Collaborators shipped all cores to EMSL on ice within 24 hr of core collection for immediate processing. One 30-cm core was dedicated to a biotic workflow for time sensitive measurements, and the other 30-cm core was dedicated to an abiotic workflow for structural and hydraulic characterization ( Figure 1B ). To account for surface heterogeneity, three of the 10 cm cores were homogenized with the top 10 cm of the larger biotic core. A full list of data types generated by the 1000 Soils Pilot is available in SI Table 1 .

Immediately on arrival (<48 hr), we processed the biotic core by separating it into 10 cm depth intervals, which were then sieved at 4 mm. The middle section was archived, and all analyses were conducted on homogenized top (0-10 cm) and bottom (deepest 10 cm) sections as described below. We measured microbial biomass C and N via chloroform fumigation (30–32). Aliquots of K₂SO₄ extractions from the initial (background) steps of microbial biomass measurements were archived at - 20° C for later analysis of NO₃ ^- via Cd-reduction method and NH₄ ⁺ via spectroscopic determination (33, 34). Inorganic P was extracted from each soil section via Bray or Olsen extraction (depending on pH) and subsequent analysis via methods outlined in (35, 36). We collected sub-samples of soil for DNA extraction and subsequent metagenomic sequencing on the Illumina NovaSeq platform (stored at -80° C). We measured soil pH using a 1:1 soil to water ratio using a calibrated pH probe. Gravimetric water content was determined via oven drying at 60° C. Remaining soil from each section was air dried prior to the analyses below. β-glucosidase potential activity and respiration rates were measured using colorimetric assays and the CO₂ burst method. Ion concentrations from 1:10 ammonium acetate extraction (K, Ca, Mg, Na) and 1:2 soil to diethylenetriaminepentaacetic acid (DPTA) extraction solution (Zn, Mg, Cu, Fe, B, SO₄ ^2--S) extraction were measured with Inductively coupled plasma mass spectrometry (ICP-MS), while SOM concentration was determined by total organic carbon-total nitrogen (TOC-TN) analysis and composition from liquid chromatography mass spectrometry (LC-MS), and Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS) analysis. Details on SOM characterization are provided in the SI.

The abiotic core was used to determine soil structure and hydraulic properties. We used X-ray computed tomography (XCT) to measure soil pore size, connectivity, distribution, and volume, as detailed below. In addition, soil hydraulic properties were measured using a series of instruments by The Meter Group (KSat, Hyprop, WP4C, Meter Group, Inc., USA).

Detailed methods for data collection and analysis for XCT and FTICR-MS can be found in the SI.

As the frequency of wildfires continue to increase due to anthropogenic climate change and land management practices, it is critical to understand how fire disturbances affect belowground biogeochemistry (18, 37). Wildfires alter belowground processes that generally result in decreased structural stability, microbial respiration and biomass, and SOM concentration (20–22, 24, 25, 38, 39). For instance, Certini et al. (40) reported wildfire significantly decreased total soil C, which was associated with increased lignin degradation. Others have reported the selective loss of SOM oxygen containing functional groups and/or with less condensed structures in response to wildfire (38). Changes in SOM concentration and composition are coincident with structural and hydrologic changes including decreased water infiltration and increased soil erosion (23). Because belowground C storage and bioavailability are dramatically altered by wildfire, it is important to understand the nature of wildfire effects on SOM chemistry and accessibility. Below, we describe a case study to demonstrate the utility of high-resolution data for addressing this need.

Our primary intent in this paper is to stimulate discussion and interest in large-scale collaborative efforts by introducing MONet. To facilitate this discussion, we present initial results from three of the first cores in the 1000 Soils Pilot, collected in November 2021 in partnership with the Confederate Tribes of Warm Springs in Warm Springs, OR, USA. This region typically experiences cold, snowy winters followed by warm dry summers with a mean annual precipitation of 1778 mm and a mean annual precipitation of 8.3°C. Soil samples were collected from plots dominated by warm and moist type grand fir (A. grandis) and snowbrush (C. velutinus). The unburned (UB plots) had forb and shrub ground cover, mature living conifers, and surface litter present (>1mm) and had no recent burn history in the last 20 years. The moderate burn frequency (MB) plots contained shrubs and dead-standing trees but very little surface litter (<1 mm) and had moderate severity burns in 2014 and 2020. The high frequency burn (HB) plots consisted of a low shrub ground cover, no live or dead standing trees, and very little surface litter (<1 mm) and had moderate severity burns in 2003, 2014 and 2020.

This subset of data is intended as a proof of concept for the utility of high-resolution data in understanding impacts of fire burn histories on soil physio-chemical properties and microbial functions. Though the dataset is small and does not allow for robust statistics (n = 3), it provides a wealth of information that is unfeasible with traditional biogeochemical approaches. With these differences in mind, we compared FTICR-MS and XCT across burn histories and soil depth to show differences in molecular SOM composition and soil structure.

We first compared soil biogeochemical properties including, β-glucosidase potential activity (BG), microbial biomass C (MBC) and N (MBN), and respiration at 24 and 96 hr for differences among soil sections and burn histories ( Figures 2A, B ). Elevated microbial biomass concentrations, respiration rates, and potential enzyme activities in the unburned site are consistent with increased biological availability when compared to the burned sites (39). We found the highest respiration rate at UB followed by the MB, with the lowest respiration at HB ( Figure 2B ). This is consistent with other literature, showing the sterilizing effects of high frequency fires on microbial respiration (26, 41). Across the first 24 hours after re-wet, the rate of CO₂ produced was greater in the top 10 cm vs. the lowest 10 cm across all cores. However, after 96 hours the rate of CO₂ production in soils from the bottom 10 cm was greater than the top 10 cm for all sites (39). Similarly, β-glucosidase potential activity and microbial biomass C and N were highest at UB ( SI Figure 1A and Figure 2B respectively). Lower microbial activity at the burned sites may be related to chemical recalcitrance or physical protection of SOM related to the fires (26, 41). For all the sites, respiration increased between 24 and 96 hours in the bottom 10 cm, possibly reflecting less bioavailable SOM in deeper soils than in surface soils (42). Corresponding changes in soil chemistry are shown in SI Figure 2 .

To examine pore space heterogeneity across soil sections and burn histories, we used XCT to interrogate the top and bottom 10 cm of each soil core. We compared total porosity, pore connectivity, and mean pore volume between the top and bottom sections of cores from UB, MB, and HB. Wildfire can impact soil structure at least to 10 cm depth depending on its duration, frequency, and intensity (22). For UB and HB, we observed higher porosity and pore connectivity in the top sections when compared to the bottom sections. This is likely related to the higher abundance of root systems in the top 10 cm. However, there was minimal difference between total porosity or connectivity between the top and bottom section of the MB site, possibly due to a large root section resulting in a larger mean pore size ( SI Figure 5 and SI Table 2 ). Across all cores, total porosity was more variable in the top than bottom section (8.82-20.08% vs 6.52-8.49%; SI Table 2 ). Pore connectivity was highest in the top section at UB (96.41%) and lowest in the bottom section at UB (70.12%); potentially indicating a homogenizing effect of wildfire on pore network structure at MB and HB. Overall, these data suggest that wildfire may decrease pore connectivity and total porosity in surface soils, likely contributing to increased hydrophobicity and soil erosion, while the deeper soils are more insulated from wildfire effects.

Soil organic matter is derived from chemical, physical, and biological processes that decompose plant and animal residues into microbial metabolites and other byproducts (1, 5, 43–45). This generates a heterogenous mixture of compounds in SOM pools. We used FTICR-MS to distinguish differences in SOM chemistry along depth profiles and across the three wildfire histories. There was a clear separation in water-extractable SOM composition between the top and bottom sections of each core as well as across burn regimes, as indicated by principal component analysis ( Figure 2C ). Additionally, the bottom sections exhibited more similar SOM pools (i.e., tighter clustering) than the top sections, possibly signifying stronger impacts of wildfire on surficial soils.

Since fire results in the oxidation of SOM to carbon dioxide (CO₂), we hypothesized that the extent of SOM oxidation, as indicated by the nominal oxidation state of carbon (NOSC), would be related to burn regime. In the top section, the average NOSC of SOM increased with stronger burn regimes (-0.34, -0.29, and -0.11 respectively for UB, MB, and HB), indicating more oxidization, while there was a binary effect of burning on NOSC in the bottom section ( Figure 2D ). MB and HB had similar mean NOSC that was higher than UB (-0.16 vs -0.35). The presence of more oxidized SOM in burned soils, is likely related to the incomplete oxidation of SOM and the addition of charred materials from wildfire (46–48). While limited to three sites, changes in NOSC could be an indicator of previous fire exposure, with the NOSC becoming more oxidized with increased high frequency fire exposure that persists even into deeper soil horizons.

We also analyzed FTICR-MS spectra to understand potential biochemical transformations and microbial pathways involved in SOM decomposition. Commonly inferred biochemical transformations such as the addition of methyl chain (-CH₂) groups were abundant at all sites and depths. We also note that one of the most abundant transformations observed was the conversion of carboxyl groups to CO₂ (greater than 1400 occurrences at each site and depth). The transformation of carboxyl groups to CO₂ likely results in combustion byproduct associated with incomplete oxidation of SOM during fire exposure. Within each core, the oxidation and/or addition of CO₂ accounted for a greater proportion of transformations at depth when compared to the top section at the same site. This is potentially related to the translocation of water-soluble oxidized OM to deeper soil sections (49).

Additionally, we mapped compounds detected by FTICR-MS to microbial reference pathways in KEGG to infer mechanisms of SOM decomposition. Across all samples, nearly 50% of all mapped metabolites were detected in a set of eight metabolic pathways. These included ‘biosynthesis of secondary metabolites’ (map01110, 11.7%), ‘metabolic pathways’ (map01100, 8.4%), ‘biosynthesis of antibiotics’ (map01130, 7.5%), and ‘microbial metabolism in diverse environments’ (map01120, 4.9%) among other pathways ( SI Table 4 ). Notably, the degradation of aromatic compounds (map012220) and the polycyclic aromatic hydrocarbons (PAH) degradation (map00624) were also among pathways with the highest number of mapped metabolites (16th and 19th most metabolites). The degradation of aromatic compounds was the most common pathway in the core with high frequency burns (HB). The degradation of PAHs was most common in the MB site; 1.61% of metabolites mapped to metabolic pathways in the moderate frequency burned core were associated with the degradation of PAHs in comparison to <1% of mapped metabolites in the unburned core.

Overall, our results were consistent with literature describing wildfire impacts on soils and provided greater understanding of molecular changes in soil organic matter composition and structure than is possible with typical measurements. In soils with recent fire history, we observed decreases in soil respiration, microbial biomass, and potential enzyme activity (20, 21, 24, 26, 39, 50). This was associated with shifts in SOM composition, with higher rates of oxidation apparent in surface soils. Deeper soils also showed increased NOSC as a function of previous fire history. By inferring biogeochemical transformations from SOM pool composition, we also show the possible conversion of carboxyl groups to CO₂ that is likely related to the oxidizing effects of wildfire. Using metabolic pathway mapping, we observed that the degradation of aromatic compounds and polycyclic aromatic hydrocarbons may also be associated with wildfire. Finally, wildfire appeared to induce structural changes including decreased pore connectivity and total porosity in surface soils, possibly contributing to hydrophobicity and decreased soil infiltration which can result in greater soil erosion (20, 23). The combined molecular and microstructural changes observed in this use case provide important insight and considerations in the management and modelling of forest ecosystems post-wildfire. As the MONet database continues to expand, we hope to use this information to model wildfire impacts on belowground C cycling at the regional to CONUS scale.