Species-Specific Effects of Litter Management on Soil Respiration Dynamics in Urban Green Spaces: Implications for Carbon Cycling and Climate Regulation

Abstract

The disposal of urban tree litter as waste has significant implications for material cycles, energy flows, and global climate change within urban ecosystems. However, the species-specific contributions of urban trees to atmospheric CO₂ emissions through soil respiration (R_S) remain poorly understood. This study investigates the effects of litter management on R_S dynamics in urban green spaces, focusing on six common species (Mangifera indica, Ficus microcarpa, Cinnamomum camphora, Bauhinia purpurea, Triadica sebifera, and Celtis sinensis) in Fuzhou, China. Three litter treatments—litter retention (CK), litter removal (RL), and litter doubling (DL)—were established to monitor monthly R_S fluctuations. Results indicate that DL significantly increased R_S rates, while RL reduced them. The increase in R_S due to litter addition was more pronounced than the decrease caused by litter removal for most species. R_S rates exhibited a unimodal seasonal pattern, peaking in summer. Furthermore, litter treatments influenced the temperature sensitivity coefficient (Q₁₀), with F. microcarpa showing the highest average Q₁₀ (4.16) and M. indica the lowest (1.88). This study underscores the critical role of litter input in modulating R_S in urban green spaces and highlights the joint but asymmetric effects of soil temperature and moisture on R_S dynamics.

1. Introduction

Uncertainties in urban carbon cycling are closely linked to global climate change—including elevated atmospheric CO₂, warming, and nitrogen deposition—and anthropogenic activities such as urbanization, litter removal, and pollution [1,2,3]. While rising CO₂ levels and warming may enhance ecosystem productivity and aboveground litter inputs [4], rapid urbanization disrupts natural litter recycling. For instance, most urban tree litter is discarded as waste rather than returned to soils, altering soil respiration (R_S), the combined CO₂ flux from root respiration (autotrophic), and microbial decomposition (heterotrophic) [5]. Litter acts as a critical conduit for energy and nutrient transfer between above- and belowground systems. It not only constitutes the primary carbon input to soils but also fuels decomposer activity, thereby regulating terrestrial ecosystems’ carbon balance and nutrient cycling [6,7,8]. Changes in litter inputs directly modulate R_S rates, with cascading impacts on urban forest carbon dynamics.

The global soil-to-atmosphere carbon flux, comprising approximately 68 Pg C·yr⁻¹ from autotrophic and heterotrophic respiration alongside 50 Pg C·yr⁻¹ from litter decomposition [9], underscores R_S as a pivotal yet vulnerable component of terrestrial carbon cycling. Manipulation experiments have elucidated the sensitivity of R_S to litter dynamics across natural ecosystems. In subtropical China, Yan et al. [10] demonstrated 39%, 22%, and 24% reductions in soil CO₂ efflux following litter removal in camphor, pine, and mixed forests, respectively. Mediterranean oak forests exhibited 21.9% R_S dependency on aboveground litter inputs [11], while Zhuang et al. [12] documented 22% R_S enhancement through litter addition, contrasting with negligible effects of removal. These responses are spatiotemporally heterogeneous, governed by synergistic interactions among hydrothermal gradients [11], soil biota composition [13], and litter stoichiometry [14] as meta-analyzed by Wang et al. [7].

The study of R_S is of great significance in urban green spaces, as it directly relates to carbon cycling, ecosystem functions, and climate change mitigation. R_S is the main pathway for soil to release CO₂ into the atmosphere, and its dynamic changes reflect the strength of urban green spaces’ carbon sequestration capacity [15]. Research has shown that the R_S rate of urban green spaces is jointly regulated by vegetation type, litter input, and soil environmental factors and is a key indicator for evaluating the carbon balance of urban ecosystems [16]. In the context of global climate change, urban green spaces play an important role in mitigating urban heat island effects and enhancing carbon sequestration potential by regulating atmospheric CO₂ concentration through R_S [17]. Therefore, in-depth research on R_S mechanisms and their driving factors is of great significance for optimizing urban green space management, enhancing ecosystem service functions, and addressing climate change.

However, R_S exhibits significant spatial heterogeneity in different ecosystems, and its rate and dynamics are jointly influenced by vegetation type, climatic conditions, and soil characteristics. For example, R_S in forest ecosystems is typically higher than in grasslands and farmland, mainly attributed to higher litter input and root activity [18]. Wetland ecosystems have relatively low R_S rates due to anaerobic environment limitations on organic matter decomposition [19]. In contrast, research on R_S in urban green spaces is relatively scarce, but their unique environmental pressures (such as human interference, pollutant inputs, and heat island effects) may significantly alter the driving mechanisms of R_S. In addition, the role of urban green spaces in the global carbon cycle is increasingly prominent, and their R_S dynamics not only affect regional carbon balance, but also directly relate to the service functions of urban ecosystems and the mitigation potential of climate change [20]. Therefore, strengthening the research on R_S in urban green spaces is of great significance for understanding the carbon cycle mechanism in the urbanization process, optimizing green space management strategies, and enhancing the sustainability of urban ecosystems [16].

Existing research indicates that R_S in urban environments is jointly regulated by vegetation, soil properties, and human activities. Vegetation type and litter input are key factors affecting R_S, and the quality and decomposition rate of litter from different tree species significantly affect soil CO₂ emissions [16]. Soil properties, such as organic matter content, pH value, and texture, also regulate R_S by altering microbial activity and substrate availability. In addition, human activities such as irrigation, fertilization, and trampling further affect R_S dynamics by altering soil temperature, moisture, and nutrient status. For example, frequent irrigation in urban green spaces may increase soil moisture, thereby promoting microbial activity and increasing R_S rate [20]. However, human interference such as pollutant input and soil compaction may also inhibit R_S and reduce the carbon sequestration capacity of urban green spaces [17].

The study of R_S in urban environments provides an important scientific basis for ecological planning, carbon sequestration, and sustainable urban development. By 2024, China will commence the construction of over 6200 “pocket parks” and build more than 7300 km of urban greenways in the urban green spaces [21]. By quantifying R_S dynamics and their driving factors, research can reveal the carbon sequestration potential of urban green spaces, providing guidance for optimizing green space layout and vegetation configuration [16]. For example, selecting tree species with high carbon sequestration efficiency and improving soil management practices can enhance the carbon sequestration capacity of urban green spaces [20]. In addition, R_S research can help evaluate the impact of human activities such as irrigation, fertilization, and trampling on the carbon cycle of urban ecosystems, in order to develop sustainable urban management strategies. In the context of addressing climate change, urban green spaces serve as important carbon sinks and cooling spaces, and their R_S research provides critical support for enhancing the resilience and sustainability of urban ecosystems [17]. Therefore, a deeper understanding of R_S mechanisms not only helps achieve urban carbon neutrality goals, but also promotes the improvement of ecological service functions and urban living environments.

Notably, while forest and grassland R_S-litter relationships are well characterized, urban ecosystems remain conspicuously underexplored despite their unique ecological drivers. Urban green spaces deliver critical regulatory services including stormwater mitigation, heat island alleviation, and particulate matter filtration [22], yet face intensifying spatial constraints under urbanization [23]. This paradox highlights an urgent need to quantify how anthropogenic litter management—ranging from hyper-cleaning to organic accumulation—modulates urban R_S. Such knowledge gaps hinder predictive modeling of urban carbon fluxes and evidence-based optimization of vegetation ecosystem services.

We address this challenge through a controlled manipulation experiment in Fuzhou City, China, focusing on six dominant street tree species (Mangifera indica, Ficus microcarpa, Cinnamomum camphora, Bauhinia purpurea, Triadica sebifera, and Celtis sinensis). Our study pioneers three novel dimensions:

• Fine-scale temporal dynamics: Monthly tracking of soil respiration (R_S), temperature (T), and moisture (M) responses to litter removal (RL, simulating municipal cleaning) and doubling (DL, mimicking urban litter accumulation).
• Thermal sensitivity analysis: Quantification of Q₁₀ > coefficients to assess how litter alterations affect temperature-R_S > coupling in engineered urban soils.
• Multivariate interaction modeling: Decoupling synergistic effects of T > M > -litter inputs on R_S > through structural equation modeling.

By bridging the critical disjuncture between natural ecosystem theory and urban management practices, this work advances mechanistic understanding of anthropogenic carbon cycling while informing climate-resilient urban green space design.

2. Materials and Methods

2.1. Study Area Characterization

The main reason we chose these tree species as the research objects is their quantity, cultivated area, and distribution. F. microcarpa is the City Tree of Fuzhou, with a local planting history of over 1000 years and a total of over 100,000 trees planted throughout major streets, parks, and communities. The number of C. camphora also exceeds 100,000, accounting for about 25% of roadside trees. T. sebifera accounts for about 5% of urban green spaces and is included as a key tree species in the “Colorful Project” of Fuzhou City, along with B. purpurea. And, C. sinensis and M. indica trees are representative local and exotic tree species in Fuzhou. Although there is no specific statistical quantity, they are mainly distributed in urban parks, streets, residential areas, and suburban orchards [24].

The experimental site is located within the Jinshan Campus of Fujian Agriculture and Forestry University (26°05′ N, 119°13′ E), a representative urban greenspace in Fuzhou City’s western suburban transition zone. This 234-hectare area occupies a unique geomorphological position within the Minjiang River alluvial basin, hydrologically bounded by the Minjiang and Wulong Rivers and topographically constrained by flanking mountain ranges.

Climatically, the region exhibits a humid subtropical monsoon regime (Köppen climate classification Cfa), characterized by pronounced seasonality with hot-humid summers (July–August mean: 24–29 °C; record maximum 42.3 °C) and mild winters (January–February mean: 6–10 °C; record minimum −1.2 °C). The extended growing season is evidenced by a 326-day frost-free period and mean annual precipitation of 1340 mm, peaking during the East Asian summer monsoon (May–September) (Fuzhou Meteorological Bureau, 2022). Such hydrothermal conditions support the development of secondary subtropical, evergreen broad-leaved forests dominated by Castanopsis spp. and Lithocarpus spp. [25], with a 51.77% average canopy cover facilitating distinct microclimate buffering effects (Fuzhou Municipal People’s Government, 2023).

This site was strategically selected due to the following:

• Urban–rural ecotone representativeness: Transitional zone between metropolitan Fuzhou (population > 7.6 million) and peri-urban agricultural landscapes;
• Anthropogenic gradient: Subject to intermediate levels of human disturbance compared to central urban parks and protected forest reserves;
• Vegetation homogeneity: Monospecific tree stands of six target species (Mangifera indica, Ficus microcarpa, etc.) established through municipal greening programs since 2005.

2.2. Methods

2.2.1. Sample Plot Setup

The experiment was initiated in December 2020 in Fuzhou City, China. Six dominant tree species—Mangifera indica, Ficus microcarpa, Cinnamomum camphora, Bauhinia purpurea, Triadica sebifera, and Celtis sinensis—were selected based on their ecological prevalence and representativeness in subtropical urban forests [25]; additional criteria included the following: canopy structure, leaf phenology, and litter production rates. The basic information of experimental trees is shown in

Table 1. The basic information of experimental trees.

Species	Height/m	DBH/cm	Crown/m	Clear Length/m
Mangifera indica	9.08 ± 0.41	34.67 ± 2.35	6.38 ± 1.10	2.20 ± 0.25
Ficus microcarpa	10.35 ± 1.36	25.47 ± 3.12	7.54 ± 2.05	2.64 ± 0.08
Cinnamomum camphora	8.13 ± 0.95	20.00 ± 1.65	5.42 ± 0.96	1.80 ± 0.14
Bauhinia purpurea	9.73 ± 2.33	19.67 ± 3.10	6.50 ± 1.21	2.13 ± 0.10
Celtis sinensis	6.99 ± 1.11	21.30 ± 2.73	6.00 ± 1.05	2.07 ± 0.55
Triadica sebifera	10.66 ± 2.16	26.33 ± 3.52	6.67 ± 1.25	2.47 ± 0.99

Note: DBH means Diameter of Breath Height. The data in the table are the mean ± standard deviation (N = 30).

. To minimize spatial heterogeneity, all study plots were established in areas with homogeneous soil type (red clay loam), slope (<5°), and light exposure (full sunlight) [26].

For each species, three independent 10 m × 10 m plots were established as biological replicates, with a minimum buffer zone of 15 m between adjacent plots to avoid edge effects. Prior to treatment application, all plots underwent a 30-day stabilization period to mitigate soil disturbance from initial setup activities [27]. Within each plot, three cylindrical PVC soil respiration (R_S) collars (inner diameter: 20 cm; height: 7.5 cm) were installed randomly beneath the tree canopy. Collars were sharpened at the base to ensure minimal soil compaction during insertion and were embedded 2.5 cm into the soil, leaving 5 cm exposed aboveground. Three litter manipulation treatments were applied to 1 m × 1 m subplots:

• Control (CK): Natural litter layer maintained without disturbance.
• Litter Removal (RL): All aboveground litter within the subplot was manually removed using non-metallic tools to avoid soil scraping, and a vertically suspended nylon mesh (0.5 m height; 1 mm aperture) was installed to intercept incoming litterfall [28].
• Double Litter (DL): Ambient litter was retained, and additional litter collected from adjacent RL subplots was evenly distributed within a 1 m radius around the collar. Litter mass was quantified using dry-weight equivalents (g m⁻²) to standardize treatment intensity [29].

2.2.2. R_S and Soil Temperature and Moisture Measurement

Soil respiration (R_S) was measured monthly from January to December 2021 using a portable soil carbon flux analyzer (LI-8100A, LI-COR Biosciences, Lincoln, NE, USA). To minimize diurnal variability, measurements were conducted between 09:00 and 11:00 local time on consecutive rain-free days [30]. During each campaign, soil temperature (T; °C) and soil moisture (M; %) were simultaneously recorded at 5 cm depth using the analyzer’s integrated probe. Ensuring consistency, the measurements for each tree species were conducted within the same time frame each month, totaling 12 field measurements annually.

For quality control, each collar was measured three times consecutively over a 9 min interval (30 s pre-purge followed by three 3 min observation cycles). Outliers caused by transient environmental noise (e.g., animal activity) were excluded using the manufacturer’s post-processing software (LI-8100A SoilFluxPro v.1.5.1).

2.2.3. Modeling R_S Responses to Soil Temperature and Moisture

Four widely validated models were employed to quantify the relationships between R_S (dependent variable) and soil temperature (T) and moisture (M) (independent variables) [31]. Rates were employed to examine the treatment effects on the R_S rate and soil microclimate parameters (5 cm depth) across species. When significant differences were detected (p < 0.05), post hoc pairwise comparisons were performed using Fisher’s least significant difference (LSD) method with Bonferroni-adjusted α levels to control family-wise error rate. The effect sizes were quantified through partial eta-squared for ANOVA and Cohen’s d for pairwise contrasts [32].

Temperature sensitivity is often represented by the relative changes in respiration at a temperature of 10 °C, which is called the Q₁₀ value [33]. According to the above exponential equation, the Q₁₀ value can be calculated as

Q₁₀ = e^10b

(1)

where e is a natural constant and b is a constant representing the rate of change in the reaction rate for every 10 °C increase in temperature.

Results were visualized through Origin 2018 (OriginLab Corp., Northampton, MA, USA) using publication-ready formats: Continuous variables were presented as mean ± standard error (SEM) with 95% confidence intervals, while categorical comparisons utilized clustered column charts with explicit significance annotations (letters or asterisks) [34]. All graphical elements adhered to colorblind-friendly palettes and minimum line width standards.

3. Results

3.1. Monthly Dynamics of R_S with Removal and Addition of Litter

Throughout the monthly observations, R_S dynamics among the six tree species exhibited a consistent pattern under the treatments of litter retention, removal, and addition. All species followed a similar trend, characterized by an initial increase followed by a decrease. Notably, the R_S dynamics of each species were distinctively represented by a single-peak curve, peaking primarily during June, July, and August, and reaching their lowest points in January and December (Figure 1).

Over the entire observation period, the annual mean values of R_S rates ranged from 1.86 to 4.02 µmol·m⁻²·s⁻¹ for the RL treatment, 2.58 to 5.38 µmol·m⁻²·s⁻¹ for the CK treatment, and 3.18 to 6.71 µmol·m⁻²·s⁻¹ for the DL treatment. The R_S rates were highest under the DL treatment, followed by CK and RL, indicating that litter addition enhanced the R_S rate, while litter removal had the opposite effect.

3.2. Monthly Dynamics of Soil Temperature and Moisture with Removal and Addition of Litter

Under the three varying litter treatments, the soil temperature at a depth of 5 cm peaked for each tree species in June and July, with the lowest temperatures recorded in December. Among the tree species, C. camphora exhibited the highest soil temperature at this depth under all litter treatments. Overall, the soil temperatures of the six species followed the order DL > CK > RL. Specifically, litter addition led to increases in soil temperatures of 1.81, 0.75, 0.33, 0.80, 0.32, and 0.20 degrees Celsius for M. indica, F. microcarpa, C. camphora, B. purpurea, C. sinensis, and T. sebifera, respectively (Figure 2).

The monthly fluctuations in soil moisture followed a similar pattern to soil temperature, peaking between June and September. Summer and fall months saw significantly higher soil moisture levels compared to spring and winter. However, the change in soil moisture did not align with increasing little inputs. Only T. sebifera exhibited a similar pattern of DL > CK > RL in soil moisture as was observed for soil temperature among the six species. The other species displayed varying and inconsistent patterns (Figure 2).

In terms of the annual average moisture of the 5 cm soil layer under different apoplastic treatments, the top three tree species were C. camphora, F. microcarpa, and B. purpurea (Figure 2).

3.3. Effects of Litter Removal and Addition on R_S Contributions

The impact of various litter treatments on R_S differed significantly, as the addition and removal of litter did not have an equal effect on R_S. Neither did they alter R_S in a proportional manner (Figure 3). In the treatments where litter was added, the percentage changes in R_S ranged from 1.32% to 47.44% for M. indica; 13.23% to 76.97% for C. camphora; 11.87% to 83.18% for Ficus microcarpa; 10.80% to 74.07% for B. purpurea; 4.19% to 86.38% for Celtis sinensis; and 4.72% to 79.37% for T. sebifera. On the other hand, in the treatments where litter was removed, the percentage changes in R_S ranged from 6.73% to 45.71% for M. indica; 7.30% to 73.53% for C. camphora; 9.21% to 59.87% for F. microcarpa; 3.73% to 48.37% for B. purpurea; 10.10% to 48.99% for C. sinensis; and 2.22% to 60.09% for T. sebifera. Overall, the increase in R_S with the addition of litter was more significant than the decrease with the removal of litter, and most tree species were more responsive to litter changes during the spring season.

3.4. Effect of Litter Removal and Addition on Temperature Sensitivity Coefficient Q₁₀

The regression analysis results depicted in Figure 4 revealed that the temperature sensitivity coefficient, Q₁₀, for the six tree species subjected to litter removal and addition treatments ranged from 1.90 to 4.82. Specifically, for M. indica, the Q₁₀ values were DL (1.95) > RL (1.90) > CK (1.79). For C. camphora, the Q₁₀ values were 4.04, 2.99, and 2.88 for RL, DL, and CK, respectively, while for F. microcarpa, the values were 4.66, 4.07, and 3.76, respectively, both species exhibiting a trend of RL > DL > CK. B. purpurea showed Q₁₀ values of 2.23, 1.90, and 2.15. T. sebifera’s Q₁₀ values were 4.81, 3.29, and 3.70, while Celtis sinensis’s values were 3.13, 2.19, and 2.49. Across all species, a pattern of RL > CK > DL was observed. Overall, the impact of litter removal on Q₁₀ was greater than litter addition among the six tree species. Notably, F. microcarpa exhibited the highest average Q₁₀ value (4.16) under different treatments, indicating its high sensitivity to temperature. Conversely, M. indica had the lowest average Q₁₀ value (1.88), indicating its relatively low sensitivity.

3.5. Correlation of R_S with Temperature and Moisture with Removal and Addition of Litter

The R_S rate exhibited strong linear correlations and exponential relationships with both 5 cm soil temperature and moisture. Notably, the R_S rate displayed significant linear and exponential relationships with soil temperature (p < 0.05), with most relationships being highly significant (p < 0.01). Overall, the linear model outperformed the exponential model in terms of fit (Figure 4). Among the six tree species, 5 cm soil temperature accounted for 82.56%, 85.42%, 82.35%, 64.77%, 80.34%, and 64.71% of the variation in R_S rate under the litter removal treatment. By contrast, litter addition treatment explained 87.78%, 80.27%, 62.74%, 47.94%, 77.58%, and 75.33% of the variance. The control group accounted for 82.77%, 82.91%, 76.86%, 52.43%, 76.74%, and 70.43% of the variance. The proportion of soil temperature explaining R_S changes was affected by litter removal or addition. However, the linear and exponential relationships between R_S rate and soil moisture were not statistically significant (p > 0.05), and the R² values obtained from fitting equations between R_S and soil moisture were significantly lower than those of temperature (Figure 5).

We performed a regression analysis to assess the relationship between soil temperature and moisture data at a depth of 5 cm and R_S rate. The composite model derived from this analysis effectively captures the influence of these factors on R_S, as evident from

Table 2. The multiple regression equations between R_S and soil temperature (T) and soil moisture (M).

Species	Treatment	Multiple Regression Equations	R2
Mangifera indica	CK	RS = −0.9079 + 0.1540T + 0.0025M	0.7898
	RL	RS = −0.5441 + 0.1250T − 0.0058M	0.7909
	DL	RS = −1.6878 + 0.2271T − 0.0126M	0.8592
Cinnamomum camphora	CK	RS = −8.6875 + 0.5531T − 0.0317M	0.8218
	RL	RS = −7.8850 + 0.4409T − 0.0066M	0.8238
	DL	RS = −13.2848 + 0.8194T − 0.0500M	0.8081
Ficus microcarpa	CK	RS = −4.6105 + 0.4673T − 0.0569M	0.8095
	RL	RS = −4.0007 + 0.3580T − 0.0365M	0.8498
	DL	RS = −7.5529 + 0.6962T − 0.0742M	0.6074
Bauhinia purpurea	CK	RS = −7.7607 + 0.2186T − 0.0501M	0.8185
	RL	RS = −0.8113 + 0.1456T − 0.0226M	0.7009
	DL	RS = −0.3306 + 0.2450T − 0.0646M	0.8120
Triadica sebifera	CK	RS = −6.5042 + 0.4144T − 0.0233M	0.6524
	RL	RS = −4.7890 + 0.2564T + 0.0497M	0.6583
	DL	RS = −6.2573 + 0.4402T − 0.0270M	0.7167
Celtis sinensis	CK	RS = −3.3435 + 0.3702T − 0.0509M	0.8241
	RL	RS = −3.8145 + 0.3290T − 0.0341M	0.8261
	DL	RS = −2.6455 + 0.3717T − 0.0381M	0.7899

. Notably, the correlation coefficients of most composite models approach or exceed 0.80, indicating that the combined effects of 5 cm soil temperature and moisture can account for 80% or more of the variations in R_S.

4. Discussion

4.1. R_S in Response to Changes in Litter Inputs

As a primary contributor to soil carbon pools, alterations in the quantity and composition of litter inputs significantly impact R_S [35,36,37]. This study found that R_S rates under varying litter treatments exhibited a pattern of DL > CK > RL throughout the measurement period. This implied that the urban R_S could be intensified (weakened) by augmented (diminished) litter inputs, echoing prior research conducted in diverse forest ecosystems [38,39]. It also aligned with the speculation that urbanization processes would result in a decline in R_S [40]. This occurred because the role of litter in R_S is a intricate biological process. Litter addition stimulates an increase in soil microorganisms, leading to a priming effect, which elevates the mineralization rate of soil organic carbon and subsequently boosts R_S [41]. Conversely, litter removal diminishes the influx of soil carbon and soil microorganism numbers in the surface layer, leading to a loss of the micro-environmental protection provided by litter. This reduction in microbial activity diminishes the carbon source and microbial support, ultimately leading to a decrease in R_S [7,42].

There exist distinct disparities in the functions of urban and non-urban ecosystems [43]. This study reveals that the annual mean R_S rates of six distinct tree species, under varying litter treatments, ranged from 1.86 to 4.02 µmol·m⁻²·s⁻¹ (RL), 2.58–5.38 µmol·m⁻²·s⁻¹ (CK), and 3.18–6.71 µmol·m⁻²·s⁻¹ (DL). These observations are notably higher compared to prior investigations exploring the impact of litter input alterations on R_S in non-urban settings [44,45,46]. Kaye et al. [47] conducted a comparison between urban and non-urban green spaces in Collinsburg, USA, uncovering that the urban ecosystem exhibited a 2.5 to 5-fold increase in R_S and underground carbon distribution compared to non-urban systems. Similarly, Groffman et al. [48] reported significantly elevated R_S in urban areas compared to rural regions in a study conducted across urban and rural gradient forests in the United States. This elevated respiration can be attributed to the abundant anthropogenic CO₂ emissions in urban areas, coupled with the intricate structure of urban ecosystems. Additionally, the heat island effect prevalent in urban environments leads to warmer soil temperatures compared to suburban areas. To counter the escalating atmospheric CO₂ concentration, the CO₂ flux from urban green space soil surpasses that of agricultural and natural soils, further elevating R_S in urban green spaces compared to other ecosystems, regardless of litter treatments [49,50,51].

In this study, we discovered that removing and adding litter to soil can significantly alter R_S rates. However, the magnitude of these changes varied depending on the type of tree species and the specific litter treatments applied. Specifically, under litter removal, R_S decreased by 25.78%, 25.73%, 34.08%, 26.50%, 29.05%, and 34.95% for M. indica, C. camphora, F. microcarpa, B. purpurea, C. sinensis, and T. sebifera, respectively. Conversely, litter addition increased R_S by 25.31%, 37.11%, 47.85%, 29.33%, 42.69%, and 26.48% for these species. Notably, the average annual increase in R_S for the remaining four species was higher than the average annual decrease, except for M. indica and T. sebifera.

These findings suggest that litter has a positive priming effect on R_S. This is likely due to the fact that litter addition enriches soil carbon sources, stimulating the decomposition of existing organic matter [52]. The observed variability in R_S responses can be attributed to several factors: (1) the varying time frames of soil microbial activity in response to different litter inputs, leading to differences in the duration of soil stimulation and subsequent R_S rates [53,54]; (2) litter input slowing down CO₂ release from R_S, with species-specific differences in litter quantity and a balancing act between this shielding effect and CO₂ release from litter decomposition [39]; (3) species-specific differences in litter size, thickness, and input, which can affect litter turnover rates [55]; and (4) physiological differences among plant species, such as the deeper root systems and frequent activities of F. microcarpa, which may enhance R_S rates.

It is evident that R_S is influenced by numerous factors, including plant species, microbial activity, and litter input, making it a complex biological process. Consequently, there is a nonlinear correlation between litter input and R_S, with increases or decreases in R_S not always synchronous with litter input [39,56]. In this study, annual mean R_S values ranged from 25.31% to 47.85% with litter removal and addition, exceeding the average litter contribution to forests in China (20.2%) [57], and varying widely compared to karst primary and secondary forests (17.41%–26.20%) [58]. These values were similar to the range observed in different litter treatments of mixed P. massoniana and C. camphora forests in Hunan Province (24.30%–39.60%) [59] but lower than the range reported in maple forests (20.85%–71.31%) [60]. These findings suggest that the litter species studied had a higher contribution to R_S and that there is spatial variability in this contribution. This spatial variability may be attributed to differences in hydrothermal conditions across latitudes [61], which can affect litter influence mechanisms, as well as species-specific growth characteristics, root development, and soil temperature and moisture, which can all impact R_S and contribute to this variability [46]. In addition, R_S is primarily driven by microbial decomposition of organic matter, with its rate regulated by the chemical composition of litter (for example, an increased lignin/cellulose ratio inhibits decomposition) [62]. Small soil fauna such as termites and earthworms accelerate the mineralization of refractory substances through physical fragmentation, gut enzyme activity, and microbial symbiosis (synergistic effects), with their metabolic contribution accounting for approximately 5%–20% of soil CO₂ flux [63].

4.2. Response of R_S to Temperature and Moisture Under Different Litter Treatments

Soil respiration, a crucial component of the terrestrial ecosystem carbon cycle, is influenced by numerous factors involved in this cycle. Specifically, soil autotrophic respiration primarily relies on the vertical transport of plant photosynthesis products, which is determined by plant root biomass and rhizosphere symbiont type. Conversely, soil heterotrophic respiration is constrained by soil temperature, moisture, substrate availability, and microbial community structure. This dependence arises from the capacity of soil microorganisms to decompose soil organic matter and plant litter [64,65]. With the escalating urbanization, the heat island effect further induces temperature elevations and alterations in the spatial and temporal patterns of precipitation. This trend increases litter production and disrupts the decomposition, transformation, and supply of soil organic substrates. Additionally, urban greening management, often motivated by aesthetic considerations, can introduce or eliminate litter, thereby exerting a range of impacts on soil temperature and moisture [66].

Litter can regulate soil temperature and moisture by enhancing soil water permeability, minimizing surface runoff and evaporation rates, and encouraging soil microorganisms to degrade litter, thereby influencing R_S [67]. In this study, the monthly R_S rates of six tree species across different litter treatments exhibited a distinct single-peak pattern, peaking significantly during summer and fall compared to spring and winter. Concurrently, the changes in soil temperature and moisture at a depth of 5 cm followed a similar trend. This pattern is primarily attributed to seasonal factors. Summer and fall temperatures are higher, coinciding with the growing season for plants. During this period, plants engage in robust photosynthesis, funneling abundant photosynthetic products to their underground components. This stimulates frequent root system activity and robust growth, which in turn enhances soil autotrophic respiration. As winter approaches, plants enter a dormant state, resulting in reduced root respiration and microbial activities, thereby limiting R_S [68]. Additionally, the Fuzhou region experiences significant rainfall in summer and fall, replenishing soil moisture and fostering microbial activities. This environment promotes soil heterotrophic respiration, further elevating R_S rates [55].

Previous studies have confirmed that temperature and moisture are crucial factors influencing R_S [69,70]. Specifically, the correlation between temperature and R_S is stronger than that between moisture and R_S [71]. This study found a significant correlation between R_S and soil temperature across various litter treatments. However, most correlations with soil moisture were non-significant. Furthermore, the R² value for moisture derived from the fitting equation of R_S with soil temperature and moisture was notably lower than that for temperature. This indicates that soil temperature is a better predictor of R_S rate than soil moisture. This might be due to the subtropical monsoon climate of Fuzhou, which experiences limited weather extremes and results in less variable soil moisture. Under optimal moisture conditions, plant growth, development, and soil microbial activities are not significantly constrained by moisture [72]. Conversely, in drought and waterlogged regions, moisture becomes a primary limiting factor for R_S. The mechanism of soil moisture on R_S is highly complex, as both excessively high and low soil moisture levels can directly affect root activities and microbial physiological activity and indirectly impact the soil substrate and oxygen diffusion. This leads to significant variability in soil moisture under different climatic conditions [72,73,74].

The measurement of soil temperature’s impact on R_S primarily relies on the temperature sensitivity coefficient, Q₁₀, which quantifies the multiplicative increase in R_S rate for every 10 °C rise in soil temperature [72]. The greater the Q₁₀ value, the more sensitive R_S responds to temperature changes, which is often used as a key indicator of the relationship between global climate change and carbon cycling in terrestrial ecosystems [75]. In this study, we observed distinct trends in Q₁₀ values among various tree species under different litter treatments. Mangifera indica exhibited a pattern of DL > RL > CK, with an average Q₁₀ value of 1.88, indicating low sensitivity. This low sensitivity persisted even after litter addition, possibly due to substrate utilization offsetting the negative impact of activated carbon on Q₁₀ [76]. Conversely, the Q₁₀ values of the other five tree species with litter removal treatments were higher than those of the control and litter addition treatments. This could be attributed to the positive correlation between Q₁₀ and substrate inertness [2], or due to the removal of litter leaving the surface bare, decreasing the soil’s resilience to external environments and thus elevating Q₁₀ values [77]. The Q₁₀ values in this study ranged from 1.90 to 4.82, aligning with R_S values of 1.33 to 5.53 reported for Chinese forests [57]. Notably, soil temperatures at a depth of 5 cm explained 47.94% to 87.78% of R_S variations under different litter treatments, confirming temperature as the primary regulator of R_S in the study area, surpassing the impact of moisture as a standalone factor.

Most studies have indicated that soil temperature and moisture have a synergistic effect on R_S [78]. In the present study, the R_S of M. indica, C. camphora, and T. sebifera was primarily influenced by temperature as a single factor, rather than by a combination of both temperature and moisture. Conversely, F. microcarpa, B. purpurea, and C. sinensis exhibited a stronger relationship with the two-factor combination. This variance may stem from the distinct physiological and ecological traits of the species, leading to disparities in their respective response mechanisms to temperature and moisture. Nevertheless, further exploration into the precise response mechanisms is imperative.

4.3. Methodological Considerations and Future Directions

While our study elucidates urban R_S drivers, three limitations warrant attention:

Partitioning uncertainty: Lack of autotrophic/heterotrophic R_S differentiation overestimates litter contribution [64]. Isotopic (¹⁴C) or trenching approaches are needed.

Temporal resolution: Monthly measurements may miss diurnal hysteresis effects. Continuous eddy covariance integration is recommended.

Urban gradient effects: Single-city focus limits extrapolation. Multi-city meta-analyses incorporating impervious surface coverage and pollution gradients are crucial.

Future work should prioritize the following:

(i)

Mechanistic links between urban tree functional traits and RS sensitivity;

(ii)

Long-term monitoring of legacy effects from episodic disturbances (e.g., heatwaves and pollution pulses);

(iii)

Integration with belowground C flux networks to constrain urban carbon budgets.

5. Conclusions

This study reveals the regulatory mechanism of multidimensional ecological functions of litter in urban ecosystems, especially in carbon cycle dynamics, microclimate regulation, and maintenance of ecosystem stability, which have important value. From the perspective of carbon cycling, litter input significantly increased R_S rate, with an annual average of 3.18–6.71 µmol·m⁻²·s⁻¹ under double litter treatment (DL), which increased by 23.3% and 71.0% compared to CK and RL groups, respectively. This indicates that litter, as an organic carbon sink, accelerates carbon turnover through microbial decomposition processes and enhances carbon exchange flux at the soil atmosphere interface. It is worth noting that the promotion effect of litter addition on R_S is particularly significant in spring (with some tree species having a response rate as high as 86.38%), revealing the seasonal characteristics of litter dynamics and plant growth cycles synergistically driving carbon cycling.

At the level of climate regulation function, litter significantly regulates the soil thermal environment through physical coverage effect. DL treatment increases the average temperature of the 5 cm soil layer by 0.20–1.81 °C, forming a seasonal temperature buffer layer. This thermal effect not only prolongs the microbial activity period and promotes organic matter decomposition, but more importantly, enhances the climate resilience of urban soil ecosystems by slowing down extreme temperature fluctuations. Although litter has tree species specificity in regulating soil moisture, species such as C. camphora exhibit a higher annual average of humidity maintenance ability under litter treatment, suggesting that specific tree species’ litter may optimize soil water retention performance through structural characteristics.

The differential response of Q₁₀ further highlights the ecological stability function of litter. Removing litter increased the Q₁₀ values of most tree species, indicating that the lack of litter exacerbated the sensitivity of R_S to temperature fluctuations, and the coverage of litter reduces Q₁₀ values through thermal buffering (such as the DL treatment of M. indica being 1.95), proving that it can alleviate the amplification effect of climate change on soil carbon release. This regulatory effect is closely related to the optimization of the microenvironment at the litter soil interface. The composite model shows that the combination of temperature and moisture can explain more than 80% of R_S variation, confirming that litter stabilizes carbon metabolism processes by improving the surface soil physical environment.

The results have practical implications for urban ecological management: retaining or moderately increasing litter can effectively enhance the carbon sequestration function of urban green spaces, while alleviating urban heat island phenomena through thermal buffering effects. The tree species selection strategy should prioritize species with high litter yield and strong climate regulation ability (such as C. camphora and F. microcarpa) to maximize the ecological benefits of litter. It is recommended to establish a differentiated litter management plan in urban green space management, combined with seasonal, dynamic, optimization control measures, in order to comprehensively enhance the service functions and climate adaptation capabilities of urban ecosystems. Among policies supporting urban ecological carbon management, tree species selection based on chemical regulation of litter (such as high-lignin tree species) can enhance soil carbon sequestration and provide a mechanistic basis for carbon sink assessments in urban greening and tree planting initiatives.