform.

2.4. Statistical Analyses

Statistical analyses were performed using the JMP ver. 13 software (SAS Institute, Tokyo, Japan). The dairy farm effect was assessed using unidirectional ANOVA, and multiple comparisons using Tukey’s test. The significance of the different sample means was assessed at p < 0.05. In addition, the correlation between the bacterial community and environmental parameters was examined using Majorbio Cloud, an open and accessible platform.

3. Result and Discussion

3.1. Chemical and Microbial Composition of Pre-Ensiling Alfalfa

Table 1. Chemical composition and microbial composition of pre-ensiling alfalfa.

Region	Cold–Rainy			Warm–Rainy			Cold–Dry			Warm–Dry
Bunker Silo	A	B	C	D	E	F	G	H	I	J	K	L	SEM	p-Value
DM (g/kg)	236 fg	244 fg	274 cd	286 bc	314 a	293 b	262 de	229 g	264 de	295 b	273 cd	321 a	6.36	<0.01
pH	6.01 abc	5.93 bcd	6.13 ab	5.89 cd	5.93 cd	6.14 ab	5.83 d	5.93 cd	6.03 abc	6.15 a	5.83 d	6.03 abc	0.06	<0.01
WSC (g/kg DM)	43.2 d	46.2 bc	41.9 d	45.1 c	39.7 e	42.1 d	47.9 ab	48.3 a	46.2 bc	38.9 e	43.2 d	38.7 e	0.60	<0.01
CP (g/kg DM)	212 de	197 f	218 bcd	228 ab	231 a	216 cd	193 fg	217 bcd	184 g	214 cde	204 ef	225 abc	3.81	<0.01
NDF (g/kg DM)	365 abc	372 ab	368 bc	359 bc	383 a	374 ab	356 bc	349 c	362 bc	367 abc	349 c	372 ab	6.43	<0.05
ADF (g/kg DM)	263 a	257 a	248 ab	242 bcd	256 ab	263 a	238 cd	241 bcd	258 a	253 abc	237 d	258 a	5.22	<0.01
LAB (log cfu/g)	5.24 cd	6.24 a	4.18 f	5.38 c	5.25 cd	5.69 b	4.83 e	5.17 d	5.39 c	5.82 b	5.32 cd	6.15 a	0.06	<0.01
Yeasts (log cfu/g)	4.29 e	4.89 d	4.21 e	4.93 d	5.83 a	5.28 c	3.84 f	3.58 g	4.25 e	5.39 bc	4.85 d	5.51 b	0.07	<0.01
Enterobacter (log cfu/g)	4.29 e	4.83 a	4.58 bc	4.87 a	4.49 cd	4.73 ab	3.83 f	3.73 f	4.28 e	4.37 de	4.52 cd	4.63 bc	0.06	<0.01

DM, dry matter; WSC, water-soluble carbohydrates; CP, crude protein; NDF, neutral detergent fiber; ADF, acid-detergent fiber; LAB, lactic acid bacteria. Different letters in the same row indicate significant differences (p < 0.05). SEM, standard error of the mean.

shows pre-ensiling alfalfa’s microbial and chemical composition in various regions. To ensure consistency, all alfalfa samples were collected during the flowering stage, as the silage microbial and chemical compositions are influenced by the growth phase and harvest time [17]. In this study, the DM content of all fresh materials ranged from 229 to 321 g/kg, with the lowest and highest contents noted in the fresh materials of H and L alfalfa, respectively (p < 0.01). The DM of fresh material is a crucial determinant of the fermentation type of silage [18]. The DM contents of A, B, and G alfalfa were 236, 244, and 241 g/kg, respectively. One possible reason is that these alfalfa samples were harvested on the same growth date [19]. Furthermore, the DM content of the Algonquin and Vernal alfalfa cultivars was significantly lower (p < 0.01) in Northeast China than in Southwest China [20]. Hence, this result indicates that the DM content is probably influenced by high humidity and low temperature.

The WSC content is an essential factor in the fermentation process, aiding LAB growth to yield lactic acid, which determines the rapid decrease in pH during the ensiling process and is essential for high-quality silage. The WSC content exceeded 38 g/kg DM in all samples, in this study, with a maximum (p < 0.01) of 48.3 g/kg of DM observed in the H sample, which contradicts previous findings [21]. The chemical composition differed between regions and was affected by the alfalfa cultivar. The CP content ranged from 184 and 231 g/kg DM, similar to the results of a previous study [5].

The microbial numbers and compositions differ between the pre-ensiling and the ensilage processes [22]. The total number of microorganisms in pre-ensiling crop ranged from 10² to 10⁸ cfu/g, with various bacteria (Lactobacillus, Bacillus, Clostridium, Acetobacter) and other microorganisms (Aspergillus, Filobasidium, Neurospora) at the genus level [23]. Most microbes cause nutrient degradation in silage, with a detrimental-to-beneficial microbial ratio approaching 10:1 [1]. The LAB count on fresh alfalfa is less than 5 log10 cfu/g of FM [5]. In contrast, the highest LAB count is typically observed during the flowering phase of specific seasons, such as the first, second, and third cuts of grass and alfalfa [24]. The LAB quantities exceeded 2 log10 cfu/g FM for all samples. Samples B and L exhibited the highest (p < 0.01) LAB counts, exceeding 6 log10 cfu/g of FM. All samples contained both yeasts and Enterobacteria. Enterobacteria and yeasts exhibited a high number, ranging between 4 and 5 log10 cfu/g of FM, except for the G and H samples. The LAB, Enterobacteria, and yeast counts were lower than those for fresh alfalfa collected in southern Japan in April [9] and southern China in May [25]. These findings indicate that high humidity and low temperature can suppress microbial growth in fresh alfalfa.

3.2. Fermentation Products of Alfalfa Silages Produced in Farm Bunker Silos

Table 2. pH, NH₃-N and fermentation products of alfalfa silages produced in farm bunker silos in Inner Mongolia, China.

Region	Cold-Rainy			Warm-Rainy			Cold-Dry			Warm-Dry
Bunker Silo	A	B	C	D	E	F	G	H	I	J	K	L	SEM	p-Value
pH	5.74 a	5.52 b	4.84 cd	4.75 cd	4.38 e	4.82 cd	4.72 d	5.63 ab	4.86 c	4.71 d	4.87 c	4.47 e	0.05	<0.01
NH3-N (g/kg DM)	6.38 c	7.13 a	3.47 e	3.98 d	2.86 g	3.13 f	2.64 g	6.78 b	3.17 ef	2.74 g	2.83 g	2.29 h	0.08	<0.01
Lactic acid (g/kg DM)	16.75 h	17.84 g	23.74 b	13.87 k	14.74 i	21.38 e	22.64 d	14.25 j	19.75 f	24.21 a	23.17 c	24.27 a	0.07	<0.01
Acetic acid (g/kg DM)	12.31 h	13.43 g	18.96 c	24.86 a	23.67 b	17.34 e	16.45 f	23.75 b	18.54 d	16.45 f	17.53 e	16.28 f	0.09	<0.01
Butyric acid (g/kg DM)	8.03 c	9.14 b	2.36 d	1.86 g	1.74 g	1.86 g	2.31 d	10.36 a	2.27 d	2.14 e	2.07 e	2.03 ef	0.06	<0.01
2,3-Butanediod (g/kg DM)	2.07 a	1.78 d	1.54 f	1.86 c	1.24 g	1.75 d	1.95 b	1.76 d	2.08 a	1.84 c	1.92 b	1.69 e	0.01	<0.01
1,2-Propanediod (g/kg DM)	1.09 i	2.74 c	6.85 b	7.23 a	6.85 b	1.93 e	1.32 gh	2.54 d	1.49 g	1.25 hi	1.74 f	1.46 g	0.06	<0.01

NH₃-N, ammonia nitrogen, different letters in the same row indicate significant differences (p < 0.05). SEM, standard error of the mean.

shows the pH changes and concentrations of NH₃-N, butyric acid, acetic acid, lactic acid, 1,2-propanediol, and 2,3-butanediol in the silage during fermentation. Because appropriate water activity is necessary for LAB in silage fermentation to grow and proliferate, the DM level affects the fermentation products during ensilage [26]. In the current study, differences in DM content significantly affected silage pH after 60 d of ensiling. In addition, samples A, B, and H exhibited significantly higher (p < 0.01) pH values and butyric acid and NH₃-N concentrations than the other samples. Since the microbiota in these silages included Clostridium at the genus level after 60 d of ensiling, the poor fermentation quality in A, B, and H silages was likely caused by an unsuccessful fermentation process. The lactic acid content in samples collected from an alfalfa bunker silo in Northeast China ranged from 14.25 to 24.27 g/kg of DM [27]. Nonetheless, in contrast to a previous report on alfalfa silage collected from warm–rainy regions in Inner Mongolia, the samples in this study had higher acetic acid levels. Although the examined alfalfa bunker silo was in the warm–rainy region of Inner Mongolia, the acetic acid content varied depending on the season and on environmental factors. Samples D and E had the highest (p < 0.01) concentrations of 1, 2-propanediol and acetic acid, whereas no alfalfa bunker silo contained propionic acid. This could be related to the higher number of heterofermentative LAB and acetic acid-producing bacteria found in samples D and E after 60 d of ensiling.

3.3. Bacterial Diversity of Fresh Material and Silage

High-throughput amplicon sequencing was used to sequence the 16S rRNA gene in 72 samples (36 fresh material and 36 silage), yielding 10,971,127 superior-quality sequence reads in total, with an average of 152,377 reads per sample. These sequence reads were clustered into 3172 OTUs with 97% sequence similarity (Figure 2A,D). Good’s coverage estimates for all samples were approximately 0.99, implying that most bacterial communities could be observed.

OTUs and Chao and Shannon indices were used to assess the bacterial abundance and diversity in silage and fresh materials (Figure 2). During anaerobic fermentation, the fresh material microflora is replaced by LAB, which represents one of the standards for high-quality silage [18]. Thus, successful silage fermentation results in a drastic decline in bacterial diversity. Except for silages taken from farms B, C, and H, the Shannon, OTU, and Chao indices decreased in all samples compared to fresh material. The total number of OTUs in silages (Figure 2D) was higher than that in fresh materials (Figure 2A), which contradicts previous findings [28], because different environmental factors (precipitation, temperature, humidity, and location) affect epiphytic bacteria in fresh material. There were 823 shared OTUs for the main bacteria from the fresh material. The number of unique OTUs in the diverse farms varied from 68 for C fresh material to 452 for G fresh material (Figure 2A). The diversity of the main bacteria consisting of 2349 OTUs in all silages was found after ensilage for 60 d. Except for samples A, B, and H, the Clostridium OTUs decreased in all samples ensiled for 60 d. According to preliminary findings, similar bacterial communities in high-quality alfalfa silage shared the main bacteria and had fewer OTUs with Clostridium. Similar results were obtained for alfalfa silage in laboratory silos [5].

The principal coordinate analysis of the bacterial communities of silage and fresh material is shown in Figure 3. Samples A, B, and F were separated from the other samples with respect to fresh material, as shown in Figure 3a. After 60 d of ensilage, samples A, B, and H had a high butyric acid content and were significantly separated from the other samples (Figure 3b). Li et al. [5] found that after 56 d of ensiling, alfalfa silage with Clostridium fermentation was separated from other silages in the principal coordinate analysis.

3.4. Bacterial Community Dynamics in Fresh Material and Silage

Natural silage fermentation is widely recognized to rely on epiphytic microbes, particularly LAB, under anaerobic conditions [29]. In a previous study, we discovered diverse bacterial microbiota and succession in before- and after-ensiling alfalfa [9]. The composition of the bacterial community plays an essential role in silage fermentation, and understanding the complex ensilage process is required to determine the composition of bacterial communities.

The phylum-level bacterial microflora in silage and fresh material is depicted in Figure 4, with Proteobacteria, Bacteroidetes, Cyanobacteria, Firmicutes, and Actinobacteria being the dominant phyla in the fresh material (Figure 4a). However, after 60 d of ensiling, Proteobacteria and Firmicutes were the dominant phyla in all samples, accounting for more than 90% of the total examined sequences (Figure 4b). As shown in Figure 5a, the dominant genera in the fresh material were Pantoae and Pseudomonas, which were found in pre-ensiled crops, such as Leymus chinensis [23]. Although some microorganisms died during the ensiling process, Lactobacillus was the most dominant genus in all samples, except for sample B, after 60 d of ensilage (Figure 5b). Similarly, Li et al. [21] reported that laboratory-scale alfalfa silage contained 50% Lactobacillus after 60 days of ensiling. Samples A, B, and H showed higher Clostridium richness but lower Lactobacillus richness than the other samples after 60 days of ensiling, suggesting that a low DM content in silage is a possible condition for the activity of Clostridium, which can cause spoilage in high-moisture silage. The samples D and E had higher levels of Acetobacter than the other samples. However, the presence of acetic acid bacteria is undesirable in silages because Acetobacter consumes lactic and acetic acids and leads to aerobic degradation, which has been observed in whole-crop corn silage using denaturing-gradient gel electrophoresis [10]. However, upon silage aerobic exposure, Acetobacter can oxidize ethanol to acetic acid [28]. Therefore, acetic acid has been recommended to suppress the growth of detrimental microbes, such as yeasts and molds, and the aerobic spoilage of silage [30].

The species-level bacterial communities of the alfalfa silage are shown in Figure 6. Lactobacillus acetotolerans, Lactobacillus plantarum, Weissella paramesenteroides, Lactobacillus curvatus, Pediococcus pentosaceus, and Acetobacter pasteurianus were the dominant bacterial flora in the alfalfa silage after 60 d of fermentation. In addition, a previous study identified L. acetotolerans in whole crop corn and corn stover silage collected from bunker silos in China [31]. Three strains of L. plantarum were isolated from native grass silage. Inoculation with L. plantarum increased the level of lactic acid and decreased the pH and concentration of NH₃-N in Leymus chinensis silage [32]. W. paramesenteroides FG 5, a LAB strain isolated from grass silage, failed to improve the fermentation quality of Italian ryegrass and alfalfa silages during silage fermentation, according to Cai et al. [33]. The relative A. pasteurianus richness of silages D and E was higher than that of other silages in alfalfa silage with ensilage for 60 d. High A. pasteurianus richness is usually reported in laboratory silos and is regularly detected in bunker-made corn silages from Japan [34,35]. It is well established that both A. pasteurianus and L. acetotolerans can tolerate 30 g/L of acetic acid [36]. Li and Nishino [34] reported that though bunker-made corn silage had a high acetic acid content, none of the farmers acknowledged a problem in the silage, which is in agreement this study. However, Gerlach et al. [37] discovered that acetic acid could lower silage consumption in cattle. Therefore, it is worth examining how acetic acid bacteria grow and survive in different silage environments. Although high acetic acid and 1, 2-propanediol contents were observed in samples D and E, L. buchneri was not found in the bacterial community. This result contradicts the findings of Li and Nishino [34]. After 60 d of ensilage, samples A, B, and H had a higher relative richness of C. estertheticium and C. sensu than the other alfalfa silage samples. This could explain the higher concentrations of butyric acid and NH₃-N in silages A, B, and H than in other silages. Clostridium species of silage develop from aerobic bacterial spores located on fresh material because of contamination from air, soil, and farmyard manure [1]. However, this process in silage fermentation is obviously quite variable, reflecting differences in environmental parameters, including climate, management practice, geographical location, and type of fertilization [38]. Therefore, the microbiota composition is likely influenced and altered by environmental differences and silage conditions during the fermentation process.

3.5. Association of the Bacterial Microflora with the Fermentation Products

Ensilage, as a natural preservation process, involves interactions between the major bacteria and their metabolites. The LAB primarily affect the lactic acid content of silage during fermentation. Clostridium has been linked to the production of butyric acid and NH₃-N, whereas Enterobacteria can ferment water-soluble carbohydrates into ethanol, lactic acid, acetic acid, and 2,3-butanediol [1]. After 60 d of ensiling, the Spearman’s correlations of the main bacteria with pH and the concentrations of lactic, acetic, and butyric acids, 2,3-butanediol, 1,2-propanediol, and NH₃–N were also estimated (Figure 7). Based on Spearman’s correlation analysis, the lactic acid concentration was highly positively correlated (p < 0.01) with L. plantarum (R = 0.443), W. paramesenteroides (R = 0.625), L. acetotolerans (R = 0.528), P. borvungensis (R = 0.433), and Lactobacillus brevis (R = 0.443), whereas a significantly positive association of acetic acid (p < 0.01) with A. pasteurianus (R = 0.393) was found. The pH and NH₃-N and butyric acid concentrations were positively correlated (p < 0.01) with C. estertheticium, with correlation coefficients of 0.554, 0.403, and 0.458, respectively. In contrast, negative associations of pH and NH₃-N and butyric acid concentrations (p < 0.01) with L. acetotolerans were found, showing distinct correlation values of −0.642, −0.796, and −0.649, respectively. This finding is consistent with previous reports [18], indicating that acetic, butyric, and lactic acid bacteria in alfalfa bunker silos have a unique relationship with their metabolites.

3.6. Correlations between Bacterial Communities and Environmental Factors

The bacterial community of silage is influenced by various environmental factors, such as altitude, longitude, latitude, humidity, temperature, and precipitation. In order to examine the correlations between bacterial communities and environmental factors in fresh material, a Spearman’s correlation heatmap was generated (Figure 8a). In Inner Mongolia, China, the majority of alfalfa silage is produced in the summer months from July to September, which are characterized by high temperatures, high humidity, and heavy rainfall. These conditions are unfavorable for farmers, as heavy precipitation during harvesting and high temperatures during fermentation negatively impact the aerobic stability and quality of the silage [39]. In this study, the average precipitation was positively associated (p < 0.01) with the relative richness in Curtobacterium (R = 0.330), Sphingobacterium (R = 0.414), Serratia (R = 0.364), and Stenotrophomonas (R = 0.371) of fresh material. This result indicates a positive influence of mean precipitation on epiphytic bacteria in fresh material. Excessive rainfall prolongs the production time of alfalfa silage by promoting the reproduction of epiphytic bacteria in fresh material. This may explain the poor quality of the silage samples A, B, and H. We reinvestigated the farms and discovered that these silages took much longer to fill and pack bunker silos than the other silages. We found positive correlations between latitude and longitude and the relative richness in Curtobacterium (R = 0.459 and 0.446) and Sphingobacterium (R = 0.218 and 0.254). In addition, the temperature correlated with Pantoea relative abundance at p < 0.01 (R = 0.621) and Pseudomonas relative abundance at p < 0.01 (R = 0.459). This indicates that average precipitation, temperature, longitude, and latitude significantly contributed to the bacterial communities in fresh material.

Slow silo filling during bunker-made silage can increase the butyric and acetic acid concentrations because of detrimental climatic factors, increasing the richness of unwanted microbes such as yeast and mold. These microbes degrade silage quality in the absence of additives. Consequently, previous studies have reported that lactic acid bacteria inoculations outperform enzymes and chemical additives in preventing Clostridial fermentation in high-moisture silage because they are convenient, safe, and do not pollute the environment [40]. Therefore, we suggest that short-term filling and inoculation with LAB could help to improve the nutritive value and fermentation quality in the presence of climate factors with a negative impact.

According to Spearman’s correlation analyses after ensilage for 60 d, environmental humidity was positively correlated (p < 0.05) with Pseudomonas (R = 0.414), Paenibacillus (R = 0.443), and Pedobacter (R = 0.432) (Figure 8b). In contrast, environmental humidity was negatively correlated (p < 0.05) with Weissella (R = −0.433) and Lactobacillus (R = −0.371). Latitude and longitude had positive relationships (p < 0.05) with Paenibacillus (R = 0.362 and 0.351), Bacillus (R = 0.387 and 0.342), and Paeniclostridium (R = 0.383 and 0.374). The average precipitation was positively associated (p < 0.01) with Pseudomonas (R = 0.478) and Pedobacter (R = 0.487). The environmental temperature was positively associated (p < 0.05) with Pseudomonas (R = 0.624), Acetobacter (R = 0.489), and Lactobacillus (R = 0.404). In contrast, it was highly negatively associated (p < 0.05) with Bacillus (R = −0.378), Paeniclostridium (R = −0.445), Turicibacter (R = −0.341), Romboutsia (R = −0.342), and Clostridium (R = −0.387). The environmental temperature is a crucial factor influencing silage. The optimal temperature for silage fermentation is generally 20 °C–30 °C. Gulfam et al. [41] found that adverse effects on silage preservation were exerted by ensiling at high temperatures (>35 °C), leading to poor silage quality and easy aerobic spoilage. Alfalfa silage was collected from 12 farms between late October and early November. The low temperature (mean temperature, −2–9 °C) may significantly influence the bacterial community composition during fermentation. After fermentation for 60 d, the A, B, and H silages, whose temperatures were lower than those of the other silages, exhibited higher NH₃-N content and pH and more Clostridial fermentation. Clostridium sensu and Clostridium estertheticium survived a more extended period in silages A, B, and H, probably because of delayed acidification and a lower pH drop rate at low temperatures. As demonstrated by the present study, alterations in the diversity of the bacterial microbiota occurred during fermentation at low temperatures.

4. Conclusions

This study is the first to define Spearman’s correlation between bacterial microflora and environmental parameters in 72 samples collected from bunker silos at 12 farms in 4 different regions of Inner Mongolia. All alfalfa silages fermented well, except for the A, B, and H silages during 60 d of ensiling. After fermentation, the dominant species were L. acetotolerans and L. plantarum, which directly led to increased lactic acid levels in the silages. C. sensu and C. estertheticium promoted Clostridial fermentation in silages A, B, and H with low DM. The composition of epiphytic bacteria in the fresh material was significantly affected by precipitation, temperature, longitude, and latitude. Low temperatures were found to have a significant impact on the bacterial communities and fermentation products after silage fermentation. Moreover, the present study’s findings demonstrate that Clostridium species could be inhibited by a high DM content (>26%) at low temperatures, thereby reducing Clostridial fermentation in alfalfa silage produced in farm bunker silos.