Data sources
We searched for published articles reporting lignin phenols and amino sugars in forest soils using Web of Science (http://webofknowledge.com) and China National Knowledge Infrastructure (https://www.cnki.net). Searched phrases included combinations of the terms ‘lignin phenol’ or ‘amino sugar’ or ‘microbial residues’ or ‘microbial necromass’, and ‘forest’. We also screened previous meta-analyses on the relevant topics such as those by Liang et al.13 and Ni et al.44 that focused on microbial necromass only. We carefully checked all results published before December 2021, and collected the data from tables or extracted from figures using the GetData Graph Digitizer (version 2.26, http://www.getdata-graph-digitizer.com).
In order to reduce deviations among different studies, the following rules were followed during data collection. First, we only included data collected using similar extraction methods to ensure compatibility. For example, lignin phenols were detected by cupric oxide (CuO) oxidation using gas chromatography-mass spectroscopy (GC-MS) analysis45. CuO-oxidation products were composed of vanillyl (V) units, syringyl (S) units and cinnamyl (C) units25. Amino sugars were hydrolysed following the method by Zhang and Amelung43. Quantifiable amino sugars mainly included glucosamine, galactosamine, mannosamine and muramic acid. Glucosamine came primarily from fungal cell walls, while muramic acid was considered to come only from bacterial cell wall46. Second, we only included data from natural forests and excluded indoor culture experiments. For studies that included the experimental treatments such as N addition, warming, rainfall changes and others, we only extracted the data from the control. Third, we only used data from surface mineral soils (0-20cm) or A horizon but excluded those from organic horizons. Forth, where multiple samples were taken from the same stand at different times, we averaged the values for the different sampling time to avoid pseudo replication. Finally, a total of 85 articles were compiled in a global dataset of 126 individual sties for lignin phenols and 137 individual sites for amino sugars in forest soils (Fig. 1).
In order to better understand the effects of forest types (coniferous forest, mixed forest, and broad-leaved forest) on lignin phenols and amino sugars, we recorded forest types or dominant tree species for classification. We also collected additional site-specific information, mainly location (place name, latitude and longitude), climate factor data [mean annual temperature (MAT), mean annual precipitation (MAP), and aridity index], NPP (net primary production) and soil physical-chemical properties (soil C/N ratio, soil pH and clay content). When the environmental variables were not available in the articles, we used data from public databases of global environmental variables (Table S1).
We quantified the contents of lignin phenols and amino sugars based on their respective SOC-normalized values (mg/g SOC) to indicate the contribution to SOC. The total contents of lignin phenols and amino sugars are calculated by the sum of V, S and C monomers and the sum of glucosamine (GluN, the biomarker of fungal necromass), galactosamine, mannosamine and muramic acid (MurA, the biomarker of bacterial necromass) monomers, respectively. Moreover, we calculated lignin monomer ratios (S/V, C/V, (Ad/Al)v and (Ad/Al)s) to indicate the oxidation and degradation of lignin in soils10. Some studies only reported fungal and bacterial necromass C, which we converted to the contents of glucosamine (mg/g soil) and muramic acid (mg/g soil) using the following formulas:
Bacterial residual C (mg/g soil) = Muramic acid (mg/g soil) ×45
Fungal residual C (mg/g soil) = (Glucosamine (mg/g soil) / 179.17–2 × Muramic acid (mg/g soil) / 251.23) × 179.19 × 9
Statistical analysis
Before the data analysis, all data and residuals were tested for normality. Differences of lignin phenols and amino sugars contents among the three forest types were examined using one-way ANOVA for normally distributed data or Kruskal-Wallis test for non-normally distributed data. The significance of the relationships between lignin phenols or amino sugars and SOC or environmental parameters was tested using Pearson correlation coefficient. If data was nonnormally distributed, Spearman correlation was used instead. Multiple regression analyses with stepwise variable selection method were performed to select an optimal model though AIC values to explain the relationships of MAT, MAP, aridity index, NPP, soil pH, soil C/N and clay content with soil lignin phenols and amino sugars, respectively. Moreover, we further calculated the relative importance of predictor variables to analyze the key control factors. These statistical analyses were performed using the ‘stats’, ‘relaimpo’, ‘ggplot2’ and ‘ggpubr’ packages of R version 4.0.5.