We assessed SOC and soil organic nitrogen (SON) concentrations and stocks, and the likelihood of SOC and SON degradation by microbes, by analyzing bulk soil samples at 10 cm intervals. We determined SOC and SON on subsamples (∼ 75 mg) via an elemental analyzer (vario MACRO cube, Elementar, Ronkonkoma, NY). We used SOC and SON concentration measurements to calculate each subsample's C:N ratio. To determine stocks of SOC in each horizon, we multiplied SOC concentrations by soil bulk density obtained in each horizon. Bulk density was measured using a three-dimensional laser scanner (3D Scanner Ultra HD, NextEngine, Inc., Santa Monica, CA) following Rossi et al. (2008).

We measured extractable DOC to estimate organic C that can be relatively easily mobilized and transported out of the soil profiles; note that this differs from DOC measured in soil porewater using lysimeters, and instead represents a salt-extractable pool. We analyzed soil samples at 10 cm intervals to auger refusal collected at each site during the growing season. Soil samples were extracted within 3 months of collection date. A total of 7.5 g of soil at field moisture was extracted with 30 mL of simulated rainwater (Laegdsmand et al., 1999). The extracted soil solutions were comprised of 47.9 µM NaNO3, 4.69 µM KCl, 23.81 µM CaCl2 × 2 H2O, 12.09 µM MgSO4 × 7 H2O, and 18.24 µM (NH4)2SO4, and adjusted to a pH of 4.2 ± 0.5 using HCl. Samples were placed on a shaker table for 30 min and centrifuged at 80 Hz (s-1) for 15 min. Samples were filtered through 0.45 µm nylon syringe filters and 50 mL acid-washed syringes. Filtered samples were stored in 10 mL centrifuge tubes, frozen and shipped overnight in a cooler with dry ice to the University of Kansas. DOC was analyzed from the thawed samples using a violet-pink Mn (III) pyrophosphate solution and a microplate reader (Biotek, UT).

To better understand the potential for microbial activity in these soils, we quantified microbial biomass C by horizon from pits dug in the summer of 2022 (Brookes et al., 1985). We exposed 5 g of each soil sample to chloroform for 24 h. To these fumigated sub-samples and to 5 g of unfumigated sub-samples we added 20 mL of 0.5 M K2SO4 and shook for 30–40 min at 220 rpm. These samples were filtered through a 0.45 µm polyethersulfone filter and their DOC concentration was determined via colorimetry (Bartlett and Ross, 1988) on a Synergy HT microplate reader (Agilent, USA). To assess the degree to which soil microbial communities were generating exo-enzymes that catalyze soil organic matter decay and thus can provide assimilable C- and N-rich compounds, we quantified potential activity rates of two such enzymes. We measured activity of β-glucosidase and N-acetyl-β-D-glucosaminidase, herein referred to as BGase and NAGase, which are linked to microbial C (BGase), and N and C (NAGase) acquisition (Sinsabaugh and Moorhead, 1994; Allison, 2014; Stone et al., 2014), using 4-methylumbelliferyl β-D-glucopyranoside (for BGase) and 4-methylumbelliferyl N-acetyl-β-D-glucosaminide (for NAGase) fluorescent tags. These tags were added to slurries made from approximately 1 g of soil and pH-adjusted 50 mM sodium acetate. We pipetted the blended sample into the desired substrate and incubated all plates at 25 °C for 18 h. Fluorescence from a Synergy HT plate reader (Agilent, USA) was used as a proxy for each enzyme's capacity to cleave monomers from the respective molecules undergoing decay (DeForest, 2009; German et al., 2011).

We quantified salt-extractable NO3- because of its importance as a biotically available form of N, and also because of its status as a readily leachable ion. As such, it can serve as an indicator of each soil's capacity to undergo elemental loss in surface soil with hydrologic fluxes, and provides a valuable point of comparison to DOC values. We extracted ∼ 10 g (fresh weight) of each soil sample with 0.5 M K2SO4 and repeated the shaking and filtering steps described above for microbial biomass C. Extracts were analyzed for NO3- (Synergy HT, Agilent, USA) using Shand et al. (2008), a microplate-based approach that relies on hydrazine sulfate and sulfanilamide to generate a color intensity directly related to NO3- concentration.

We also quantified soil organic matter δ15N, given these signatures' value as an indicator of the degree to which soil microbes have processed soil organic matter (Nadelhoffer and Fry 1988; Billings and Richter 2006). Sub-samples of each soil were dried, ground to fine powder, and weighed into a tin capsule for analysis. Values of δ15N were obtained at the Kansas State University Stable Isotope Lab, where an Elementar EA vario PYRO cube linked to an Elementar GeovisION isotope ratio mass spectrometer determine N concentration and δ15N, respectively.

To better assess possible differences in the chemical and physical controls on SOC stability, we also measured pH, effective cation exchange capacity (ECEC), soil texture, and wet aggregate size distribution (ASD). We focused on pH as it is known to strongly control microbial communities and MAOC (Kleber et al., 2015). The soil pH was determined in a 1:1 H2O soil slurry (Soil Survey Staff, 2023b). We focused on ECEC because ECEC has a high positive correlation with SOC, clay content, and aluminum and iron oxides (Solly et al., 2020), which are highly correlated with the formation of MAOC (Kleber et al., 2015). ECEC was determined by summing Ca, Mg, and K extracted using a Mehlich-3 solution (Culman, 2019). Mehlich-3 extraction was used instead of an ammonium acetate extraction because the soil samples had a pH of < 7.5 and there is very little to no calcium carbonate. In these conditions Mehlich-3 and ammonium acetate extractions yield similar ECEC values (Rutter et al., 2022).

We examined soil texture at each pit for several reasons. First, the total amount of clay is important to MAOC, and second, texture is known to impact the distribution and connectivity of pores. This connectivity influences how easily oxygen can diffuse into a soil profile and thus processes such as microbial respiration (Schjønning et al., 1999; Moldrup et al., 2013), and further regulates water and solute transport down-profile. Soil texture was analyzed on pit samples collected from 2020–2021 using a laser diffraction unit (Bettersizer S3, Bettersize Instruments, Dandong, Liaoning, China). Five grams of soil was sieved to 2 mm, and organic matter was removed by treating samples with 30 % hydrogen peroxide. A total of 10 mL of 10 % sodium hexametaphosphate (HMP) was added to the solution to prevent flocculation. The soil solution was pipetted into the Bettersizer until obscuration levels were between 14–20. We set clay–silt and silt–sand boundaries to be 6.6 and 60.33 µm, respectively (Makó et al., 2017).

We quantified aggregate size distributions as one key metric of soil structure. Aggregate-size distributions were measured on each soil horizon following Nimmo and Perkins (2002). Briefly, around 25 g of the largest air-dried aggregates were fully saturated with a Dickson apparatus (Dickson et al., 1991), and placed on a Yoder device where sieves (#4, 10, 17, 70) and soil samples were raised and lowered in the water 2.8 cm per stroke at a rate of 36 strokes per minute for 10 min. Following this agitation in water, the sieves with their respective aggregates were placed in a drying oven at 105 °C for 12 h. The soil material remaining on each sieve was dispersed with 200 mL of 2 gL-1 HMP, mixed for 10 min, passed through the sieve again, and oven dried at 105 °C for 2 h. Weights were recorded and mass fractions of water-stable aggregates were then calculated. Sieves divided aggregates into five classes: aggregates > 4.76 mm, aggregates between 2 and 4.76 mm, aggregates between 0.21 and 1 mm, and aggregates less than 0.21 mm. To simplify our analysis, we agglomerated these into three classes following Souza et al. (2023): fine aggregates (< 0.21 mm), intermediate aggregates (0.21–4.76 mm), and coarse aggregates (> 4.76 mm). A weighted geometric mean aggregate diameter (GMD) was calculated for each triplicate using the mass fractions of each aggregate-size class; the mean and standard deviation were calculated from these triplicate values to represent the aggregate diameter of each sample. The GMD values were divided by SOC content and the resulting values were used to characterize the propensity of C to form aggregates.

To determine the relationship between roots and C stability and transport, we measured the fraction of soil volume containing fine and coarse roots throughout the soil profiles using images collected from all 10 pits (e.g., 2020/2021 and 2022). We used ImageJ (Schneider et al., 2012) to overlay each image with a 1 cm × 1 cm grid. We then manually checked each 1 cm × 1 cm grid cell for the presence of a fine root (diameter < 1 mm) or coarse root (diameter > 1 mm) and noted these presence/absence scores for each grid cell. Our focus is the soil volume containing roots and thus directly influenced by roots. As such, only presence/absence and not count data were recorded, and in any cell containing both fine and coarse roots the presence of only the coarse root(s) was recorded given their greater volume (Billings et al., 2018). These measures are thus a conservative measure of direct root influence on soil volumes, derived at the centimeter scale for soil pedons. Centimeter-scale cell presence/absence data were transformed into the fraction of each 1 cm thick layer containing roots.

Soil sensor arrays were installed in the first set of pits (2020/2021) at the completion of sampling. Sensors were installed at depths of 15, 45, and 110 cm (or deepest depth; Table ) to monitor soil temperature (°C) and volumetric water content (VWC; EC-5, Meter Group, Pullman, WA), matric potential (kPa; Teros 21, Meter Group, Pullman, WA), O2 concentration (%; IB201806, Apogee Instruments, Logan, UT), and CO2 concentration (ppm; F0275476, Eosense, Dartmouth, Canada). Data were collected every 30 min for moisture, matric potential, and temperature, and hourly for O2 and CO2 given the power requirements. We focus on CO2 and O2 as they are indicators of soil microbial and root biotic activities, including heterotrophic respiration. Microbial activity directly and indirectly affects the formation of MAOC, SOC stabilization, and microaggregation (Dohnalkova et al., 2022). We used sensor data to investigate additional environmental controls on C dynamics. We converted O2 from millivolt readings to % by adding calibrated values to the millivolt value of O2. Each calibrated value was specific to the sensor installed and determined using atmospheric concentrations prior to installation. To focus on the growing season, we selected data from 15 June–29 August, which was 14 d before the first sample was collected (29 June) and ending 14 d after the last sample was collected (15 August). AS reflects 2021 data, SS reflects average daily 2021 and 2022, and ESG, EAG, and SG reflect 2022 data. These differences were because pits were installed with sensors in different years and some of the instrumentation had power outages and other unforeseen issues. We averaged daily temperature and VWC by week, and examined average and standard deviation of the O2 and CO2 over the growing season.

Spatial replicates controlling for all ecosystem-scale factors were not feasible in this study. Instead, we advance our understanding of SOC stability by examining a more diverse suite of biotic and abiotic ecosystem characteristics than is often the case in SOC-focused work. Our work begins to unravel the complex interactions among cover-type characteristics, soil properties, and hydrologic settings in SOC dynamics. We used Wilcoxon rank sum tests to determine if differences between aspen and spruce concentrations of SOC, DOC, total soil nitrogen and nitrate, and ratios of DOC to SOC were significant. We used linear mixed effects (LME) methods via the R package lme4 (Bates et al., 2015) to assess the influence of vegetation type, depth, and their interaction (N=5) on soil abiotic conditions, various forms of soil nitrogen and C and δ15N, ASD, and root abundances. We tested whether variables were normally distributed using the Shapiro–Wilk test and transformed the data to achieve a normal distribution if they were non-normal. The soil chemical properties of SOC, DOC, DOC:SOC, and ECEC were log transformed, while C:N data were transformed with the function x1/3. Root fractions and soil solution pH did not require transformation to meet model assumptions. We assessed whether vegetation type exerted a meaningful influence on the previously mentioned variables by constructing four models. The two simplest models included only vegetation type or depth, both as fixed effects. A third model included those fixed effects additively (e.g., vegetation + depth), and a fourth model included their interaction. We resolved the lack of independence of soil depth within each pedon by incorporating site identifiers as a random effect term in the model. We then tested the normality of the model residuals using the Shapiro–Wilk test. For all models that passed this test, we compared the model fits using analysis of variance and visually examined model residual errors for homogeneity of variance; the best model fit was selected based on the lowest Akaike information criterion following Hauser et al. (2020). We interpret the results of these LME models conservatively, given the low number of replicate sites for each land cover type. We could not perform an LME model on microbial biomass and enzyme data due to the relatively limited number of samples. This limited our ability to include vegetation*depth interactions in those models.

Clay, silt, and sand content at the aspen sites (AS and EAG; Fig. S1 in the Supplement) and one of the conifer sites (ESG) exhibited little variation with depth (average 33.1 % clay and 18.8 % sand), while the two other conifer sites had a greater sand and lower silt and clay content, particularly at depths greater than 25 cm (SS and SG; Fig. S1). Cation exchange capacity was similar among the aspen and conifer sites, with averages of 7.2 ± 4.9 and 8.2 ± 6.8 (meqper100gsoil), respectively, with elevated values at the surface that declined with depth (Fig. S2).

We were able to access and describe soil profiles to approximately 100 cm (Fig. , Table S1 in the Supplement). All sites had weak to moderately strong subangular blocky structure throughout the soil profile, and most sites had weak to moderately strong granular structure in A and upper B horizons. Dendritic tubular pores, interpreted to be abandoned root channels, were present throughout the soil profile of aspen sites, while they were less common in the conifer soil profiles. Aspen sites exhibited faint organic stains and organoargillans (i.e., dark, organic stained clay films) throughout the soil profile, while conifer sites had clay bridges and krotovina throughout the soil profile. The krotovina suggest greater bioturbation under conifer than aspen. Both vegetation types exhibited ferriargillans (i.e., clay coats that include Fe oxides), clay films, and charcoal, although ferriargillans and clay films were more prominent under conifer. Clay bridges, organoargillans, and ferriargillans indicate illuviation. Lithologic discontinuities were identified in SS, ESG, and EAG, indicating colluvial inputs into these footslope pedons.

Soils at both aspen sites (AS and EAG) are Ustic Haplocryolls with thick, SOC-enriched surface horizons (mollic epipedons) and showing evidence of incipient subsoil development in the form of moderately thick cambic horizons. Soils under conifer sites are Typic Haplocryepts (SS and SG) and Eutric Haplocryalfs (ESG). Although surface horizons under conifer were not as well developed (ochric epipedons), the subsurface showed similar incipient pedogenesis in the form of cambic horizons for SS and SG and greater development in the case of ESG, where an argillic horizon was identified between 19 and 90 cm below the mineral surface.

To understand how soil abiotic conditions are linked to SOC forms and processing pathways, we focused our analysis of soil temperature and moisture during the growing season (June–August; Fig. ). As expected, soil temperature increased at all sites as the growing season progressed, peaking in mid to late July, with the warmest temperatures observed near the surface and lower variability observed at depth. We also observed that the aspen sites (AS and EAG), which are on south-facing slopes, are warmer than conifer sites with an average surface soil (15 cm deep) temperature of 14.3 ± 1.2 and 10.4 ± 1.0 °C, respectively, during the growing season. Aspen sites were generally drier than conifer. The average volumetric water content in the surface soils (15 cm deep) at aspen sites was 0.15 ± 0.05 and the average volumetric water content at spruce sites was 0.24 ± 0.05 cm³ cm⁻³.

We also examined soil pH. Across the entire soil profile, pH was similar at the aspen and spruce sites (5.6 ± 0.3 and 5.3 ± 0.4, respectively) but their depth trends differed with spruce soils having slightly more acidic pH near the surface compared to the aspen (Fig. S2). This trend reversed at approximately 60 cm, where the aspen soils became slightly more acidic compared to the conifer soils.

Across all sites, SOC concentrations ranged from 46.0–62.6 mgg-1 near the surface (5 cm deep) to 4.8 to 29.0 mgg-1 at depth (Fig. a). SOC concentrations were generally higher under aspen compared to spruce sites (p < 0.0001; Fig. a), but LME models also suggest that the best-fit model included a significant interaction between vegetation and depth (p < 0.001), suggesting that SOC declines with depth for both vegetation types but to a greater extent under spruce compared to aspen. Stocks of SOC for all depth intervals ranged between 0.01 and 1.31 kgm-2 (Fig. b); these values did not exhibit consistent declines with depth or clear differences across cover type.

In contrast to SOC concentrations, DOC was generally higher under the spruce stands compared to aspen. Similar to SOC, a model including a significant interaction between vegetation and depth was the best predictor of DOC values (p < 0.001), likely reflecting variable DOC values at different depths in both vegetation types (Fig. c). The DOC:SOC ratio also exhibited a significant interaction between vegetation type and depth (Fig. d; p = 0.0007). As with DOC, this significant interaction likely reflects variable ratio values for each cover type across depths.

Total soil nitrogen ranged from 0.2 mgg-1 at depth to 4.63 mgg-1 near the surface. A model including an interaction between vegetation type and depth was the best fit (p = 0.003; Fig. a). Aspen values were greater than those in spruce-dominated soils at all depths; the significant interaction implies that the decline with depth was greater in spruce soils. Nitrate concentrations averaged 214 ± 323 ngg-1 near the surface and 2.3 ± 3.8 ngg-1 at depth. Nitrate was elevated under the aspen compared to the spruce sites (Fig. b), and values under both vegetation types varied with depth. A model that included both depth and vegetation type with no interaction was a meaningfully better predictor of nitrate concentrations than either depth or vegetation alone (p = 0.035), and including a depth–vegetation interaction did not improve model fit. Aspen soil C:N averaged 10.9 ± 1.1 and remained fairly constant with depth (Fig. c). The spruce sites showed greater variation with depth with a similar mean value of 19.3 in the top 20 cm but widely variable values at the deepest points, ranging from 4.6 to 28.7 (Fig. c). Including the interaction between vegetation type and soil depth improved model fit (p = 0.0008), indicating that C:N varied more with depth in spruce soils than in aspen soils, where values stayed fairly constant. The lowest C:N value, found at depth in one of the spruce forest, suggests that the soil organic matter there has been heavily processed by microbes (Ziegler et al., 2017).

δ15N signatures showed less distinct depth trends compared to the total nitrogen and nitrate, mirroring the relative lack of clear depth trends in C:N. Though variation across sites limited our ability to find statistically significant differences across vegetation types or a significant influence of depth, δ15N of soil organic matter in spruce plots tended to be lower than that of aspen (Fig. d), hinting that soil N has undergone more microbial processing (Nadelhoffer and Fry 1988; Billings and Richter 2006) under aspens compared to under conifers. This interpretation is consistent with the mean spruce C:N values being greater than those in aspen forests (Fig. c).

The mean geometric diameter of soil aggregates was generally smaller under aspen compared to spruce (Fig. a). Aspen aggregates tended to be finer, with fewer intermediate and coarse aggregates compared to the spruce soil (Fig. b–d). LME models indicated that vegetation–depth interactions were the most meaningful in driving all three aggregate size classes (p < 0.001). To explore the propensity of SOC to promote and stabilize aggregation in these soils, we standardized aggregates by SOC concentrations (Fig. ; Souza et al., 2023). LME models indicate that an interaction between vegetation type and depth contributes to variation in the fine aggregate:SOC data (Fig. a; p < 0.0001), suggesting that in the bottom third of the profile, SOC in spruce soils also contributes to fine aggregate formation. Adding an interaction between vegetation and depth to the model improved model fit for coarse aggregate:SOC values (p < 0.0001), highlighting a higher coarse aggregate:SOC ratio in the spruce sites that varied little with depth compared to aspen soils (Fig. b) and suggesting that SOC in spruce-dominated soils exhibits a higher propensity to form coarse aggregates throughout the soil profile.

Our data provide multiple lines of evidence that SOC protection, and thus C fate, in these montane forests is largely controlled by biotic processes linked to soil mineral material. Here, greater values of total N, nitrate, BGase, and δ15N, and lower C:N under aspen compared to spruce (Figs.  and ), suggest a greater degree of microbially processed organic matter under the aspen stands, where greater SOC contents were measured (Fig. ). These data hint that the microbial community under aspen stands functions in a manner consistent with the MEMS framework (Cotrufo et al., 2013), transforming relatively labile leaf litter (e.g., under aspen) into byproducts more readily stabilized within soil profiles to a greater extent than appears to occur with slower-turnover litterfall (e.g., spruce). Differences in litterfall composition and thus decay rates across aspen and conifer species have been widely reported, with generally lower lignin and higher nitrogen content in aspen litter (Moore et al., 2006). Our inference about litterfall differences promoting microbial byproduct stabilization is consistent with findings from across western Canada, where investigators observe relatively more active microbial communities under aspen compared to paired spruce stands throughout a growing season (Norris et al., 2016). Specifically, one interpretation of these C:N, δ15N, exo-enzyme, SOC, and DOC data at our sites is that tree species-specific composition of litterfall appears to have prompted greater microbial activity (Fig. ), likely promoting greater contributions of microbial necromass to the SOC pool. This, in turn, may promote greater SOC retention in aspen-dominated soils; though investigation of specific necromass-derived compounds in these soils (e.g., Liang et al., 2019) is beyond the scope of this work, it represents a valuable way forward to test this inference.

The smaller aggregate sizes in aspen-dominated soils further support the notion that SOC stability is enhanced by higher microbial activity and increased necromass production rates. SOC is better protected and has generally longer mean residence times in smaller aggregates than larger aggregates (Six and Jastrow, 2002; Six et al., 2004). Literature hints that the larger-size aggregates (Fig. c) and greater propensity for C to form large aggregates (Fig. b) observed in the spruce-dominated soils at our sites may be due to a greater abundance of particulate organic matter in spruce compared to aspen forest soils (Cotrufo et al., 2015, 2019); this may be the case since spruce litterfall is more difficult to decompose. Taken together, these lines of evidence are consistent with aspen-dominated forests harboring SOC pools that tend to promote relatively small aggregate formation that can preserve SOC to a greater extent – especially at depth, where MAOC tends to dominate SOC pools (Jackson et al., 2017).

Our soil data also indicate that SOC pools in aspen soils are more strongly dominated by MAOC compared to those in spruce soils. The greater abundances of smaller aggregates and total soil nitrogen and nitrate concentrations, and lower C:N ratios (Fig. a–c), in aspen compared to conifer soils are consistent with relatively greater MAOC than POC concentrations (Kögel-Knabner et al., 2008; Ye et al., 2018; Sokol et al., 2022). Combined with the lower DOC:SOC ratio in aspen-dominated soils, these data suggest that a greater fraction of SOC in aspen-dominated soils is mineral-bound and relatively difficult to transform into microbially available pools of DOC. We interpret these data to suggest that microbially mediated transformations of SOC promote differences in the abundance of MAOC and the physical structure of soil aggregates that lead to differences in the SOC protection.

Roots can influence SOC stability through their promotion of both physical and chemical protection (Sullivan et al., 2022b). Specifically, roots can play an important role in the formation and breakdown of soil aggregates (Oades, 1984; Singer et al., 1992; Le Bissonnais, 1996; Attou et al., 1998), they can create biopores that can support the transport of DOC in deeper soil layers (Angers and Caron, 1998; Zhang et al., 2015; Lucas et al., 2019), and root exudates can prime microbial activity, enhance decomposition, and support the formation of MAOC (Jilling et al., 2021; Fossum et al., 2022). Our data revealed little direct correspondence of root abundance with SOC. However, per unit root abundance, spruce soils appear to harbor more DOC compared to aspen (Fig. ). This pattern – especially evident in total and fine root abundance – suggests that DOC moves more readily through spruce soil profiles, potentially leading to greater DOC losses to stream water compared to aspen-dominated soils. A complementary explanation would be that there are differences in the amount of DOC exudation by roots between the two species, and indeed such differences in exudation rates have been hypothesized in the literature (Buck and St. Clair, 2012; Boča et al., 2020). We might expect that greater exudation would lead to a greater increase in the MAOC pool and enhanced C stability (Even and Cotrufo, 2024), which could explain the lower values of DOC relative to SOC observed under aspen.

South-facing slopes tend to be warmer and drier than north-facing slopes in the Northern Hemisphere (Burnett et al., 2008), and thus they can prompt more microbial decomposition of SOC. In our study aspen cover occurs where soil temperatures are warmer (Fig. ). As such, it is possible that the exo-enzymatic signals of generally greater microbial activity in aspen-dominated soils compared to spruce-dominated soils (Fig. ) are prompted more by enhanced soil temperatures than by differences in aspen and spruce organic matter characteristics, and that enhanced soil temperatures also contribute to smaller soil aggregates, perhaps also due to greater microbial activity. We note that the volumetric fraction of soil moisture was also lower in the aspen, particularly at the shallowest soils, but that aspen soils appear to stay sufficiently moist (0.10–0.20 under aspen vs. 0.20–0.30 under the spruce) to support microbial activity responses to the higher temperatures. Consistent with this idea, soil CO2 and O2 concentrations generally suggest that microbial activities in the warmer, aspen-dominated soils are greater than in the cooler, spruce-dominated soils. Cooler, wetter conditions of the spruce-dominated soils, particularly following snow melt, may prompt a deeper infiltration of moisture and DOC down the soil profile, leading to the elevated DOC/root biomass observed under the spruce stands. While disentangling the impact of elevated soil temperatures from that of the chemical composition of organic inputs from aspen trees within the soil profile is difficult, soil nitrogen and δ15N data are consistent with the idea that litterfall chemistry, and not just temperatures, promoted greater microbial activities in the aspen-dominated soils. We suggest that investigating the comprehensive, integrated effects of warmer, aspen-dominated sites on SOC dynamics compared to cooler, spruce-dominated sites offers a straightforward approach to assessing landscape-scale transitions in watershed C dynamics.