Climate change represents a significant challenge in the 21st century, primarily driven by unprecedented increases in greenhouse gases (GHG), such as CO2, CH4, and N2O1. Agricultural ecosystems are one of the world’s largest contributors to GHG, accounting for approximately 24% of global GHG emissions2. Changes in land-use and cropping practices significantly influence GHG emissions from agroecosystems, thereby contributing to global warming3. Agroforestry, which integrates trees or shrubs with crops or livestock, is a sustainable land-use practice with substantial potential for creating carbon sinks and reducing agricultural GHG emissions, thereby mitigating climate change4. For instance, agroforestry can reduce CH4 emissions by 44–64% compared to coffee monocultures5. Conversion of 20% (126 million hectares) of global unproductive land into agroforestry system has the potential to reduce emissions by 3.4 ± 1.7 × 109 t CO2 eq·y− 1 6. Moreover, the yield-scaled global warming potential (GWP) for coffee production was lower in agroforestry systems (0.52–0.58 kg CO2 eq·kg− 1) than in monoculture production (0.72 kg CO2 eq·kg− 1)5. Consequently, it is crucial to develop sustainable agroforestry practices that reduce GHG emissions, while maintaining crop yield and enhancing soil carbon sequestration6.

The rubber tree (Hevea brasiliensis) is native to the rainforests of the Amazon region, serving as the primary source of natural rubber and essential for economic and industrial development7. Projections indicate that rubber plantations in Southeast Asia will expand by an additional 8.5 m hectares over the next decennium8. However, the rubber monoculture practice, prevalent since the early 1900s, has led to environmental challenges, including deforestation, soil erosion, biodiversity loss, depleted carbon stock9, weakened soilʼs function as a methane sink10, and increased GHG emissions11,12,13. Integrating rubber plantations into agroforestry systems offers numerous benefits. By optimizing land use and fostering the growth of diverse plant species alongside rubber trees, these systems enhance productivity14. These systems also play a key role in carbon sequestration, helping combat climate change15. Furthermore, rubber-based agroforestry systems have played a significant role in providing a potential source of income and mitigating smallholders’ vulnerability to market fluctuations given the volatility of global rubber prices. Consequently, local governments and farmers have been increasingly interested in expanding rubber-based agroforestry systems16. In light of these considerations, adopting a rubber-based agroforestry system emerges as a crucial strategy for advancing both sustainable agricultural practices and environmental development within rubber plantations17.

China accounts for around 50% of global natural rubber consumption and concentrates explicitly on diversification initiatives in major rubber-producing areas such as Yunnan and Hainan Province18. Over the past decade, the adoption of a double-row planting system has facilitated the development of diverse rubber-based agroforestry practices19,20. Previous research on rubber-based agroforestry systems has predominantly emphasized on enhancing rubber yield21, the physiological traits of both rubber and associated crops22, soil physicochemical properties23, soil enzymes, microbial processes, and nutrient cycling20. For instance, a study conducted in Southwest China revealed that 13-year-old rubber-coffee intercropping notably increased soil carbon and nitrogen stocks compared to rubber monocultures, whereas rubber-tea and rubber-cocoa intercropping did not produce significant effects. This difference was primarily attributed to the beneficial impact of coffee litter and root inputs on soil organic carbon (SOC) pools24.

With the intensification of global climate change, researchers are increasingly interested in studying GHG emissions from soils in rubber plantations25,26. Previous research highlights that rubber plantation soils can significantly contribute to the release of carbon dioxide (CO₂), nitrous oxide (N₂O), and methane (CH₄) under specific conditions27,28,29. These GHG emissions are influenced by multiple factors such as soil temperature, moisture, vegetation types, and land management practices. Furthermore, these emissions are intricately tied to the biochemical processes and environmental conditions within rubber plantation soils30,31. Studies have also demonstrated considerable variability in GHG emission patterns based on differing habitat conditions. For example, CO₂ emissions in rubber plantation soils are primarily driven by soil respiration, which includes microbial decomposition of organic matter and root respiration25. In contrast, N₂O emissions are associated mainly with the processes of nitrification and denitrification, which are regulated by the availability of nitrogen in the form of ammonium and nitrate31. Studies have indicated that compared to conventional plantations, agroforestry systems reduce GHG concentrations in the atmosphere32. Shelter belts and other agroforestry practices can reduce CH4 emissions compared to adjacent cropped fields33. However, there is a lack of studies examining the effects of rubber-based agroforestry systems on GHG, particularly N2O, CO2, and CH4.

Soil enzyme activities play a vital role in the breakdown and synthesis of organic matter, serving as a key indicator of soil health and quality34. These activities result from crop root secretions and the stimulation of soil microorganisms, which are vital for decomposing organic matter and nutrient cycling35. Understanding these enzymes can help clarify how agroforestry systems improve soil fertility. It has been observed that intercropping increased soil enzyme activity36. Nevertheless, some studies have reported that while intercropping increases crop yields and soil nutrient availability, the activities of soil enzymes show little difference compared to monocultures37. Additionally, the responses of soil enzyme activities to intercropping seem to be specific to the type of enzyme36,37. Higher activities of enzymes like β-glycosidase (BG) and acid phosphatase (AcP) significantly affected N2O emissions. According to Shah et al.38 soil moisture, dissolved organic carbon (DOC), and pH were identified as the most important factors for CH4 and N2O emissions with soil properties having direct and indirect positive effects, while enzyme activities having positive effect. According to Piotrowska-Długosz et al.39 decreased carbon and nutrient availability with soil depth led to reduced microbial abundance and enzyme activity, affecting emission patterns through the soil profile.

Hainan Province ranks as China’s second-largest area for natural rubber cultivation, the rubber plantation areas cover over 5.4 × 105 hectares20. Rubber-based agroforestry area exceeds 5.1 × 104 hectares in Hainan Province and are distributed across 17 counties19. Several well-established rubber-based agroforestry system management models have been developed, such as rubber trees intercropping with tropical fruits, spices, beverage crops, and medicinal plants19. Previous studies have addressed GHG emissions from rubber plantation soils27,40. For example, Yang et al.27 investigated GHG emission patterns and drivers in rubber monoculture plantations. Xian et al.40 analyzed diurnal GHG emission variations in monoculture rubber plantations and intercropping systems with Citrus medica var. sarcodactylis. Nonetheless, there have been minimal investigations conducted on GHG emission patterns and their underlying mechanisms in various rubber-based agroforestry systems. In this study, six representative rubber-based agroforestry systems were selected, encompassing the intercropping patterns of rubber trees with arboreal, shrub, and herbaceous species in Danzhou, Hainan, China. Using a controlled laboratory incubation experiment comparing GHG emissions from various rubber-based agroforestry systems at two soil depths (0–10 cm and 10–20 cm). It also investigated the relationship between soil physiochemical properties, enzyme activities, and GHG emissions. The study aims to (1) assess the impact of agroforestry on soil health parameters and their influence on GHG emissions, and (2) evaluate the overall effect of integrating rubber trees and diverse crops on GWP. We hypothesize that (1) significant differences exist between rubber monoculture and agroforestry systems, with agroforestry potentially enhancing soil health and altering GHG emissions, and (2) soil enzyme activity plays a key role, as it influences nutrient cycling and organic matter decomposition, which in turn affects GHG emissions. The primary objective of this experiment was to provide a basic reference for understanding GHG emission mechanisms and optimizing rubber-based agroforestry practices in tropical regions.

Materials and methods

Site description

The research site was situated in Danzhou, located in the northwestern of Hainan Island (109° 28′ 30′′ E, 19° 32′ 47′′ N). The region experiences a tropical monsoon climate. The yearly temperature ranges from 21.5 ℃ to 28.5 ℃. The climate is characterized by two distinct seasons: a rainy season from May to October, followed by a dry season during the remaining months. The region receives an average annual precipitation of approximately 1961.1 mm, with relative humidity consistently exceeding 80% year-round. The soils in the study area are classified as having a silty clay loam texture41.

The experiment took place in a double-row rubber plantation with inter-row spacing of 20 m, intra-row spacing of 4 m, and tree spacing of 2 m, resulting in a density of 420 trees per hectare20. The rubber plantation, established in March 2002 using clone 7-20-59, began tapping in August 2010. From April through December each year, trees were tapped every 3 to 4 days, producing an impressive annual latex yield ranging between 707 kg.ha⁻¹ and 1,904 kg.ha⁻¹ from 2010 to 201820. The initial soil (0–20 cm) properties before the start of the experiment characterized by pH 4.35, soil organic matter content 12.50 g·kg⁻¹, total nitrogen content 0.40 g·kg⁻¹, available nitrogen content 19.70 g·kg⁻¹, and availabe phosporus content 19.08 g·kg⁻¹.

Based on a preliminary investigation in this region, six representative patterns in the agroforestry system were selected for this study. These patterns included Hevea brasiliensis paired with Alpinia oxyphylla (AOM), Alpinia katsumadai (AKH), Coffea arabica (CAA), Theobroma cacao (TCA), Cinnamomum cassia (CCA), and Pandanus amaryllifolius (PAR) and a rubber monoculture as control (RM). Each treatment had three replicates. Intercrops were established between the main rows of rubber trees, spaced approximately 12 to 16 m apart, with all intercrops planted in 2015. Field management practices, such as fertilization and pruning, were applied uniformly across all treatments, and no irrigation was used. In the rubber monoculture system, each tree received an annual application of 2.0 kg of synthetic fertilizer (a compound with an N: P2O5 ratio of 23:13:11) and 10 kg of organic fertilizer. Fertilizers were applied in ditches, with organic fertilizer incorporated as a basal treatment in January, while synthetic nitrogen fertilizer was applied in April, July, and September at a 2:1:1 ratio. In the agroforestry systems, fertilizer rates were 450 kg N ha⁻¹.yr⁻¹, 300 kg P2O5 ha⁻¹.yr⁻¹, and 300 kg K2O ha⁻¹.yr⁻¹20.

Soil sampling and analysis

In October 2023, soil samples from double row rubber plantation were collected from two soil depths (0–10 cm and 10–20 cm) using an auger from 6 agroforestry systems. The samples were transported to the laboratory in ice bags. A portion of the samples was immediately stored at 4 °C to measure DOC concentration, ammonium nitrogen (NH4+-N), nitrate nitrogen (NO3−-N), and enzyme activities, as well as to conduct an incubation experiment for GHG emissions analysis. The remaining samples were air-dried, ground, and sieved through a 0.149 mm mesh for physicochemical analysis.

Soil pH was determined by soil-to-water suspension at a ratio of 1:2.5, and the measurement was done with a pH meter (PHS-3E, INESA, China). SOC content was analyzed using the dichromate oxidation method42. The methods described by Lu (1999)43 were used to analyze soil total nitrogen (TN), total phosphorus (TP), total potassium (TK), available phosphorus (AP), and available potassium (AK). Soil NH4+-N and NO3−-N were extracted with a 2 M KCl solution, shaken for 30 min, and analyzed using a continuous flow injection analyzer. Soil DOC was assessed using methods44. Enzyme activities, including BG, catalase (CAT), polyphenol oxidase (POX), L-Leucine aminopeptidase (LAP), AcP, and β-1, 4-N-acetylGlycosaminidase activity (NAG), were measured fluorometrically using the method DeForest (2009) devised in polystyrene 96-well and 300-ml microplates45.

Laboratory incubation

This experiment, which was conducted over 63 days, focused on measuring the concentration of CO2, N2O, and CH4. Each soil sample weighing 150 g was placed in 250 ml glass flasks. The soil moisture was adjusted to 25% of the maximum water-filled pore space (WFPS) by adding deionized water. The flasks were pre-incubation at 30 ℃ for 7 days to achieve equilibrium. After pre-incubation, 100 g of soil was transferred to culture flasks. In addition to the 42 samples, 3 blank flasks without soil were used as controls. Incubation was performed at 30 ℃ for 63 days, with gas sampling occurring on the 1st, 3rd, 5th, 7th, 10th, 13th, 20th, 30th, 40th, and 63rd day. The flasks were sealed using butyl rubber stoppers, and the gas in the headspace was thoroughly mixed by withdrawing and rejecting it three times with a syringe. Following this, a syringe was used to withdraw 20 ml of gas and then transferred to a vacuumed vial with a capacity of 12 ml for analysis. The gas samples were analyzed for N2O, CO2, and CH4 concentrations using a gas chromatograph (Agilent 7890 A, Agilent Ltd, USA). After each sampling, the culture flasks were opened for one hour for aeration, and any lost water was replenished based on changes in flask weight.

GHG emissions

The following equation calculated GHG emissions from the soil:

$$\:F=\rho\:\frac{\varDelta\:C}{\varDelta\:t}\:\times\:\frac{273}{273+T}\times\:\:\frac{V}{W}$$
(1)

Where F is the gas emission flux, and the unit is ng (kg·h)−1; ρ is the density of gas in the standard state, the unit is kg.m− 3; △C/△t is the changing rate of gas content in serum vial; V is the effective space volume of the upper part of the serum vial (unit: m3); T is the ambient temperature, the unit is ℃; m is the dry base weight of the soil, the unit is kg.

Cumulative gas emissions are calculated according to the following formula:

$$\:C={\sum\:}_{i=1}^{n}(\frac{Fi+Fi+1}{2}\:\times\:D\times\:24)$$
(2)

Where C is the cumulative emission during gas culture, and the units of cumulative emission of N2O and CH4 are µg·kg− 1 while CO2 is mg·kg− 1, respectively; D is the number of days between samples, expressed in d; i is the number of samples.

Total GWP was computed with the equation provided by Eq. 346:

$$\:\text{G}\text{W}\text{P}=FC{O}_{2}\hspace{0.17em}+\hspace{0.17em}FC{H}_{4\hspace{0.17em}-\hspace{0.17em}C\:}\times\:\frac{16}{12}\:\times\:\:28\hspace{0.17em}+\hspace{0.17em}F{N}_{2}{O}_{-\hspace{0.17em}N}\times\:\:\frac{44}{28}\:\times\:\:265$$
(3)

Where GWP is the global warming potential of GHG in mg CO2-eq·kg− 1; FCO2 is CO2 cumulative emissions in mg CO2·kg− 1; FCH4-C is the cumulative emissions CH4 in mg CH4-C·kg− 1; and FN2O is N2O cumulative emissions in mg N2O-N kg− 1.

Statistical analysis

Preceding to performing the ANOVA (analysis of variance), the Shapiro-Wilk test was used to verify the normality of all data. Data that were not normally distributed were transformed using the natural logarithm. A two-way ANOVA was then applied to assess the impact of agroforestry systems and soil layers on soil enzyme activities and physicochemical properties. Additionally, one-way ANOVA was utilized to determine the differences in CO2, CH4, and N2O emissions from soils across various agroforestry systems, with the least significant differences determined at a 5% significance level. Graphs were made using OriginPro 2024 software (Origin Lab Corporation, USA).

Results

Soil physicochemical properties

As shown in Table 1, the soil pH of all agroforestry systems but CCA significantly lower than that of the RM. Compared with RM, all agroforestry systems, except AOM, increased SOC content in the topsoil layer, with only TCA boosting it in the subsoil. DOC content increased in the topsoil for AOM, AKH, CAA, and TCA, but declined with PAR, while in the subsoil, all agroforestry systems enhanced DOC compared to RM. TN significantly increased in AKH, CCA, and PAR in the topsoil, while AKH showed a significant decrease in the subsoil compared to RM. Compared with RM, CAA, and TCA treatments substantially increased the TP content in both soil layers, all agroforestry systems increased the AP (except for AKH in the subsoil layer) and AK (except for CCA) contents, but decreased the TK content in both soil layers. NO3−-N levels increased in both layers for AOM and in the topsoil for AKH, NH4+-N increased in AOM and AKH in the topsoil, but decreased in the subsoil for all agroforestry systems (except for TCA), compared to RM (Table 1).

Table 1 Physicochemical parameters of soil in the topsoil (0–10 cm) layer and subsurface (10–20 cm) layer across seven different planting types.
Full size table

Soil enzyme activities

In both soil layers, NAG activity significantly decreased in PAR, whereas in the subsoil layer, its activity significantly increased by 17.87% in TCA compared to RM (Fig. 1a). BG activity increased in both soil layers under the AKH, CAA, CCA, and AOM systems, but in the TCA system, this increase was observed only in the topsoil compared to RM (Fig. 1b). CAT activity was significantly higher in RM in both soil layers compared to AOM, CAA, TCA, CCA, and PAR (Fig. 1c). POX activity significantly increased in TCA and CCA in both soil layers compared to RM (Fig. 1d). AcP activity decreased substantially in the topsoil of PAR compared to RM (Fig. 1e). LAP activity increased in the topsoil in AOM, AKH, CAA, TCA, CCA, and PAR by 39.40%, 39.34%, 22.31%, 30.04%, 26.31%, and 28.41%, respectively. Similarly, AOM, AKH, and CCA also showed significant increases in the subsoil compared to RM (Fig. 1f).

Fig. 1
figure 1

Soil enzymes in topsoil (0–10 cm) layer and subsoil (10–20 cm) layer under seven planting types. Values in the bar chart are mean ± SD (n = 3). AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea arabica; TCA, Hevea brasiliensis-Theobroma cacao; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don.; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; RM, rubber monoculture. (a), β-1, 4-N-acetylGlycosaminidase activity; (b), β-1, 4-glucosidase activity; (c), Catalase activity; (d), Polyphenol oxidase activity; (e), Acid phosphatase activity (f), L-Leucine aminopeptidase. Significant variations among the treatments in soil layers are indicated by different uppercase letters (p < 0.05), while significant variations between soil layers within the same treatment are indicated by different lowercase letters (p < 0.05).

Full size image

GHG emissions characteristics and GWP

Agroforestry systems and RM affected soil CO2 emission rates, and the CO2 emission rates progressively increased from Day 1 till the end of the incubation for all soils studied in both soil layers (Fig. 2a, c). In the topsoil layer, the CO2 emission rates were significantly lower in PAR treatment throughout the incubation, while in CAA, the emissions rates were substantially higher on days 7, 10, 13, 20, 30, 40, and 63 compared to RM (Fig. 2a). In the subsoil layer, the CO2 emissions rates in the CAA were higher, whereas in CCA, TCA, and PAR systems were lower than those in RM (Fig. 2c). Similarly, compared to RM, the cumulative CO2 emissions in PAR significantly decreased by 29.4% while increased intensely by 52.9% in CAA in the topsoil layer (Fig. 2b); AOM significantly decreased by 28.47% while significantly increased by 28.84%, 51.68%, and 15.31% in CAA, CCA and PAR in the subsoil layer (Fig. 2d).

Fig. 2
figure 2

The CO2 emissions overtime and cumulative CO2 emissions from the soil during the 63-day laboratory incubation in the topsoil (0–10 cm) layer and subsoil (10–20 cm) layer under seven planting types. Values in the bar graph are mean ± SD (n = 3). (a) The CO2 emission in the topsoil layer; (b) the cumulative CO2 emissions in the topsoil layer during the laboratory incubation period; (c) the CO2 emission in the subsoil layer; (d) the cumulative CO2 emissions in the subsoil layer. AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea arabica; TCA, Hevea brasiliensis-Theobroma cacao; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don.; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; RM, rubber monoculture. The letters above the column represent significant differences among treatments at p < 0.05.

Full size image

In both soil layers, all the agroforestry systems were a net sink of CH4 (Fig. 3). Within the agroforestry systems, CH4 uptake rates remained steady from Day 1 to Day 7, then increased sharply till Day 20 for all treatments except CAA, which continued to increased till Day 30, then declined till Day 40, and then again increased till Day 63 in the topsoil layer (Fig. 3a). In the subsoil layer, the CH4 uptake rates increased from Day 1 to Day 30, then slightly decreased till Day 40, except for PAR and RM, and then again increase till Day 63 (Fig. 3c). In the topsoil layer, within agroforestry system, CCA had the maximum CH4 sink capacity, while AOM, AKH and TCA have the minimum (Fig. 3b). In the subsoil, the net CH4 sink in all agroforestry systems was lower than that of RM system (Fig. 3d).

Fig. 3
figure 3

The CH4 emissions over time and cumulative CH4 emissions from the soil during the 63-day laboratory incubation in topsoil (0–10 cm) layer and subsoil (10–20 cm) layer under seven planting types. Values in the bar graph are mean ± SD (n = 3). (a) The CH4 emission topsoil layer; (b) the cumulative CH4 emissions in topsoil layer during the laboratory incubation period; (c) the CH4 emission in subsoil layer; (d) the cumulative CH4 emissions in subsoil layer. AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea arabica; TCA, Hevea brasiliensis-Theobroma cacao; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don.; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; RM, rubber monoculture. The letters above the column represent significant differences among treatments at p < 0.05.

Full size image

All treatments affected soil N2O emissions in both soil layer, and emissions of N2O continued to increase from Day 1 till the end of incubation; all treatments were net sources of N2O emissions (Fig. 4a, c). In the topsoil, cumulative N2O emissions were significantly higher in CAA treatment while notably decreased by 41.11%, 27.76%, and 56.16% in AKH, TCA, and PAR, respectively, compared to RM (Fig. 4b). In the subsoil, the cumulative emissions of N2O were significantly increased in all treatments, except AOM, compared to RM (Fig. 4d).

Fig. 4
figure 4

The N2O emissions overtime and cumulative N2O emissions from the soil during the 63-day laboratory incubation in a topsoil (0–10 cm) layer and subsoil (10–20 cm) layer under seven planting types. Values in bar graph are mean ± SD (n = 3). (a) The N2O emission rates over time in the topsoil layer; (b) the cumulative N2O emissions from soil in the topsoil layer during the laboratory incubation period; (c) the N2O emission rates over time in the subsoil layer; (d) the cumulative N2O emissions from soil in subsoil layer. AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea arabica; TCA, Hevea brasiliensis-Theobroma cacao; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don.; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; RM, rubber monoculture. The letters above the column represent significant differences among treatments at p < 0.05.

Full size image

The Global Warming Potential (GWP) ranged from 355.0 to 2399.3 mg CO2-eq kg− 1 across the different systems, and all treatments exhibited significantly higher GWP in the topsoil than in the subsoil (Fig. 5; Table 2). In the topsoil, CAA exhibits the highest GWP, while AKH, TCA, and PAR had lower potential compared to RM (Fig. 5a), while in subsoil, all agroforestry systems except AOM had higher GWP compared to RM (Fig. 5b). Overall, at the 0–20 cm soil depth, the GWP in the CAA and CCA systems significantly increased by 36.78% and 7.18%, respectively, and the AOM, AKH, TCA, and PAR systems significantly decreased GWP by 6.61%, 24.96%, 14.24%, and 41.01%, respectively, compared with the RM system (Table 2).

Fig. 5
figure 5

GWP of soil during the 63-day laboratory incubation in topsoil (0–10 cm) layer and subsoil (10–20 cm) layer under seven planting types. Values in the bar graph are mean ± SD (n = 3). (a) The GWP in the topsoil layer; (b) the GWP in the subsoil layer. AOM, Hevea brasiliensis-Alpinia oxyphylla Miq; AKH, Hevea brasiliensis-Alpinia katsumadai Hayata; CAA, Hevea brasiliensis-Coffea arabica; TCA, Hevea brasiliensis-Theobroma cacao; CCA, Hevea brasiliensis-Cinnamomum cassia (L.) D. Don.; PAR, Hevea brasiliensis-Pandanus amaryllifolius Roxb; RM, rubber monoculture. Significant variations among the treatments in soil layers are indicated by different uppercase letters (p < 0.05), while significant variations between soil layers within the same treatment are indicated by different lowercase letters (p < 0.05).

Full size image
Table 2 Global warming potential (mg­CO2-eq·kg− 1) in different agroforestry system soils.
Full size table

Correlation analysis

The pH, SOC, DOC, TP, AP, AK, and NO3−-N contents were positively correlated with GWP and GHG emissions, except for pH with Cm-CH4. DOC, AP, and AK contents strongly correlated with GHG emissions and GWP. In contrast, TN, TK, and NH4+-N contents were negatively correlated with GWP and GHG emissions, except for NH4+-N with Cm-CO2 (Fig. 6). Enzyme activities such as CAT, POX, and LAP were generally negatively correlated with GWP and GHG emissions, except for LAP with Cm-CH4. Conversely, BG, and AcP activities were positively correlated with GWP and GHG emissions. NAG activity was positively correlated with GWP and Cm-N2O but negatively correlated with Cm-CO2 and Cm-CH4 (Fig. 6).

Fig. 6
figure 6

Pearson correlation between GWP, GHG, soil physicochemical properties and enzyme activities. GWP, global warming potential; Cm-CO2, cumulative carbon dioxide; Cm-CH4, cumulative methane; Cm-N2O, cumulative nitrous oxide; SOC, soil organic carbon; DOC, dissolved organic carbon; TN, soil total nitrogen; TP, soil total phosphorous; TK, soil total potassium; AP, Available phosphorous; AK, Available potassium; NH4+ -N, ammonium nitrogen; NO3− -N, nitrate nitrogen; NAG, β-1,4-N-acetylGlycosaminidase; BG, β-1,4-glucosidas; CAT, catalas; POX, Polyphenol oxidas; AcP, Acid phosphatas; LAP, L-Leucine aminopeptidase. * P < 0.05. ** P < 0.01. *** P < 0.001.

Full size image

Discussion

Agroforestry systems effects on soil physicochemical properties

Our findings show a notable increase in SOC, DOC, TN, TP, AP, and AK, particularly in the topsoil of agroforestry systems compared to rubber monocultures (Table 1). This suggests that agroforestry systems enhance soil microbial activity, This indicates that agroforestry systems increased soil microbial activity, accelerating the breakdown and conversion of organic matter, thereby releasing available nitrogen to satisfy the nutritional demands of both rubber trees and intercrops. The primary objective of intercropping is to enhance the soil carbon pool and increase plant diversity aboveground, which results in higher litterfall and root biomass, ultimately contributing to the enrichment of SOC24. The increase in DOC content in both topsoil and subsoil is consistent with the findings46. This increase primarily occurs by directly incorporating fresh organic matter into the soil47. The incorporation of organic residues not only improves soil physical properties, stimulates microbial and enzyme activity, fosters the formation and stabilization of soil aggregates, enhances soil structure, and promotes SOC sequestration48. Our research indicated a higher SOC in surface soil than in subsurface layers, which can be attributed to increased litter accumulation on the surface49. Soil labile carbon pools play vital roles in nutrient retention and provide energy for microbial activity50. They also increase nutrient content, improve nutrient utilization, and promote nutrient cycling in the soil51. These effects can be attributed to community composition, biomass changes, organic matter inputs, and soil microclimate structure52. During this study, the soil nutrients (TN, TP, AP, and AK) measured in agroforestry systems were increased in surface soils, which was consistent with the findings53. Changes in the physicochemical properties of soil caused by intercropping might be due to differences in the distribution of plant roots and litter cover in agroforestry systems54. Soil pH can affect microbial cells’ acid-base balance and regulate soil nutrient utilization55. In this study, soil pH was lower in the AOM, AKH, and TCA agroforestry systems than RM, suggesting that agroforestry may improve soil pH and maintain soil productivity56. Key biological functions of plant roots, such as water and nutrient uptake, respiration, and exudation, alter soil pH55,56.

Agroforestry systems effects on enzyme activities

This study reveals that rubber-based agroforestry systems, when compared to rubber monocultures, show increased BG and LAP enzyme activities. These findings are consistent with observations from other intercropping systems57, where heightened enzyme activity is often linked to improved soil health. This alignment underscores the potential of agroforestry to enhance ecosystem functions, offering a sustainable alternative to traditional rubber monoculture practices. This increase is likely due to the increased diversity of root exudates and organic inputs to the soil from different plant species. Intercropping systems that include legumes can increase LAP activity. Legumes contribute to increased nitrogen inputs through biological nitrogen fixation, which can stimulate microbial activity and enzyme production. Plants that produce carbon-rich root exudates can stimulate microbial activity and enzyme production, which is relevant for BG, which is involved in carbon cycling58. There was a strong positive correlation between NAG activity and NH4+-N, suggesting that NH4+-N significantly influences NAG activity. In order to properly understand the role of enzymes in the nutrient cycle, it is important to understand the role of enzymes in the nutrient cycle like BG is involved in the carbon cycle, NAG and LAP participate in the nitrogen cycle, and AcP is part of the phosphorus cycle59. Plants with more extensive root systems or root hairs can promote enzyme activities. For instance, root hairs have been shown to promote LAP activity in Hordeum vulgare60. Enzyme activities correlate positively with DOC, which is consistent with the findings20. Extracellular enzyme activities are considered good indicators of soil carbon decomposition. The relationship between extracellular enzyme activities and DOC suggests that these enzymes play a crucial role in breaking down organic matter and releasing DOC into the soil solution., whereas POX and CAT show different patterns61. CAT activity is important for soil microbial activity and oxidative metabolism, which relate to CO2 and N2O emissions62.

Agroforestry systems effects on GHG

Our results suggested that cumulative CO2 emissions significantly decreased in PAR while increasing intensely in CAA compared to RM (Fig. 2a). However, in subsoil, it decreased in AOM while significantly increased in CAA, CCA, and PAR compared to RM (Fig. 2b). The correlation analysis demonstrated a significant positive relationship between CO2 emissions and DOC levels in the soil. This is consistent with findings from other intercropping systems, where increased DOC concentrations are closely associated with higher CO2 fluxes63. These findings indicate that DOC plays a pivotal role in regulating soil carbon dynamics and GHG emissions within agroforestry systems. The correlation between carbon fluxes and fluctuating levels of labile soil carbon pools highlights the influence of these pools on CO2 and CH4 emissions. Studies have shown that intercropping systems enhance these labile organic fractions, primarily through the addition of carbon-rich compounds, such as root exudates and organic residues, which are key substrates for microbial processes driving CO2 and CH4 emissions. The introduction of fresh, degradable carbon compounds into the soil provides a readily available substrate for microbial decomposition, thereby stimulating soil respiration and increasing CO2 fluxes. This process potentially leads to greater carbon losses, particularly under specific intercropping conditions where microbial activity and carbon turnover are intensified. Moreover, research suggests that the duration of intercropping plays a significant role in shaping carbon dynamics, with prolonged intercropping leading to a gradual increase in CO2 fluxes. This underscores the complexity of carbon cycling in intercropped agroforestry systems, where the balance between carbon sequestration and emission is tightly linked to the interactions between plant inputs and microbial activity61,62,63. Soil CO2 emissions stem from various sources such as root respiration, plant residue decomposition, rhizosphere microbial respiration, and decomposition of soil organic matter by microbes64. Under laboratory conditions without vegetation, CO2 emissions predominantly represent microbial respiration during the decomposition of SOC65. The observed positive correlation between CO2 emissions and DOC indicates that higher levels of labile carbon enhance microbial activity, leading to increased respiration66. This suggests that as DOC availability rises, microbial decomposition processes are accelerated, resulting in greater CO2 emissions due to the breakdown of easily degradable carbon compounds. This relationship underscores the critical role of labile carbon in driving microbial respiration and carbon fluxes in soil systems.

In both soil layers, all agroforestry systems functioned as net sinks for CH4, demonstrating their capacity to mitigate methane emissions. SOC plays a crucial role in regulating CH4 dynamics within these systems. Higher SOC levels create favorable conditions for microbial communities, particularly by promoting methanotrophic activity, which facilitates methane oxidation and contributes to CH4 mitigation. Furthermore, elevated SOC can influence the microbial community structure, fostering methanogenic dominance and functional diversity. This microbial unevenness may enhance the system’s ability to sequester CH4, suggesting that SOC enrichment in intercropped soils is key to reducing methane emissions and promoting carbon stability. CH4, a significant greenhouse gas, is typically produced through the anaerobic decomposition of organic carbon67. In our study, soils were incubated at 25% WFPS under aerobic conditions, which are typically not conducive to methane production. Two key factors influence methane uptake in soils under these conditions. First, the availability of organic carbon is critical in regulating methanotrophic activity, as it provides the necessary substrates for these methane-oxidizing microbes29,67. Second, the degree of soil aeration and gas diffusion plays an essential role in controlling the transport of atmospheric CH4 and O2 into the soil, which are vital for methanotrophic processes68. Effective gas exchange ensures that sufficient methane and oxygen are available for microbial consumption, highlighting the importance of both carbon availability and soil physical properties in methane mitigation.

The cumulative emissions of N2O were significantly higher in CAA treatment while they were lower in AKH, TCA, and PAR compared to RM however in the subsoil, the cumulative emissions of N2O were significantly increased in all treatments except AOM compared to RM. The NH4+-N content in soil minerals is critical for controlling the rate of N2O production, which mainly arises from nitrification and denitrification processes69,70. Intercropping practices optimize the use of resources, including light and nutrients, by filling niche spaces that remain unutilized in traditional monoculture systems71. In this situation, soil mineral N tends to be retained due to interspecific competition for limited N resources, resulting in relatively lower N2O emissions, just like in AKH, CAA, TCA, and PAR agroforestry systems compared to RM (Fig. 5b). However, the response of N2O emissions to intercropping varies depending on different experimental conditions, with soil pH playing a significant role. Previous studies based on data analysis have highlighted soil pH as the primary factor influencing soil N2O emissions, with acidic soils as emission hotspots72. Long-term studies indicate that intercropping has a minimal effect on soil pH71, likely because soil pH influences the microbial community composition, and acidic conditions can inhibit N2O-reductase, thereby restricting complete denitrification73,74. Soil carbon availability acts as an electron donor for denitrification, and intercropping has increased soil labile carbon73. When applied to soils with high SOC content, intercropping may further boost carbon resources and enhance N2O emissions. Additionally, the ongoing increase in soil carbon storage through intercropping affects the soil C: N ratio, which could improve nitrogen immobilization rates and reduce nitrogen loss75.

Conclusions

The study of the impact of rubber-based agroforestry systems on soil health and GHG emissions provides significant insights into the global warming of various agroforestry approaches. These findings demonstrate that different agroforestry systems influence soil physicochemical properties, enzyme activities, and GHG emissions in distinct ways. SOC significantly increased in all agroforestry systems except AOM in topsoil layer. Notably TN and TP significantly increased in the topsoil across several agroforestry systems, with AP rising in both layers. All rubber plantations acted as net CO₂ and N₂O resources, and net CH₄ sinks in the 0–20 cm soil depth. Compared with the RM, the CAA and CCA systems significantly increased the cumulative CO2 and N2O emissions and GWP, whereas these parameters were significantly decreased in the AOM, AKH, TCA, and PAR systems. The soil DOC concentration was the primary factor influencing GHG emissions and GWP. Intercropping rubber trees with arboreal and herbaceous species is recommended over shrub species, considering their beneficial effects in reducing soil GHG emissions and GWP for the sustainable development of rubber plantations on Hainan Island. This study offers valuable guidance for policymakers and farmers seeking more sustainable agroforestry practices. Future studies should explore long-term effects and include additional environmental factors, such as biodiversity and soil conservation, to further optimize agroforestry systems for improved soil health and reduced GWP.