Soil communities are essential to ecosystem functioning, yet the impact of reducing soil biota on root-associated communities, tree performance, and greenhouse gas (GHG) fluxes remains unclear. This study examines how different size fractions of soil biota from young and mature forests influence Alnus glutinosa performance, root-associated community composition, and GHG fluxes. We conducted a mesocosm experiment using soil community fractions (wet sieving through 250, 20, 11, and 3 µm) from young and mature forest developmental stages as inocula. The results indicate that the root-associated community composition was shaped by forest developmental stage but not by the size of the community fractions. Inoculation with the largest size fraction from mature forests negatively affected tree growth, likely due to increased competition between the plants and soil biota. In addition, GHG fluxes were not significantly impacted by either size fraction or forest developmental stage despite the different community composition supplied. Overall, our research indicates that A. glutinosa strongly selects the composition of the root-associated community, despite differences in the initial inoculum, and this composition varies depending on the stage of ecosystem development, impacting the performance of the trees but not GHG fluxes.

To study the effect of a young and mature forest soil community, forests spanning two distinct developmental stages were selected. The forests were categorized based on their time of planting as either “young” (2010 – 2015) or “mature” (1880 – 1927). Within each developmental stage, six individual forests were sampled (6 young and 6 mature), each with their own soil community, characteristic of each developmental stage. Each individual forest acted as a replicate for its corresponding development stage. All sampled forests were located in the province of Drenthe in the Netherlands, had sandy soil and were initially planted with oak (Quercus robur and Quercus petraea).  In each forest stand, soil was collected 15 cm deep from multiple points around the stand (February 2022). Each point was at least 10 m apart from the previous one and the soil from all points was placed in one bag to make one homogenate per forest stand. Additionally, three soil cores (150 cm³, 12 cm deep) were collected from each stand to measure soil properties. In total, we collected 12 bags of soil (one from each forest) and 36 soil cores. The soil from each bag was initially sieved through a 2 cm mesh to further homogenize the soil and remove large debris and stones.

The soil from the cores (150 cm³, 129.5 ± 0.76 g) was oven dried at 40 ℃ for 96 hours to measure soil moisture, pH (1:2 soil:water ratio) and the bulk density of the soil (Schofield et al., 1955). The loss on ignition (LOI) method was used to measure the soil organic matter of the soil by first drying the soil at 105 ℃ for 48 hours and then heating at 550 ℃ inside a muffle furnace for four hours (Heiri et al., 2001). The soil heated at 550 ℃ was then used to determine the sand/silt fractions using a HAVER EML 200 Premium automated shaker (Oelde, Germany). Soil texture was classified into gravel (>2 mm), Sand (2 - 0.045 mm) and Silt and Clay (<0.045 mm). Black alder (Alnus glutinosa, (L.) Gaertnr) was selected as the species to grow in the inoculated sterile soils of our experiments. A. glutinosa trees were present in all of the sampled sites despite a clear dominance of oak trees (Quercus robur and Quercus petraea). Seeds of native alder were acquired from the Nature Agency Staatsbosbeheer located in the Netherlands. The seeds were germinated in a nursery using commercial potting soil where they were grown for 3 weeks and subjected to a watering regime of three times a week.

In order to assess the impact of soil community size on A. glutinosa tree performance, we prepared four community size fractions (250, 20, 11 and 3 µm) from each of the forest soil samples, under sterile conditions, following methods from studies that have reliably proven that with reducing size fraction, the soil community simplifies (e.g. Wagg et al., 2014; Wang et al., 2019; Li et al., 2020). We prepared phosphate buffer (1 g KH₂PO₄ in 1 L DI H₂O, pH 6.5) and mixed it with soil (1:2, w/w) in four batches, each followed by sieving through 1 mm and 250 µm meshes. One-quarter of the filtrate was stored at 4°C (the 250 µm fraction), and the rest was further processed. To prevent clogging of the filter papers, the filtrate was passed through 63 and 45 µm sieves to remove more soil particles, and then through 20, 11 and 3 µm Whatman filter papers to generate the remaining fractions. These steps were done using a Buchner funnel and a LABOPORT N816.3 KT.18 vacuum pump (KNF, Utrecht, the Netherlands). After each step, portions of the recovered filtrate (1/3^rd at 20 µm, 1/2 at 11 µm and the rest at 3 µm) were stored at 4°C, representing the respective size fractions.

To prepare the inocula, each fraction stock was thawed and centrifuged to decant the supernatant (LB medium + glycerol). For each of the four fractions of each forest, the pellets were then resuspended in KH₂PO₄ buffer and transferred to a flask that was filled up to 200 ml with KH₂PO₄ buffer. The same procedure was followed for each of the four fractions of all twelve forests, resulting in a total of 48 liquid inocula with each individual forest acting as a replicate in its forest development stage (six replicates for young and six replicates for mature). Each of these inocula was randomly applied to 3.15 L (15x15x20cm) pots that were filled with gamma-sterilized grassland soil (Salonius et al., 1967). Finally, ten pots with gamma-sterilized grassland soil received just the KH₂PO₄ buffer without any community fraction inocula and another ten pots with gamma-sterilized grassland soil were left completely untreated and received neither buffer nor inocula bringing the total number of pots to 68.Two additional experiments were performed using the same setup but supplemented with either bactericide or fungicide with the aim of disentangling the effects of fungi and bacteria from each fraction.

After receiving the inocula, the pots were placed in a climate room with relative humidity 70%, light regime of 16h:8h (light:dark), air temperature of 20 ℃ (light) and 18 ℃ (dark). The pots were then left to settle for four days while the soil was kept moist.

On the fifth day after receiving the inocula, one four-week-old A. glutinosa seedling, germinated in autoclaved potting soil, of similar stem height (~ 2 cm) and number of leaves (at least two) was planted in each pot. The seeds were initially surface-sterilized with bleach solution (14%) to minimize surface contaminants (vertically shaking at 200 rpm for 10 minutes) before being thoroughly rinsed with autoclaved MQ H₂O to remove residual bleach. Seedlings were watered three times per week. Each pot was fully saturated with water during each watering event. During the first week, four seedlings died and these were immediately replaced. After 18 weeks, stem height (cm) and diameter (mm) were measured. Hereafter, CO₂, N₂O and CH₄ gas fluxes were measured for each tree, see below. When measuring GHG fluxes, the 11 µm size fraction was excluded from the measurements. Since access to the necessary equipment was limited, this size fraction was excluded as it was deemed to be the most redundant based on stem height measurements from the last week before the harvest.

Two days prior to the harvest of the pots, CO₂, N₂O and CH₄ gas fluxes were measured for each of the trees (pot + tree system). The measurements were conducted using an UGGA GLA-915-0011 (Los Gatos Research, Inc., Mountain View, CA, USA) gas analyzer instrument for CO₂ and CH₄ fluxes and a LICOR LI-7820 (LI-COR Environmental., Lincoln, Nebraska USA) instrument for N₂O fluxes. Each pot was placed in a 30 cm diameter PVC base on top of which we fitted a 100x30 cm (70 L) plexiglass chamber covered with double-layer black plastic bag to prevent photosynthesis and the fluxes were measured over a 5-minute enclosure time. A fan was placed at the top of the chamber, to ensure mixing during flux measurements. Between each flux measurement, the chamber was removed from the base and placed on its side so that it was flushed with atmospheric air.  After each flux measurement soil moisture and temperature were measured in each pot using a Delta T probe HH2 moisture meter (Delta-T Devices Ltd, Cambridge, United Kingdom) and a HANNA checktemp1 (HANNA instruments, Nieuwegein, Netherlands) instrument, respectively. These two instruments were thoroughly washed with ethanol in between samples to avoid cross contamination.

Two days later, all plants were harvested. After removing the soil, roots were thoroughly rinsed with running tap water, and the root nodule density was recorded. Root subsamples were then collected for both morphological characterization (preserved in water at 4 ℃) and DNA analysis (preserved at -20 ℃). Root scanning was conducted using an Epson Perfection V850 Pro scanner at 1200 dpi, and the WinRHIZO software (Regent Instruments, Quebec, QC, Canada) was employed to determine root length, specific root length, and the percentage of fine roots (defined as those with a diameter < 0.3 mm). The subsample was then oven-dried for 96 hours (40 ℃).  After removing the leaves, the stems were separated from the roots and all of them were oven-dried, similar to the root subsamples. The dry weight of the stems, leaves, roots and root subsamples was then recorded and this was used to calculate the above and belowground dry biomass of the trees.

After the plants were unpotted, the soil from each pot was thoroughly mixed to homogenize the top and bottom layer and subsamples of ~ 100 g of fresh soil were kept from each pot and stored at 4 ℃ for a week. The subsamples were sent to the university of Copenhagen to measure soil respiration. Roughly 50 g of each sample was placed in an incubator and all samples were normalized to 10% soil moisture. The temperature sensitivity of soil heterotrophic respiration (Q10) rate was measured for moisture-normalized samples (10% gravimetric water content) using a respirometer (Nordgren, 1988). Shortly, the soil sample was placed in a sealed cup and the respired CO₂ produced from the soil dissolved in a 0.1 M KOH-solution placed inside the cup. The change in electric conductivity of the KOH-solution is directly proportional to the CO₂ production. Data was recorded automatically on a computer.  Soil CO₂ respiration was measured at 9 temperature intervals from 5 ℃ to 25 ℃ with increments of +2.5 ℃.