Changes in soil CH₄ fluxes from the conversion of tropical peat swamp forests: a meta-analysis

Abstract

Currently, 25% of all deforestation in insular Southeast Asia occurs in peat swamp forests. When peatlands are deforested, drainage ditches are often constructed to lower the ground-water level, which may result in changes in peat methane (CH₄) fluxes. Our aim was to evaluate how tropical peat swamp forest conversion affected these fluxes. Average CH₄ fluxes, water table depths and the relationship between both were established for common land uses (LU) of Southeast Asia. We also compared peat CH₄ fluxes before and after land-use change (LUC) using a meta-analysis. Average soil CH₄ emissions amounted to 28.6 ± 9.7 and 107.6 ± 60.2 kg C ha⁻¹ y⁻¹ in virgin peat swamp forests and rice fields, respectively. In the other land uses the emissions (9.5 ± 6.1 kg C ha⁻¹ y⁻¹) were significantly lower. Methane fluxes displayed an exponential response to water table depth changes across LU ranging from dry-drained to wet-undrained situations. The significant overall effect size of LUC on CH₄ emissions (−0.4 ± 0.2) indicated a small decrease of CH₄ emissions with peat swamp forests conversion to another land use, including rice cultivation. It's important to stress, however, that the overall decrease in CH₄ emission from peat swamp forests conversion would never offset the simultaneous increase in soil CO₂ emissions due to accelerated peat decomposition.

Keywords:

• carbon
• climate change
• greenhouse gas
• land-use change
• Southeast Asia

1. Introduction

Despite covering only about 0.25% of the Earth's land surface, tropical peatlands contain around 3% of the global soil carbon (C) stocks and at least 20% of global peat C (Page et al. 2004; Page and Banks 2007). The largest area of tropical peatlands occurs in Southeast Asia, where they are found in Indonesia, Malaysia, Brunei, and Thailand (Rieley and Ahmad-Shah 1996). In their natural state, lowland tropical peatlands support a growth of swamp forest overlying peat deposits up to 20 m thick (Page et al. 1999). Peat soils are characterized by high C contents over the full depth of the peat deposit and very low bulk densities (<0.2 g cm⁻³) (Andriesse 1988). Anaerobic conditions in the soil limit decomposition of the litter, leading to peat accumulation and also channel a small fraction of the excess organic matter (OM) into methane (CH₄) (Jauhiainen et al. 2005), which is a greenhouse gas (GHG) with a global warming potential (GWP) 25 times stronger than carbon dioxide (CO₂) (Forster et al. 2007) over a 100 year time horizon.

Between 2000 and 2010, 25% of all deforestation in insular Southeast Asia occurred on peatlands (Miettinen et al. 2011). This deforestation was driven by wood production and demand for land on which to establish small- and large-scale agriculture including oil palm and timber plantations (Hooijer et al. 2006). Accessible peat swamp forests of Southeast Asia are often logged, legally or not. Cutting canals in order to extract the logged wood is a widespread practice in the region that results in a subsidence of the peat dome (Kool et al. 2006), a fall of the water tables and consequently induces changes in the carbon cycle. When peatlands are reclaimed for other uses, the forest is cleared and the land is prepared for cultivation, often using fire. Drainage ditches are often constructed to lower the ground-water level and the soil may be compacted with heavy machinery to allow anchorage of trees and to increase the bearing capacity of the soil (Andriesse 1988; Wösten et al. 1997). Increased aeration due to drainage may result in decreased CH₄ production and increased CH₄ consumption in peat soils (Melling et al. 2005). Large areas of tropical peatlands will continue to be cleared to establish oil palm and Acacia plantations (Barr 2001; Miettinen 2004; Hooijer et al. 2006; Germer and Sauerborn 2008) and to a lesser extent, sago palm and rubber.

Our aim was to evaluate how land-use change (LUC) in tropical peat swamps of Southeast Asia affected soil fluxes of CH₄. As a first step, we reviewed studies on soil CH₄ fluxes of seven land-use (LU) types prevalent in Southeast Asia: virgin peat swamp forest, drained forest, fire-damaged forest, mixed croplands & shrublands, rice fields, oil palm, and sago palm plantations. From this compilation, we examined relationships between soil CH₄ fluxes and environmental variables and proceeded to a meta-analysis in order to compare CH₄ fluxes before and after LUC.

2. Materials and methods

2.1. Data collection, calculation, and presentation

We collected data from peer-reviewed publications in scientific journals. In most cases, the fluxes were measured monthly during one or two years (Table 1) and reported as annual budgets. Whenever the fluxes were reported per unit hour, these were extrapolated to a full year by considering 365 days within a year. Mean annual CH₄ fluxes and associated water table depths were calculated for the different LU treatments. We also explored relationships between soil fluxes of CH₄ and two important abiotic factors, soil temperature and water table level. Given the high variation in the responses of soil CH₄ fluxes to LUC among sites, we used a meta-analysis statistical approach to compare CH₄ fluxes before and after LUC. We used data from studies with paired observations on the same site. Six bibliographic references were included in the meta-analysis, considering 16 case studies of conversion from a virgin peat swamp forest to another LU. In the meta-analysis, the control treatment was the virgin peat swamp forest and the “other LU” treatment included all other LU than virgin peat swamp forest.

Table 1. Annual soil fluxes of CH₄ (standard deviation SD and n number of average values used to calculate annual fluxes) in different land uses (LU) on tropical peatlands.

					(kg C ha^−1 y^−1)
References	Country	Sf^a	Sl^b (months)	LU	LUT^c	CH4	SD	N	Altitude (m)	Rainfall (mm y^−1)	Ta^d (°C)	Ts^e (°C)	WT^f (cm)
Furukawa et al. (2005)*	Indonesia	1	12	Forest	F	63.1	55.7	3	10	2582	26.7	–	3.3
Hadi et al. (2005)*	Indonesia	12	11	Forest	F	44.7	79.6	11	3	2500	26.5	–	0
Hirano et al. (2009)*	Indonesia	12	12	Forest	F	10.6	5.5	3	30	1852	26.7	–	–
Hirano et al. (2009)*	Indonesia	12	12	Forest	F	12.8	6.0	3	30	2292	26.4	–	–
Hirano et al. (2009)*	Indonesia	12	12	Forest	F	13.0	6.1	3	30	2560	25.9	–	–
Inubushi et al. (1998)*	Malaysia	12	24	Forest	F	72.3	40.1	4	–	–	–	–	–
Inubushi et al. (2003)*	Indonesia	12	12	Forest	F	12.0	4.0	3	3	3133	26.5	–	−10
Melling et al. (2005)*	Malaysia	12	12	Forest	F	0.2	0.4	12	–	2163	27.2	25.7	−45.3
Furukawa et al. (2005)*	Indonesia	12	12	Drained forest	DF	8.8	9.1	12	10	2582	26.7	25.5	−18.1
Furukawa et al. (2005)*	Indonesia	12	12	Cassava	C&S	25.4	54.6	12	10	2582	26.7	27.1	−23.5
Furukawa et al. (2005)*	Indonesia	12	12	Upland rice	R	27.2	24.3	12	10	2582	26.7	27.2	−13
Furukawa et al. (2005)*	Indonesia	12	12	Lowland rice	R	371.4	752.6	12	10	2582	26.7	27.1	5.2
Furukawa et al. (2005)*	Indonesia	1	24	Coconut	C&S	−0.9	0.9	2	10	2582	26.7	–	−43
Furukawa et al. (2005)*	Indonesia	1	24	Pineapple	C&S	0.0	0.0	2	10	2582	26.7	–	−27
Furukawa et al. (2005)*	Indonesia	1	12	Pineapple	C&S	0.9	4.6	3	10	2582	26.7	–	−50
Hadi et al. (2005)*	Indonesia	12	12	Rice	R	196.0	291.7	11	3	2500	26.5	–	−2
Hadi et al. (2005)*	Indonesia	12	12	Rice–soybean	R	26.0	43.7	11	3	2500	26.5	–	–
Hirano et al. (2009)*	Indonesia	12	12	Drained forest	DF	8.3	3.2	3	30	2560	25.9	–	–
Hirano et al. (2009)*	Indonesia	12	12	Drained forest	DF	9.4	3.8	3	30	2331	26.3	–	–
Hirano et al. (2009)	Indonesia	12	12	Abandoned cropland	C&S	0.8	–	–	30	2331	26.3	–	–
Inubushi et al. (1998)*	Malaysia	12	24	Sago palm	S	90.7	53.9	10	–	–	–	–	–
Inubushi et al. (2003)*	Indonesia	12	12	Abandoned cropland	R	6.0	7.0	3	3	3133	26.5	–	−15
Inubushi et al. (2003)*	Indonesia	12	12	Abandoned rice	R	19.0	5.0	3	3	3133	26.5	–	2.1
Melling et al. (2005)*	Malaysia	12	12	Sago palm	S	0.2	2.3	12	–	2928	32.1	27.8	−27.4
Melling et al. (2005)*	Malaysia	12	12	Oil palm	OP	−0.2	0.5	12	–	2471	30.5	27.8	−60.2
Jauhiainen et al. (2008)	Indonesia	12	12	Drained forest	DF	−1.6	0.6	4	30	2560	25.9	–	−47
Jauhiainen et al. (2008)	Indonesia	12	12	Drained forest	DF	−2.8	0.7	4	30	2331	26.3	–	−43
Jauhiainen et al. (2008)	Indonesia	12	12	Burnt forest	FF	1.5	1.0	3	30	2560	25.9	–	33
Jauhiainen et al. (2008)	Indonesia	12	12	Burnt forest	FF	2.1	1.1	3	30	2331	26.3	–	−21
References followed by * were included in the meta-analysis. ^aSampling frequency per year. ^bLength of the experiment. ^cLand-use type classification used for the calculation of mean annual CH4 fluxes (F: virgin peat swamp forest, FF: fire-damaged forest, C&S: mixed croplands and shrublands, DF: drained forest, R: rice fields, OP: oil palm plantation, S: sago palm plantation). ^dAir temperature. ^eSoil temperature. ^fWater table depth.

2.2. Statistics analysis

Statistical analysis was performed using the software InfoStat (2004), with a probability level of 0.05 to test the significance of the treatments effects. For multiple comparisons between LU types, the non parametric Kruskal–Wallis test was performed since CH₄ fluxes were no-normally distributed. The distribution of CH₄ fluxes was tested using the Shapiro–Wilks test.

Meta-analysis was used to evaluate the response of CH₄ fluxes to LUC in tropical peatlands. Only the studies comparing a virgin peat swamp forest to another land use on the same site were included in the analysis. The magnitude of the effect of LUC on CH₄ fluxes was evaluated using the Hedges's g metric (bias-corrected standardized mean difference) as defined by Borenstein et al. (2009). Positive Hedge's g values indicated that LUC increased the value of the variable with respect to that in virgin peat swamp forest; negative values indicated that LUC decreased the value of the variable. According to Borenstein et al. (2009), Hedge's g values of 0.2 or less indicate a small effect size; values around 0.5 indicate a medium effect and 0.8 or above indicate a large effect size. The overall effect size was calculated using a random effects model which allows that the true effect could vary from study to study (Borenstein et al. 2009), rather than using a fixed effect model for which the true effect size is assumed to be shared by all the included studies. A t-test was used to assess the significance of individual and overall LUC effect sizes on soil fluxes of CH₄. The meta-analysis was performed with the software Comprehensive Meta Analysis version 2.2.048 (Biostat Inc., New Jersey, USA).

Throughout the manuscript uncertainties, estimates are reported as standard errors except those of CH₄ fluxes from individual studies, which are reported as standard deviations for the purpose of the meta-analysis.

3. Results and discussion

Mean rates of soil CH₄ emissions in virgin peat swamp forests from all available studies (Table 1) ranged from 0.2 to 72.3 kg C ha⁻¹ y⁻¹, with a mean rate across all sites of 28.6 ± 9.7 kg C ha⁻¹ y⁻¹ (n = 8). In rice fields, emissions of CH₄ amounted to an average value of 107.6 ± 60.2 kg C ha⁻¹ y⁻¹ (n = 6). In the other LUs, mean soil CH₄ fluxes (9.5 ± 6.1 kg C ha⁻¹ y⁻¹, n = 15) were significantly lower (P = 0.0033) than in virgin peat swamp forests and rice fields. Mean soil temperature was higher in the other LUs (27.1°C) than in the virgin peat swamp forest (25.7°C) and mean water table level was the highest in the rice fields (−4.5 ± 4.0 cm), followed by the virgin peat swamp forest (−13 ± 11.1 cm) and other LUs (−35.8 ± 4.1 cm). Methane fluxes displayed an exponential response to water table depth changes across LU (Figure 1) indicating increased emissions with decreased drainage. The effect size of LUC on emissions of CH₄ (Figure 2) ranged from −2.2 ± 0.7 (conversion of virgin peat swamp forest to a drained forest) to 1.2 ± 0.7 (conversion of virgin peat swamp forest to a rice field). Positive effect sizes (i.e. increased CH₄ fluxes) were associated with conversions to rice fields and a sago palm plantation, corresponding in most cases to increased water table levels. The overall effect size was −0.4 ± 0.2 and was significantly different from zero (P = 0.0387) indicating a small decrease of CH₄ emissions with the conversion of virgin peat swamp forests to another LU, including rice cultivation.

Figure 1. Annual peat CH₄ fluxes in Southeast Asia as a function of the water table depth across land-uses ranging from dry-drained to wet-undrained situations. The solid curve indicates the relationship based on measurements (solid diamonds). The equation of the relationship, the R ² of the linear regression between observed and predicted annual peat CH₄ fluxes, and the number of observations (n) are specified in the top left corner.

Figure 2. Mean effect size (Hedges's g) and standard error of individual (open circle) and overall (closed circle) land-use change on soil emissions of CH₄ in tropical peatlands. The land-use change types include peat swamp forest conversion to croplands & shrublands (F to C&S), drained forest (F to DF), oil palm plantation (F to OP), rice fields (F to R), and sago palm plantation (F to S).

Our study confirms that soil water table depth is one of the key factors governing soil emissions of CH₄ as observed elsewhere in northern (Lai 2009) and tropical (Jauhiainen et al. 2005) peatlands. The overall decrease in soil CH₄ emissions resulting from LUC in tropical peatlands may, however, also be associated with changes in other factors such as soil temperature and vegetation type. Currently, available data on CH₄ fluxes in tropical peatlands are yet insufficient to proceed to such an analysis. In order to assess how deforestation of tropical peat swamps may affect climate change, it is necessary to compare the overall decrease in soil CH₄ emission observed here to the associated changes in soil nitrous oxide (N₂O) and heterotrophic CO₂ emissions. Autotrophic soil CO₂ emissions also called root respiration are already integrated in biomass C stocks assessments generally used for quantifying CO₂ losses from the vegetation. Very few studies have quantified all three GHG in changing land uses on tropical peat and almost none has partitioned soil respiration into autotrophic and heterotrophic components. Therefore, much more research is needed in this area. Using data from Melling et al. (2005), the conversion of a virgin peat swamp forest into an oil palm plantation on peat fertilized at a rate of 100 kg of N ha⁻¹ y⁻¹ would represent a decrease in soil CH₄ fluxes of 0.33 kg C-CH₄ ha⁻¹ y⁻¹ or 0.01 Mg of CO₂eq ha⁻¹ y⁻¹ (applying the CH₄ GWP of 25 over a time horizon of 100 y (Forster et al. 2007)). The same LU conversion leads to an increase in soil N₂O emissions of 0.5 kg N-N₂O ha⁻¹ y⁻¹ (Melling et al. 2007) or 0.23 Mg of CO₂eq ha⁻¹ y⁻¹ (applying the N₂O GWP of 298 over a time horizon of 100 y (Forster et al. 2007)). With increased soil N₂O emissions 20 times larger than the respective decreased soil CH₄ emissions, when both expressed in CO₂ equivalent, it's clear that converting the peat swamp forest into the oil palm plantation is harmful for the climate. A comparison between mean annual soil CO₂ and CH₄ emissions in six LU types prevalent in Southeast Asia demonstrated that the decrease in CH₄ emissions arising from deforestation of peat swamp forest was negligible compared with the corresponding release of CO₂ into the atmosphere due to intense decomposition of the peat in aerobic conditions (Hergoualc'h and Verchot 2011). Finally, there is evidence of significant CH₄ emissions from ditches in northern drained peatlands. Such measurements are still inexistent in the tropics. Emissions from this source can approach or even exceed those from the soil in undrained peatlands (Roulet and Moore 1995) and should, therefore, be taken into account in GHG inventories.

4. Conclusion

The fate of tropical peat swamp forests is a major concern within the framework of climate change because of the high amount of carbon they currently store and could carry on storing, and the consequences of LUC for GHG release into the atmosphere. These ecosystems are, therefore, an important issue for climate change mitigation mechanisms, such as reducing emissions from deforestation and forest degradation (REDD). Quantifying GHG emissions from LUC requires studies on all three GHG (N₂O, CH₄, and CO₂) achieved simultaneously in changing LU on tropical peat. The meta-analysis achieved with the few studies available in the literature indicated a small overall decrease of soil CH₄ fluxes due to conversion of virgin peat swamp forests. Such an analysis on the response of soil N₂O emissions to LUC in tropical peatlands is still inexistent and would be very useful. Finally, it is important to underline that the overall decrease in CH₄ emission from peat swamp forests conversion would never offset the corresponding increase in soil CO₂ emissions due to accelerated decomposition.

Acknowledgments

This work was generously supported by the contributions of the governments of Australia (Grant Agreement # 46167), Finland (Grant Agreement HELM023-29) and the European Commission (REDD-ALERT project, Grant Agreement # 226310) to the Center for International Forestry Research. We would like to thank all researchers that kindly provided material and/or explanation used for this work. The constructive comments of the anonymous reviewers greatly improved the paper and are much appreciated.