A versatile gas flux chamber reveals high tree stem CH₄ emissions in Amazonian peatland

Highlights

• New flexible, light-weight flux chamber for campaigns of diverse stem sizes & shapes.

• Steady-state stem gas flux method can provide higher accuracy than non-steady state.

• Results are from the first multiyear tree-flux study in Amazonian peatlands.

• Mauritia flexuosa CH₄ emissions are comparable to highest known tree stem emissions.

• Dominant in Amazonian peatlands, M. flexuosa stems are likely a major source of CH₄.

Abstract

Tree stems can be a major source of CH₄, altering the impact forests have on atmospheric radiative forcing. In the tropics stem flux monitoring has been limited, nevertheless, available field studies have observed high variability of stem CH₄ flux rates between species and individuals and over the surface of individual tree stems. To evaluate the sources of variation and controls of CH₄ stem fluxes, methods supporting large sampling campaigns are needed. We designed a portable, flexible, easily installed chamber that produces replicable flux measurements across broad species diversity. The chamber creates a strong seal on a range of stem sizes and surface roughnesses and extends radially, integrating surface flux variation so that small hot spots are not missed. Working in forested Amazonian peatlands, we used this chamber to measure stem fluxes on a variety of palms and trees with stems ranging from 10 cm to 85 cm DBH, both rough and smooth barked species. We compared a non-steady state, headspace recirculation method to a steady-state flow-through method and found the flow-through method likely yields more accurate estimates for high emissions. Our novel stem flux measurements from Amazonian peatlands reveal that stem CH₄ emissions of the dominant palm (M. flexuosa) contrast markedly against dominant hardwood species, with palms emitting CH₄ at much higher rates (up to 84 mg-C m⁻² h⁻¹ at 0.23 m stem height) that compare to the highest published rates of tree stem emission. Flux rates from M. flexuosa demonstrated exponential decay with stem height, rates at 0.5 m were 3 to 10 times higher than at 1.4 m. This pattern underscores the importance of quantifying CH₄ flux rates over the lower 2 m of stem in order to accurately quantify stem CH₄ fluxes. Considering their broad distribution and high density in Amazonian peatlands, M. flexuosa stems may be a major source of atmospheric CH₄.

Introduction

Tropical peatlands are widely distributed in forested regions across the Americas (480,000 km² mainly in Amazonia), South East Asia (220,000 km²), and Africa (190,000 km²) according to latest estimates (Gumbricht et al., 2017; Xu et al., 2018; Ribeiro et al., 2020). In the Amazon basin, peatlands are concentrated in the subsiding Pastaza-Marañón Foreland Basin (PMFB), a 120,000 km² region estimated to hold more than 3 Gt C in peatlands (Lähteenoja et al., 2012) including large areas of palm-dominated (palm swamp, 27,700 km²) and tree-dominated (pole forest, 3700 km²) sites (Draper et al., 2014). Peatland plant diversity in the PMFB is low within a site (alpha) but high across sites (beta; Draper et al., 2018), leading to forest stands where one or two species can dominate plant-associated biogeochemical processes. However, the effects of plant type (palm or tree) on carbon cycling and greenhouse gas fluxes has received little attention, particularly in the case of CH₄ emissions. Studies of CH₄ emissions from Amazon peatlands have centered in soil flux (Teh et al., 2017; Winton et al., 2017; van Lent et al., 2019; Finn et al., 2020; Hergoualc'h et al., 2020) but none has reported the contribution of vegetation to CH₄ flux.

Because palm and tree stems have not been included in peatland CH₄ measurements, a potentially large source of Amazon CH₄ may have been overlooked. In the PMFB peatlands stem density can be greater than 1700 stems ha⁻¹ in pole forests and 600 stems ha⁻¹ in palm swamps (Draper et al., 2018). Recent studies have revealed that tree stems can be CH₄ sources in ecosystems representing much of the world's forested area, including, tropical (Pangala et al., 2013; Jeffrey et al., 2020), temperate (Pitz and Megonigal, 2017; Barba et al., 2019), and boreal (Machacova et al., 2016) climates, and both flooded (Pitz et al., 2018; Schindler et al., 2020) and upland (Pitz and Megonigal, 2017; Plain et al., 2019) hydrologic regimes. But nowhere have emissions been as large as Amazon floodplains where tree stems contribute 3% of global CH₄ emissions (Pangala et al., 2017). While the soils of these floodplains are seasonally flooded and primarily act as the source of tree stem emissions during the inundation period, Amazonian peatlands can be saturated most of the year, resulting in high potential CH₄ production in the soil (Kelly et al., 2014; Finn et al., 2020) as a source to the stems all year long. Additionally, studies of Indonesian (Pangala et al., 2013) and Malaysian (Wong et al., 2018) tropical peat swamps found that soil fluxes only appear to account for 40% and 13% of the observed ecosystem CH₄ emission, respectively, suggesting that tree stems fluxes may be the primary CH₄ source. Therefore, it is likely that stems are also a significant CH₄ source in the understudied Amazonian peatlands.

To determine if tree stems are an important CH₄ source in Amazonian peatlands, stem CH₄ fluxes should be measured over variation in factors that are known to have strong impacts on stem fluxes in other ecosystems, including species, season, and stem height. There is consistent, significant variation in stem CH₄ fluxes between species in both tropical (Pangala et al., 2013; Sjögersten et al., 2020) and temperate (Pitz et al., 2018) climates. Also, CH₄ stem fluxes can vary substantially over season due to the range of environmental properties with seasonal dynamics and this variation can confound comparisons of species or individuals. Flux measurements made over multiple seasons are needed to make robust comparisons of differences between individuals and species. Many studies only contain measurements of stem flux at a single height between 10 cm and 150 cm up the stem (e.g., Terazawa et al., 2015; Machacova et al., 2016; Pitz et al., 2018; Plain et al., 2019), however, there is growing evidence of the importance of characterizing a height profile of stem fluxes starting at the base of the stem where fluxes are often the highest (Pangala et al., 2013; Wang et al., 2016; Pitz and Megonigal, 2017; Schindler et al., 2020). For example, fluxes at 20 cm height can exceed the CH₄ emission rate of fluxes at 100 cm height by a factor of 1000 or more (Jeffrey et al., 2020). In order to analyze fluxes over all these factors, measurements must be collected from enough stems to provide a dataset with high statistical power. Such large research projects require methods optimized for speed and cost.

Chamber flux methods have been adapted by tree flux researchers from soil gas flux methodology (Parkin and Venterea, 2010) and chamber design has a dramatic effect on the accuracy and uncertainty of soil trace gas flux measurements, causing variation in flux estimates of at least 50% (Pihlatie et al., 2013). Stem chamber designs are being actively developed, tested, and improved with a variety of designs currently employed ranging considerably in size, installation speed, and materials used. Stem flux chambers generally fall into one of four types: 1, small, simple chambers made from a piece of wide rigid tubing (often polyvinylchloride) that contact a limited area of the stem (Marthews et al., 2014; Jeffrey et al., 2020), 2, large rigid box chambers that encompass the whole stem over a length of 20+ cm (Gauci et al., 2010; Pangala et al., 2013; Wang et al., 2016), 3, rigid partial boxes that require a curved edge to seal against the stem and cover more area than type 1 but less than type 2 (Terazawa et al., 2015; Pitz and Megonigal, 2017), and 4, chambers constructed of a sheet of flexible material for the chamber ‘lid’, allowing for a close fit within a few cm of the stem, these can be provide partial coverage of the stem (Machacova et al., 2016; Siegenthaler et al., 2016) or can encompass the stem radially. Type 1 chambers are highly portable, inexpensive, and fit many stem morphologies but their small area coverage requires that many flux measurements must be made in order to characterize a single stem's flux accurately. Type 2 chambers integrate a large stem area in a single measurement but requires significant customization during installation, are difficult to use on large stems, and are not highly portable or inexpensive. Type 3 chambers share the limitations of types 1 and 2. Type 4 chambers integrate over a large area of a stem, can install quickly on a range of stem types, and are inexpensive and portable. However, the narrow headspace gap, and thus the chamber volume, of type 4 chambers can be unpredictable on undulating stem surfaces so these chambers may not perform well over a wide range of stem morphologies. As these examples show, existing chamber designs were generally for a specific application or research question and most are not ideally suited to collecting fluxes from a large number of trees. If researchers aim to collect thousands of fluxes each season, they will need a robust, simple approach that is repeatable and accessible globally. We believe the following criteria are necessary chamber properties to enable collection of fluxes on the scale needed to characterize tree stem greenhouse gas fluxes:

• Portable and lightweight for ease of transport.

• Easy and quick to install for speed and efficiency of sampling.

• Provides coverage around the stem to capture radial flux variability.

• Flexible structure to fit trees of many diameters.

• Forms a strong seal on both smooth and rough bark species.

• Consistent or predictable surface area and volume so that dimensional measurements are not needed for each installation.

• Moderate or high ratios of surface area to volume to detect low fluxes.

• Constructed of widely available, inexpensive materials so many can be produced for large campaigns and in developing countries.

Soil trace gas flux research has been heavily dominated by the use of non-steady-state methods for decades due to the high sensitivity and lower variability compared to steady-state approaches (Livingston and Hutchinson, 1995; Hutchinson and Livingston, 2001) while leaf surface exchange has been dominated by steady-state approaches, such as flow-through methods (Field et al., 1989). Flow-through methods can have advantages of quicker results and lower impact on the diffusion gradients whose preservation is important for accurate flux estimation. In a typical flow-through technique a closed chamber has a consistent supply of known concentration gas added to the headspace that mixes with the headspace gas, that headspace then vents to a detector (LiCor 1998), only headspace equilibrium (steady-state) concentration, flow rate, and inlet gas concentration are needed to calculate a flux rate. Modern high-frequency laser spectroscopes are well-suited for flow-through measurements.

We designed and constructed a novel flexible, robust, lightweight (< 500 g) chamber capable of rapid installation on a range of stem sizes using inexpensive materials that are available widely. The dimensions of this design allow it to naturally integrate over radial variation in stem fluxes and maintain a high flux sensitivity. We utilized this chamber on a variety of tree and palm stems with DBH between 10 cm and 85 cm and a range of stem roughness and surface morphology. To estimate fluxes, we utilized both a headspace re-circulation (standard non-steady state static chamber flux) and a flow-through (steady state static chamber flux) method to test the efficacy and variability of the flow-through method for tree fluxes. This is the first published study testing a steady state (flow-through) method on tree stems, other studies have used re-circulation methods or other non-steady state approaches.

We applied this novel chamber design to our measurement campaigns over multiple years to characterize the stem fluxes of dominant plant species in Amazonian peatlands. As we report the first CH₄ stem fluxes from the trees and palms of this ecosystem, we sought to assess the main sources of flux variation that describe the magnitude of stem influence on ecosystem CH₄ flux. Specifically, we worked to address these questions:

• Considering the fundamental differences in palm and tree stem architecture, do fluxes from these groups differ?

• How consistent are stem fluxes over seasons?

• How are fluxes from peatland plant stems affected by vertical location on the stem?