Forest health, heart rot disease, and their impact on the source of carbon‐based greenhouse gas fluxes
Chathuranga K. Senevirathne, Alan Huff, Debit Datta, Nathan G. Swenson, Adrian V. Rocha
TL;DR
Heart rot disease in maple trees increases methane emissions from stems, affecting carbon cycling and greenhouse gas dynamics in forests.
Contribution
The study reveals that heart rot shifts tree stems from methane sinks to sources, with implications for atmospheric greenhouse gas dynamics.
Findings
Heart rot increases stem CH4 emissions but does not affect CO2 or soil gas fluxes.
Severe heart rot causes bark fractures, enhancing CH4 diffusion and creating emission hotspots.
Methanogens are present in all stems, with CH4 produced in heartwood and CO2 in sapwood.
Abstract
Forest health is critical for sustaining ecosystem services like carbon sequestration. Heart rot, a widespread disease in upland northern hardwood forests, may affect greenhouse gas (CO2 and CH4) fluxes, but its impacts remain poorly measured.Using non‐destructive tomography and direct gas flux measurements, we quantified the effects of heart rot on sugar maple (Acer saccharum Marshall) stems and surrounding soils.Heart rot increased CH4 emissions from stems but did not affect CO2 fluxes from stems or soils, nor CH4 fluxes from soils. All stems emitted CO2 and CH4, while soils absorbed CH4 and emitted CO2. Stem CH4 fluxes strongly correlated with decay severity, but CO2 fluxes did not. CH4 was produced in the heartwood, CO2 in the sapwood, and methanogens were present in all stems. Severe heart rot often caused bark fractures, enhancing CH4 diffusion to the atmosphere and creating…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8- —Division of Environmental Biology10.13039/100000155
- —Advanced Exploration Systems10.13039/100016389
- —Merrilee Clark Redmond Endowment for Excellence, University of Notre Dame Environmental Research Center (UNDERC)
- —Bernard J. Hank Family Endowment, University of Notre Dame Environmental Research Center (UNDERC)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAtmospheric and Environmental Gas Dynamics · Soil Carbon and Nitrogen Dynamics · Plant responses to elevated CO2
Introduction
The increasing rate of carbon‐based greenhouse gas accumulation in the atmosphere has highlighted a need to understand their sources and sinks (IPCC, 2013; Sonwani & Saxena, 2022). Forests are atmospheric carbon‐dioxide (CO_2_) sinks when photosynthesis exceeds both autotrophic respiration from vegetation and heterotrophic respiration from the soil microbial community (Pan et al., 2011; Xu et al., 2025). Some forests are important sources of methane (CH_4_) to the atmosphere via soil and canopy methanogenesis (Covey et al., 2012; Wang et al., 2016; Feng et al., 2023; Forster et al., 2023; Barba et al., 2024; Leung et al., 2026), a potent greenhouse gas with a global warming potential c. 28 times greater than CO_2_ over a 100‐yr time scale (IPCC, 2013). Minimizing forest emissions of carbon‐based greenhouse gases remains a high priority for climate mitigation, and some have proposed that this goal could be attained by focusing on forest health (Trumbore et al., 2015). Healthy forests with low rates of disease are thought to maximize the sink strength of forests by reducing autotrophic and heterotrophic carbon (C) sources and enhancing or maintaining photosynthetic C sinks (Hu et al., 2018; Pan et al., 2024). However, the role of disease in regulating forest carbon‐based greenhouse gas fluxes has yet to be fully explored and remains a large uncertainty in predicting future climate (Covey & Megonigal, 2019).
Although there remains much nuance in characterizing and quantifying forest health (O’Laughlin et al., 1994), it is clear tree disease has a negative impact on physiological processes that regulate both forest CO_2_ and CH_4_ fluxes (Meyer, 1998; Visakorpi et al., 2018; Brodribb et al., 2019). Disease can cause localized stem damage or whole tree mortality that decreases C gains through photosynthesis through the loss of leaf area and leaf chlorophyll content (Ryan, 1991; Talebzadeh & Valeo, 2022). However, the impacts of disease on forest C losses are less understood. On the one hand, C losses from autotrophic respiration may decrease in response to lower root and shoot carbohydrate concentrations from lower photosynthetic rates. On the other hand, C losses from heterotrophic respiration may increase through increased C substrates from increased branch mortality and litterfall (Zhao et al., 2017; Feng et al., 2022, 2024). Increases in C substrates may stimulate methanogen activity, thereby increasing CH_4_ fluxes, especially in anoxic environments, such as waterlogged soils (Conrad, 2009; IPCC, 2013; Gauci et al., 2024). Whether forest diseases increase or decrease carbon‐based greenhouse gas fluxes remains uncertain, largely due to the limited number of case studies and model systems, as well as a poor understanding of how disease alters autotrophic and heterotrophic respiration.
Here we investigate heart rot disease as a case study on disease impacts on soil and tree carbon‐based greenhouse gas fluxes. Heart rot is a common fungal disease in northern hardwood forests that results in internal decay and degradation of the central core (i.e. heartwood) of the trees (Boyce, 1961; Arhipova et al., 2011; Gonzalez et al., 2025). Although the relationship between heart rot disease and carbon‐based greenhouse gas fluxes has never been directly studied, several factors suggest a potential connection (Mukhin & Voronin, 2009; Covey & Megonigal, 2019). The internal decomposition of wood could either increase or decrease CO_2_ fluxes, depending on the extent of decay and the availability of organic material for microbial breakdown. Additionally, decayed stem cavities may create connections with soil‐derived CO_2_ and CH_4_ to escape through the stems, similar to aerenchyma‐mediated gas transport in wetland plants (Megonigal & Guenther, 2008; Pitz & Megonigal, 2017). This soil derived C flux could be particularly significant for CH_4_, especially in anoxic waterlogged or saturated soils (Pangala et al., 2013). Anoxic conditions required for CH_4_ production are also found in a tree's heartwood, where heart rot occurs, suggesting a potential link between heart rot and stem CH_4_ fluxes. However, attributing changes in carbon‐based greenhouse gas fluxes to heart rot disease requires further investigation, as current knowledge of tree derived CH_4_ fluxes is still emerging and remains limited.
We investigated heart rot disease impacts on carbon‐based greenhouse gas fluxes from stems and soils in sugar maple (Acer saccharum Marshall) in an upland northern hardwood forest. Sugar maple is a prominent species in upland northern hardwood forests and is a species that is highly susceptible to heart rot disease (Frederick, 1973). Heart rot and its extent within the stem (i.e. severity) was determined with non‐destructive sonic tomography surveys (Brazee et al., 2011; Gilbert et al., 2016, 2025), while carbon‐based greenhouse gas fluxes were measured on tree stems and nearby soils using an enclosed static chamber connected to a portable H_2_O, CH_4_, and CO_2_ gas analyzer. We also conducted several other measurements and experiments to determine the point source of CH_4_ and CO_2_ production within the stems' bark, sapwood, and heartwood region, and to validate the presence of methanogenic bacteria within the CH_4_ production point source. We hypothesized that heart rot disease affects carbon‐based greenhouse gas fluxes from tree stems by promoting heartwood degradation and decomposition, while having minimal impact on soil fluxes due to its localized presence within the stem. We further hypothesized that the magnitude of stem greenhouse gas fluxes would vary with the severity of heart rot infection. Addressing these hypotheses is important for improving our understanding of the relationship between forest disease and carbon‐based greenhouse gas fluxes.
Materials and Methods
Site description and study design
This study was conducted at the University of Notre Dame Environmental Research Center (UNDERC), located on the border of Wisconsin and the Upper Peninsula of Michigan, USA (Jones, 2017). UNDERC soils are predominantly mineral and well‐drained overlain by organic surface horizons (Parsley, 2016). UNDERC has a humid continental climate type with an average annual temperature of 4.3°C and an annual precipitation of 50–100 cm that occurs as both snow in winter and rain in summer. UNDERC is a secondary‐growth temperate mixed deciduous forest with a substantial presence of evergreen species. The deciduous hardwoods include sugar maple (A. saccharum Marshall), red maple (Acer rubrum L.), aspen (Populus tremuloides Michx. and Populus grandidentata Michx.), and birch (Betula alleghaniensis Britt. and Betula papyrifera Marsh.). Evergreen trees in the upland areas consist of conifers such as balsam fir (Abies balsamea (L.) Mill.) and hemlock (Tsuga canadensis (L.) Carr.). We chose sugar maple for this study because the study area is dominated by sugar maple (Whitney, 1999), and sugar maple comprises c. 20% of the vegetation at UNDERC with c. 2–10% of sugar maples infected with heart rot (Gonzalez et al., 2025).
We employed a novel experimental design and state‐of‐the‐art instrumentation to assess the impact of heart rot on carbon‐based greenhouse gas fluxes from tree stems. Using a portable CH_4_/CO_2_/H_2_O gas analyzer, we measured CH_4_ and CO_2_ gas fluxes from tree stems and soils at the base of each sampled tree. These measurements aimed to help resolve the long‐standing question of whether stem‐emitted CH_4_ originates from internal stem processes or is transported from the soil. A key challenge in this study was identifying trees with heart rot and assessing the severity of decay, as heart rot is often not externally visible in living trees (Heineman et al., 2015; Jha, 2020). To address this, we used sonic tomography – a non‐destructive, advanced imaging technique – to detect internal decay caused by heart rot. We selected three study locations c. 1 km apart within the property to ensure spatial separation and minimize the influence of local environmental variability on CH_4_ flux measurements. At each location, trees with heart rot were paired with nearby (within 30 m) trees with no tomographic evidence of decay. This ensured that comparisons between heart rot and non‐heart rot infected trees were balanced across the three study sites, minimizing variation in topographic elevation and soil properties such as soil type and volumetric water content (VWC). Sonic tomography was conducted at the beginning of the growing season, and gas fluxes were measured over three periods: mid‐June, mid‐July, and October.
Tomography measurements
In this study, we used sonic tomography to determine the extent of heart rot within the main tree stem at breast height. Sonic tomography has been widely used to quantify heart rot and assess its severity (Wu et al., 2018; Son et al., 2021). Because sound waves travel faster through intact, healthy wood than through decayed or hollow areas affected by heart rot, data collected from an array of sonic tomography sensors placed around a tree can be used to non‐destructively map and visualize a tree's internal wood structure (Supporting Information Fig. S1a,b) through the measurement of sound wave velocities (Gilbert et al., 2016; Rust, 2021). The ArborSonic 3D Acoustic Tomograph (Agfalva, Hungary), used in this study, consists of 10 SD02 piezo transducers connected to a data logger, amplifier box, and power source (Soge et al., 2020). For each sampled tree, ten SD02 Piezo transducers were inserted c. 2 cm into the stem at evenly spaced intervals around the stem's circumference (Fig. S1c). Following installation, each transducer was tapped three times with a metal hammer and sound wave velocity and signal strengths were recorded by the instrument for tomographic analysis using ArborSonic 3D software (Gilbert et al., 2016). The software converts the sound velocities recorded by the sensors into a cross‐sectional map showing the internal decay percentage, based on user‐defined settings such as the measured tree diameter, bark thickness, and wood density. Since a single species was measured, bark thickness and wood density were held constant across all trees and were values typical for sugar maple. Trees with a stem decay percentage below 5% were classified as trees without heart rot, while those with higher decay percentages were categorized as trees with heart rot.
Gas exchange measurements
Stem and soil carbon‐based greenhouse gas fluxes were measured in situ in mid‐June, mid‐July, and early‐October 2024, using the enclosed static chamber method. This method involves installing collars that enable repeated enclosed chamber measurements at the same location. For stem flux measurements, we installed 10.2 cm diameter, c. 9 cm high PVC pipes in either 4 or 8 cardinal directions – depending on tree diameter – to cover most of the stem surface at the same locations used for tomography measurements. Each collar was shaped with a jigsaw to fit the stem's curvature after lightly smoothing the bark using a bark scraper, then attached and sealed with silicone sealant (Fig. S2). We tested the silicone sealant in the lab for CH_4_ and CO_2_ fluxes and found it to be inert. For soil flux measurements, PVC pipes were inserted c. 2–5 cm into the soil in the four cardinal directions at the base of each tree where stem collars were installed (Fig. S3). Following collar installation at the start of the growing season, the height of each collar was measured at four points to enable the calculation of chamber volume used for flux quantification.
A PVC closet flange with knock out fit snuggly into the PVC collars and served as a customized enclosed static chamber. The knockout section of the closet flange served as the top of the chamber, where two 0.64 cm and two 0.32 cm holes, spaced c. 2.5 cm apart, were drilled to accommodate a pressure equalization port, a chamber temperature sensor, and Bev‐A‐Line tubing. Each component atop the chamber was secured using a stainless‐steel bulkhead adapter fitted to each drilled hole and further sealed using silicon glue. Pressure differentials within the chamber and surface are a known source of error for static chamber measurements (Lund et al., 1999; Levy et al., 2011). Hence, a hollow 0.32 cm metal tube, c. 15 cm long and coiled into 2 cm diameter loops, was attached to a bulkhead to maintain pressure equalization and ensure a closed system during measurements. Temperature inside the chamber was monitored at 1 Hz using a thermocouple connected to a SUPCO Data Logger (Model: SL500TC) mounted on the outside of the chamber. Two 2‐m‐long, 0.64 cm diameter Bev‐A‐Line tubes connected the chamber to the gas analyzer's input and output ports, maintaining a closed system (Fig. S4; Table S1).
A LI‐COR LI‐7810 CH_4_/CO_2_/H_2_O Trace Gas Analyzer measured CO_2_ and H_2_O concentrations in parts per million (ppm) and CH_4_ concentrations in parts per billion (ppb) at a frequency of 1 Hz for 180 s inside the chamber to calculate fluxes. Data were recorded on a Bluetooth tablet (Dell Latitude 7220, Windows 11) connected to the gas analyzer. Before flux measurements, collars were cleaned and sprayed with 70% alcohol to sterilize the bark surface. For each measurement, the chamber was placed into the collar creating a completely enclosed chamber where fluxes could be calculated using the temporal change in their concentration over time (Fig. S5). CO_2_ and CH_4_ fluxes were calculated with Eqn 1 (Welles et al., 2001).
F is either the CH_4_ flux in nmol m^−2^ s^−1^ or CO_2_ flux in μmol m^−2^ s^−1^, V is the chamber volume (cm^3^), P 0 is the atmospheric air pressure (kPa), W 0 is the initial water vapor mole fraction (mmol mol^−1^) inside the chamber, a measure of the dry air content in the chamber and is used to correct the measured CH_4_ and CO_2_ mole fractions to dry‐air values (Hooper et al., 2002), R is the universal gas constant (8.314 Pa m^3^ K^−1^ mol^−1^), S is the surface area of the measurement (81.1 cm^2^), T 0 is the initial air temperature inside the chamber (K) measured by the SUPCO Data Logger and thermocouple from the average of the first 23 s of valid measurements, and dC/dt is the initial rate of change of either CH_4_ or CO_2_ mole fractions in the chamber (nmol mol^−1^ s^−1^ or μmol mol^−1^ s^−1^). Data were quality controlled using diagnostic codes provided by the trace gas analyzer. Measurements flagged for sensor errors were excluded from further analyses. Flux calculations began 105 s after the chamber was sealed to ensure equilibration and adequate gas mixing within the chamber. A linear regression was fit to the water vapor dilution–corrected CH_4_ or CO_2_ mole fraction time series from the gas analyzer, and the slope of the regression was used to quantify dC/dt for each gas.
Collar CO_2_ and CH_4_ fluxes were scaled to the level of the individual stems sampled. For stem fluxes, we calculated an area‐weighted average flux across all collars across the tree to account for variability in fluxes across the stem. The area‐weighted average stem flux (F S) was calculated using Eqn 2, which integrated fluxes from all collars around the stem circumference (C) and divided by the total circumferential area of the sampled stem. The total circumferential area of the sampled stem was determined by multiplying the stem circumference (D S) by the collar diameter (D C).
The integral in Eqn 2 was approximated with the Newton‐Cotes formula (i.e. trapezoidal rule), where dC was determined from measurements of collar spacing around the stem. The SD of the area‐weighted average stem flux was calculated as the SD of fluxes from all collars around each stem. Herein, area‐weighted average stem fluxes will be referred to as stem fluxes, whereas fluxes from individual collars will be referred to as stem collar fluxes. For soil fluxes, the mean and SD of the four collar flux measurements around each tree was calculated for each sampled time period, along with the VWC and soil temperature.
Identifying point sources of carbon‐based greenhouse gas fluxes within stems with the drilling flux experiment
Point sources of carbon‐based greenhouse gas fluxes in stems without heart rot were identified using an incremental drilling method followed by flux measurements, referred to here as the drilling flux experiment (Fig. S6). This incremental approach allowed us to determine relative differences in fluxes across stem cross section that we used to designate point sources of gas flux. A large change in the relative flux with drilling indicated regions in the stem of high gas production. The drilling flux experiment was conducted in October on four trees that had stem chambers (collars) installed earlier in the growing season. Before drilling, a baseline flux measurement was recorded from the intact stem to characterize pre‐disturbance fluxes. A 2‐thread Haglöf 12″ increment borer (Långsele, Sweden) was then used to drill a single hole into the stem at c. 3 cm intervals. At each depth, the wood core was extracted, saved for later lab analysis, and a flux measurement was taken immediately. To prevent microbial contamination, the borer was sterilized between each drilling interval using ethanol and flame‐sterilized with a propane torch for c. 10 s (Arnold et al., 2024). To account for differences in absolute flux rates across individual trees, flux values from incremental drilling were normalized to the minimum and maximum flux values within each stem (see Statistical Analyses section below for further information). In the laboratory, collected wood cores were examined to determine the boundaries and depths of the bark, sapwood, and heartwood using a ruler, guided by visual and tactile assessments of changes in color, texture, and moisture content (Wang et al., 2016).
Detection of methanogens in heartwood
To aid in the interpretation of CH_4_ fluxes and identify potential point sources, we analyzed the microbial and bacterial community within the heartwood of selected tree stems, with a particular focus on methanogens. Heartwood was targeted based on previous studies showing that methanogen populations are typically most abundant in the anoxic interior of the stem (Yip et al., 2019; Arnold et al., 2024, 2025). In October 2024, heartwood samples were collected from nine stems – each from a different tree. Three of these stems showed no signs of heart rot, while six exhibited varying degrees of decay, as indicated by sonic tomography. All selected trees had similar diameters at breast height. From each tree, a single core was extracted at 137 cm height using a 2‐thread Haglöf 12″ increment borer (Långsele, Sweden), drilled perpendicularly into the stem with the goal of reaching the pith. The target coring depth was based on each tree's radius. For stems with fractures, cores were taken from the side opposite the fracture to avoid structurally compromised tissue. The increment borer was sterilized with hot water between groups of trees and flame‐sterilized with ethanol between individual samples.
Immediately after extraction, cores were sealed in sterile aluminum foil, placed in polyethylene bags, and stored on dry ice for transport to the lab. In the lab, c. 0.6 g of heartwood tissue per sample was weighed using an electronic balance on a decontaminated surface. Samples were preserved using GeneCount® qKit preservation kits, following the protocols provided and shipped to Luminultra, a biotechnology company specializing in microbial community analysis and sequencing, (Luminultra Technologies, Fredericton, NB, Canada). Once received, DNA was extracted and purified by a commercially available magnetic bead‐based purification method (GeneCount qKit Purify Auto; LuminUltra Technologies). Total prokaryote concentration was determined by quantitative polymerase chain reaction with a SYBR green assay using the following cycling conditions: 95°C for 3 min, 1 cycle, 95°C for 20 s. followed by 60°C for 45 s, repeated for 40 cycles. Next Generation Sequencing for microbial community analysis was performed by next generation sequencing of the V4 variable region of the 16S rRNA gene (Primers 515F and 806R) using and run on an Illumina MiSeq Sequencer with v2 500 cycle reagent kit (Illumina, San Diego, CA, USA). Bioinformatics processing of the sequencing data was performed using mothur (v.1.36.1) and taxonomic identification was completed using the most current SILVA SSU ARB database. Methanogen cell counts were converted to relative abundance by dividing the methanogen count by the total prokaryotes cell count within each sample.
Statistical analyses
Statistical significance was evaluated at the 95% confidence level. Both parametric and non‐parametric tests were used to assess significant differences among heart rot, non‐heart rot, and other tree and soil categories. Parametric tests were applied when the data met the assumptions of normality, whereas non‐parametric tests were used when those assumptions were not met. Normality was assessed using the Shapiro–Wilk test and tests for homogeneity of variances. In cases where data were heavily skewed, a log transformation was applied to improve visualization and meet the assumptions of normality. For pairwise comparisons, Mann–Whitney U tests were used, whereas multiple group comparisons were conducted using the Kruskal–Wallis one‐way ANOVA on ranks, followed by Dunn's post hoc test. Chi‐squared tests were used to determine whether observed variables conformed to a specified expected distribution. Both linear and non‐linear three‐parameter exponential models were fitted using least squares regression. For certain analyses, flux data from individual collars (f) were normalized (F N) using Eqn 3 with values scaled between the observed maximum (f max) and minimum (f min) values from collars around the stem or from incremental drilling flux experiment measurements over time.
This normalization enabled meaningful comparisons of cross‐sectional and cross‐stem circumference flux patterns across stems by scaling the flux measurements so that the maximum flux within each stem was set to 1 and the minimum to 0 (Mazziotta & Pareto, 2022). Statistical analyses were conducted in MATLAB.
Results
CH_4_ fluxes significantly differed among stems with and without heart rot and soils, whereas CO_2_ fluxes were statistically similar among stems with and without heart rot (Fig. 1; Table S2). For stem and soil pairs, there was no significant correlation between soil and stem CO_2_ and CH_4_ fluxes (all P‐values: >0.05). Tree stems and soils were a source of CO_2_ to the atmosphere, whereas stems were a source of CH_4_, and soils were a sink of CH_4_. There was no significant difference between soil CH_4_ or CO_2_ fluxes at the bases of trees with or without heart rot. On a per unit area basis, stems with heart rot emitted more CH_4_ than soils assimilated. The median stem with heart rot CH_4_ flux was 6.0 nmol m^−2^ s^−1^, whereas the median soil CH_4_ flux was −2.5 nmol m^−2^ s^−1^. Both stem CH_4_ and CO_2_ fluxes exhibited a right‐skewed distribution, especially in stems with heart rot (Shapiro–Wilk test: P‐value: < 0.05). Median CO_2_ fluxes were not significantly different between stems with or without heart rot, whereas the median CH_4_ flux was seven times higher in stems with heart rot and statistically significant (P‐value: 0.001). For example, the median CH_4_ flux of stems with heart rot was 6.0 nmol m^−2^ s^−1^, whereas the median CH_4_ flux of stems without heart rot was 0.9 nmol m^−2^ s^−1^. By contrast, the median CO_2_ flux of stems with heart rot was 15.8 μmol m^−2^ s^−1^, whereas the median CO_2_ flux of stems without heart rot was 15.0 μmol m^−2^ s^−1^.
Violin plots of collar CH4 fluxes (nmol m−2 s−1) (a) and CO2 fluxes (μmol m−2 s−1) (b) from sugar maple (Acer saccharum Marshall) stems and soils at tree bases. The violins represent stems with heart rot (stem HR; yellow violin), stems without (i.e. no) heart rot (stem NHR; purple violin), soil at the bases of heart rot trees (soil HR; blue violin), soil at the bases of no heart rot trees (soil NHR). Box boundaries within each violin represent the 25th and 75th percentiles, the line inside the box indicates the median, and circles represent outliers. Different letters indicate statistically significant differences among stem groups at the 95% confidence level (Kruskal–Wallis test followed by Dunn’s post hoc test, P < 0.05).
We analyzed the relationship between stem flux and within stem flux variability (i.e. the SD of all collar fluxes on an individual stem) across all sampled trees (Fig. 2). Stem flux and within stem flux variability spanned two orders of magnitude and were log transformed to meet assumptions of normality. For CH_4_, stems with heart rot had higher stem flux and within stem flux variability across collars than stems without heart rot. For CO_2_, stems with heart rot had similar within stem flux variability and higher stem fluxes than in stems without heart rot with the exception of several flux outliers associated with stems with heart rot. Within stem CH_4_ flux variability was strongly and positively correlated with stem CH_4_ flux (R ^2^ = 0.9, P‐value: < 0.001), whereas within stem CO_2_ flux variability was less correlated with stem CO_2_ fluxes (R ^2^ = 0.4) but still statistically significant (P‐value: < 0.001). This indicates that high stem fluxes are associated with greater variability across collars on an individual stem, suggesting that localized hotspots of elevated flux, particularly for CH_4_ are driving the overall stem fluxes.
Relationships between stem collar flux variability and stem CH4 (a) and CO2 (b) fluxes in sugar maple (Acer saccharum Marshall). Blue and yellow circles represent heart rot stems and no heart rot stems, respectively. Both axes were log‐transformed. Black lines depict the fitted linear regression models. Model fit is quantified using R 2, and the statistical significance of the regression is indicated using P‐values.
Further analysis of the collar and stem fluxes along with field observations revealed that CH_4_ flux ‘hotspots’ were associated with the presence of bark fractures (Fig. S7). Bark fractures are outwardly visible separations of wood along the tangential–longitudinal plane of the tree stem. This includes separations occurring in both the bark and ring shakes (i.e. splits along the annual growth rings of the stem) (Kubler, 1983). To test the hypothesis that bark fractures caused CH_4_ or CO_2_ flux hotspots, we normalized the collar fluxes from each tree and plotted them as a function of the distance from the fracture (Fig. 3). The highest measured CH_4_ and CO_2_ fluxes on a given stem were measured on collars associated with the presence of bark fractures. On average, normalized CH_4_ fluxes were c. 9 times greater on bark fractures than on the rest of the tree, whereas normalized CO_2_ fluxes were c. 2 times greater on bark fractures than on the rest of the tree. The effect of bark fractures on stem CH_4_ fluxes was particularly strong, and CH_4_ fluxes declined exponentially with increasing distance from bark fractures (R ^2^ = 0.9, P‐value: < 0.0001). By contrast, CO_2_ fluxes did not exhibit such a significant exponential decrease with distance (P‐value: 1.0), indicating a lesser impact of bark fractures on CO_2_ fluxes compared to CH_4_ fluxes.
Relationships between the normalized stem CH4 (a) and CO2 (b) fluxes and the circumferential distance from the bark fracture to each measurement collar in sugar maple (Acer saccharum Marshall). Error bars represent the SD across stems, and the line represents the best fit of the data with a three‐parameter exponential decay model. Model fit is quantified using R 2, and overall significance is indicated by P‐values from the non‐linear regression.
Heart rot severity and carbon‐based greenhouse gas flux
Stem decay percentage (i.e. heart rot severity) was positively correlated with stem CH_4_ fluxes (R ^2^ = 0.3, P‐value: < 0.001), but not with CO_2_ fluxes (Fig. 4). Stem decay percentage ranged from no decay to a maximum of 68% of the cross‐sectional area, whereas both CH_4_ and CO_2_ fluxes ranged several orders of magnitude. Although the highest CH_4_ and CO_2_ fluxes coincided with high stem decay severity, there was a large variability in CH_4_ and CO_2_ fluxes especially for CH_4_ fluxes at stem decay percentages > 40% that appeared to be associated with stems with and without fractures. Closer examination revealed statistically significant differences in CH_4_ fluxes among no heart rot stems, heart rot stems, and stems with both heart rot and fractures (P‐value:< 0.001) (Fig. 4a inset), whereas CO_2_ fluxes did not significantly differ among these three categories (P‐value: 0.6) (Fig. 4b inset). CH_4_ fluxes from stems with both heart rot and fractures were c. 10 times higher than those from no heart rot stems and c. 3 times higher than those from stems with heart rot alone.
Relationships between CH4 (a) and CO2 (b) fluxes and heart rot severity as measured by stem decay percentage, in no heart rot stems (NHR; yellow circles), heart rot stems (HR; cyan triangles), and heart rot stems with bark fractures (HRF; dark blue circles) of sugar maple (Acer saccharum Marshall). Stem gas fluxes are log‐transformed, and black line depicts the fitted linear regression model. Model fit is quantified using R 2, and the significance of the regression is indicated by P‐value. Inset violin plots illustrate the distribution of untransformed stem fluxes for each stem category, whereas different letters indicate statistically significant differences at the 95% confidence level.
Due to the importance of bark fractures in regulating CH_4_ fluxes and their occurrence at high stem decay severity, we tested the hypothesis that greater stem decay severity is associated with an increased incidence of bark fractures in the trees sampled for this experiment. Our results support this hypothesis, as we observed a clear increase in the proportion of trees exhibiting bark fractures across stem decay classes. Only 4% of stems exhibited bark fractures in the 0–25% decay class, 12% of stems exhibited bark fractures in the 26–50% decay class, and 19% of stems exhibited bark fractures in the 51–75% decay class (Fig. 5). This represented a threefold increase in bark fracture occurrence between stems without decay and those with moderate decay severity and a fivefold increase between stems without decay and those with high decay severity. A chi‐squared test of independence confirmed that the association between stem decay category and increased proportion of stems with heart rot is significant (χ^2^ = 22.9, P‐value: < 0.001), indicating that stems with higher decay are more likely to have fractures.
Percentage of fractured stems across different stem decay percentage bin classes in sugar maple (Acer saccharum Marshall) trees. A chi‐squared test indicates that stem decay percentage class and bark fracture occurrence are significantly associated.
Point sources of stem CH4
and CO2 fluxes
The point source for CH_4_ and CO_2_ fluxes measured at the atmosphere–bark interface differed for the two carbon‐based greenhouse gases (Fig. 6). Point sources of CH_4_ and CO_2_ fluxes were assessed by conducting a series of flux measurements following incremental drilling into the stems of four healthy trees. All trees had similar diameters at breast height [38.4 ± 7.6 (μ ± σ cm)] and showed consistent patterns in carbon‐based greenhouse gas fluxes across the stem cross‐section. The point source for CH_4_ was located in the heartwood where the relative CH_4_ fluxes were highest, whereas the point source for CO_2_ was in the sapwood where relative CO_2_ fluxes were highest. For CH_4_, fluxes were lowest in the bark and sapwood portions of the stem, dramatically increased at the sapwood heartwood interface, and reached a peak inside of the heartwood. On average, normalized CH_4_ fluxes in the heartwood were c. 6 times higher than in the sapwood and c. 5386 times higher than in the bark. By contrast, CO_2_ fluxes varied across stem tissues, increasing at the bark–sapwood interface, remaining high throughout the sapwood, and declining at the sapwood–heartwood interface, eventually reaching its lowest value in the deep heartwood. Normalized CO_2_ fluxes in the sapwood were c. 2.5 times higher than in the heartwood and were c. 1.2 times higher than in the bark.
Normalized CH4 (a) and CO2 (b) fluxes within (i.e. bark, gray region; sapwood, blue region; and heartwood, pink region) no heart rot stems from the drilling flux experiment. Different colored lines represent the four individual sugar maple (Acer saccharum Marshall) stems sampled. Fluxes were normalized based on the minimum and maximum flux values from the complete drilling progression of a single tree.
As we observed elevated CH_4_ fluxes in the heartwood of non‐heart rot trees, we aimed to test whether methanogens are consistently associated with heartwood across three tree categories – non‐heart rot, heart rot without bark fractures, and heart rot with bark fractures – and whether heart rot influences methanogen composition. We detected methanogens in the heartwood of all three stem categories, suggesting that heartwood has the potential to produce CH_4_ regardless of internal decay status (Fig. 7). Differences in methanogen relative abundance among stems with both heart rot and fractures, heart rot only, and no heart rot trees were not statistically significant at the 95% confidence level, despite an eightfold higher mean in trees with both heart rot and fractures. Likely, the reason for this lack of significance was high variability among the relative abundance measurements, especially for stems with heart rot and bark fractures. For example, the variability in the relative abundance of methanogens in the heart rot stems with bark fractures was c. 7–10 times as high as the variability found in stems with and without heart rot.
Percent methanogen abundance in the heartwood of heart rot fractured, heart rot, and no heart rot stems of sugar maple (Acer saccharum Marshall). The same letter on each bar indicates that differences in methanogen percentages among stem groups are not statistically significant. Group means were compared using one‐way ANOVA after confirming normality (Shapiro‐Wilk) and equal variance (Brown‐Forsythe). Error bars represent the SD across stems.
Discussion
This is the first study we are aware of to investigate the impacts of heart rot disease on tree stem and soil CO_2_ and CH_4_ fluxes, providing a case study to better understand the role of tree health in carbon‐based greenhouse gas fluxes. Fig. 8 summarizes the key findings from this study. We found that CO_2_ production occurred in the sapwood because of autotrophic respiration, whereas CH_4_ production took place in the heartwood due to the presence of methanogens. We also found that the impact of heart rot on carbon‐based greenhouse gas fluxes differed in stems and soils and differed based on the carbon‐based greenhouse gases (CO_2_ or CH_4_) (Fig. 1). Soil fluxes were unaffected by the presence of heart rot in nearby trees. All trees, regardless of heart rot presence, emitted both CH_4_ and CO_2_ from their stems. Stems with heart rot exhibited higher CH_4_ fluxes than those without, whereas CO_2_ fluxes were similar across all stems and not influenced by the presence of heart rot. These findings are in line with previous work showing CH_4_ emissions from stems and high CH_4_ concentrations in the heartwood (Wang et al., 2016; Covey & Megonigal, 2019; Epron & Mochidome, 2024), but contradict recent work showing CH_4_ uptake by methanotrophs on the bark of trees (Jeffrey et al., 2021). In the remaining sections, we explain the physiological mechanisms underlying these observations and discuss their significance for understanding the relationship between forest health and carbon‐based greenhouse gas fluxes.
Schematic diagram summarizing the key results from this study for heart rot stems, no heart rot stems, heart rot stems with bark fractures of sugar maple (Acer saccharum Marshall), and soil. Arrows are orange for CH4 fluxes, green for CO2 fluxes, and their size is proportional to the magnitude of the flux. Stem cross sections indicate the relative amount of stem decay next to each stem category, the point source of the two gases within the stem in the most left‐hand cross section, and the relative abundance of methanogens in the heartwood for each stem category.
Why was there no impact of heart rot on soil CH4
and CO2 fluxes?
Our results showed that heart rot did not affect soil CH_4_ or CO_2_ fluxes (Fig. 1), likely because the decay is confined within the tree stem and does not contribute sufficient carbon to the surrounding soil to stimulate fluxes (Hilman & Angert, 2016). Such stimulation would typically occur through positive soil priming, where the addition of fresh carbon inputs – such as decayed heartwood – enhances microbial respiration (Schoenholtz et al., 2000; Conrad, 2009; IPCC, 2013). Since heart rot is restricted to the interior heartwood and often occurs in localized sections along the stem, it is not surprising that it did not influence soil CH_4_ and CO_2_ fluxes at short timescales. However, over longer timescales, heart rot near the base of the tree or structural weakening from heart rot may lead to treefall (Boyce, 1961; Arhipova et al., 2011; Yatsko et al., 2025) and potentially stimulate these fluxes at broader spatial and temporal scales. Further research at the ecosystem level is needed to understand the larger scale implications of heart rot for soil greenhouse gas emissions and forest carbon budgets.
What was the source of CH4
fluxes from tree stems?
Reports of CH_4_ fluxes from plants are relatively recent (Megonigal & Guenther, 2008; Wang et al., 2016, 2021), and since then several studies have debated over the source of this CH_4_ (Barba et al., 2019; Vargas & Barba, 2019). Some studies suggest that stem CH_4_ fluxes result from upward transport of CH_4_ produced by methanogens in anaerobic soils (Rusch & Rennenberg, 1998; Barba et al., 2024), whereas others argue that CH_4_ is generated internally within the tree through biological processes (Wang et al., 2016). In this study, we leveraged the presence of aerobic upland soils to test the hypotheses that CH_4_ production occurs in stems rather than soils, and found strong support for internal CH_4_ production within trees. Specifically, we observed that soils acted as a source of CO_2_ to the atmosphere but as a sink for atmospheric CH_4_. This pattern is consistent with aerobic upland forest soils, where oxidation promotes CO_2_ production and methane‐oxidizing bacteria (methanotrophs) consume CH_4_ (Megonigal & Guenther, 2008; Conrad, 2009). CH_4_ fluxes in the surrounding soils did not differ between trees with and without heart rot, even though CH_4_ fluxes from the stems themselves differed significantly, ruling out a soil contribution to stem CH_4_ fluxes. Methane fluxes were also unlikely to be derived from the waterlogged soils or a high water table as the water table is 5–13 m below the studied upland forests and soil VWC measurements were below the 47% threshold in which soils begin to generate CH_4_ (Kaiser et al., 2018; Evans et al., 2021; Korkiakoski et al., 2022). Together, these observations support the conclusion that CH_4_ is produced internally within stems rather than transported upward from the soil, as soil fluxes cannot explain the differences observed in stem CH_4_ emissions across stems with and without heart rot. Further evidence for internal stem production comes from the presence of methanogens in the heartwood. We found direct evidence of active methanogens in this tissue, where previous studies have reported elevated CH_4_ concentrations (Hietala et al., 2015; Yip et al., 2019) and where our stem drilling experiments detected the highest CH_4_ fluxes (Fig. 6). Finally, we can rule out termites inside the tree as a source of CH_4_, as CH_4_ was emitted from both stems with and without heart rot, and termites generally consume decaying wood (Leponce et al., 1996; Brune & Ohkuma, 2011; Ito, 2023). Furthermore, no visual evidence of termite infestation (e.g. saw dust, frass, or termites) was found in or around any of the sampled trees.
Why do stems with heart rot exhibit higher fluxes for CH4
, but not for CO2 ?
Understanding carbon‐based greenhouse gas fluxes from tree stems requires examining their regulating mechanisms. The main drivers are the site and source of gas production within the stem and the resistance to diffusion as gases move through stem tissues into the atmosphere. CH_4_ is produced by methanogenic archaea inhabiting the anoxic heartwood of stems, both with and without heart rot (Figs 7, 8) (Boddy, 2001; Teskey et al., 2008; Covey & Megonigal, 2019). By contrast, CO_2_ is primarily generated through autotrophic respiration in the living sapwood. Diffusion resistance differs between the two gases because of their distinct sites of production. CH_4_ faces greater resistance than CO_2_ because it originates in the heartwood and must diffuse through heartwood, sapwood, and bark before reaching the atmosphere, whereas CO_2_ is produced in the sapwood and only diffuses through sapwood and bark. These production and diffusion processes establish the foundational background for interpreting stem fluxes. Assessing the effects of heart rot requires additional consideration of its indirect impacts on the internal stem environment.
It is important to note that heart rot fungi in living trees do not produce CH_4_ (Mukhin & Voronin, 2008; Covey et al., 2012; Hietala et al., 2015). Instead, heart rot likely enhances CH_4_ fluxes by affecting methanogen production through changes in habitat quality and influencing gas diffusion within the stem through alterations in wood structure. Heart rot primarily affects the heartwood, whereas the sapwood is less impacted because of its higher moisture content and active defensive compounds. In some cases, heart rot can compromise the sapwood, allowing water to saturate the remaining heartwood to form ‘wetwood’. These saturated, anoxic conditions provide favorable habitats for methanogens and promote CH_4_ production (Covey & Megonigal, 2019; Epron et al., 2023). Although we could not statistically demonstrate higher methanogen abundance in stems with heart rot, likely owing to high variability and limited sample size, stems with severe heart rot and fractures did exhibit the highest percent abundance among the three stem categories (Fig. 7).
Heart rot may also influence gas diffusion by weakening stem structure. Indeed, we found that greater heart rot severity was associated with an increased likelihood of stem and bark fractures (Fig. 5). This aligns with previous work showing that internal decay reduces structural integrity, making stems more prone to damage and bark fracturing (Boyce, 1961; Arhipova et al., 2011). Such fractures likely enhance gas diffusion, especially through bark fractures, because diffusion through air faces minimal resistance. Both CH_4_ and CO_2_ fluxes are expected to increase under these conditions. However, we observed stronger evidence for CH_4_ hotspot emissions than for CO_2_, which we attribute to differences in gas solubility in stem water, particularly in the water‐rich sapwood. CO_2_ is more soluble in water, allowing it to move readily through moist tissues such as sapwood (Sorz & Hietz, 2006). By contrast, CH_4_ is less soluble and more hydrophobic, encountering greater resistance in saturated or dense wood – especially when diffusing through both heartwood and sapwood (Anttila et al., 2024). These interacting factors – decay severity, structural damage, internal moisture, and gas‐specific diffusion properties – complicate our understanding of CH_4_ and CO_2_ fluxes from stems. Consequently, future research should focus on disentangling the relative contributions of production and diffusion processes to stem carbon‐based greenhouse gas emissions.
What are the study's implications for understanding the role of heart rot on these fluxes?
Our study found that the impact of heart rot disease on carbon‐based greenhouse gas fluxes from both soil and trees is nuanced and attributable to the severity of stem decay, especially for stem CH_4_ fluxes. Heart rot creates a range of localized effects within the tree stem, including internal decay and changes in wood structure that can lead to bark fracturing (Figs 5, 8). These fractures facilitate the release of CH_4_ from anoxic internal heartwood environments to the atmosphere, whereas CO_2_ fluxes remain largely unaffected (Fig. 3). These complex interactions between heart rot and bark fracturing are analogous to the ‘forest death spiral’ hypothesis (Manion, 1991), which suggests that heart rot initiates a cycle of structural degradation that weakens the tree and, in our case, leads to increased CH_4_ fluxes through bark fracturing. Our findings highlight the need for future research on disease‐related greenhouse gas fluxes to account for multiple interacting factors – including tree structure, decay processes, and stem gas transport pathways – when evaluating the broader ecological impacts of forest disease.
Our results also have important implications for measuring and scaling CH_4_ fluxes from forests to the ecosystem level. We found that bark fractures create localized CH_4_ ‘hotspots’ around the stem, contributing to substantial spatial variability in fluxes at the tree scale (Fig. 2) (Vantellingen & Thomas, 2021; Rey‐Sanchez et al., 2022). These hotspots can strongly influence total stem fluxes, meaning that the placement of measurements – either directly over a bark fracture or away from one – can lead to over‐ or underestimation of overall fluxes due to CH_4_ flux variability. While hotspots and hot moments of greenhouse gas fluxes from soils are well documented (Le Mer & Roger, 2001; Megonigal & Guenther, 2008), this study is the first to identify CH_4_ flux hotspots in forest stems and to describe their underlying mechanism. Future research should explore the temporal variability of these hotspots to determine whether they also include ‘hot moments’ (i.e. brief periods of unusually high flux). To improve accuracy, sampling strategies should be adapted to account for both spatial and temporal variability in tree‐based greenhouse gas fluxes in forests impacted by heart rot and structural stem damage such as bark fractures, particularly for CH_4_ fluxes.
Last, our findings have important implications for understanding the role of forests in the global CH_4_ budget. Upland forests are typically considered CH_4_ sinks due to their well‐drained, unsaturated soils that consume atmospheric CH_4_ (Ridgwell et al., 1999; Pangala et al., 2017; Covey & Megonigal, 2019). However, the high stem CH_4_ fluxes observed in this study – driven by heart rot induced bark fractures – reveal a previously unrecognized source of CH_4_ that may offset this sink strength. If similar emission mechanisms are widespread across other tree species and forest ecosystems, these unaccounted fluxes could represent a significant contribution to overall forest greenhouse gas budgets. Accurately quantifying these budgets will require a better understanding of the spatial and temporal variability of stem fluxes as well as robust methods for upscaling from individual trees to the ecosystem level. Additionally, climate change may influence the prevalence and severity of heart rot disease as well as the distribution of tree species that are particularly susceptible to it. These changes could further alter forest CH_4_ dynamics in the future. Improving our understanding of these processes is essential for constraining forest greenhouse gas budgets – past, present, and future – and should be a key focus of ongoing and future ecological research.
Competing interests
None declared.
Author contributions
CKS and AVR conceived and executed the research, interpreted data, and wrote the manuscript. NGS, AH and DD collected field data and edited the manuscript. All authors gave final approval for paper publication.
Disclaimer
The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.
Supporting information
Fig. S1 Tree‐stem tomographic survey of sugar maple (Acer saccharum Marshall). Fig. S2 Gluing and sealing the shaped stem collar (PVC pieces c. 9 cm in length) to the stem at 150 cm height from the tree base. Fig. S3 Field setup for measuring soil gas flux using a closed chamber system attached to an LI‐COR LI‐7810 CH4/CO_2_/H_2_O Trace Gas Analyzer. Fig. S4 Customized enclosed static chamber attached to a tree collar. Fig. S5 Field setup for measuring stem gas fluxes using a closed chamber system attached to an LI‐COR LI‐7810 CH4/CO2/H_2_O Trace Gas Analyzer. Fig. S6 Incremental stem drilling experiment setup. Fig. S7 A collar placed on a bark fracture. Table S1 List of parts used in the flux chamber design, with corresponding part numbers and vendors. Table S2 Observed stem and soil CH_4_ and CO_2_ fluxes.Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Anttila J , Tikkasalo O‐P , Hölttä T , Lintunen A , Vainio E , Leppä K , Haikarainen I , Koivula H , Ghasemi Falk H , Kohl L et al. 2024. Model of methane transport in tree stems: case study of sap flow and radial diffusion. Plant, Cell & Environment 47: 140–155.10.1111/pce.1471837712449 · doi ↗ · pubmed ↗
- 2Arhipova N , Gaitnieks T , Donis J , Stenlid J , Vasaitis R . 2011. Decay, yield loss and associated fungi in stands of grey alder (Alnus incana) in Latvia. Forestry 84: 337–348.
- 3Arnold W , Gewirtzman J , Raymond PA , Bradford MA , Butler C , Peccia J . 2024. A method for sampling the living wood microbiome. Methods in Ecology and Evolution 15: 1084–1096.
- 4Arnold W , Gewirtzman J , Raymond PA , Duguid MC , Brodersen CR , Brown C , Norbraten N , Wood QTV , Bradford MA , Peccia J . 2025. A diverse and distinct microbiome inside living trees. Nature 644: 1039–1048.40770104 10.1038/s 41586-025-09316-0 · doi ↗ · pubmed ↗
- 5Barba J , Bradford MA , Brewer PE , Bruhn D , Covey K , van Haren J , Megonigal JP , Mikkelsen TN , Pangala SR , Pihlatie M . 2019. Methane emissions from tree stems: a new frontier in the global carbon cycle. New Phytologist 222: 18–28.30394559 10.1111/nph.15582 · doi ↗ · pubmed ↗
- 6Barba J , Brewer PE , Pangala SR , Machacova K . 2024. Methane emissions from tree stems – current knowledge and challenges: an introduction to a Virtual Issue. New Phytologist 241: 1377–1380.38267825 10.1111/nph.19512 · doi ↗ · pubmed ↗
- 7Boddy L . 2001. Fungal community ecology and wood decomposition processes in angiosperms: from standing tree to complete decay of coarse woody debris. Ecological Bulletins 49: 43–56.
- 8Boyce JS . 1961. Forest pathology. 3rd Ed. New York, NY: Mc Graw‐Hill Book Co.Inc.