Introduction

Methane (CH4) ranks as the second most significant anthropogenic greenhouse gas, following carbon dioxide (CO2), and its presence in the atmosphere has shown a swift and notable rise during recent decades1. After atmospheric consumption by OH radicals, soil methanotrophs represent the second most substantial CH4 sink, ranging between 20 and 45 teragrams (Tg) of CH4 per year2. Forested upland soil exhibits CH4 uptake of 13.88 (9.18–17.7) Tg per year, constituting roughly 43% of the aggregate soil CH4 consumption2,3,4. However, the CH4 consumption in forest soil is not always stable. Changes in temperature and hydrology may cause forest soil to transition from a CH4 sink to a CH4 source5,6. Poplar (Populus spp.) is globally widespread and known for its rapid growth and effortless hybridization within and between species7. Globally, poplar trees cover an extensive 31.4 million hectares8, with a notable presence covering 8.25 million hectares in China9. Remarkably, poplar plantations emerge as a distinctive forest type susceptible to temperature and hydrology variations10. Therefore, tracking the soil CH4 flux in these extensively dispersed and fast-growing plantation forests is crucial for precise estimation of the CH4 budget within the forest ecosystem.

Soil CH4 exchange dynamics are shaped by two contrasting biological processes: anaerobic methanogenesis and aerobic methanotroph consumption11,12. Nonetheless, substantial ambiguity exists regarding the scale and temporal variations in CH4 emissions from forests. The net CH4 flux results from a complex interaction of factors affecting diffusion and soil oxygen levels13, including soil texture, structure, temperature, and moisture14,15. Several investigations have proposed that variations in soil temperature, soil air-filled porosity, and soil physical properties (e.g., texture) have the potential to influence the extent of CH4 uptake. However, these factors typically cannot induce a shift in the direction of CH4 flux from a sink to a source16,17. Soil gas diffusion depends on soil texture and structure, which subsequently impacts the activity and scale of soil microbial communities engaged in gas diffusion or CH4 transformation18,19. Sandy soils are generally more efficient at oxidizing CH4 than silty soils due to their better ability to diffuse gases20,21. In contrast, the gas diffusion capacity of clay loam is poorer than that of sandy loam, making it easier to form an anaerobic environment conducive to CH4 production22. Nevertheless, only a small amount of research has been conducted on the impact of soil texture on CH4 flux. Therefore, more on-site measurements are crucial for narrowing our knowledge gap.

The contribution of forest CH4 flux to the greenhouse gas budget remains surrounded by substantial uncertainty23. Micro-meteorological assessments of CH4 fluxes within forested ecosystems have revealed that certain upland canopy environments exhibit a net CH4 uptake24, whereas others function as sources of CH4 emissions over annual cycles or specific time intervals25,26. This variability could arise from intermittent CH4 emissions from soil in humid regions, potentially enabling canopy-level CH4 emissions23. The primary source of these uncertainties lies in the pronounced spatiotemporal heterogeneity of CH4 flux27. The temporal dynamics of soil-atmosphere greenhouse gas exchange exhibit variations following daily and seasonal patterns28,29. These patterns are mainly modulated by shifts in soil moisture content and temperature30. Fluctuations in precipitation and air temperature across space and time modify soil moisture and temperature dynamics, consequently influencing soil CH4 flux31.

The extrapolation of CH4 flux data from traditional static chambers to the stand level suggests that upscaling CH4 flux using the static chamber approach does not align with ecosystem-scale CH4 flux observed through eddy covariance (EC) methods23,32. Monitoring CH4 fluxes at the ecosystem scale using EC techniques has comprehensive advantages, including long-term time series, large spatial scale, and consideration of the relationship between CH4 fluxes and changes in environmental factors33. However, the EC technique is highly dependent on weather conditions and may experience gas leakage due to topography, along with limited precision in tracking minimal CH4 flux levels34,35. Static chamber measurements are an effective method for monitoring low throughput and can capture CH4 emissions at “hot spots” and “hot moments”. Although CH4 fluxes have been assessed across global forest ecosystems, only a few studies extensively track the diurnal, seasonal, and interannual changes in CH4 emissions at the soil surface level. This constraint obstructs our comprehension of the temporal changes in soil CH4 fluxes and the spatiotemporal patterns of CH4 fluxes arising from subtle shifts in environmental factors such as temperature, precipitation, and soil moisture.

Our primary objectives were to explore the temporal and spatial patterns governing soil CH4 flux at the ground scale in poplar plantations, as well as to identify the relationship between key environmental drivers and soil CH4 flux. We hypothesized that: (1) the soil CH4 flux in poplar plantations exhibits distinct diurnal and seasonal patterns, and (2) CH4 flux in soils with sandy loam texture would be lower in magnitude compared to soils with clay loam texture during the same periods.

Results

Patterns of temperature, precipitation, and soil water content in two poplar forests

Throughout 2019–2020, trends in Ta, precipitation, Ts, and SWC exhibited consistent patterns in both the Sihong and Dongtai poplar plantations (Fig. 1a,b). In 2019, the highest temperatures in the Sihong and Dongtai poplar plantations were recorded in July (28.0 ℃ for Ta and 29.4 ℃ for Ts) and August (27.0 ℃ for Ta and 27.5 ℃ for Ts), respectively. In 2020, the peak temperatures in both plantations occurred in August, with Sihong reaching 29.1 ℃ for Ta and 28.6 ℃ for Ts, and Dongtai reaching 28.9 ℃ for Ta and 28.2 ℃ for Ts. Precipitation in both plantations exhibited a unimodal pattern. In 2019, the peak monthly cumulative precipitation occurred in August, with 145 mm in Sihong and 166 mm in Dongtai. In 2020, the peak monthly cumulative precipitation in both plantations occurred in July for Sihong (357 mm) and June for Dongtai (319 mm).

Figure 1
figure 1

Monthly averages of air temperature, cumulative precipitation, soil temperature, and soil water content in Sihong (a) and Dongtai (b) poplar plantations. SH, Sihong; DT, Dongtai.Ta, Ts, and SWC represent air and soil temperature and soil water content.

Full size image

Throughout 2019–2020, the SWC in the Dongtai poplar plantation consistently exceeded that in the Sihong poplar plantation, likely due to slightly higher precipitation levels in Dongtai compared to Sihong. Generally, the SWC in both poplar plantations exhibited a seasonal trend, with lower levels during summer and higher levels during winter. Notably, during the summer of 2020, increased rainfall resulted in the SWC reaching its peak in July for both the Sihong and Dongtai poplar plantations, recording values of 0.29 and 0.40 cm3 cm–3, respectively.

Diurnal variations of soil CH4 flux in two poplar forests

In the Sihong poplar plantation, during the early growing season (March 31 to April 1) and the rapid growing season (May 29 to 30), the soil exhibited CH4 emissions throughout the day (ranging from 0.001 to 0.11 mg m–2 h–1), with peak emission rates observed at around 12:30 (0.06 and 0.11 mg m–2 h–1), followed by a gradual decline. From 18:00 until the next day at 08:00, the emission rate remained stable (Fig. 2a,b). During the late peak growing season (August 29 to 30), the soil emitted CH4 throughout the day (0.001 to 0.05 mg m–2 h–1), peaking at around 12:00, and then exhibited a fluctuating and decreasing trend (Fig. 2c). During the non-growing season (December 15 to 16), the soil did not emit CH4 but instead showed CH4 absorption throughout the day (–0.05 to –0.0001 mg m–2 h–1), with fluctuations and a peak absorption rate at 12:30 (Fig. 2d).

Figure 2
figure 2

Diurnal variation of soil CH4 fluxes in the Sihong (a–d) and Dongtai (e–f) poplar plantations at different growing periods in 2019. Positive and negative values of CH4 flux represent the CH4 emission and uptake, respectively. The triangles represent the flux values marked at 2-hour intervals. SH, Sihong; DT, Dongtai.

Full size image

Due to logistical challenges related to transportation and weather conditions, we were only able to obtain effective diurnal variation data for the Dongtai poplar plantation during the rapid growing season (May 13 to 14) and the non-growing season (December 10 to 11) (Fig. 2e,f). During the rapid growing season, the soil CH4 emission rate increased from the early morning and reached a peak of 0.07 mg m–2 h–1 at 14:30. It then gradually decreased until 18:00 and remained relatively stable until the next day at 08:00. Notably, the soil exhibited CH4 emissions during the daytime (0.0007 to 0.07 mg m–2 h–1), while showing fluctuations between emission and absorption during the nighttime (–0.02 to 0.02 mg m–2 h–1). During the non-growing season, the soil absorbed CH4 throughout the day (–0.04 to –0.0007 mg m–2 h–1), with fluctuations trend, reaching a peak absorption in the morning before sunrise.

Seasonal variations of soil CH4 flux in two poplar forests with different soil textures

From April 2019 to December 2020, we conducted 31 CH4 sampling events at the Sihong poplar plantation, with a daily average soil CH4 flux of 1.56 ± 0.50 mg m–2 d–1 (Fig. 3). In the Dongtai poplar plantation, we conducted 15 sampling events, resulting in a daily average soil CH4 flux of 0.70 ± 0.39 mg m–2 d–1. Both plantations exhibited CH4 emissions during the growing season and a weak CH4 uptake during the non-growing season. From April to December 2019, the soil CH4 flux in both plantations fluctuated slightly and showed a decreasing trend over time. However, in 2020, the soil CH4 flux in both plantations displayed a unimodal seasonal variation pattern, with a peak occurring between July and September.

Figure 3
figure 3

Daily average soil CH4 flux at different sampling days during the study period April 2019–December 2020 in the Sihong and Dongtai poplar plantations. Positive and negative data of CH4 flux stand for the CH4 emission and uptake, respectively. Error bars represent the standard error. Data from February 1 to April 30, 2020, was not gathered due to the COVID-19 pandemic. SH, Sihong; DT, Dongtai.

Full size image

A t-test revealed that the soil CH4 fluxes in the Sihong poplar plantation significantly exceeded those of the Dongtai poplar plantation from August to October and December 2019, as well as July to August and November 2020 (Fig. 4). Repeated measure analysis indicated that, in the Sihong poplar plantation, the soil CH4 fluxes in April, May, July, and September 2019 were significantly higher than those in June and August. Furthermore, soil CH4 fluxes in June and August were significantly elevated compared to October and November, with July and August 2020 also displaying significantly higher CH4 fluxes compared to January, May, June, and September to December. In the Dongtai poplar plantation, the soil CH4 fluxes from May to July 2019 were significantly higher than those from August to December. Moreover, soil CH4 fluxes between June and September 2020 exhibited a significant increase compared to January and the period from October to December.

Figure 4
figure 4

Spatiotemporal differences of monthly average soil CH4 flux in the Sihong and Dongtai poplar plantations. Positive and negative data of CH4 flux stand for the CH4 emission and uptake, respectively. The capital and lower-case letters on the error bars represent respectively the significant difference (at P ≤ 0.05) of soil CH4 fluxes in the Sihong and Dongtai poplar plantations at different months. Asterisks indicate the significant difference in soil CH4 fluxes between the Sihong and Dongtai poplar plantations in the same month (* 0.01 < P ≤ 0.05, ** 0.001 < P ≤ 0.01, * P ≤ 0.001). Error bars represent the standard error. Data from February 1 to April 30, 2020, was not gathered due to the COVID-19 pandemic. SH, Sihong; DT, Dongtai.

Full size image

Relationships between soil CH4 fluxes and environmental factors in two poplar forests

To identify the dominant driving factors influencing changes in CH4 fluxes between the soil and the atmosphere, we used a GLMM to test the effects of environmental factors on CH4 fluxes. The result indicated that soil CH4 fluxes significantly increased with rising Ts (P = 0.010) and SWC (P < 0.001) (Table 1; Fig. 5). However, precipitation did not have a significant impact on soil CH4 fluxes (P = 0.634).

Table 1 Statistical summary of generalized linear mixed model analysis of the effect of climatic and soil factors as fixed effects and soil static chambers as random effect on soil CH4 fluxes.
Full size table
Figure 5
figure 5

The effect of soil temperature (a) and soil water content (b) on the soil CH4 flux. The light gray shading indicates the 95% confidence interval. Ts, soil temperature; SWC, soil water content.

Full size image

Discussion

In this investigation, diurnal patterns displayed a single emission peak during the growing season, specifically in its early stages (March and May), occurring between 12:30 and 14:30. During the non-growing season (December), there was a solitary uptake peak detected at 12:30. (Fig. 2). These findings support our first hypothesis regarding diurnal variation trends. However, our results differ from previous observations in upland forests, which generally indicated an uptake peak around noon during the summer3.

Figure 6
figure 6

Layout of experimental design for the Sihong (a) and Dongtai poplar plantations (b), static chamber base installation (c), static chamber seal (d), and diurnal variation sampling of soil CH4 flux (e). SH, Sihong; DT, Dongtai.

Full size image
Table 3 Stand characteristics and soil properties (0–20 cm) in the two investigated poplar plantations (mean ± SD, n = 3).
Full size table

Soil CH4 gas collection and measurement

In March 2019, three random sampling points were selected within each block to collect soil CH4 gas samples in both plantations, respectively. Three cylindrical static chambers constructed from polyvinyl chloride (PVC) were randomly installed in each block in the Sihong and Dongtai poplar plantations, resulting in a total of 9 static chambers installed in each plantation (Fig. 6a,b). The static chamber system includes a base (30 cm inner diameter, 35 cm outer diameter, and 15 cm height) and a chamber (30 cm in both diameter and height). The base was inserted vertically into the ground to a depth of 10 cm, leaving 5 cm above ground, thereby creating a closed ring with distinct inner and outer diameters (Fig. 6c). During CH4 gas sampling, water was added to the rings to establish an effective seal between the base and the chamber (Fig. 6d). A stand (15 cm height) equipped with a temperature recorder (DS1923; Wdsen Electronic Technology Co. Ltd, Shanghai, China) was installed in the center of the base during the gas sampling period for calculating CH4 fluxes. Gas sampling was carried out at the center of the chamber top through a rubber stopper with a hole, using a syringe (18-gauge needle, 15 cm long) connected to a T-joint.

Specifically, from April 2019 to December 2020, gas sampling was conducted respectively for 9 chambers at the Sihong and Dongtai poplar plantations between 8:00 and 12:00 am on days without rain or snow. During each sampling day, we gathered 30 mL gas samples at intervals of 0, 20, 40, and 60 minutes and subsequently transferred them into pre-evacuated gas bottles for storage. At the Sihong poplar plantation, gas sampling frequency ranged from 1 to 4 times per month, resulting in a total of 31 on-site sampling activities and the collection of 1,116 soil gas samples. Due to traffic and accessibility constraints at the Dongtai poplar plantation, gas sampling was conducted once per month, resulting in 15 on-site sampling activities and the collection of 540 soil gas samples. Data collection was suspended from February to April 2020 owing to the impact of the COVID-19 pandemic.

Gas samples were transported to the lab and subsequently examined for CH4 concentration using a gas chromatograph (GC; 7890B, Agilent Technologies, Inc., Palo Alto, CA, USA). The GC used in the study was equipped with a pair of Porapak Q columns (each measuring 1.83 meters in length, with a 2 mm inner diameter and 80/100 mesh) alongside a flame ionization detector (FID). The column oven and detector were operated at temperatures of 60 ℃ and 250 ℃, respectively. We employed ultra-high purity nitrogen (N2, 99.999%) as the carrier gas, flowing at a rate of 30 mL min–1. Moreover, hydrogen (H2) and high-purity air (99.999%) served as the fuel and auxiliary fuel gases for the FID, with flow rates set at 40- and 400-mL min–1, respectively. The instrument was calibrated with a standard gas (10.2 ppm) both before and after each measurement. The formula for calculating soil CH4 flux51 (F, mg m–2 d–1) is as follows:

$$F=\rho \times \frac{V}{A}\times \frac{\Delta C}{\Delta t}\times \frac{273.15}{(273.15+T)}$$
(1)

Here, ρ represents the CH4 density under standard conditions (g L–1); V represents the volume of the static chamber (cm3); A denotes the surface area encompassed by the static chamber (cm2); and T stands for the temperature inside the static chamber during sampling moment (℃). Additionally, ∆C/∆t represents the linear slope of CH4 concentration change within the chamber over time (ppm min–1), with an R2 ≥ 0.9 considered valid for further analysis. The limit of detection (LOD) of CH4 flux for the GC system was 0.49 ppm, calculated using methodologies provided by Minamikawa et al.52.

Moreover, air temperature (Ta) was measured using a platinum resistance thermometer (PRT) installed on a flux (CO2 and H2O) tower. Precipitation was recorded with a tip** bucket rain gauge installed at the base of the flux tower. For missing meteorological data, we gap-filled using data from the National Meteorology Information Center (NMIC) of China. Soil temperature (Ts) and SWC were measured using thermocouples and time domain reflectometry (TDR, Trime-EZ, IMKO) probes, respectively, installed at a depth of 10 cm.

Measurement of diurnal variation of soil CH4

In 2019, we employed a swiftly deployable chamber system utilizing laser technology to measure diurnal variations in soil CH4 flux, using the Los Gatos Research (LGR; ABB, Canada) instrument. The recording frequency was set to 2 Hz (Fig. 6e). Monitoring took place during different growth stages of the poplar, including the early growing season (March–April), the rapid growing season (May–June), the late peak growing season (August–September), and the non-growing season (November–December). The measurement events were conducted on days without rain or snow in different seasons. Considering the complexities of field sampling, including the daytime heat, nighttime cold, and threats from certain animals (e.g., snakes), we selected a static chamber at each site to measure the diurnal variation of soil CH4 flux. The sampling started at around 08:00 am and ended the following day at around 08:00 am, with the static chamber sampled for 15 min each time. The CH4 concentration change rate per unit of time was calculated using the concentration values during the stable 10-minute period after excluding the first 3 min and the last 2 min of CH4 concentration values, with an R2 ≥ 0.9 considered valid for further calculating fluxes. There should be at least a 15 min interval between each measurement activity to ensure sufficient battery power and prevent the instruments from overheating and shutting down. The minimum detection limit (MDL) of the ultraportable LGR CH4 analyzer is 0.002 ppm, with an accuracy of less than 1%53.

Statistical analysis

A repeated measures analysis of variance was used to examine variations in soil CH4 flux across different periods. Subsequently, a T-test was performed to evaluate the difference in soil CH4 flux between the Sihong and Dongtai poplar plantations for the same month. When the number of repetitions differed between the two groups, specific data processing steps were taken to ensure the validity of the comparison. For the Sihong site, where CH4 flux measurements were conducted two or four times in a month, the average of these replicates was calculated for each month and considered as a single monthly CH4 flux value. For the Dongtai site, where CH4 flux measurements were conducted only once per month, this single measurement served as the monthly CH4 flux value. A generalized linear mixed model (GLMM) was used to assess the effect of climate and soil factors on soil CH4 flux, incorporating precipitation, Ts, SWC, and sampling site (SH and DT poplar forests) as fixed effects and sampling points (static chambers) as a random effect. All numerical variables were standardized (subtracting the mean and then dividing by the standard deviation) to enhance the likelihood of model convergence54. We used the R package "performance" for model diagnostics to verify the normality of residuals and the multicollinearity of variables55. Variables displaying correlation coefficients exceeding 0.7 were eliminated from the analysis56. Unless otherwise stated, the statistical analysis was conducted at a significance level of 0.05. All statistical analyses were conducted using R software version 4.3.057.