Ecosystem‐Dependent Responses of Microbial Necromass Carbon to Warming Are Governed by the Balance Between Necromass Formation and Decomposition

ABSTRACT

Microbial necromass carbon (MNC) is increasingly recognized as a major contributor to persistent soil organic carbon (SOC), yet its response to climate warming and the underlying regulatory mechanisms remain poorly understood. Here, we conducted a global meta‐analysis to assess how total MNC, fungal necromass carbon (FNC), and bacterial necromass carbon (BNC) respond to warming and to identify the key drivers. Overall, warming had no significant net effect on total MNC, FNC, or BNC across all observations, although publication bias‐corrected analyses revealed a significant 8.6% increase in FNC. However, warming effects varied substantially among ecosystems, increasing MNC accumulation in permafrost (+25.4%), grassland (+8.2%), and cropland (+9.9%) soils, while decreasing it in forests (−12.4%) and showing no significant effect in wetlands. Warming effects were further influenced by warming method, soil depth, climatic conditions, and initial soil properties. Meta‐regression analyses showed that warming‐induced changes in microbial biomass were the strongest predictor of MNC responses, highlighting the central role of microbial growth and turnover in regulating necromass dynamics. Moreover, shifts in nutrient availability, soil pH, and extracellular enzyme activities significantly influenced the balance between necromass production and decomposition. Positive coupling between MNC and SOC responses suggests that microbial necromass formation represents an important mechanism linking microbial processes to soil carbon persistence under warming. Overall, our findings demonstrate that warming affects MNC by altering microbial traits and nutrient availability, thereby regulating the balance between necromass production and decomposition. Ecosystem‐specific conditions further determine the magnitude and direction of these responses. These findings highlight the need to incorporate microbial necromass dynamics into predictions of soil carbon‐climate feedbacks under future warming scenarios.

Warming alters microbial necromass carbon across ecosystems by shifting microbial traits, nutrient availability, and enzyme activities, thereby regulating the balance between necromass formation and decomposition.

1. Introduction

Soil organic carbon (SOC) represents the largest terrestrial carbon pool, storing approximately three times more carbon than the combined vegetation and atmospheric carbon pools (Simpson et al. 2007; Schmidt et al. 2011). Even minor fluctuations in SOC storage can elevate atmospheric CO₂ concentrations and trigger positive climate feedbacks (Fontaine et al. 2007; Melillo et al. 2017). In turn, climate warming alters soil carbon sequestration and its stability, creating bidirectional interactions between SOC dynamics and climate change. Under a high greenhouse gas emission scenario, global mean temperature is projected to rise by 4.4°C by the end of the century relative to pre‐industrial levels (IPCC 2023), with profound implications for soil carbon dynamics (Mora et al. 2013). Therefore, it is imperative to elucidate the mechanisms linking soil warming to the carbon cycle at the global scale, which is crucial for accurately evaluating soil‐atmosphere climate feedbacks under future warming (Walker et al. 2018; Nottingham et al. 2022).

Traditional soil carbon sequestration theories have long regarded plant root and litter residues as the primary source of stable SOC (Kallenbach et al. 2016), while largely overlooking the critical contribution of microbial‐derived carbon (Liang et al. 2017). Accumulating evidence highlights that continuous microbial growth, turnover, and mortality are key processes maintaining long‐term SOC stability. Microbes decompose plant litter and soil organic matter through extracellular enzymatic hydrolysis and oxidative reactions (Lehmann et al. 2007; Liang et al. 2019; Tian et al. 2021; Zhu et al. 2020; Qin and Zhou 2025). Meanwhile, microbes assimilate plant‐derived substrates as energy sources to support microbial biomass growth and physiological maintenance. Upon microbial cell death, microbial residues accumulate in soils and form microbial necromass carbon (MNC) (Liang et al. 2017; Whalen et al. 2022). Increasing evidence has demonstrated that MNC accounts for 27%–82% of total SOC, dominating stable carbon formation and far exceeding the carbon contribution of living microbial biomass (Liang et al. 2015; Khan et al. 2016; Fan et al. 2021). As a core stable carbon fraction in soil, MNC plays a central role in maintaining terrestrial carbon sequestration (Liang et al. 2019; Hu et al. 2024). Nevertheless, the mechanisms underlying MNC formation and stabilization, as well as how its contribution to SOC stocks shifts under ongoing climate change, remain poorly understood.

Climate warming regulates the production, decomposition, and preservation of MNC by modulating soil microbial metabolism, plant litter inputs, soil physicochemical properties, and nutrient availability (Cai et al. 2023; Jia et al. 2022; Hu et al. 2023; Lyu et al. 2025). Microbial necromass is primarily generated through microbial growth, turnover, and cell death, with residues subsequently being stabilized through associations with soil minerals and aggregates (Liang et al. 2015; Hu et al. 2024). Conversely, necromass can be decomposed and recycled by living microorganisms through extracellular enzyme‐mediated depolymerization and reutilization, making its persistence dependent on the balance between formation and decomposition processes (Liu et al. 2021). Based on Arrhenius enzyme kinetics, warming stimulates soil microbial metabolic activity and accelerates plant residue degradation, thereby enhancing substrate assimilation and subsequent MNC formation (Davidson et al. 2011; Liu et al. 2025). These biogeochemical processes are further modulated by environmental constraints, including soil moisture and nutrient limitation (Zhao et al. 2022; Liu et al. 2025). From the perspective of microbial metabolic strategies, warming reshapes microbial traits by regulating microbial carbon use efficiency (CUE), growth turnover rate and community abundance (Tajmel et al. 2023; Tian et al. 2024; Wang, Ni, et al. 2026; Pan et al. 2026). This shifts microbial communities toward a high‐yield metabolic strategy and further promotes MNC accumulation. However, under nutrient‐limited conditions, microbes may decompose indigenous MNC to acquire nitrogen resources, resulting in substantial MNC turnover and carbon loss (Jia et al. 2017; Buckeridge et al. 2022; Hu et al. 2023; Ma et al. 2023). Ultimately, net MNC accumulation is determined by the dynamic balance between MNC production and decomposition. Notably, fungi and bacteria exhibit distinct functional traits and adaptive capacities to environmental changes and soil depth stratification, leading to divergent accumulation patterns of fungal and bacterial necromass carbon. Bacterial necromass carbon (BNC) is relatively labile and susceptible to decomposition and reuse under substrate limitation, whereas fungal necromass carbon (FNC) possesses higher biochemical stability and environmental resilience (He et al. 2011; Song et al. 2023). Warming tends to selectively enhance FNC or BNC accumulation, alter the FNC/BNC ratio (Cai et al. 2023; Tian et al. 2025), and further reshape SOC stability.

Despite growing research attention, MNC responses to experimental warming remain highly controversial, with warming reported to exert stimulatory (Cai et al. 2023; Chang et al. 2021), inhibitory (Ding et al. 2020; Zhao et al. 2022; Sun et al. 2025), or neutral effects (Tian et al. 2025). For example, warming below 2°C increased MNC in alpine meadows, whereas such positive effects gradually diminished under warming magnitudes above 2°C (Cai et al. 2023). In contrast, warming decreased MNC by 11.0% in temperate forest soils (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024) and by 21.8% in wetland soils (Sun et al. 2025), exacerbating soil microbial carbon loss. In other cases, warming showed no detectable impacts on topsoil MNC and FNC, but substantially increased BNC content by 81.3% (Tian et al. 2025). Such inconsistent findings reveal large uncertainties in ecosystem‐ and soil depth‐dependent MNC responses to variations in warming magnitude. These discrepancies likely arise because warming simultaneously regulates microbial necromass formation and decomposition through changes in microbial biomass and extracellular enzyme activities, while these processes are further modified by ecosystem type, soil depth, and experimental warming regimes (Qin et al. 2024; Liu, Lin, and Li 2024; Liu, Tian, et al. 2024; Wang, Zhao, et al. 2026). In addition, differences in methodological approaches for estimating microbial necromass among studies may further contribute to the reported variability (Hu et al. 2024). This uncertainty is particularly evident regarding the functional differentiation between FNC and BNC under climate warming. Collectively, these inconsistent findings limit our mechanistic understanding of global SOC responses to climate change.

To address these knowledge gaps, this study conducted a global meta‐analysis of field warming experiments. Using amino sugars as biomarkers and adopting the revised MNC conversion coefficients proposed by Hu et al. (2024), we quantified the effects of warming on MNC, FNC, and BNC. The objectives were to: (1) reveal the global‐scale divergent responses of MNC to experimental warming; and (2) identify key factors regulating MNC responses to experimental warming. We hypothesized that: (1) warming induces ecosystem‐dependent and heterogeneous responses of MNC across global soils; and (2) warming‐induced changes in microbial biomass and nutrient availability are primary regulators of MNC dynamics by governing the balance between necromass production and decomposition. This study aims to provide a mechanistic understanding of MNC responses to warming and offer valuable insights for accurately evaluating soil carbon sequestration potential across global terrestrial ecosystems.

2. Materials and Methods

2.1. Data Collection

Data were collected from peer‐reviewed articles published before May 7, 2026 on Web of Science (http://apps.webofknowledge.com/). Studies were included if they reported the effects of warming on MNC or amino sugar contents. The keywords used for literature search were “microbial necromass or fungal necromass or bacterial necromass or microbial residue* or fungal residue* or bacterial residue* or amino sugar*”, and “warming or elevated temperature” and “soil”. To avoid bias, all publications were screened according to the following criteria: (1) Contents of MNC, BNC and FNC were directly reported, or the contents of glucosamine (GlcN) and muramic (MurA) were available for the indirect calculation of BNC and FNC. (2) Only field in situ warming experiments were included, while laboratory incubation studies were excluded. (3) Each study contained paired control and warming treatments, with all plots established under consistent abiotic and biotic conditions. Finally, a total of 166 paired observations from 31 publications were compiled into the final dataset to explore the warming effects on MNC accumulation.

For each paper, raw data regarding MNC, FNC, BNC, GlcN, and MurA were either directly extracted from tables or digitized from figures with GetData Graph Digitizer (http://getdata‐graph‐digitizer.com). Mean values, standard deviations (SD), and sample replicates of all target variables were retrieved from eligible studies. For studies that only provided standard error (SE), SD values were back‐calculated using the formula SD = SE × $\sqrt{n}$. In a small number of studies, neither SD nor SE was reported. For these cases, the missing SD values were estimated as 1/10 of the corresponding mean values (He et al. 2024).

All MNC data in the dataset were uniformly recalculated using the revised conversion coefficients proposed by Hu et al. (2024). Among the 31 studies included in our dataset, 28 studies originally estimated MNC using conventional conversion coefficients, 2 studies had already adopted the revised conversion coefficients, and 1 study only reported amino sugar concentrations without converting them into MNC. The conventional conversion methods used in previous studies are provided in the Supporting Information. Specifically, MNC values originally derived from conventional conversion formulas were recalculated using the updated parameters. Correspondingly, FNC and BNC concentrations were quantified from soil GlcN and MurA contents following the revised calculation framework described by Hu et al. (2024):

$$\mathrm{BNC}=\text{MurA}\times 31.3$$

$$\mathrm{FNC}=\left(\frac{\text{GlcN}}{179.17}-1.63\times \frac{\text{MurA}}{251.23}\right)\times 179.17\times 10.8$$

where 179.17 and 251.23 are the molecular weights of GlcN and MurA, respectively. The revised molar ratio of bacterial cells is 1.63; the conversion factor from GlcN to FNC is 10.8, and the conversion factor from MurA to BNC is 31.3. Total MNC was calculated as the sum of BNC and FNC.

2.2. Experimental and Environmental Variables

Additional contextual information was also extracted from each study, including experimental duration, warming magnitude, warming methods, and geographic location (longitude, latitude, and altitude). We also collected climatic conditions including mean annual temperature (MAT) and mean annual precipitation (MAP), along with soil properties including soil moisture (SM), pH, SOC, total nitrogen (TN), total phosphorus (TP), C/N ratio, NO₃ ⁻–N, NH₄ ⁺–N, dissolved organic carbon (DOC), and soil available phosphorus (SVP). Microbial and enzymatic variables were also compiled, including the abundances of bacterial, fungal, and total microbial PLFAs, as well as the fungi to bacteria (F/B) ratio. Soil enzymes included β‐glucosidase (BG), N‐acetyl‐β‐glucosaminidase (NAG), acid phosphatase (AP), alkaline phosphatase (ALP), leucine aminopeptidase (LAP), cellobiohydrolase (CBH), β‐xylanase (XYL), and phenol oxidase (POX). To further explore the regulatory mechanisms of warming on MNC, all data were grouped based on multiple categorical variables, including ecosystem type, warming method, soil depth (topsoil, 0–20 cm; subsoil, > 20 cm), initial soil pH (acidic, < 6.6; neutral, 6.6–7.3; alkaline, > 7.3; Shakoor et al. 2021), and initial soil SOC content (low, < 15 g kg⁻¹; high, > 15 g kg⁻¹).

2.3. Data Analysis

The natural log response ratio (lnRR) was used to quantify the warming effects on MNC (Hedges et al. 1999). The lnRR was computed as follows:

$$\text{lnRR}=\ln \left(\frac{X_t}{X_c}\right)$$

where X _t and X _c represent the mean values of the target variable in the treatment and control groups, respectively. The variance of lnRR (v) was calculated as follows:

$$v=\frac{{S_c}^2}{{n_c{X}_c}^2}+\frac{{S_t}^2}{{n_t{X}_t}^2}$$

where S _c and S _t are the standard deviations, and n _c and n _t are the sample sizes for control and treatment, respectively. The weighted effect size (lnRR₊₊) was calculated as follows:

$$\ln {\mathrm{RR}}_{++}=\frac{\sum \left(w_i\times {\mathrm{RR}}_i\right)}{\sum w_i}$$

where w _i is the weight for each observation, and was calculated as:

$$w_i=\frac{1}{v_i}$$

The standard error of ln RR₊₊ and 95% confidence intervals (CIs) was calculated with the following equations:

$$\mathrm{s}\left(\ln {\mathrm{RR}}_{++}\right)=\sqrt{1/\sum w_i}$$

$$95\%\mathrm{CI}=\ln {\mathrm{RR}}_{++}\pm 1.96\times s\left(\ln {\mathrm{RR}}_{++}\right)$$

The metafor package was used to estimate the weighted effect size and 95% CIs (Viechtbauer 2010). A significant warming effect was identified when the 95% CIs did not cross zero. Funnel plot and Egger's regression test were performed to evaluate potential publication bias (Figure S1). A symmetrical funnel plot and a non‐significant Egger's regression test were considered indicative of limited publication bias. Trim‐and‐fill correction was therefore applied to evaluate the robustness of the overall results. When the adjusted effect sizes showed little difference from the original estimates after imputing potentially missing studies, the results were considered robust against potential publication bias. Results showed that the effect sizes of warming on MNC and BNC remained unchanged before and after correction, demonstrating that these findings were statistically robust and insensitive to potential publication bias. In contrast, the warming effect on FNC became statistically significant after adjustment (Table S1), indicating that the corrected result is more reliable. The glmulti package was employed for model selection to assess the relative importance of predictive variables regulating the warming response of MNC. The relative importance of each predictor was defined as the sum of Akaike weights across all models containing the given predictor. A threshold of 0.8 was adopted to distinguish essential predictors from non‐essential ones (Terrer et al. 2016). Meta‐regression was further conducted to explore the relationships between MNC responses and potential explanatory variables.

3. Results

3.1. Overall Effects of Warming on MNC

Among all individual effect sizes of field warming on total MNC, 22.9% were significantly positive, 15.7% were significantly negative, and 61.4% showed no statistical significance (Figure 1a). The distribution of effect sizes for FNC and BNC exhibited a similar pattern (Figure 1b,c). While the overall effect of warming on total MNC, FNC, and BNC was non‐significant across all observations, trim‐and‐fill correction for publication bias indicated a significant 8.6% (p < 0.01) increase in FNC under warming (Figure 1). In addition, considerable heterogeneity was detected among the observations, with Qt values of 9313.70, 744.72, and 1243.39, respectively (all p < 0.001). Such high heterogeneity necessitates further subgroup and meta‐regression analyses to identify the underlying influencing factors.

FIGURE 1.

Distribution of effect sizes of field warming on total MNC (a), FNC (b), and BNC (c). Yellow circles represent significantly positive effect sizes, blue circles indicate significantly negative effect sizes, and gray circles denote non‐significant effect sizes. Green circles represent the cumulative effect size of all observations.

3.2. Factors Affecting the Response of MNC to Warming

Subgroup analysis revealed that ecosystem type, warming method, soil depth, initial soil pH, and SOC content all modulated the effect of warming on MNC (Figure 2). Specifically, warming increased total MNC by 25.4% in permafrost, 8.2% in grassland, and 9.9% in cropland ecosystems. In contrast, warming reduced MNC by 12.4% in forest soils, while no significant responses were observed in wetland soils (Figure 2). In terms of warming methods, open top chambers (OTCs) and infrared heaters both promoted MNC, whereas heating cables suppressed it by 12.0%. In addition, warming promoted total MNC and FNC in subsoil, but exerted no significant effect on BNC in either topsoil or subsoil. Furthermore, the positive effect on MNC was more pronounced in soils with alkaline initial pH or low SOC content (Figure 2).

FIGURE 2.

Effects of warming on total MNC (a), FNC (b), and BNC (c) as related to ecosystem types, warming methods, soil depths, and initial soil pH and SOC. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon.

Meta‐regression analysis showed that MAT and MAP significantly modulated the response of MNC to warming (Figure S3). Specifically, the warming effects on both total MNC and FNC declined significantly with increasing MAT and MAP (all Q _m values were significant at p < 0.001). In contrast, BNC showed no significant response to MAT (Figure S3c), but decreased with increasing MAP (p < 0.001). Further meta‐regression identified strong spatial gradients in the responses of MNC to warming (Figure S4). Longitude showed significant positive relationships with the warming response of MNC, whereas latitude had no detectable effect. Altitude emerged as a strong predictor of warming effects, showing significant positive correlations with the responses of MNC (p < 0.001). In addition, neither warming magnitude nor experimental duration significantly modulated the warming responses of MNC (Figure S5).

3.3. Relationships Between Warming‐Induced Changes in Soil Variables and MNC

Meta‐regression showed that soil pH exhibited pronounced regulatory effects on MNC responses (Figure S7). The responses of total MNC and FNC both increased significantly with increasing pH (p < 0.001). In addition, soil moisture variation could not explain the alterations in MNC under warming (Figure S8). For soil nutrients, DOC responses were positively correlated with both total MNC and FNC responses (Figure 3a–c). Similarly, NO₃ ⁻ responses were positively correlated with total MNC and FNC responses (Figure S9). In contrast, NH₄ ⁺ responses showed a negative relationship with BNC responses. TN and SAP responses showed a positive correlation with BNC responses (Figure 3). Extracellular enzyme activities were also important modulators (Figure 4). BG responses were negatively correlated with BNC responses (p < 0.001), but not with total MNC or FNC. In contrast, NAG responses were significantly negatively correlated with total MNC, FNC, and BNC responses. AP responses showed a significant negative correlation only with BNC responses (p < 0.01).

FIGURE 3.

Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of DOC (a–c), TN (d–f), and SAP (g–i). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; DOC, dissolved organic carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon; SAP, soil available phosphorus; TN, total nitrogen. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

FIGURE 4.

Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of BG responses (a–c), NAG responses (d–f), and AP responses (g–i). The size of each circle represents the weight of the corresponding observations. AP, acid phosphatase; BG, β‐glucosidase; BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon; NAG, β‐1,4‐N‐acetyl‐glucosaminidase. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

3.4. Relationships Between Living Microbial Biomass, MNC, and SOC Responses Under Warming

Model selection analysis revealed that the responses of SOC and MBC were identified as the primary factors regulating total MNC responses to warming (Figure 5a). Similarly, the response of SOC was identified as a key driver of BNC variation under warming (Figure 5c). These findings suggest strong coupling among MBC, MNC, and SOC. Meta‐regression further revealed clear associations between MNC responses and shifts in MBC and SOC under warming. Significant positive relationships were observed between the response of MBC and MNC (Figure 5d–f), indicating that warming‐induced changes in MBC were strong predictors of shifts in MNC. More importantly, total MNC and BNC showed positive correlations with SOC responses (Figure 5; both p < 0.001). FNC responses showed only a marginally significant positive association with SOC (Q _m = 3.466, p = 0.06) (Figure 5h).

FIGURE 5.

Model‐averaged importance of the predictors for the effects of warming on total MNC (a), FNC (b), and BNC (c). Cutoff is set at 0.8 to differentiate among the most important predictors. Importance is based on the sum of Akaike weights obtained from the model selection using the modified Akaike information criteria. Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of MBC responses (d–f), and SOC responses (g–i). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MBC, microbial biomass carbon; MNC, microbial necromass carbon; SOC, soil organic carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

3.5. Effects of Warming on Plant Biomass, Soil Properties, Microbial Traits, and Enzyme Activity

Warming significantly increased BGB by 5.9%, whereas it exerted no notable influence on AGB (Figure S10). Overall, warming induced pronounced changes in soil properties. Specifically, warming significantly reduced soil moisture by 7.5% and decreased soil pH, TN, and TP. In contrast, DOC, NO₃ ⁻–N, and NH₄ ⁺–N contents were significantly elevated by 10.3%, 23.5%, and 18.4%, respectively (Figure S10). No significant response of SOC to experimental warming was observed (Figure S10). In addition, warming markedly increased the abundance of total microbial PLFAs. Meanwhile, soil extracellular enzyme activities including BG, LAP, ALP, and XYL were also significantly enhanced under warming (Figure S10).

4. Discussion

4.1. Warming Alters MNC Dynamics by Shifting the Balance Between Necromass Formation and Decomposition

The highly variable responses of MNC to warming across ecosystems suggest that warming does not exert a uniform control on necromass accumulation (Ding et al. 2020; Cai et al. 2023; Xiao et al. 2026). Instead, our results support a unifying conceptual framework in which MNC dynamics are governed by the balance between necromass formation and decomposition. Within this framework, warming simultaneously stimulates necromass formation and decomposition (Liang et al. 2019; Ding et al. 2020; Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). Consequently, the direction of MNC responses depends on the relative strength of these opposing processes. This framework provides a mechanistic explanation for the divergent ecosystem responses observed in this study. More importantly, because MNC constitutes a dominant component of persistent SOC (Zhou et al. 2023), shifts in this balance may represent a critical mechanism linking microbial responses to long‐term carbon persistence under warming (Lyu et al. 2025).

Microbial necromass formation is fundamentally driven by microbial growth and turnover (Zhou et al. 2023; Lyu et al. 2025). Warming frequently stimulates both AGB and BGB, thereby increasing carbon inputs to soils and enhancing substrate availability for microbial communities (Reyes‐Fox et al. 2014; Qin et al. 2024). Increased plant carbon allocation to roots and rhizosphere processes can further stimulate microbial growth and turnover, ultimately promoting necromass formation. Consistent with this mechanism, model selection and meta‐regression analyses identified microbial biomass responses as the strongest predictor of warming‐induced MNC changes (Figure 5). In addition, warming significantly increased total microbial PLFAs (Figure S10), suggesting enhanced microbial growth and turnover. In contrast, warming also stimulated extracellular enzyme activities associated with nutrient acquisition (Figure S10). Meta‐regression further revealed significant negative relationships between MNC responses and the responses of key hydrolytic enzymes (Figure 4), indicating that warming also accelerates MNC decomposition and recycling. Consequently, the net response of MNC depends on the relative strength of microbial growth‐driven necromass production and enzyme‐mediated necromass turnover (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024).

The differential responses of fungal and bacterial necromass after publication bias correction further support this framework. Fungal necromass contains relatively recalcitrant compounds such as chitin and melanized cell‐wall structures that confer greater biochemical stability and environmental persistence (He et al. 2011; Song et al. 2023; Yao et al. 2025). In contrast, bacterial necromass exhibits faster turnover and is more readily recycled under resource limitation. These differences also reflect contrasting ecological strategies under warming. Fungi generally adopt a resource‐acquisition strategy that facilitates the exploitation of complex substrates and supports necromass formation, while the biochemical stability of fungal residues promotes their long‐term persistence (Strickland and Rousk 2010; Liang et al. 2019; Fu et al. 2025). In contrast, bacteria typically follow a more opportunistic strategy characterized by rapid growth and turnover, resulting in faster necromass recycling and potentially limiting net BNC accumulation under warming (Liang et al. 2019; Wang and Kuzyakov 2024; Ullah et al. 2026). Therefore, the relatively stronger response of FNC may be attributed to the combined effects of enhanced fungal necromass formation and greater residue persistence, whereas rapid bacterial necromass turnover may constrain BNC accumulation. These differences likely explain the divergent responses of FNC and BNC observed among ecosystems and warming scenarios.

4.2. Microbial Traits and Nutrient Availability Regulate MNC Dynamics Under Warming

Microbial biomass emerged as the strongest predictor of warming‐induced MNC responses, suggesting that microbial traits play a central role in determining necromass accumulation (Hu et al. 2022; Zhou et al. 2024; Wu et al. 2025). Because microbial biomass reflects integrated microbial processes, including carbon allocation, carbon use efficiency, growth turnover, and community composition, increases in microbial biomass under warming are likely to enhance necromass production (Shao et al. 2021; Zeng et al. 2022). In addition, previous studies have shown that pH strongly influences microbial community composition, microbial carbon use efficiency, and energy allocation strategies (Rousk et al. 2010; Hill and Jones 2019; Shi et al. 2025). The positive relationship between warming‐induced pH shifts and MNC responses suggests that pH‐mediated changes in microbial physiology may indirectly regulate both necromass production and decomposition.

Nutrient availability further determines whether warming‐induced increases in microbial activity favor necromass production or decomposition. Carbon, nitrogen, and phosphorus are fundamental resources supporting microbial growth and biomass synthesis (Kuypers et al. 2018; Walton et al. 2023; Hemkemeyer et al. 2021). Accordingly, warming‐induced increases in DOC, TN, and SAP were positively associated with MNC responses (Figure 3), indicating that greater substrate and nutrient availability promote microbial growth and subsequent necromass production (Wang et al. 2021; Feng et al. 2023; Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). Under these conditions, enhanced microbial growth likely increases necromass inputs to soil, thereby contributing to greater MNC accumulation. In contrast, nutrient limitation may intensify microbial competition for resources and stimulate recycling of existing necromass pools (Whalen et al. 2022; Buckeridge et al. 2022), thereby accelerating decomposition and reducing net necromass accumulation. This interpretation is consistent with previous studies showing that nutrient constraints can shift microbial resource allocation from biomass production toward nutrient acquisition and necromass recycling (Ma et al. 2023; Hu et al. 2023). Additionally, warming‐induced changes in inorganic nitrogen availability influenced fungal and bacterial necromass differently. Positive relationships between FNC and NO₃ ⁻ responses and negative relationships between BNC and NH₄ ⁺ responses suggest that nitrogen transformations under warming may selectively favor FNC accumulation while constraining BNC. Such divergent responses likely reflect fundamental differences in nutrient acquisition strategies and resource‐use efficiency between fungi and bacteria.

Although warming significantly reduced soil moisture (Figure S10), moisture responses explained little variation in MNC responses across studies (Figure S8). This suggests that soil moisture primarily influences MNC indirectly through its effects on microbial activity, nutrient diffusion, and substrate accessibility rather than serving as a dominant direct control (Wang et al. 2022; Hu et al. 2023). Overall, these results indicate that microbial traits and nutrient availability jointly regulate microbial growth, residue turnover, and net necromass accumulation under warming.

4.3. Ecosystem‐Dependent Responses Reflect Differences in the Balance Between Necromass Formation and Decomposition

One of the most important findings of this study was the strong ecosystem dependence of MNC responses to warming. Warming increased MNC accumulation in permafrost, grassland, and cropland ecosystems, but reduced MNC in forests (Figure 2). These contrasting responses indicate that ecosystem‐dependent MNC dynamics are not driven by warming alone, but by differences in how warming alters the relative strength of necromass formation and decomposition across ecosystems.

In permafrost ecosystems, warming alleviates strong thermal constraints on plant productivity and microbial activity, thereby increasing substrate availability and stimulating microbial growth (Li et al. 2022). Under these conditions, warming‐induced increases in necromass production appear to exceed increases in decomposition, resulting in net MNC accumulation. Similar mechanisms likely operate in grassland and cropland ecosystems, where warming frequently enhances plant carbon inputs and microbial biomass production, thereby promoting necromass formation (Zhang et al. 2014; Wang et al. 2021). Previous studies have demonstrated that warming‐induced increases in plant productivity are often more pronounced in cold ecosystems, providing greater carbon inputs to support microbial growth and necromass production (Reyes‐Fox et al. 2014; Dobson and Zarnetske 2025; Dang et al. 2025). In contrast, forest canopy growth is often constrained by light, nutrient availability, and atmospheric conditions, limiting the extent to which soil warming enhances plant C inputs (Lim et al. 2019; Dang et al. 2025). Under such conditions, additional plant‐derived carbon inputs may provide only marginal benefits to microbial growth, while warming‐induced increases in extracellular enzyme activity can accelerate necromass decomposition (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). Unlike other ecosystems, wetlands are strongly regulated by hydrological and redox conditions. Although warming may enhance microbial necromass formation by increasing plant‐derived carbon inputs, it may also stimulate Fe(III) reduction, destabilizing Fe‐associated organic carbon and accelerating microbial necromass decomposition (Feng et al. 2025; Sun et al. 2025). These opposing processes may offset each other, resulting in no significant net change in MNC.

These contrasting responses indicate that ecosystem‐dependent MNC dynamics are fundamentally governed by differences in how warming modifies the balance between necromass formation and decomposition (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). This finding suggests that ecosystem type acts as a higher‐order regulator determining whether warming‐induced increases in necromass formation exceed decomposition losses or vice versa. Because microbial necromass represents a major precursor of persistent SOC (Liang et al. 2019; Zhou et al. 2023; Hu et al. 2024), ecosystem‐specific shifts in this balance may ultimately contribute to the divergent SOC responses to warming frequently observed among biomes.

4.4. Environmental and Experimental Controls on MNC Responses to Warming

Soil depth further modified warming responses. Compared with topsoils, subsoils generally exhibit lower microbial activity, lower substrate availability, and stronger physicochemical protection through aggregation and mineral associations (Wang et al. 2020; He et al. 2022; Chen et al. 2023; Huang et al. 2023). Warming may alleviate biological limitations by stimulating root growth and increasing carbon inputs to deeper soil layers (Panchal et al. 2022; Qiao et al. 2025), thereby supporting microbial growth and promoting necromass production in subsoils. However, increased carbon inputs alone are unlikely to fully explain the greater MNC accumulation observed in subsoils. Instead, the relatively slow turnover of organic matter in deeper soil layers, resulting from lower oxygen availability and reduced microbial activity, decreases necromass decomposition and recycling (Fontaine et al. 2007). At the same time, stronger mineral associations and aggregate protection enhance the physical stabilization of newly formed necromass, increasing its persistence in soil (Kallenbach et al. 2016; Liang et al. 2019). Consequently, the stronger warming‐induced MNC accumulation in subsoils is likely driven by the combined effects of enhanced carbon inputs, reduced decomposition, and stronger stabilization, with the latter two processes playing a particularly important role in promoting net necromass preservation.

Differences among warming methodologies also contributed to response variability (Figure 2). Open‐top chambers typically provide moderate warming while largely preserving soil structure, whereas heating cables may induce stronger drying effects and alter soil physical conditions. These differences influence plant productivity, microbial activity, substrate availability, and stabilization processes, thereby affecting MNC accumulation (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024; Luo et al. 2026). Climatic conditions exerted additional controls on MNC responses (Lyu et al. 2025). The negative relationships between MNC responses and both MAT and MAP indicate that warming effects were strongest in cold and relatively dry environments. In these ecosystems, warming alleviates thermal constraints on plant productivity and microbial growth, thereby increasing necromass production (Chang et al. 2021; Zhang et al. 2022). Notably, warming magnitude showed no significant effect on MNC responses (Figure S5), suggesting that baseline climate may be more important than warming intensity in regulating MNC sensitivity. Cold and dry ecosystems with stronger climatic constraints are more responsive to warming‐induced increases in microbial growth and necromass formation, whereas warmer and wetter ecosystems show weaker MNC responses. Collectively, these findings indicate that environmental conditions and experimental warming approaches regulate microbial necromass accumulation by altering substrate supply, microbial activity, and stabilization processes.

Taken together, the ecosystem‐specific responses, environmental controls, and experimental factors identified in this study support the conceptual framework proposed in Figure 6. Within this framework, warming regulates microbial necromass dynamics through coordinated effects on microbial growth, nutrient availability, and extracellular enzyme activities. The net response of MNC to warming depends on the balance between necromass formation and decomposition (Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). Consequently, warming‐induced changes in microbial necromass may alter the capacity of soils to stabilize carbon and maintain persistent SOC pools across ecosystems.

FIGURE 6.

Graphical summary of the key findings with respect to the effects of warming on microbial necromass carbon accumulation. Upward and downward arrows indicate positive and negative effects, respectively, while a horizontal line represents no change. Question marks represent that we lack sufficient data to test this hypothesis. AP, acid phosphatase; BG, β‐glucosidase; BNC, bacterial necromass carbon; CUE, microbial carbon use efficiency; DOC, dissolved organic carbon; FNC, fungal necromass carbon; MBC, microbial biomass carbon; MNC, microbial necromass carbon; NAG, β‐1,4‐N‐acetyl‐glucosaminidase; SAP, soil available phosphorus; TN, total nitrogen; TP, total phosphorus.

4.5. Implications for Soil Carbon Persistence Under Future Warming

Our findings have important implications for understanding soil carbon persistence under climate change. Although warming is widely recognized for accelerating SOC decomposition (MacDougall et al. 2012; Nottingham et al. 2020), our results demonstrate that warming can simultaneously enhance MNC accumulation in many ecosystems. This suggests that microbial necromass represents an important pathway contributing to SOC persistence (Ma et al. 2022; Fu et al. 2025). The contribution of MNC to long‐term SOC stabilization likely depends on its subsequent transformation and incorporation into more persistent carbon pools, particularly mineral‐associated organic carbon (Kallenbach et al. 2016; Liang et al. 2019). Microbial necromass can be protected through interactions with clay minerals and reactive metal oxides, as well as through physical occlusion within soil aggregates, thereby extending carbon residence time. Importantly, the positive coupling between MNC and SOC responses suggests that MNC serves as an important intermediary linking microbial physiological responses with SOC dynamics under warming (Zhou et al. 2023; Hu et al. 2023; Lyu et al. 2025). Within this framework, warming influences SOC persistence through its effects on microbial growth, nutrient availability, extracellular enzyme activities, and the resulting changes in microbial necromass turnover (Xiong et al. 2016; Ding et al. 2019; Liu, Lin, and Li 2024; Liu, Tian, et al. 2024). Consequently, ecosystem differences in SOC responses to warming may partly arise from differences in MNC dynamics, helping explain why warming can simultaneously stimulate microbial activity yet enhance carbon stabilization in some ecosystems. These findings highlight MNC as an important source of uncertainty in future soil carbon‐climate feedbacks. Future soil carbon models should therefore consider not only microbial necromass production and decomposition, but also the subsequent stabilization processes that regulate its transition into persistent SOC pools. Incorporating MNC dynamics and the formation‐decomposition balance framework into soil carbon models may substantially improve predictions of soil carbon responses under future warming scenarios.

4.6. Limitations and Implications for Future Research

While this study provides valuable insights into warming effects on MNC pools at the global scale, several limitations should be acknowledged. First, although this synthesis included studies from multiple regions worldwide, the geographic distribution of observations remained uneven, with some climatic regions being underrepresented. Future research should incorporate more diverse climate zones and geographic regions to improve global representativeness. Second, most forest warming experiments focus primarily on soil heating beneath the canopy, while largely neglecting canopy‐mediated changes in plant carbon inputs under warming. It is therefore necessary to integrate canopy dynamics and adopt more holistic warming approaches to better capture ecosystem carbon cycling responses in forests. Additionally, the relatively small sample size from permafrost ecosystems may constrain the robustness and generalizability of the corresponding results. Expanding experimental coverage in these regions is essential to strengthen future conclusions. Furthermore, future research should move beyond single‐factor warming experiments and investigate the combined effects of multiple global change factors, such as warming, nitrogen deposition, and altered precipitation regimes on MNC dynamics. These multifactor experiments are critical for revealing how concurrent environmental changes regulate microbial necromass formation and decomposition processes under more realistic climate scenarios. Overall, improving our understanding of how climate change alters the balance between necromass formation and decomposition will be critical for reducing uncertainty in future soil carbon–climate feedback predictions.

5. Conclusions

This study demonstrates that warming exerts ecosystem‐dependent effects on MNC accumulation, increasing MNC in permafrost, grassland, and cropland ecosystems while reducing it in forests. These contrasting responses reflect shifts in the balance between necromass formation and decomposition, with warming‐induced increases in microbial biomass enhancing necromass production and elevated extracellular enzyme activities accelerating necromass turnover and decomposition. The direction and magnitude of MNC responses depend on the relative strength of these opposing processes. Microbial traits and nutrient availability emerged as key regulators of microbial necromass dynamics. Ecosystem type, climatic conditions, soil depth, and warming methodology further modify this balance, explaining the strong ecosystem dependence of MNC responses. The positive coupling between MNC and SOC responses highlights microbial necromass formation as an important pathway contributing to soil carbon persistence under climate warming. These findings highlight the necessity of incorporating microbial traits, necromass turnover processes, and ecosystem‐specific controls into soil carbon models to improve predictions of soil carbon‐climate feedbacks under future warming scenarios.

Author Contributions

Xinxin Jing: investigation, validation, methodology. Jieyu Gao: investigation, methodology, conceptualization. Peng Chen: validation, formal analysis, data curation. Chenhao Lyu: conceptualization, funding acquisition, visualization, methodology, writing – original draft, writing – review and editing, software. Wenzhi Liu: conceptualization, funding acquisition, project administration, resources. Zhiguo Li: conceptualization, supervision, resources, investigation. Luping Ye: investigation, formal analysis, data curation. Yi Liu: writing – review and editing, resources, funding acquisition, supervision.

Funding

This study was supported by the National Natural Science Foundation of China (No. 42407483 and U24A20641), and the Natural Science Foundation of Hubei Province (2024AFD373 and 2025AFD442).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Text S1: Conventional calculation method for microbial necromass carbon estimation.

Table S1: Effect sizes of warming on total MNC, FNC, and BNC before and after trim‐and‐fill correction for publication bias.

Table S2: Best models for the responses of total MNC, FNC, and BNC to warming based on model selection.

Figure S1: PRISMA flow diagram for publication inclusion.

Figure S2: Funnel plots for the response of total MNC, FNC, and BNC to warming before and after trim‐and‐fill correction for publication bias. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon.

Figure S3: Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of MAT (a–c), and MAP (d–f). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S4: Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of longitude (a–c), latitude (d–f), and altitude (g–i). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S5: Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of warming magnitude (a–c), and duration (d–f). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S6: Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of AGB responses (a–c), and BGB responses(d–f). The size of each circle represents the weight of the corresponding observations in the meta‐analysis. AGB, aboveground biomass; BGB, belowgroundbiomass; BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, micorbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S7: Meta‐regression of warming effects on total MNC (a), FNC (b), and BNC (c) as functions of soil pH responses. The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, micorbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S8: Meta‐regression of warming effects on total MNC (a), FNC (b), and BNC (c) as functions of soil moisture responses. The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon; SM, soil moisture. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S9: Meta‐regression of warming effects on total MNC, FNC, and BNC as functions of NO₃ ⁻ (a–c) and NH₄ ⁺ (d–f) responses. The size of each circle represents the weight of the corresponding observations in the meta‐analysis. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S10: Effects of warming on plant biomass, soil properties, microbial traits, and soil enzyme activity. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses. AGB, aboveground biomass; ALP, alkaline phosphatase; AP, acid phosphatase; BG, β‐glucosidase; BGB, belowground biomass; CBH, cellobiohydrolase; DOC, dissolved organic carbon; LAP, leucine aminopeptidase; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; NAG, β‐1,4‐N‐acetyl‐glucosaminidase; POX, phenol oxidase; SAP, soil available phosphorus; SOC, soil organic carbon; TN, total nitrogen; TP, total phosphorus; XYL, β‐xylanase.

Figure S11: Effects of warming on SOC as related to ecosystem types, warming methods, soil depths, and initial soil pH and SOC. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses.

Figure S12: Effects of warming on the contribution of total MNC (a), FNC (b), and BNC (c) to SOC as related to ecosystem types, warming methods, soil depths, and initial soil pH and SOC. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon.

Figure S13: Effects of warming on the ratio of FNC/BNC as related to ecosystem types, warming methods, soil depths, and initial soil pH and SOC. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses. BNC, bacterial necromass carbon; FNC, fungal necromass carbon.

Figure S14: Meta‐regression of warming effects on soil pH as functions of soil moisture. The size of each circle represents the weight of the corresponding observations in the meta‐analysis. Q _m represents the test statistic for the significance of moderators in the meta‐regression model; a significant Q _m (p < 0.05) indicates a significant moderating effect.

Figure S15: Effects of warming on total MNC (a), FNC (b), and BNC (c) calculated using the original conversion methods reported in individual studies, as related to ecosystem types, warming methods, soil depths, and initial soil pH and SOC. Error bars represent 95% confidence intervals. Numbers in parentheses indicate the number of observations. Solid symbols denote significant effects, whereas hollow symbols represent non‐significant responses. BNC, bacterial necromass carbon; FNC, fungal necromass carbon; MNC, microbial necromass carbon.

Figure S16: Global distribution of study sites included in this meta‐analysis. The color of each point indicates the ecosystem type at the corresponding site.

Data Availability Statement

The data that support the findings of this study are openly available on figshare at https://doi.org/10.6084/m9.figshare.32586663.