**4. Discussion**

In this study, ASE and AUR were the main organic C fractions for all species, which is consistent with the study results of Yang et al. [29]. Consistent with our hypothesis, a large amount of labile O-alkyl C was detected in the newly shed foliar litter, so that it could be sufficiently consumed by a large microbial population in the subsequent decomposition process, which has been commonly reported [3,11]. Moreover, the functional C group contents in newly shed foliar litter varied with the plant species, and labile O-alkyl C was mostly detected in the fir, spruce and willow litter materials, whereas rosa litter contained abundant recalcitrant aliphatic C. Therefore, this result suggested that the initial C quality might be species-dependent, which is consistent with our hypothesis. This might occur because rosa litter could highly protect the leaf surface from leaching as a result of the abundant recalcitrant aliphatic C and higher hydrophobicity indices (Table 1) [11,18], while the initial C fractions in fir, spruce and willow litter might potentially contribute to C turnover in alpine forests. Moreover, in accordance with the results of other studies on litter decomposition, the initial C quality in foliar litter was controlled by DOC, N and Mn (Figure 5) [1,4,6], which illustrated the key role of the litter quality in the formation and preservation of litter-derived C in soils. Methoxyl C, aromatic C and phenolic C in newly shed foliar litter significantly regulated C loss during litter decomposition, which indicated that the variation in the C quality of newly shed litter was considered a critical factor influencing litter decomposition and greatly contributed to soil C sequestration in alpine forests.

The plant species significantly regulated the initial C quality of newly shed foliar litter. In our study, the initial C quality of newly shed fir, spruce and willow foliar litter was dominated by labile carbohydrates because strong signals were detected in the O-alkyl C regions [35,42]. Furthermore, previous studies found that the rapid decrease in O-alkyl C attributed to polysaccharide materials at the initial decomposition stage generally occurred due to the preferential degradation of labile cellulosic compounds [2,43]. Hence, the result whereby O-alkyl C release in rosa litter was higher than that in fir, spruce and willow litter in this study might mostly occur because carbohydrates in O-alkyl C in rosa litter might be easier to degrade than those in fir, spruce and willow litter. Moreover, xylose is considered the dominant hemicellulose monosaccharide in broad-leaved tissues, and its degradation occurs at least one year after broad-leaved litter decomposition. Nevertheless, arabinose and galactose, the main dominant hemicellulose monosaccharides in coniferous tissues, immediately degrade upon litter formation [27,44]. Therefore, the ASE content in rosa and willow litter was higher than that in fir and spruce litter. Our study also found that the aliphatic C content and A/O-A and HB/HI ratios in rosa litter were higher than those in the litter of the other species, suggesting that the C structure in rosa litter was mainly assigned to aliphatic C and was more complex than that in the litter of the other species [20,45]. The results are consistent with the studies of Hishinuma et al. [2] and De Marco et al. [26] but inconsistent with the studies of Ono et al. [31] and Ono et al. [25], which could be explained by the variations in biosynthetic lipid processes among plant species. In addition, the less complex chemical structure determined by the indices of A/O-A and HB/HI in coniferous species versus broadleaved species might also be attributable to the plant species [45,46]. Thus, rosa litter might be more resistant to decomposition. In addition, the abundance of guaiacyl lignins and stable tannins in coniferous fir and spruce and broadleaf willow and rosa might lead to differences in the proportions of aromatic C and phenolic C among the various litter types [25,45]. In general, the contents of ASE and aliphatic C might be the key C quality factors controlling the chemical structure of newly shed foliar litter, particularly that of broad-leaved rosa in this study.

**Figure 5.** Results of stepwise regression analysis of the functional C groups ((**a**): aliphatic C, (**b**): methoxyl C, (**c**): O-alkyl C, (**d**): di-O-alkyl C, (**e**): aromatic C, (**f**): phenolic C, and (**g**): carboxyl (C)) using the litter quality (C: carbon; DOC: dissolved organic carbon; N: nitrogen; P: phosphorus; Mn: manganese; WSE: water-soluble extractives; OSE: organic-soluble extractives; ASE: acid-soluble extractives; AUR: acid-unhydrolyzable residues). The dominant variables are emphasized in red in the regression model. The values represent standardized coefficients, and the \* indicates significant effects of the variables on the functional C fractions in a given regression model \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001. For example, aliphatic C was significantly impacted by P (*p* < 0.05), DOC (*p* < 0.01) and N (*p* < 0.001) in the stepwise regression testing the effects of litter chemical quality on aliphatic C (**a**).

The litter quality also dramatically determined the C quality of newly shed foliar litter, and this might lead to seasonal variations in the C quality together with environmental conditions. Our analysis results showed that the N content also dominated the recalcitrant C quality of newly shed foliar litter (Figure 5), which is consistent with the decomposition process. This might be due to foliar litter releasing a large amount of nutrients, such as N, at the early stage of decomposition under the notable influences of the litter quality and environmental conditions and then further stimulating microbial activity that might easily accelerate the degradation of recalcitrant substances [47–49]. This might also be the reason that stable aromatic C and phenolic C and low ratios of A/O-A and HB/HI in willow and rosa litter were detected in November. The insulating effect of snow cover at our study site during this period might promote microbial activity, thus resulting in fast decomposition [41,50]. Similarly, soluble nutrients, such as DOC, were rapidly released at the early stage of decomposition. In our study, DOC was the most significant factor influencing WSE and O-alkyl C and strongly responded to environmental conditions (Figure 5); therefore, a large amount of easily available compounds may be formed as a result of the thick litter layer and could be easily leached from foliar litter by light

snowfall in October and November of this region, thus leading to more labile WSE and O-alkyl C released from foliar litter [51]. Moreover, our analysis results showed that DOC significantly affected the recalcitrant AUR (Figure S1), aliphatic C and aromatic C, which might have resulted from DOC derived from the oxidative degradation of lignin. Watersoluble components primarily controlled the formation and persistence of recalcitrant substances in forest soils by significantly affecting microbial activity [52,53]. This might be the main reason why the temporal characteristics of the aliphatic C and AUR fraction were opposite to those of WSE and O-alkyl C during the periods of October and November. In addition, the initial Mn generally functioned as a cofactor of lignin-degrading enzymes to regulate litter decomposition and humus stability, revealing a high contribution to the control of C sequestration in the organic layer [4,53]. The significant correlations between Mn and C quality, especially O-alkyl C and O-alkyl C, in newly shed foliar litter in our study might also indicate that the initial Mn might potentially contribute to C sequestration in soils.

In accordance with our hypothesis (2), the initial C quality notably controlled litter decomposition processes and could further affect C sequestration in soils. In combination with the litter decomposition experiment of Ni et al. [41], we found that the C loss in the 4-year decomposition process was highly related to the initial aromatic C and phenolic C, which are commonly referred to as structurally complex guaiacyl- and syringyl-lignin, respectively. This result notably illustrated that the initial recalcitrant C quality mainly determined C sequestration in soils in the decomposition process, which is contrary to the results of recent studies whereby labile components were the main factor influencing the decomposition process [18,20]. Nevertheless, the initial methoxyl C, which is also commonly reported as the aromatic rings of guaiacyl and syringyl units in lignin and could be structurally stable [11,36], was positively related to the C loss of decomposing litter in our study. This might be due to the region of initial methoxyl C also arising from other sources including proteins and amino acids [14].
