We sourced two different types of commercial horticulture peat which contained sphagnum peat moss, perlite, limestone, and wetting agents. A total of 220 g of oven-dry equivalent peat-based growing media (peat hereafter) was placed into 40 different pots (30.5 cm diameter, 20 cm height) of 5 L volume. The peat had an initial gravimetric moisture content of 43% ( ± 5.49), pH of 5.71 ( ± 0.23), and a C:N ratio of 40 ( ± 4.33). CO₂ was measured from pots (n = 12 for the two types of peat) with only peat to determine HR in the absence of plants.

Lettuce (Lactuca sativa) and petunia (Petunia × hybrida) seedlings were obtained from Jolly Farms, New Brunswick, Canada, and transplanted individually in pots (n = 28, 7 replicates for two plant types grown in two different types of peat) on 3 March 2022. The experiment was set up in a controlled growth chamber zone at McGill University Phytotron. The growth chambers were set at a temperature of approximately 23°C, 75% relative humidity, a diurnal light schedule of 16 h, photosynthetic photon flux density (PPFD) of 300 μmol m⁻² s⁻¹, and under ambient CO₂ conditions (~420–450 ppm). The growth period for the lettuce and petunia plants was 90 and 120 days, respectively. Plants grown in peat and fallow peat were fertilized with 100 mL of a 1,000 mg L⁻¹ of 20–20–20 water-soluble NPK fertilizer (The Scotts Miracle-Gro, Ohio, United States) every 2 weeks.

We conducted direct CO₂ measurements in the pots using transparent chambers manually. Chambers of 20 L volume were fitted with a fan to allow for adequate mixing and were placed on water-filled saucers to ensure that the chambers were air-tight. Measurements on CO₂ exchange on the transparent chambers in full light represented NEE, and dark measurements with covered (opaque) chambers represented ER. Measurements using covered chambers without plants represented HR from the soil. For plant setups, we assumed that AR = ER − HR.

In all cases, we measured CO₂ concentrations in the chamber every second over a period of approximately 5 min, using an SBA-5 CO₂ gas analyzer (PP Systems, USA). Then, we calculated CO₂ flux rates from the rates of change in concentration within the headspace volume for the given surface area extent of the pot. Flux rates are expressed as g CO₂-C m⁻² day⁻¹. We adopted the convention that a positive NEE value represents a net emission of C and a negative value represents net uptake of C.

We measured temperature and moisture in each pot after taking the CO₂ measurements. Temperature and soil moisture were measured at a depth of 12–15 cm of peat. Soil moisture was measured using an MP406 soil moisture sensor, ICT International, Australia. Pots were watered every week to a volumetric water content of ~30% (v/v) after taking the CO₂ measurements. In addition, we monitored the plant growth index by measuring the height and width of the plant. We complemented dimension measurements with images of plants together with a reference to a known measurement. The number of pixels in the reference was then used to calculate the areal extent and the plant biomass using a Photoshop application. At the end of the experiment, we carried out destructive sampling, washed the plant roots, measured the dry mass of aboveground and belowground plant biomass, and converted the mass to C by a 50% conversion.

On day 50 of the experiment, in subsamples (n = 3 each for fallow peat, lettuce, and petunia grown in peat substrates), we collected gas samples in a closed chamber to determine the δ¹³C (V-PDB) signature of CO₂ and CH₄ and N₂O emissions. A 25-mL sample was taken at 0, 10, 20, 30, 40, and 50 min. Five milliliters of the sample was used to measure CO₂, CH₄, and N₂O concentrations on a Shimadzu GC-2014 gas chromatograph equipped with a methanizer and flame ionization detector. Nitrogen was the carrier gas. The SRI column temperature was 70°C, and the flame ionization detector (FID) temperature was 110 110°C. Three to five standards of 5,000, 5, and 20 ppm of CO₂, CH₄, and N₂O, respectively, were run through the GC before, during, and after the sampling period. Methane and N₂O emission rates were calculated from the rates of change in concentration within the headspace, expressed in mg m⁻² day⁻¹.

For δ¹³C determination, the remaining 20 mL of the sample was run through a G2201-i CRDS Isotopic Analyzer system (Picarro, Santa Clara, CA). During each sampling period, two replicate CO₂ standards of 850 ppm and −28.5‰ VPDB and an ambient air sample were run through the instrument. Measurements on the standards had a standard error of <0.4‰ throughout the sampling period. The isotopic analyzer system was calibrated prior to the measurement period with two additional isotopic standards (100% CO₂) with δ¹³C values of −15.6 and −43.2‰ VPDB (Stix et al., 2017). δ¹³C of emitted CO₂ was calculated using Keeling plots, where intercepts were accepted only when the regression coefficient was >0.9 (Keeling, 1958; Pataki et al., 2003).

On the final day of sampling, we used the same closed chambers to collect the emitted CO₂ for ¹⁴CO₂ isotope analysis. An opaque 20-L chamber was placed on the water-filled saucer and allowed to accumulate CO₂ for 5 to 30 h based on CO₂ emission rates, to obtain sufficient mass of C to allow ¹⁴C measurement. The longer hours for collection were to allow for fallow peat samples to accumulate at least 2 mg C to be collected in 1 L gas jars from the headspace. A minimum of 2 mg C was required for ¹⁴C analysis of CO₂ in the air.

After the period, a pump with a low flow rate was used to collect 1–2 L of gas. All the gas samples were sent to the AMS Laboratory, University of Ottawa, to be processed for ¹⁴C analysis.

Radiocarbon analyses were performed on an Ionplus AG MICADAS (Mini Carbon Dating System). ^12,13,14C+1 ions were measured at 200 kV terminal voltage with He stripping. Data were processed using the BATS data reduction software as described by Wacker et al. (2010). The fraction modern carbon, F¹⁴C, was calculated as the ratio of the sample ¹⁴C/¹²C to the standard ¹⁴C/¹²C (Ox-II) measured in the same data block. Both ¹⁴C/¹²C ratios were background-corrected, and the result was corrected for spectrometer and sample preparation fractionation using the online AMS-measured ¹³C/¹²C ratio and was normalized to δ ¹³C (PDB). Radiocarbon ages were calculated as −8033 ln (F¹⁴C) and reported in percent modern carbon (pMC) as described by Stuiver and Polach (1977).

To determine the radiocarbon signature of the respired CO₂ and to partition it into old and new C, we first corrected for the background CO₂ concentration and the background F¹⁴C signature following Wang et al. (2021) and used the following equation to calculate the F¹⁴C value of the respired CO₂.

Where CO_2ms and CO_2bac are the CO₂ concentrations at the start and the end of the chamber closure. F¹⁴C_ms is the measured signature of the emitted CO₂. F¹⁴C_bac is the signature of background CO₂. For the background signature, we used the mean value of −9‰ (pMC = 1.0017) for the year 2022 from Niwot Ridge station (Levin et al., 2023).

For the plants grown in peat-based media, using the isotope signature, we divided the total respiration into AR and HR using a two-carbon source model. The measured F¹⁴C-CO₂ from peat and F¹⁴C of the background, representing the signature of the plants, were used to calculate the fraction of respiration from peat and from plants using the equation below (Wang et al., 2021):

Where f_plant and f_peat are the relative contributions by plant and peat to total ecosystem respiration measured in the setups with plants.

Statistical analyses were performed using the R Statistical Software version 4.0.2 (R Core Team, 2021). We use linear models to understand the influence of environmental variables (biomass, temperature, moisture) on the fluxes measured. Fit of the models was checked using the distribution of the residuals and p-values of the model. Comparison among the treatments was done using ANOVA or a t-test. Mean and standard error were reported, and an alpha of 0.05 was used to establish statistical significance.

Values of δ13C− CO₂ ranged from −22.08‰ to −28.21‰, with an average value of −24.38‰ (Figure 6). The Kruskal−Wallis test showed that the values did not statistically differ between peat, lettuce, and petunia (K = 1.80, df = 2, p = 0.4).

Measurement of CH₄ and N₂O fluxes showed a large range from source to sink, but they did not differ by treatment. The average CH₄ flux was 0.55 mg m⁻² day⁻¹ ( ± 4.66) and did not differ between the treatments (K = 1.37, df = 2, p = 0.5) (Figure 7). The average N₂O measurement for all the setups was 2.6 ( ± 3.47) mg m⁻² day⁻¹, which did not differ significantly between the treatments (K = 2.8, df = 2, p = 0.2) (Figure 7).

Heterotrophic respiration values measured in the fallow peat ranged between 0.05 and 0.7 g CO₂-C m⁻² day⁻¹, combining both biotic respiration as well as abiotic dissolution of limestone added to raise the pH value of the substrate (Biasi et al., 2008). Sharma et al. (2024) showed that the contribution to CO₂ from limestone is largest at the onset of addition and the contribution decreases over time. Potentially reflective of diminishing limestone contribution to CO₂ and removal of the effect of an initial disturbance, CO₂ values from peat show a decline after approximately the 70th day. The δ¹³C signature of emitted CO₂ on day 86 was −24‰, demonstrating a small low contribution from carbonates (−0.03‰) and a high contribution to total CO₂ flux from the biotic peat source (−27.5‰). Based on the original mass of peat added to each pot and an assumed C content of 50%, the average emission rate over the period of the experiment yields a 1-year decomposition rate of 6.6% ± 3.1% of mass loss. The measured value in this study is very close to the 5.4% ± 1.1% mass loss from lab incubations of horticultural peat (Sharma et al., 2024). The slightly higher mass loss in this study could be because the limestone contribution has not been separated, as was done in the previous incubation study. For instance, considering that there is no limestone contribution after day 70 and taking the average emissions after day 70, the extrapolated yearly mass loss would be approximately 3.31% ± 1.62%.

Previous peat incubations have shown that temperature and moisture conditions are key controllers of peat respiration (Scanlon and Moore, 2000; Blodau et al., 2004). Though our experiment was run in a controlled environment at ~23°C, minor fluctuations in temperature among the sampling plots and moisture explained 14% of the variability measured in HR values. When extrapolating the HR values, from our study that contained 2.8 kg dry peat per m² and emitted on average 0.28 ± 0.15 g CO₂-C m^–2 day⁻¹, to 32 km² of CEA of mushroom and ornamental production in Canada (assuming similar dry peat mass per m² and not accounting the leftover peat mass after use), we estimate 3.27 ± 1.75 kt of CO₂ emissions per year from the peat-based growing substrate.

The average NEE values were −1.17 ( ± 0.85) g CO₂-C m⁻² day⁻¹ for lettuce and −0.62 ( ± 1.72) g CO₂-C m⁻² day⁻¹ for petunia, while the average ER values for lettuce and petunia were 2.09 ( ± 0.79) and 2.96 ( ± 1.39) g CO₂-C m⁻² day⁻¹. During the initial phase of growth, as the plants accumulated biomass, both ER and NEE exhibited their highest rates and then gradually declined as the plants matured and established stable growth conditions. NEE and ER measurements were a combination of plant respiration, which is dependent on the stage of plant growth (Van Iersel, 2003), soil respiration, and limestone-derived CO₂. The biomass measurements, temperature, and moisture measurements could not always explain the large portion of the variability in the NEE and ER values.

In our study, the δ13C− CO₂ signatures did not differ between the three possible sources, suggesting that by this time the limestone contribution to total flux was insignificant. However, lacking continuous measurements of the δ13C− CO₂ signature, we could not calculate the contribution of limestone-based emission through time.

The radiocarbon age of peat was relatively older than the contemporary radiocarbon age (1737 years BP, pMC = 0.81), and so was the CO₂ respired from peat-only setups (pMC = 0.87). Even in fallow peat setups, the ¹⁴C of the CO₂ emitted was always more modern than that of the solid peat, which indicates younger fractions of peat within the bulk peat are preferentially decomposed relative to older peat fractions. Similar results have been shown by previous studies (Biasi et al., 2013; Bader et al., 2018). The higher modern C fraction in setups with plants (pMC = 0.98) clearly shows that the respiration values are dominated by the living plants (pMC = 1.0017) and, to a lesser degree, the older C of the peat substrate. While partitioning the total flux into their sources, we showed that plants’ respiration fraction differed between plant species, as peat respiration contributed 10% and 18% to the total respiration in lettuce and petunia, respectively. Previous studies have also pointed out that the relative contribution depends on plant functional type and abiotic factors (Rankin et al., 2022). In natural peatlands, Rankin et al. (2023) and Hicks Pries et al. (2013) found the AR contribution to ER to be approximately 75% and between 40% and 70%, respectively. The larger proportions of AR (82%–90%) observed in our study are likely due to the much smaller mass of peat in the experimental pots that contributes to the absolute HR being lower as compared to the amount of peat under a peatland.

Our data also demonstrate that the priming effect, caused by the roots increasing heterotrophic respiration, was minor or positive, as shown from the separation of ER using radiocarbon signatures. In studies on peatlands, HR enhancement (Bader et al., 2018; Basiliko et al., 2012) was suppressed, as well as neutral (Estop-Aragonés et al., 2022; Bader et al., 2018; Wild et al., 2023) effects due to priming have been shown. Experiments in a laboratory mimicking root exudates have shown that the positive priming effect depends on the compound added (Wild et al., 2023), soil type (Bader et al., 2018; Wild et al., 2023), and other factors. The differences in root exudates and structure could be a possible reason for the difference in priming between lettuce and petunias in our study. Petunia roots are known to form a symbiotic association with arbuscular mycorrhizal fungi (Reddy et al., 2009; Druege and Franken, 2019). These fungi form intracellular structures by penetrating the individual cells in the root cortex and play a role in supplying nutrients to host plants (Reddy et al., 2009). At the same time, arbuscular mycorrhizal fungi are also known to be substantial contributors to total ecosystem flux that rapidly return plant-derived C to the atmosphere (Nottingham et al., 2010). However, validating this would require future studies that look at both respiration components and root structures. Additionally, it is important to note that we collected only the CO₂ samples for radiocarbon measurements toward the end of the experiment because of the large cost of the isotope analysis. Monitoring the extent of the priming effect throughout the experimental period could have provided a more nuanced picture of the priming effect in horticultural crop production.

In the experiment period, the plants assimilated 168 and 225g C m⁻² in lettuce and petunia, respectively. However, most of the horticulture plants are used for food production or as ornamental plants, and their biomass is readily consumed as food or discarded to decompose at the end of the season; therefore, a large part of the biomass will be returned to the atmosphere within a short timeframe of years to decades.

Methane emissions from our study were small (0.55 ± 4.66 mg CH₄ m⁻² day⁻¹), and some uptake was also recorded. In cropped peat soils, low CH₄ emissions are well documented (Taft et al., 2017) as soils are well mixed and well aerated, creating an environment unsuitable for methanogens and supporting methanotrophic activity (Mer and Roger, 2001). The small and even negative CH₄ fluxes that we observed are similar to the results from container horticulture CH₄ measurements reported (Murphy et al., 2019, 2021; Marble et al., 2012b, 2012).

Except for two measurements, all the treatments were a source of N₂O, but emissions are lower (2.69 ± 3.47 mg N₂O m⁻² day⁻¹) than reported for vegetable crops in organic soils in Ohio with 40–133 mg N₂O m⁻² day⁻¹ (Elder and Lal, 2008), in peat soil in Finland with 14 mg N₂O m⁻² day⁻¹ (Regina et al., 2004), and in arable peat soil in the summer months in the United Kingdom ranging from 59 to 132 mg N₂O m⁻² day⁻¹ (Taft et al., 2017). Nevertheless, when comparing our values to horticultural plants grown in containers, the results we report are in general agreement with an average of 0.83 mg N₂O m⁻² day⁻¹ in peat-based substrate in annual horticultural species (Murphy et al., 2021) and with an average of 2.23 mg N₂O m⁻² day⁻¹ from pine bark and sand-based media (Marble et al., 2012b). We recognize that to track a complete picture of N₂O emissions, a larger control is required. For instance, N₂O emissions have been shown to depend on fertilizer intensity and type (Brown et al., 2025) as well as N₂O emissions depend on seasonal or cropping patterns as N₂O peaks have been reported following irrigation (Lloyd et al., 2019), cultivation, or management interventions (Elder and Lal, 2008; Regina et al., 2004).

Nonetheless, we think that these findings are important to further constrain the understanding of overall GHG impact from horticulture, the values of which are rarely compared to conventional agriculture. Using the global warming potential of emitted N₂O as 270, the average CO₂ equivalent for N₂O emissions from our study is 0.73 g CO₂-eq m⁻² day⁻¹. If these values are extrapolated over a year (i.e., 365 days), then they are roughly equivalent to what is emitted from a square meter of actively harvested peatland (Clark et al., 2023). Given that the N-fertilizer application in horticultural systems is heavy and recurrent, we show that quantifying N₂O emissions is important to ascertain the overall GHG budget of peat use in horticulture. Although the results that we present come from a limited number of samples in a limited timeframe, an initial extrapolation of N₂O emissions from using peat as a growing substrate in horticulture in Canada amounts to approximately 2.33 kt of CO₂-eq per year.