Flux measurements
The experimental set up focused on comparing vascular and non-vascular
plant communities in the restored peat fields and former drainage
ditches at the restored peatland. Six vascular plots and three
non-vascular plots were set up in the features (field, ditch),
respectively, for a total of 18 plots. Plot selection was done based
on the dominant vegetation within the respective features, with
E. vaginatum and Sphagnum
spp. chosen in the restored peat field while plots with
Typha latifolia and bare ditch areas were
selected in the former ditches. While the bare ditch plots were
initially devoid of vegetation, vascular plants did spread through the
area over the course of the season. Sprouts within the collars were
removed on a regular basis. Boardwalks were used to span the former
ditches and to traverse the restored peatland. An additional six
Sphagnum plots were created in the adjacent
undisturbed peatland, located within the same peatland complex, which
was used as a reference site.
Net CO2 and CH4 flux
measurements were carried out using the closed chamber technique on
permanently installed collars. A laser gas analyzer (LGR-UGGA, Los
Gatos Research, CA, USA) connected to a clear polycarbonate chamber
enabled simultaneous measurements of CO2 and
CH4 (and H2O)
concentration at 1 Hz. A rectangular chamber (60 x 60 x 30 cm; 0.108
m3) and collar combination was used at the
restored field plots while a cylindrical chamber (100 cm height x 26
cm diameter; 0.053 m3) and collar
combination was deployed in the former ditches, to accommodate
vertical growth of T. latifolia. We equipped the
chambers with fans to maintain a well-mixed headspace, as well as a
cooling system to prevent excessive warming during closure. NEE and
CH4 flux were calculated from the linear change
in CO2 and CH4 headspace
concentration, respectively, over a measurement period of 2 min. A
tarp was used to block incoming radiation within the chamber over a
successive closure. Gross primary productivity (GPP) was calculated
from the difference between the unshrouded measurement (NEE) and the
fully dark measurement which provided ecosystem
CO2 respiration (ER).
Gas temperature (TSAMPLE, °C) was measured at 1
Hz by the LGR-UGGA while photosynthetically active radiation (PAR;
µmol m-2
s-1) was recorded every 10 sec during
chamber closure by a quantum sensor. Following chamber deployment,
soil temperature (TSOIL) at 2, 5, 10, 15, 20,
25 and 30 cm was measured next to each collar using a digital
thermocouple temperature probe, while water table depth (WTD) was
manually measured at adjacent wells. Dataloggers (CR5000 and CR23X,
Campbell Scientific, AB, CAN) were used to record half hourly air
temperature (TAIR), and
TSOIL at multiple depths (5, 10, 20, 40, 60, 80
cm) in the restored field and former ditch locations over the
measurement season using type T thermocouples (Omega Engineering).
Paired Leveloggers and Barologgers (Model 3001, Solinst, Ontario,
Canada) determined half hourly WTD in proximity to the
TSOIL profiles.
A total of 600 chamber closures were performed over the snow-free
season of 2016. Standard chamber flux calculations were made for
linear changes in headspace CO2 and
CH4 over time. In the case where
CH4 bubbling was captured with the LGR-UGGA, a
piece-wise linear fitting routine was used to separate linear from
non-linear CH4 increase in headspace
concentration. Methane ebullition occurred repeatedly in the ditch
plots and was characterized by a sudden break in the slope of the
CH4 mixing ratio over short durations
(generally < 20 sec). The first difference of the
CH4 mixing ratio time series and standard
deviation of the first difference were used to distinguish non-linear
events. In total, 78 non-linear events passed the criteria in 2016 and
were separated out from the linear dataset. The linear slope before
and after the concentration jump was determined in order to quantify
jump magnitude as well as baseline magnitude, which theoretically
should continue during bubble events. Bubble magnitude was calculated
as the difference between the jump magnitude and baseline magnitude
and then converted to CH4 mass released (mg
CH4) using chamber volume, temperature and
pressure. The fraction of total emissions attributed to the ebullition
pathway was estimated by calculating the cumulative ebullitive and
diffusive flux over the periods where sampling took place.
Pore water sample collection and analyses
In-situ concentration of dissolved organic carbon
(DOC), DIC and dissolved CH4
(dCH4) was determined using six replicate sets
of pore water samplers installed 0.2 m and 0.8 m below the former
ditch and restored field surface, respectively, as well as at the
reference site. Pore water samplers were made of a 0.2 m length of ABS
pipe (25 mm I.D.) closed at both ends, slotted at the middle 0.1 m,
and covered in mesh to prevent clogging. Tygon tubing connected to one
end was extended above the soil surface to allow for water collection
by syringe from a stopcock. Installations occurred 30 days in advance
of sampling and temporally representative samples were obtained by
removing 60 mL of pore water from each sampler 48 hours before
sampling (Strack and Waddington, 2008). The headspace degassing
technique was used to acquire gas from the water samples. Ambient air
was drawn into the syringe in equal volume to the collected pore water
(30 mL) and the sample was degassed by shaking the sample vigorously.
Gas samples were then transferred to evacuated 12 mL sealed vials
(Exetainers, Labco, UK) and stored in a cooler for transport to McGill
University, Montreal, Canada for analysis. Gas concentrations of
CH4 and CO2 were
determined using a gas chromatograph (Mini-2, SRI Instruments,
California, USA). The remaining water sample was passed through 0.45
µm paper filters (Macherey-Nagel MN 85/90) and acidified before being
analyzed for DOC concentration on a total organic carbon analyzer
(TOC-V, Shimadzu, Maryland, USA).
Pore water sampling to determine δ13C and
acetate concentration was undertaken on DOY 163 (June 11, 2016),
200-201 (July 18-19, 2016), 216-217 (August 3-4, 2016) and 242 (August
29, 2016). The experimental set-up targeted the root zone (0.2 m) and
below the root zone in the cutover peat (0.8 m) using “sipper” sets
(rhizosphere and deep) permanently installed in the flux collars.
Sippers are 6 mm diameter stainless steel tubes with mesh-covered
holes drilled at the base and a length of Tygon tubing with a
stopcock. Sippers were flushed with a small amount of soil water prior
to slowly drawing 20 mL using a syringe. Stable carbon isotope samples
were filtered in the field through 0.1 µm in-line syringe filters
(Whatman Grade GF/D glass microfiber) and injected into 11 mL
evacuated glass vials sealed with 20 mm-thick butyl rubber septa.
Samples were duplicated and acidified in the field with 1 mL of 30%
H3PO4, and stored upside
down on ice before being express shipped to Florida State University,
Tallahassee, FL, USA. A 2-hour wait period was followed in the case of
same-day sampling for δ13C and acetate.
Duplicate acetate samples were filtered in the field through 0.1 µm
in-line syringe filters into 5 mL plastic vials and frozen prior to
being shipped to Lund University, Lund, Sweden. Acetate concentration
was additionally sampled directly from the roots of T.
latifolia and E. vaginatum plants.
This was undertaken by threading individual roots through a tiny hole
in a syringe with attached Tygon tubing and stopcock. Three roots were
sampled from for six plants of each species (36 roots total), with a
blank syringe (root hole included) placed in the vicinity of each
sampled plant (12 blanks total). Deionized water was replaced in the
root syringes 24 hours prior to sampling in order to have a temporally
representative sample. Note that δ13C and
acetate sampling in the field plots was prevented beyond June by a
water table deeper than 0.2 m and by strong resistance when drawing up
pore water from 0.8 m. Extraction was made difficult by the nature of
the cutover peat, which had low porosity caused by subsidence after
drainage.
Isotope samples were brought to ambient pressure with helium,
pressurized to one atmosphere and shaken to extract gas into the
headspace. The gas concentration and isotopic ratio in the headspace
were determined by direct injection on a gas chromatograph
combustion-interfaced isotope ratio mass spectrometer (MAT Delta V,
Finnigan, USA). We determined the dominant CH4
production pathway at the sampling points in the soil profile using
two stable isotope abundance metrics. First, acetate fermentation
(acetoclastic methanogenesis) yields CH4 whose
δ13C values fall within a typical range of
-65 and -50‰ whereas CH4 from
H2/CO2 reduction
(hydrogenotrophic methanogenesis) has δ13C
values typically between -110 and -60‰. Second, the apparent
fractionation factor for carbon (α) (see in Equation 2.8) is a measure
of the separation between CH4 and co-occurring
CO2. The factor is referred to as apparent,
because while CO2 is a precursor for
CO2 reduction, it is not an immediate precursor
for CH4 formed from acetate fermentation.
Nonetheless, variation in α is interpreted to represent variations in
CH4 production mechanism. Microbial
culture-derived α values for
H2/CO2 reduction are
found to range between 1.031 and 1.077, while α values between 1.007
and 1.027 are characteristic of acetate fermentation. In general,
values of α>1.065 and α<1.055 are characteristic of environments
dominated by H2/CO2
reduction and acetate fermentation, respectively.
Organic acid concentrations, e.g., acetic acid/acetate, were analyzed
using a high-pressure liquid chromatography tandem-ionspray mass
spectrometry system. The system consisted of a chromatography system
(ICS-2500, Dionex, Sunnyvale, California, USA) and a triple quadrupole
mass spectrometer (2000 Q-trap, Applied Biosystems, Foster City,
California, USA). Further analysis details and quality controls can be
found in Ström et al. (2012). Results are
presented in µM of acetate, given that acetate dominates at pH >
4.76. Other organic acids, namely, citric, formic, glycolic and
lactic, were also detected, but were present at insufficient amounts
to pursue further analysis.
