Analysis by next-generation sequencing of bacterial communities in Japanese cedar logs buried for 2 years for ground improvement

Optimizing the effectiveness of the log-piling method, developed in Japan to mitigate soil liquefaction during seismic activity, relies on understanding log deterioration and bacterial interactions. This study explored the role of bacterial communities in degradation of logs buried in the ground for 2 years, focusing on the relationships between bacterial composition, soil environmental factors, and wood deterioration. Next-generation sequencing was employed to analyze the bacterial communities within the logs and the surrounding soil. Soil environmental factors such as pH, total carbon, total nitrogen, and the C:N ratio were measured to evaluate their influence on bacterial community structure. Bacterial communities in the logs predominantly originated from the surrounding soil, with shifts in taxon abundance driven by environmental conditions. Depth below the soil surface had a more substantial influence on bacterial community structure than soil environmental factors. Members of the classes Alphaproteobacteria and Gammaproteobacteria, particularly the orders Rhizobiales, Burkholderiales, and Enterobacterales, were primarily responsible for the observed minor log cell wall degradation, although the wood density did not change significantly during the experimental period. Our findings suggest that bacterial activity plays a role in early wood deterioration, and further research could help to improve the use of logs in ground improvement to enhance their durability in preventing soil liquefaction.

In recent years, advancements in soil improvement techniques have emerged, based on principles from physics, chemistry, and biology [1]. While these methods show promise, they require extensive testing to assess their effectiveness in mitigating liquefaction. Liquefaction occurs when an increase in pore water pressure reduces the effective stress of the soil, leading to a loss of shear strength and causing sandy soils to behave like a liquid. The log-piling method, which involves driving logs into the ground to increase soil density and stability, has demonstrated significant potential in increasing soil resistance to liquefaction [2]. For instance, during the 1964 Niigata earthquake, Niigata Station, which had log piles as its foundation, remained undamaged. Log piles are an effective countermeasure against liquefaction, with a service life of 26–86 years depending on the wood species used [3]. Studies by civil engineers have confirmed that this method improves ground conditions not only by increasing density but also by enhancing bearing capacity, ground stiffness, and confining pressure [2, 4].

Historically, Europeans widely used wooden pile foundations in soft soils to support buildings, many of which still stand today in cities, such as Venice, Amsterdam, London, Berlin, and Stockholm [5]. This technique involved driving long wooden piles deep into the ground, sometimes reaching the base of soft sediments. These piles were placed in fully saturated or anoxic environments, where the low oxygen concentration has allowed them to endure for hundreds of years [6]. In Japan, log piling is used to improve soil and mitigate liquefaction in coastal reclaimed land areas [7] and soft soils, such as clay [8]. This technique, known as the log-piling method for liquefaction mitigation and carbon stock (LP-LiC), contributes to carbon storage and aids efforts to mitigate climate change in civil construction [8, 9].

The degree of soil stabilization or hardness is crucial for the effectiveness of liquefaction mitigation following log installation in the log-piling method [8]. Soil hardness also influences the biodeterioration of the wood and the long-term sustainability of ground compaction. Civil engineers typically assess soil hardness in the field using a portable dynamic cone penetration tool, which measures changes in the strength of log piles before and after burial. In biological studies, soil compaction can be inferred from the presence of particular types of microbial communities in the soil and logs, particularly through assessment of oxygen availability. Oxygen, along with wood moisture levels, is a key indicator of wood degradation [10, 11]. In aerobic conditions, wood decay is primarily driven by wood-decaying fungi and bacteria [12], whereas in anaerobic, waterlogged environments, erosion bacteria become the dominant agents of wood biodeterioration [10, 13]. Degradation in anaerobic conditions is much slower than that in aerobic environments. While many studies have documented patterns of pile degradation by bacteria in anaerobic conditions [5, 6, 10], fewer have focused on the bacterial community structures within these piles or the taxonomy of the organisms involved in the degradation processes [14]. Gaining a deeper understanding of the bacterial communities present in logs and the surrounding soil over time is critical for elucidating the biological processes behind wood deterioration. This knowledge could lead to more effective liquefaction mitigation strategies in log-piling methods.

In this study, to clarify the bacterial communities that form on logs in the early stage after they are placed in the ground, logs that had been experimentally piled in the ground for 2 years were pulled out, and the bacterial community structure formed inside and outside the logs was analyzed using amplicon sequencing. In addition, we explored the relationships between the bacterial community structures and biodeterioration of the logs. While traditional methods of microbial identification are time-consuming, amplicon sequencing offers a rapid and objective way to analyze microbial communities [15, 16].

Sampling site

Japanese cedar (Cryptomeria japonica) logs were used as test piles. After removing the bark, the logs were cut into cylinders approximately 4000 mm long and 140 mm in diameter. The bottoms of the logs were cut to form a 50-mm long cone (similar to the tip of a pencil). The test plot measured 2800 mm ×ばつ 2910 mm (Fig. 1) and was located in an experimental field at the Toyama Forestry and Forest Products Research Center (Toyama Prefecture, Japan), characterized by surrounding grasslands. The logs were buried vertically in the soil at 560-mm intervals (as shown in Fig. 1A), with the log heads positioned at a depth of 500 mm below ground level (Fig. 2B). In September 2022, which was 2 years after burial, five logs (logs A, B, C, D, and E) were removed from the soil and experimental samples were taken.

The soil attached to the surface of the logs was sampled at four levels: 1000, 1900, 3100, and 4000 mm below the surface of the ground (Fig. 1C). Each soil sample was collected with a spatula and placed in a plastic bag, which was immediately stored at − 25 °C until analysis. Three discs were cut from each log, specifically from the upper section (UPP, 550 mm below the surface of the ground), the middle section (MID, 2500 mm), and the lower section (LOW, 4450 mm) of the log (Fig. 1D). Wood slices of approximately 20-mm thickness were cut from the disks along the disk center and small specimens were collected from the outer sapwood for DNA extraction, microscopy observation, and density measurements. All samples were stored at − 25 °C until use.

Analysis of pH, total carbon (C), and total nitrogen (N)

The frozen soil samples were allowed to thaw to room temperature. For pH analysis, a 1 wet-gram subsample was diluted in 9 mL of deionized water, shaken, and left for 10 min before being centrifuged for 15 min. The pH was immediately measured using a Horiba Navi F-52 pH meter (Kanagawa, Japan). Another subsample was air-dried for DNA extraction, total carbon (total C), and total nitrogen (total N) analysis; the latter were determined using the dry combustion method and a Sumigraph NC-Trinity Fully automated type instrument (Sumika Chemical Analysis Service, Ltd., Osaka, Japan).

Bacterial DNA extraction and polymerase chain reaction (PCR)

Bacterial DNA was extracted from 0.3 g of soil subsamples using the Extrap Soil DNA Kit Plus v. 2 (BioDynamic Laboratory Inc., Tokyo, Japan), with slight modifications to the manufacturer’s protocol, in 1.5-mL polymer bead tubes. For log samples, a wood sample was taken from the outer side of the sapwood and ground into powder. DNA extraction was then carried out from 0.02 g of the sawdust using ISOPLANT II reagent (Nippon Gene, Tokyo, Japan), with slight modifications to the manufacturer’s instructions. Mechanical cell disruption was achieved in both soil and wood samples using a Qiagen Tissue Lyser (Qiagen, Venlo, the Netherlands) at 25 Hz for 2 min.

PCR was performed using a MightyAmp DNA Polymerase v.3 kit (Takara Bio, Shiga, Japan) with extracted DNA as the template. The V3–V4 region of the 16S ribosomal DNA (rDNA) region was targeted for bacterial sequencing analyses, and a two-step tailed PCR method was used to generate a library of amplicon sequences [17].

Next-generation sequencing analysis

The quality of the prepared libraries was assessed using a Fragment Analyzer and dsDNA 915 Reagent Kit (Agilent Technologies, Santa Clara, California, USA). Paired-end sequencing with 300-bp reads was performed on the MiSeq platform (Illumina, San Diego, California, USA) using the MiSeq Reagent Kit v. 3 (Illumina), resulting in between 15,000 and 36,000 read pairs per sample. Primer sequences were removed from the extracted reads using the fastx trimmer tool from the FASTX-Toolkit. Reads with a quality value < 20 were filtered out using Sickle v. 1.33 [18], and those with a length of < 130 bases and their corresponding pairs were discarded. Paired-end reads were merged using FLASH v. 1.2.11 [19]. Representative sequences and an amplicon sequence variant table were generated after removing chimeric and noisy reads with DADA2 [20] in Qiime2 v. 2022.8. Taxonomic classification was performed by comparing representative sequences with the Silva v. 138 database using the feature-classifier plugin in Qiime2 at 99% similarity [21]. The sequences were deposited in the DDBJ Sequence Read Archive (SRA) under the BioProject accession number PRJDB18461.

Diversity and multivariate analysis

In this study, the order was selected as the primary taxonomic level of classification because of low resolution at the family or genus level for many groups. The relative abundance, and metrics such as richness, the Shannon diversity index, and evenness were estimated. Pearson correlation coefficients among soil environmental factors were also calculated, and heatmaps along with Euclidean distances were generated, following the methodology described by Gasca-Pineda et al. [22], using v. 2.6–4 of the "vegan" R package. The impact of environmental variables on the bacterial communities at different soil sampling sites was evaluated through principal component analysis (PCA), conducted using the redundancy analysis function within the "vegan" package [23]. The PCA was applied to the Hellinger-transformed abundance data set and the environmental variables data. The differences between bacterial communities in soil and logs were visualized using non-metric multidimensional scaling (nMDS) based on Bray–Curtis similarity [24].

Information regarding the functions, environments, and roles of bacteria in ecosystems was sourced from The National Center for Biotechnology Information Taxonomy database and the List of Prokaryotic Names with Standing in Nomenclature website.

Microscopic observation

Transverse sections approximately 20-μm thick were cut from each sample piece of wood using a sliding microtome Yamato REM-700 (Yamato Koki Co. Ltd., Saitama, Japan). The sections were first stained with safranin for 3 min, then washed in distilled water for 1.5 min and 70% ethanol for 1 min. This was followed by a second staining with Astra blue and a dehydration sequence in 70%, 80%, and 100% ethanol, each step lasting 1 min. Finally, the sections were mounted on glass slides, fixed with resin, covered with a coverslip, and examined under a light microscope.

Density measurements

Wood density was determined using the method described by Fujiwara [25]. Specimens with dimensions approximately 10 mm ×ばつ 10 mm ×ばつ 10 mm were prepared, oven-dried at 55 °C for 3 days to constant weight, and then coated with a waterproof agent to prevent water absorption. The volume of each specimen was ascertained by employing the principle of Archimedes, immersing them in water and measuring the displaced volume. Wood density was calculated as

Statistical analysis

Five logs were harvested and used as replicates (n = 5). Statistical analysis was performed using OriginPro 2024b software (Lightstone Corp., Tokyo, Japan). Data with a normal distribution, such as for pH, total C, total N, richness, and the Shannon index, were analyzed for significant differences using one-way analysis of variance (ANOVA), followed by Tukey’s post-hoc test. For data that did not follow the normal distribution, such as evenness, nonparametric tests were applied. Specifically, Kruskal–Wallis ANOVA, followed by Dunn’s multiple comparison tests, was used to identify differences between groups. A significance level of P < 0.05 was set for all analyses.

Assessment of ground exposure effects on Japanese cedar logs: visual observation and analysis of soil composition

The untreated Japanese cedar logs exhibited noticeable color changes after 2 years of burial: darkening occurred from the exterior surfaces through the sapwood area (Fig. 2). The interior, however, retained its original color. The soil surrounding the logs changed from dark-brown to dark-gray as the depth increased; however, the pH of the soil did not change with soil depth (Table 1). At the time of log removal, the groundwater level was approximately 1500 mm below the surface of the ground. Soil analysis revealed peak concentrations of total C and total N at depths of 3100 and 4000 mm, with total C values of 15.9 g/kg at both depths, and total N values of 1.1 and 1.0 g/kg, respectively, at those depths. The highest C:N ratio, 22.3, was observed at a depth of 1000 mm. Pearson correlation analysis demonstrated a strong correlation between total N and total C, but weak correlations between these variables and the C:N ratio with soil pH (Fig. S1).

Analysis of bacterial diversity

Diversity indices, such as taxon richness and the Shannon index, are crucial for quantifying microbial diversity. In this study, DNA extraction was not performed on the soil sample from log D at a depth of 1000 mm because of its limited quantity. In addition, PCR results indicated an absence of DNA from the log samples A-UPP (i.e., the upper section of log A), B-UPP, and B-MID. Variability in bacterial richness at the order level was observed across the remaining sampling sites, with soil samples showing higher taxon richness and Shannon index values than samples from logs.

Soil samples from around log A exhibited higher richness values at all measured depths (values in parentheses)—1000 mm (151), 1900 mm (104), 3100 mm (117), and 4000 mm (116)—than the soil from around other logs (Fig. S2). The upper sections of logs C (63) and E (50) displayed higher richness than other samples from logs. Statistical analysis indicated no significant differences in richness between soil depths (Table 2). However, in the logs, the richness in the upper sections was significantly greater than that in the middle and lower sections. Interestingly, the richness in the soil was comparable to that observed in the upper sections of the logs. No statistically significant differences were found in either Shannon index values or evenness across the depths in soil or the sections in logs (Table 2). Values of the Shannon index at each sampling site showed minimal variation, with the values for soil samples ranging from 2.20 (D-1900) to 3.59 (A-1000), and for log samples from 0.99 (D-MID) to 2.63 (E-UPP) (Fig. S3). The evenness values in soil were relatively stable, varying between 0.40 (D-1900) and 0.65 (A-1000). The evenness values in logs showed more fluctuation, from 0.21 (D-MID) to 0.56 (E-UPP). However, despite these fluctuations, the average evenness values for both soil and logs were around 0.50.

Given that diversity, as measured by the Shannon index and evenness, showed no significant differences between the soil and logs, richness was used as the primary indicator of diversity.

Analysis of the comparative abundance of bacterial communities in soil and logs

Three dominant phyla were identified: Actinomycetota, Bacillota, and Pseudomonadota (Table 3). Actinomycetota showed high abundance in soil at a depth of 1000 mm (41% ± 8%) and low abundance in the MID (4% ± 5%) and LOW (8% ± 5%) log sections. Bacillota had their highest abundance in the LOW log sections (57% ± 15%) and their lowest abundances at depths of 1000 (3% ± 1%), 1900 (9% ± 6%), and 3100 mm (9% ± 4%) in the soil. Pseudomonadota did not show significant variation across the sampling sites, either in soil or within the logs. These findings suggest that the Actinomycetota were predominantly aerobic bacteria, while the Bacillota were mostly anaerobic. The Pseudomonadota, however, included both aerobic and anaerobic bacteria.

This pattern was corroborated at the class level, as depicted in a heatmap (Fig. S4). Actinobacteria, belonging to the phylum Actinomycetota, were abundant at the 1000-mm sampling sites, reflecting the high abundance of this phylum at this depth in the soil. Similarly, classes within Pseudomonadota, such as Alphaproteobacteria and Gammaproteobacteria, were dominant across most sampling sites, consistent with the stable relative abundance of Pseudomonadota. Alphaproteobacteria showed slightly lower abundance in samples D-4000, D-UPP, and A-MID than in other samples, while Gammaproteobacteria were less dominant in soil sample D-4000 and log sections D-UPP, A-MID, E-UPP, and E-LOW than in samples from other sites. In addition, the classes Clostridia, Desulfitobacteriia, and Negativicutes, which belong to Bacillota, were highly abundant in the LOW sections, aligning with the overall increase of Bacillota in deeper log sections.

To identify the dominant bacterial orders at each sampling site, we selected those with relative abundances exceeding 2%, grouping those below this threshold as "others" (Fig. S5). The relative abundances of bacterial orders in soil samples were generally consistent across the same depth, exhibiting a stable pattern. In contrast, within log sections from the same depth, relative abundances were more variable; sometimes these orders showed consistency within the sections, but in other instances they displayed significant fluctuations. For example, Veillonellales–Selenomonadales had relative abundance of 0.6% in C-UPP, 15% in D-UPP, and 7.6% in E-UPP.

Figure 3 presents the average relative abundance values of bacteria in soil and logs, derived from Fig. S5. Four dominant bacterial orders were identified in the soil at each depth (1000, 1900, 3100, and 4000 mm): Micrococcales, Rhizobiales, Sphingomonadales, and Burkholderiales. Except for Rhizobiales, the relative abundances of these orders decreased in the logs. Micrococcales were most abundant in the soil at 1000 mm (27.6%) and reached their highest levels in the logs in the UPP sections (5.6%). Rhizobiales peaked in the soil at 1000 mm (9.3%) and in the UPP sections of the logs (33.8%). Sphingomonadales had their highest abundance in the soil at 1000 mm (18%) and in the LOW sections of the logs (3.5%). Burkholderiales were most abundant in the soil at 4000 mm (20%) and in the LOW sections of the logs (11.3%). Other bacterial orders, such as Pseudomonadales and Xanthomonadales, exhibited fluctuations in abundance depending on depth.

Unlike the stable patterns of dominant bacterial orders in the soil, the distribution of bacterial orders in the logs varied based on the section (UPP, MID, and LOW). Three distinct bacterial community patterns were identified in logs. The first pattern was dominance in the UPP section (550 mm) and decrease with depth [in the MID (2500 mm) and LOW (4450 mm) sections]. For example, Rhizobiales showed relative abundance values of 33.8% in the UPP, 12% in the MID, and 5% in the LOW sections. The second pattern involved bacterial orders that had low abundance in the UPP sections but increased with depth. Dominant bacteria with this pattern included Desulfitobacteriales and Veillonellales–Selenomonadales, with relative abundances of 3% and 7% in the UPP, 4.8% and 12.3% in the MID, and 11% and 14% in the LOW sections, respectively. The third pattern featured bacteria that were dominant in only one section. An example of this was Enterobacteriales, which were dominant in the MID sections, with relative abundances of 1.1% in the UPP, 17.3% in the MID, and 5.5% in the LOW sections.

nMDS analysis revealed distinct differences in bacterial community composition at various depths within both the soil and the logs (Fig. S6). Furthermore, the bacterial communities in the soil were significantly different from those in the logs, indicating distinct ecological niches.

Relationship between environmental factors and soil bacterial communities

The clustering of sampling sites depicted in Fig. 4A reflects their similarity or dissimilarity in bacterial community composition; sampling sites with more similar bacterial compositions are clustered closer together, while those with greater differences are positioned farther apart. Notably, the B-4000 sample stands out because of its separation from other groups, indicating a low similarity in community composition.

A PCA plot (Fig. 4B) illustrates the relationships between bacterial communities across different sampling sites and various soil environmental factors, including total C, total N, the C:N ratio, and pH. The PCA results indicate that bacterial composition at several sampling sites was influenced to varying degrees by these environmental variables. For instance, sampling site B-1000 was moderately influenced by total N, and site A-1000 showed a lesser influence from total C. Sites C-1000 and E-1000 were both moderately influenced by total C. In addition, site D-4000 was moderately influenced by both total C and pH, while B-4000 was affected moderately by pH and the C:N ratio. Other sampling sites did not appear to be influenced by these soil environmental factors. Therefore, it can be concluded that the influence of soil environmental factors on the bacterial composition at each sampling site ranged from "no influence" to "moderate influence."

Biodeterioration analysis and wood density measurement

Observations for each sapwood sampling site were made to assess biodeterioration of the wood cell walls. Examination of the tracheid cells in the latewood area of each sample showed early signs of biodeterioration across the upper, middle, and lower sections, marked by the erosion of multiple cell walls. Representative observations can be seen in Fig. 5. Among all samples, D-MID and E-MID exhibited the most significant cell wall erosion, followed by sections C-UPP and C-LOW (Fig. S7). Notably, sample C-MID showed no signs of biodeterioration. Furthermore, no signs of fungal growth or degradation were observed in the wood cell walls of any samples. Density measurements for each sapwood sample yielded nearly identical values, ranging from 0.30 to 0.40 g/mL (Table 4), aligning with the typical density range for Japanese cedar wood of 0.30–0.42 g/mL [26].

The LP-LiC log-piling method has been extensively developed by civil engineers in Japan as a ground improvement technique to mitigate soil liquefaction during seismic activity. However, despite its effectiveness in stabilizing soil, the mechanisms and rate of log deterioration in this method remain poorly understood. Previous studies by Hara et al. [8, 27] demonstrated that soil compaction resulting from log installation influences both ground stabilization and log degradation. Ground stabilization is assessed using a portable dynamic cone penetration tool [8], while log deterioration is evaluated through visual inspection, soundness tests, or pilodyn penetration tests [27, 28]. This study focused on the role of bacterial communities in log deterioration, emphasizing their transition from soil to logs and their distribution at various depths.

Microscopic observations of the wood tissue in the log samples from the upper to the lower sections showed that the cell walls were generally intact. However, early signs of bacterial degradation, including cell wall erosion, were observed, particularly in the outer sapwood. Despite these indications, the wood density remained unchanged, suggesting minimal degradation of the logs in this study after 2 years of burial. As expected, microbial activity triggered the biodeterioration process, although the deterioration was limited, likely due to the relatively short experimental period. It is well known that erosion bacteria are the primary degraders of wooden foundations in anaerobic and waterlogged soils [10, 29]. Erosion bacteria initiate degradation by targeting the cell lumen or bordered pit regions, subsequently eroding the cell walls but leaving the middle lamella intact as a fragile network [13]. The absence of fungal activity in this study supports the observation that low oxygen concentrations inhibit fungal colonization; soft-rot fungi typically require oxygen levels > 0.3 mL/L in water [30].

To further understand the degradation processes in this study, we examined the bacterial communities in the logs and surrounding soil. Our findings indicate that the soil environment played a significant role in shaping the bacterial composition within the logs (Fig. 3). Aerobic bacterial orders, such as Micrococcales, Rhizobiales, Sphingomonadales, and Burkholderiales [31], dominated the soil. However, the relative abundances of these bacterial orders were lower in the logs than in the soil, except for that of Rhizobiales, which increased significantly in the upper log sections.

In contrast, anaerobic bacteria, particularly members of the class Clostridia and the orders Desulfitobacteriales and Veillonellales–Selenomonadales [31, 32], are present in low abundance in the soil but were significantly increased within the logs. This shift from aerobic to anaerobic bacterial communities was more pronounced in the deeper log sections (middle and lower), likely due to decreasing oxygen availability with depth. The abundance of specific taxa within the logs suggests their potential role in driving the biodeterioration of the wood while adapting to these environmental gradients.

The shift from aerobic to anaerobic bacterial communities within the logs likely reflects changes in oxygen availability at different depths. The variation in oxygenation between log sections in this study is likely driven by differences in waterlogging and fluctuations in local groundwater levels [8, 27]. The LP-LiC technique often results in only partially waterlogged logs (i.e., the middle and lower sections are frequently waterlogged, whereas the upper sections are not) [8, 28], leading to variable oxygenation. These findings align with oxygen gradient patterns observed in historical studies of log deterioration in European building foundations [5, 6, 10], where partial decreases in groundwater levels facilitated soil aeration and microbial activity, accelerating degradation [6, 29].

We compared our findings with previous research on bacterial community shifts in waterlogged logs. The study by Horisawa et al. [33] provides valuable insights, demonstrating a transition in bacterial dominance within the surrounding soil over a prolonged burial period of 55 years. Facultative bacteria, such as Rhizobiales and Burkholderiales, were predominantly found in the soil surrounding the upper log sections, whereas anaerobic bacteria were more abundant around the middle and lower sections. Within the logs, anaerobic communities—including Lachnospirales, Oscillospirales, and members of Limnochordia—were consistently present across all sections, from the upper to the lower portions. The study conducted by Horisawa et al. [33] showed that the environment around the logs was anaerobic.

In contrast, this study conditions differed, which may explain variations in bacterial community structure. These differences could be attributed to log placement depth and fluctuations in groundwater levels. In this study, log heads were positioned 500 mm below ground level, and groundwater levels fluctuated, reaching a depth of approximately 1500 mm at the time of log removal. In comparison, Horisawa et al. [33] placed log heads at a depth of 1200 mm, where groundwater remained more stable at 1300 mm below the ground surface. These conditions suggest that the logs in Horisawa et al. [33] remained fully submerged, whereas those in this study experienced periodic changes in water saturation, potentially influencing bacterial colonization and degradation patterns.

Furthermore, Horisawa et al. [33] reported minimal wood deterioration under anaerobic conditions, even after a prolonged 55-year burial period, with no significant reduction in wood density. Elam and Björdal [29] suggest that wood placed in fully saturated or anaerobic conditions experiences reduced bacterial activity, resulting in a slower and more gradual degradation process. In addition, maintaining a high groundwater level is essential, as a deeper water column above the log head significantly slows the degradation process [5]. In contrast, the proximity of the log heads to the ground surface in this study, combined with fluctuating groundwater levels, likely created a semi-oxygenated environment that influenced both bacterial community composition and the degradation process.

The fluctuation in groundwater levels observed in this study likely contributes to an oxygen gradient, which in turn influences bacterial richness and community composition. Bacterial richness was comparable between the soil and upper log sections but declined significantly in the middle and lower log sections, where low oxygen levels likely restricted bacterial diversity. These results indicate that the upper log sections are more oxygenated than the middle and lower sections, possibly because they are closer to the ground surface and not consistently waterlogged. These findings align with previous studies demonstrating that oxygen levels strongly influence bacterial activity, diversity, and community composition [34, 35]. Our nMDS analysis corroborates these findings, showing significant compositional differences between soil and log bacterial communities. Depth appears to be a key factor influencing microbial diversity, potentially having a greater impact than other environmental factors, such as pH, total C, total N, and the C:N ratio, which showed no significant effect in our analyses.

Within the oxygen gradient, facultative bacteria such as Rhizobiales, Burkholderiales, and Enterobacterales [31, 32] dominate from the upper to the lower log sections (Fig. 3), reflecting their adaptation to varying oxygen availability. Rhizobiales thrived in the semi-oxygenated upper sections, while Enterobacterales were most abundant in the middle sections, where oxygen availability declined further. In the lower log sections, Burkholderiales showed higher abundance than Rhizobiales and Enterobacterales, but Veillonellales–Selenomonadales, which are strict anaerobes, also increased in dominance and slightly exceeded Burkholderiales in abundance.

The semi-oxygenated conditions in the soil may contribute to the observed dominance of Alphaproteobacteria and Gammaproteobacteria in the logs, because these taxa are known for their metabolic adaptability to various environmental conditions, including fluctuating oxygen levels [31, 32]. Orders such as Rhizobiales, Burkholderiales, and Enterobacterales likely benefited from nitrogen-fixing capabilities that facilitated bacterial colonization [36, 37]. The surrounding environment of the experimental field, which is characterized by grasslands, may have influenced the bacterial communities found in the logs. Grassland soils harbor nitrogen-fixing bacteria, particularly from the orders Rhizobiales and Burkholderiales [37]. Their dominance in the logs suggests that they play a significant role in driving the deterioration observed in this study by actively degrading wood cell walls. Furthermore, these taxa contribute to organic matter decomposition and recycling [32, 38, 39].

Understanding the aerobic-to-anaerobic transition is essential for elucidating the microbial contributions to log deterioration in LP-LiC conditions. Our results suggest that the bacterial communities in the logs predominantly originate from the surrounding soil, with their composition adapting to the environmental gradients within the logs. The use of untreated logs with the bark removed, as in this study, facilitates bacterial penetration, supporting colonization and the establishment of distinct community structures [14]. Further research is required to clarify the specific roles of Alphaproteobacteria and Gammaproteobacteria in the early stages of degradation of such logs.

While this study provides valuable insights into the relationships between bacterial communities and log deterioration, several limitations should be addressed in future research. The 2-year burial period applied here may not capture the long-term dynamics of microbial colonization, composition, and wood degradation. In addition, environmental variables such as changes in the chemical composition of the soil, temperature, seasonal groundwater fluctuations, and soil texture were not measured but could significantly influence microbial activity. Future studies should also investigate the metabolic pathways of key bacterial taxa to elucidate their roles in wood degradation processes. Nevertheless, by integrating biological and engineering perspectives, this study may contribute to a more comprehensive understanding of the interactions between microbial communities, soil, and wood in ground improvement systems.

This study provides valuable insights into the interactions between microbial communities and log deterioration in the log-piling method over a 2-year period. The conclusions from the study are as follows:

1. 1.

   Bacterial communities in logs primarily originate from the surrounding soil, with shifts in community composition influenced by environmental factors, such as oxygen availability and the characteristics of the logs.

2. 2.

   Aerobic bacteria dominate the soil and upper sections of the logs; anaerobic bacteria become more prevalent in the middle and lower log sections, correlating with decreasing oxygen levels.

3. 3.

   Early signs of bacterial erosion were observed in the outer sapwood, resulting in minor deterioration of the logs within the 2-year burial period, without significant changes in wood density.

4. 4.

   Dominant bacterial taxa from the classes Alphaproteobacteria and Gammaproteobacteria, particularly from the orders Rhizobiales, Burkholderiales, and Enterobacterales, are likely responsible for the degradation of outer sapwood cell walls, contributing to wood deterioration.

The data sets used and analyzed in the current study are available from the corresponding author upon reasonable request. Information about sequence data files is provided in the supplemental materials.

LP-LiC:

Log-piling method for liquefaction mitigation and carbon stock

nMDS:

Non-metric multidimensional scaling

PCA:

Principal component analysis

PCR:

Polymerase chain reaction

rDNA:

Ribosomal DNA

We express our great appreciation to Mr. Chikai Igarashi of Kanematsu Sustech Corporation, Dr. Kazuhiro Shiba of Toyama Prefectural Agricultural, Forestry and Fisheries Research Center, Dr. Hideo Kato and Dr. Yoshitaka Kubojima for their cooperation in collecting samples. We are also grateful to Dr. Patricia Velez and Dr. Jaime Gasca-Pineda from the Universidad Nacional Autónoma de México, Mexico, for their instruction in R programming. Finally, we thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

This research was supported by JSPS KAKENHI grant numbers 19K06170 and JP22H02410.

Authors and Affiliations

1. Graduate School of Engineering, Kochi University of Technology, 185 Miyanokuchi, Tosayamada, Kami, Kochi, 782-8502, Japan

   Sitompul Afrida & Sakae Horisawa

2. Natural Science Cluster, Science and Technology Unit, Research and Education Faculty, Kochi University, 2-5-1 Akebono-cho, Kochi, Kochi, 780-8502, Japan

   Tadashi Hara

Contributions

Conceptualization, SH; data curation, SA and SH; validation, SA, TH, and SH; formal analysis SA and SH; investigation, SA and SH; resources, TH and SH; writing—original draft preparation, SA; writing—review and editing, SA and SH; visualization, SA and SH; supervision, TH and SH; project administration, SH; All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Sitompul Afrida.

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors have read and approved the final manuscript and consent to its publication in the Journal of Wood Science.

Competing interests

The authors declare that they have no competing interests.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions