Discussion:
Collectively, this work identifies how roots and soil microbes drive SOM loss and formation in miscanthus systems that can promote soil C sequestration and support plant productivity. Root ingrowth did not promote a net litter C loss from soil (Fig. 2 ) despite increased light POM decomposition due to the enhanced transfer of C into heavy POM (Fig. 3 ). Notably, we document the potential for roots to mobilize litter-derived N from POM without priming litter C loss (Fig. 2 , Fig. 4 ). We also identified that microbial nutrient or carbon limitation may alter how microbes grow and decompose litter-derived SOM, with more litter decomposition and less MAOM formation from litter in organically fertilized soils (Fig.5 ).
It appears that miscanthus roots can mine N from litter without stimulating corresponding litter C losses (Fig. 2 ) and can increase the C:N of litter-derived light and heavy POM (Fig.4 ). This raises the question of how miscanthus accesses N from decomposing litter without priming C losses that are commonly observed in other ecosystems (Cheng et al., 2014; Zhu et al., 2014). One plausible mechanism may be that miscanthus roots engineer their rhizosphere microbiome composition or function to preferentially decompose N-rich litter compounds like proteins, potentially by stimulating proteolytic enzyme production (Brzostek & Finzi, 2011). While the specific mechanism remains uncertain, preferential N mining from litter has important implications for miscanthus sustainability (e.g., the propensity of miscanthus to be high yielding and build soil C). The resulting increase in remaining litter C:N may make new litter-derived SOM even more resistant to further decomposition. In addition, there has been a long-standing question of how miscanthus can maintain relatively high yields with limited N inputs (Cadoux et al., 2012). Previous research has posited that high nutrient use efficiency (Beale & Long, 1997) or the promotion of N-fixing symbionts (Davis et al., 2010) sustains N nutrition by miscanthus. Overall, our results suggest that miscanthus may also meet its N nutrition by effectively mining N from litter and SOM.
Our research suggests that roots can actively support the transfer of litter derived C into more protected forms. We observed that the priming of litter decomposition from light POM was balanced by litter C incorporation in heavy POM (Fig. 3 ). The composition of heavy POM is not as well-characterized as light POM or MAOM, but this pool is commonly assumed to be composed of stable soil macro- or micro-aggregates (Lavallee et al., 2020). Aggregate occluded SOM is largely formed through root and mycorrhizal symbiont activity (Rillig & Mummey, 2006) and often consists of partially decomposed plant and microbial organic matter fragments. This pool has a higher activation energy for decomposition than low C:N compounds like those in MAOM (Williams et al., 2018) and is more protected from decomposers than free light POM (Keiluweit et al., 2017; Kögel-Knabner et al., 2008). As such, there is an opportunity to build soil carbon in high C:N, heavy POM rather than lower C:N MAOM. The N requirements of low C:N SOM retention have often been cited as a criticism to efforts to use soil C management to mitigate global change (Schlesinger & Amundson, 2019). Future research efforts that investigate how roots can build new, persistent, and high C:N SOM could help realize the potential of soil C sequestration to combat climate change.
We found that the organic fertilizer treatments had the greatest microbial biomass and litter-derived light POM decomposition, in support of our second fertilization hypothesis, but less litter C and N were incorporated into MAOM (Fig. 5 , SI Figs. 4, 5 ). On one hand, differences between fertilization treatments could arise from a shift in the microbial community structure or function with organic fertilization (Pan et al., 2014). However, other research at the site has found no significant effects of nutrient treatment on microbial diversity or mycorrhizal abundance between treatments (Kane et al. 2023,in review ). On the other hand, C vs. N limitation over microbial decomposition can regulate the rate and efficiency of SOM cycling (Averill & Waring, 2018; Schimel & Weintraub, 2003). As organic fertilization deposits both C and N, our observations could be explained by the alleviation of C limitation and induction of N limitation. In support, we observed a reduction in nitrification rates with organic fertilization relative to unfertilized plots (SI Fig. 6 ) and other research found that organic fertilization increases plot-scale microbial respiration (Kane et al., 2023, in review ). Here, microbial decomposers could increase decomposition and growth while respiring excess C and immobilizing N in living biomass rather than forming more microbially-derived MAOM (Schimel & Weintraub, 2003).
While our experiment identified several important ways living roots and soil microbes control litter decomposition and SOM formation, some mechanisms may not have been fully captured. Our experiment was designed to separate the effects of roots vs. mycorrhizal fungi on litter C and N transformations, but our data only identifies a root effect despite the presence of mycorrhizal fungal symbionts (SI Fig. 7 ). The lack of differences between fungal ingrowth and total exclusion cores could be linked to the greater dependence of AM plants on root than hyphal foraging for nutrient uptake (Chen et al., 2016). As such, our experiment may not have isolated fungal effects on litter decomposition and SOM formation. Future efforts should quantify mycorrhizal fungal ingrowth to better investigate the contribution of symbiotic fungi to root-driven SOM transformations. In addition, our observations that fertilization did not impact root biomass (SI Fig 4a ) and that there was no significant interaction between fertilization and ingrowth treatments (SI Table 2 ) do not support our first fertilization hypothesis that roots would have the greatest effect in unfertilized soils. While miscanthus root systems do not always respond to fertilization treatments(Amougou et al., 2011), this pattern may have been driven by the stand age of miscanthus in our experiment. These plots were in the third year of growth whereas older, more nutrient limited stands exhibit greater differences in root C allocation and N acquisition (Kantola et al., 2022). As such, future efforts to investigate how nutrient availability alters living root impacts on SOM formation should leverage ecosystems with longer-term fertilization history. Despite these limitations, our data has identified several important mechanisms of SOM formation in situ and provides the foundation for future efforts to study how living roots and fungi alter SOM dynamics with more sophisticated measurements, under different environmental conditions, or across different ecosystems and plant-microbe interactions.
This work has expanded our mechanistic understanding of how living roots shape ecosystem processes in agricultural systems. Our finding that miscanthus roots can simultaneously prime N release from litter without an additional C release and transfer C into a more persistent form of SOM has important implications for the sustainability of bioenergy production as well as the viability of restorative agricultural to offset carbon emissions. Overall, our work suggests that living roots can selectively mine N while sequestering soil C. This knowledge can help improve the predictive understanding of SOM cycling that is critical to meeting the goals of restorative agriculture.