Controls on plot-scale evapotranspiration from a constructed fen in the Athabasca Oil Sands Region, Alberta

Ecological Engineering 100 (2017) 199–210

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Controls on plot-scale evapotranspiration from a constructed fen in the Athabasca Oil Sands Region, Alberta S.J. Scarlett ∗ , R.M. Petrone, J.S. Price Department of Geography and Environmental Management, University of Waterloo, 200 University Ave. W. Waterloo, ON, N2L 3G1, Canada

a r t i c l e

i n f o

Article history: Received 13 April 2016 Received in revised form 2 December 2016 Accepted 16 December 2016 Keywords: Constructed peatland Oil sands reclamation Western boreal plain Evapotranspiration Ecohydrology Re-vegetation

a b s t r a c t In the Athabasca Oil Sands Region of the Western Boreal Plain mining companies now recognize the importance of reclaiming peatlands, as they cover >50% of the pre-mined regional landscape. To date, reclamation efforts have focused on marsh and open water wetlands. However, these wetland systems are not abundant in the sub-humid climate of the WBP because of their excessive evapotranspiration (ET) losses, compared to peatland wetlands. The experimental Nikanotee Fen was constructed as part of the landscape reclamation, engineered with the intent to support natural fen vegetation and hydrologic processes. The influences of vegetation and treatment types on plot-scale energy fluxes and ET demands were examined over the 2014 growing season, one year post-construction. Actual ET and environmental variables, such as water table, soil moisture, temperature and relative humidity, were monitored in thirty-one study plots comprising control, moss and seedling plots (with and without wood-strand mulch), as well as three ponds. Actual ET was measured with weighing lysimeters and used to correct equilibrium ET, calculated using the Priestley-Taylor combination method. A roaming meteorological station was instrumented to quantify and compare available energy over different cover types. Modeled plot-scale ET was validated against continuously measured eddy covariance ET values. Cumulative ET exceeded precipitation in all plot types with average rates of 4.4, 3.2, 3.9 and 2.8 mm/day from open water, control, seedling and moss plots, respectively. Mulch reduced ET rates in both seedling and moss plots, where moss-mulch plots showed the lowest rates (2.4 mm/day). The presence of mulch lowered the near-surface vapour pressure deficit, thus providing a more favorable microclimate for moss establishment by elevating near-surface relative humidity and reducing air and soil temperature by ∼2 ◦ C. Although vapour pressure deficit did not significantly control ET (p > 0.05), mulch appeared to have a greater influence on ET in moss plots than seedling plots. Water table position displayed a weak yet significant (p < 0.05) control on ET in seedling plots. As these reclamation practices have not yet been well tested in the Alberta oil sands region; this study provides recommendations to reduce evaporative stress and aid in the establishment of functioning constructed peatland ecosystems. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Oil sands mining activities in northeastern Alberta directly impact regional wetlands which cover >50% of the Western Boreal Plain (WBP), of which most are fen peatlands (Vitt et al., 1996). Until recently, marshes and open water wetlands have been the focus of oil sands reclamation studies (e.g. Rooney and Bayley, 2011), as they are able to spontaneously develop in poorly drained landscapes and

∗ Corresponding author. Present adress: Forestry and Peatlands Management, Manitoba Sustainable Development, 200 Saulteaux Cres. Box 70, Winnipeg, MB R3J 3W3, Canada. E-mail addresses: [email protected] (S.J. Scarlett), [email protected] (R.M. Petrone), [email protected] (J.S. Price). http://dx.doi.org/10.1016/j.ecoleng.2016.12.020 0925-8574/© 2016 Elsevier B.V. All rights reserved.

are hydrologically simpler than peatlands (Harris, 2007). However, these wetland systems are not naturally abundant in the sub-humid climate of the WBP due to large moisture losses though evaporation compared to local peatlands, which are natural moisture regulators and able to mitigate high evaporative demand (Devito et al., 2005; Petrone et al., 2007). Available energy (the difference between net radiation and ground heat flux) in mid-latitude peatlands is primarily controlled by site-specific microclimate, moisture conditions and vegetation composition (Lafleur et al., 2005), and thus a major driver of both evaporation and transpiration in these systems (Price, 1991; Price et al., 1998; Runkle et al., 2014). These processes are typically observed as a single process in peatlands, evapotranspiration (ET), to describe a combined water loss to the atmosphere from the sur-

200

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

face (evaporation) and vascular plants (transpiration). The latent energy flux is a large component of wetland energy and water balances and typically represents the greatest proportion of available energy (>50%; Petrone et al., 2004; Runkle et al., 2014). Observed relationships between ET and vapour pressure deficit (VPD) indicate that atmospheric moisture demand is a direct control on ET (Kucera, 1954; Admiral et al., 2006; Admiral and Lafleur, 2007). This is most evident in vascular rich peatland systems in sub humid climates, where VPD directly influences stomatal conductance and thus vascular transpiration (e.g. Phillips et al., 2015). Earlier studies, such as Ivanov (1981) and Ingram (1983), suggest water table is an important control on ET, while more recent studies (e.g. Lafleur et al., 2005; Peichl et al., 2013) have observed weak responses in ET to changes water table. Brigham et al. (1999) suggests this relationship may differ between peatland types (i.e. bogs and fens). Although there exists uncertainty surrounding the influence of water table position on ET losses, soil moisture, sustained by upward capillary flow, has been observed to directly control ET rates (Price et al., 2010; McCarter and Price, 2013). The re-establishment of peatland ecosystems in the post-mined oils sands landscape has not previously been attempted due to their hydrological complexity and long successional development (Price et al., 2010). However, the importance of these systems was recognized by Alberta’s Environmental Protection and Enhancement Act (EPEA), which mandated mining companies to test peatland reclamation (Cumulative Environmental Management Association (CEMA), 2014). The Nikanotee Fen, a pilot fen creation project, was constructed based on the conceptual and numerical model proposed by Price et al. (2010). The fen was designed with the goal of being a self-sustaining ecosystem, resilient to the normal moisture stresses of the WBP (Daly et al., 2012). While the construction of peatlands is previously untested, peatland restoration is now a common practice following drainage for land-use change and peat extraction for horticultural and energy uses in North America (e.g. Ketcheson and Price, 2011; McCarter and Price, 2013; Strack et al., 2014; Taylor and Price, 2015) and Europe (Holden et al., 2011; Haapalehto et al., 2014). Many of these studies are focused on Sphagnum-dominated peatlands in temperate climates and thus, there exists extensive literature on restoration practices in these settings (e.g. Price et al., 1998; Rochefort et al., 2003). However, Alberta’s oil sands region is located in the sub-humid climate of the WBP where potential ET exceeds precipitation in most years (Devito et al., 2005; Petrone et al., 2007; Thompson et al., 2014) and moderate-rich fens dominate the landscape (Chee and Vitt, 1989; Vitt et al., 2000). In this typically dry climate, implementing effective reclamation practices must consider evaporative stress. Previous studies of moisture and energy dynamics of restored (i.e. not constructed) peatlands identify the importance of vegetation composition and mulch cover (Price et al., 1998; Petrone et al., 2004), where mulch created a microclimate favorable for vegetation establishment and reduced moisture stress. Hares and Novak (1992) and Novak et al. (2000) documented reduced net radiation and atmospheric mixing over mulched covered surface, therefore reducing available energy and the vapour pressure deficit and thus ET. Peatland mosses are strongly affected by near surface moisture availability since they do not have roots or an internal water conducting mechanism. Instead they depend on capillary rise from the underlying peat (Rydin and Jeglum, 2006). As a result, ET from Sphagnum-dominated peatlands is typically lower than that from fens (Romanov, 1968), due to limited vertical water movement in mosses (Price, 1997; Goetz and Price, 2015). Conversely, WBP fens can have a dense vascular cover (Chee and Vitt, 1989), which increases ET through vascular transpiration (Romanov, 1968; Takagi et al., 1999; Farrick and Price, 2009). Considering the subhumid climate of the WBP, transpiration plays an important role

in the water balance of these peatlands. Therefore, differences in climate and peatland type must be considered when assessing reclamation strategies, as they can substantially influence the hydrological and thermal behavior of reclaimed peatlands. As peatland reclamation has only recently been attempted in the Alberta oil sands region, reclamation practices have not yet been well tested and have relied on knowledge from other study areas where peatland types and climate are dissimilar. On the Nikanotee fen different vegetation plots were created as part of an experimental design to help evaluate the most appropriate revegetation strategy. The effect of these plots on the hydrology of the fen is unknown. Therefore, this study aims to better understand the influences of various vegetation and treatment types, including wood-stand mulch which is untested in WBP peatland reclamation, on the energy balance and evaporative demand of a reclaimed fen in the WBP. Specifically, the objectives are to 1) quantify plot-scale surface energy fluxes and ET, and compare to site-scale fluxes, 2) determine specific plot-scale controls on ET; and 3) develop recommendations for future oil sands reclamation projects to reduce evaporative stress and aid in the establishment of a functioning fen ecosystem.

2. Methods 2.1. Study site The Nikanotee Fen (56.932◦ N, 111.417◦ W; Fig. 1) is a 2.9 ha pilot fen creation project that sits in a larger 32 ha constructed watershed within an oil sands development lease, 20 km north of Fort McMurray, Alberta. In this region of the WBP average temperatures (1981–2010) range from −17.4 ◦ C (January) to 17.1 ◦ C (July) and average annual precipitation is 419 mm, of which ∼76% is rainfall (measured at the Fort McMurray Airport ∼33 km south of the study; Environment Canada, 2014). Construction of the fen was completed in January 2013 and re-vegetated in June 2013. The peat used in the construction of this site was extracted from a local natural fen, which was scheduled for excavation for oil sands development purposes. The peat was placed to a depth of ∼2 m on the constructed site. The experimental planting was based on a factorial design that was divided into replicates of five vegetation plots, with mulched and weeding treatments (Borkenhagen, unpublished data, Fig. 1). Plot types included control (bare peat), moss, seedlings, seedling-moss and seeds. Moss, seedlings and seedling-moss plots were 17 m × 18 m, whereas control and seed plots were 8 m × 18 m. Smaller plots were chosen for control and seed treatments to reduce unplanted areas of the fen and avoid undesirable non-peatland species. Each plot was divided into four sub-plots (∼8.5 × 8.5 m), with two covered by wood-strand mulch and two weeded to remove non-peatland species. The wood-stand mulch (WoodStraw© ECM) was comprised of a wood blend split into 6.4–16 cm long and 0.5–1 cm thick stands and spread to an average cover of ∼90–95%. Moss was harvested from a donor fen, ∼8 km west of the constructed site, chosen due to presence of richfen moss species (i.e. Tomenthypnum nitens, Aulacomnium palustre, Sphagnum warnstorfii). The upper 10 cm of moss was mechanically harvested from previously disturbed cut lines, to be used in the moss-layer transfer technique (Quinty and Rochefort, 2003). The moss was hand spread on the constructed site to an average thickness of ∼1 cm. Seedlings were germinated in a greenhouse (Coast to Coast Reforestation, Smokey Lake Nursery, Alberta) and hand planted at the constructed site. Seedlings included salt-tolerant (Juncus balticus, Triglochin maritima, Calamagrostis inexpansa) and freshwater (Carex aquatilis, Betula pumila) species.

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

2.2. Environmental and microclimatic variables Data were collected during the second growing season postconstruction. Thirty-one study plots were selected of the 138 that were available for survey in 2014 (Borkenhagen, unpublished data), control (n = 6), moss (n = 6), moss-mulch (n = 7), salt-tolerant seedlings (n = 6) and salt-tolerant seedling-mulch (n = 6). All plots were in weeded treatments to remain consistent between plot types and reduce the influence of non-peatland species. Plots were chosen across a representative hydrological gradient, from dry to saturated, based on water table position. Flooded sections where the water table remained permanently above the plots surface were avoided. Freshwater seedling, seeds and seedling-moss plots were not monitored due to time and equipment constraints. In 2014, new peat was placed on the west side of the fen to fill large ponds. This area was not part of the original fen design and therefore was not included in this study. Fig. 1 represents the study site after the new peat was added, however, some ponding persisted. Plot measurements were collected 3–4 times weekly from June 2 to August 15. Water table and 0–5 cm soil moisture (; Delta-T Devices WET-Sensor) were measured to characterize differences in plot-scale hydrology. Field  measurements were corrected with a gravimetrically-determined calibration curve developed for the constructed fen peat using WET-Sensor readings taken from a sample with a known volume as it dried (Scarlett, 2016). Rel® ative humidly (RH) and air temperature (Ta ; Vaisala HUMICAP ◦ HMP42, temperature corrected at 20 C) were measured near the plot surface (under mulch in mulched plots) to monitor the nearsurface microclimate. Logging surface RH and T (Vaisala HMT337) measurements were averaged at 30-min intervals over control, moss and moss-mulch (under mulch) treatments. Seedling and

201

seedling-mulch plots did not have logging systems due to instrument limitations. Plot-scale thermal regimes were monitored using soil temperature (Tg ) profiles (Omega copper-constantan thermocouple) with manual measurements at 2.5, 5, 10, 20 and 30 cm. Stomatal resistance (Decagon SC-1 porometer) and leaf area index (LAI; AccuPAR LP-80, 1 m sensor length) were measured monthly at 23 locations in seedlings, seedling-mulch, seedling/moss and seedling/moss-mulch plots, including dominant salt-tolerant (Juncus balticus, n = 11) and freshwater (Carex aquatilis, n = 12) species. Measurements were made within a permanent 60 × 60 cm quadrat in each plot. Although overcast conditions are preferred when taking LAI readings, full sun conditions were more abundant during the study season. To insure consistency between measurement periods, LAI was measured indirectly in the field at half vegetation height under clear skies at approximately midday. While clear sky conditions can increase errors associated with leaf angle distribution (Garrigues et al., 2008), the LP-80 sensor accounts for changes in solar radiation factors, minimizing errors associated with clear conditions (Decagon Devices Inc., 2014). Triplicate porometry measurements were taken at the base, mid-height and top of one plant within each quadrat. Measurements were taken on a single leaf; open space was accounted for if the leaf did not cover the porometer’s full aperture (Phillips et al., 2015). Reported vegetation height was the average of all plants within the quadrat.

2.3. Energy fluxes and evapotranspiration A roaming meteorological station was used to account for spatial variability and compare energy fluxes between cover types. The station was moved over five cover types (Fig. 1) twice during the

Fig. 1. Map of the Nikanotee Fen in Fort McMurray, Alberta, Canada (inset). The fen meteorological station included a closed-path EC system. Ponds outlined in this map had permanent standing water during the study period. Note the outflow at the spill box to the northeast of the fen.

202

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

Table 1 Summary of daily mean energy fluxes (W/m2 ) from each plot type and the pond. QG divided into daytime and nighttime fluxes. Available energy (Q* – QG – S) was calculated with daily average QG values (not shown). In ponds, QG is reported as QW. Mulch heat storage (S) only calculated for mulched plots. Cover Type

Pond Control Moss Moss-Mulch Seedlings Seedlings-Mulch a

Q*

141 130 130 123 125 125a

QG 7:00–20:30 h

21:00–6:30 h

38 31 31 33 32 35

−26 4 1 9 −6 −2

S

QE

QH

Q* – QG – S

– – – 0.07 – 0.06

92 81 81 73 79 76

37 29 31 27 30 29

129 110 112 100 109 105

No Q* data over seedling-mulch plots; it is assumed similar to Q* over unmulched seedling plots.

season, each for 7–10 day periods, from May 28 to August 15 (80 total days). Two rotations of 7–10 days were chosen to capture a range of atmospheric conditions representative of early and late season periods (Brown et al., 2010). The station was cycled over moss-mulch, seedlings, ponds, moss and bare peat (control). This order was repeated for the second measurement period, except for bare peat, which had one measurement period due to instrument errors. Seedling-mulch was not included due to time limitations and trends were assumed similar to seedling plots based on visual similarity. Ta and RH were recorded with a HOBO Onset logger installed 1.0 m above the peat surface, housed in a perforated reflective cup to minimize heating by direct solar radiation. To measure ground heat flux, 2 heat flux plates (QG ; REBS HFT3) were installed 5 cm below the peat surface (one plate after July 13), and were paired with a Tg profile (copper-constantan thermocouple) with measurements at 2.5, 7.5 and 17.5 cm depths. Studies have suggested that heat flux plates underestimate fluxes in organic soils, specifically in peatlands, due to poor contact between the plate and peat (Halliwell and Rouse, 1987; Rouse et al., 1987). However, due to the dense, decomposed nature of the constructed fen peat, errors in the plate values were assumed to be negligible. Net radiation (Q*; Kipp & Zonen NRLite) was recorded at an average height of 0.83 m above ground surface to capture the representative cover type while also minimize shading errors associated with installing the sensor lower than 1 m (Campbell Scientific Inc., 2010). At this height, the sensor’s measurement radius would have been 8.3 m under ideal atmospheric conditions (10×sensor height), which extends beyond the sub-plot boundaries. However an error associated with the solar zenith angle reduced the sensor’s field of view by 3–5% (Campbell Scientific Inc., 2010), minimizing measurements from adjacent cover types. Since available energy were primarily driven by QG (Table 1) and plot specific calibrations were applied to calculated ET values, this likely had negligible effects on energy balance and ET results reported in this study. For pond measurements, the net radiometer was installed in the middle of the large Central Pond at 1.0 m (Fig. 1). The temperature profile of the water column (∼25 cm) was measured at 5 and 10 cm below the pond surface from a boardwalk that was built into the pond. Temperatures from these depths were used to capture the heat flux within the pond (QW ; QG plates were not used in the pond). The fixed probe depths were adjusted each site visit, if required, to reflect changes in pond elevation. The roaming station recorded all measurements at one-minute intervals, which were averaged every 30-min. To calculate continuous seasonal energy fluxes for each cover type the roaming station measurement periods were regressed against a “permanent” fen meteorological station for the same time periods. The permanent station measured Q* (Kipp and Zonen CNR4) at 2 m, Ta and RH (HOBO Onset) at 1.75 m and precipitation (Texas Electronics tipping bucket; located ∼100 m south of the fen). QG (REBS HFT3 and Hukseflux HFP01SC) was measured with plates under 4 plot types (5 cm below peat surface): moss-mulch,

seedling-mulch, seedling/moss-mulch, seed-mulch. The roaming station QG was regressed against the most representative QG site at the permanent station (i.e. roaming station moss QG was regressed against moss-mulch QG at the permanent station). As there was no roaming station over seedling-mulch, seedling-mulch QG from the permanent station was used and therefore, no regression was needed. Two regression equations, early and late season, were applied over the study season to calculate the aforementioned variables for each cover type. The split for the early and late season regressions was based on the mid-point between the two roaming meteorological station measurement periods. All regression relationships were significant (p < 0.05) with an R2 ≥ 0.70. The fen meteorological station also included an eddy covariance (EC) system. A 3-dimentional sonic anemometer (Gill Instruments Windmaster Pro) and closed-path infrared gas analyzer (IRGA) (LICOR Inc. LI7200) were placed 2.5 m above the peat surface. The EC system sampled at a frequency of 20 Hz and averaged every 30min. The IRGA was calibrated at the beginning and end of the study season. EC data were processed using EddyPro software (v5.2.1, LICOR Inc.), where ∼30% of data was lost after standard corrections were applied (Burba, 2013; Brown et al., 2010; Petrone et al., 2015), then gap filled using mean diurnal variations over 2-week moving windows (Falge et al., 2001). The surface energy balance was calculated for each cover type as follows, Q ∗ = QG + QE + QH

(1)

where Q* is the net radiative flux, QG is the ground heat flux, QE is the latent heat flux and QH is the sensible heat flux (W/m2 ). Heat storage in the mulch layer (S) was included in the QG term for mulched plots and calculated as, S = CM

T z t

(2)

where CM is the heat capacity of the mulch (J/m3 ◦ C), T is temperature (◦ C), t is time (s) and z is depth (0.02 m). The air-filled porosity of the wood-strand mulch was 0.6 (calculated gravimetrically), thus CM equals 0.6Cair + 0.4Cwood , where the heat capacity of wood is 0.82 × 106 J/m3 ◦ C. In ponds, the energy stored in the water column, QW , replaces QG in Eq. (1). QW was calculated calorimetrically, as described in Halliwell and Rouse (1987), using the pond temperature profile and the known heat capacity of water (Oke, 1987), QW = CW

T T z + ␬T t z

(3)

where CW is the heat capacity of water (4.18 × 106 J/m3 ◦ C) and ␬T is thermal diffusivity (m2 /s). Plot-scale equilibrium evapotranspiration (ETeq ) was calculated using the Priestley-Taylor available energy-based approach

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

203

30

20

Precipitation (mm)

15

20 15

10 10

Air Temperature (°C)

25

5

5

0 -5

0 152

162

172

182

192

202

212

222

Day of Year Fig. 2. Total daily precipitation (bars) and mean daily air temperature (line) from June 2 to August 15, 2014.

(Priestley and Taylor, 1972). The Priestley-Taylor equation calculates ETeq as, ET eq =

 (Q ∗ −QG ) +y

(4)

where  is the slope of the saturation vapour-pressure curve (kPa/◦ C) and ␥ is the psychrometric constant (0.00662 kPa/◦ C at 20 ◦ C) and ␭ is the latent heat of vaporization (MJ/kg) Actual evapotranspiration (ETa ) was measured 3–4 times weekly using weighing lysimeters located in each of the 31 study plots and 3 floating evaporation pans located in the SW and Central ponds (Fig. 1). Lysimeters (10 L, 19 × 28 × 19 cm, black plastic) were filled with peat and covered with the representative plot vegetation and mulch. Seedlings were transplanted into lysimeters at the time of planting in 2013 and remained healthy and visually representative of surrounding seedlings during the years of this study. Evaporation pans (53 × 35 × 15 cm) were clear plastic to minimize heating and were encompassed by flexible polyethylene foam tubes to ensure they followed pond stage fluctuations. Water loss from the evaporation pans was measured using incremented tape along the sides of the pans. Water was added to the lysimeters and evaporation pans when needed to maintain a similar water level as the surrounding environment. ETeq can be related to ETa using, ␣ = ET a /ET eq

to isolate differences between groups. Regression analyses were used to test for significant relationships between ET and environmental variables. Normality and homoscedasticity of significant relationships were checked visually using residual plots. A 95% confidence level was the significance threshold for all tests. R statistical software (R Core Team, 2013) was used for all statistical data analyses. 3. Results 3.1. Climate The study site received 132 mm of precipitation (P) from June 2 to August 15 (Fig. 2), less than the 30-year regional average (211 mm; Environment Canada, 2014). The average rainfall event was approximately 3.9 mm, with the largest event delivering 14.2 mm on July 29 (DOY 210). Average daily air temperature during the study period was 19 ◦ C, ranging from 7.9 ◦ C on June 5 (DOY 156) to 26.8 ◦ C on July 14 (DOY 195; Fig. 2). No significant difference was found between Ta (1 m) of the different plot types and over the pond (Kruskal-Wallis, p > 0.05; Table 2). 3.2. Plot-scale energy balances and evapotranspiration

(5)

where ␣ is the Priestley-Taylor coefficient of evaporability, which is the slope of the regression. Total measured ETa was compared to total modeled ETeq over the same time periods, from which ␣values were derived. All ETa /ETeq relationships had an R2 ≥ 0.47 and were significant at p < 0.01 (Scarlett, 2016). Under equilibrium conditions ␣ = 1; however, these conditions are uncommon. Therefore, ␣-adjusted ETeq is herein referred to as ET. 2.4. Statistical analysis Data sets were tested for normality using the Shapiro-Wilk test. For normally distributed datasets, an ANOVA was used to determine differences among group means, whereas the non-parametric Kruskal-Wallis one-way analysis of variance was used when a normal distribution could not be assumed. If significant effects were found, a post-hoc pairwise t-test (parametric) or Wilcoxon rank sum test (non-parametric) with a Bonferroni adjustment was used

Seasonal energy flux trends were similar between all cover types (not shown), however magnitudes differed between vegetated plots and ponds (Table 1). Q* was the highest over ponds compared to all plot types, whereas moss-mulch and seedling plots had the lowest average Q* over the season. Daytime pond QW was slightly higher on average compared to plot QG values, although nighttime QW were significantly lower (Kruskal-Wallis, p < 0.05, Table 1). Daytime QG did not differ significantly between plot types (Kruskal-Wallis, p > 0.05); however, nighttime mulched plot QG was consistently higher (i.e. less negative) than respective unmulched plots. Mulch heat storage was similar between both moss-mulch and seedling-mulch plots, yet had a negligible influence on plot QG . Although net seasonal mulch heat storage was minimal, it represented 15% and 10% of noon QG in moss-mulch and seedling-mulch plots, respectively. Available energy (Q*–QG –S) followed a trend where ponds > moss > control > seedlings > seedling-mulch > mossmulch (Table 1). Available energy over both seedling and seedling-mulch plots is similar to that over bare peat (falls on 1:1

204

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

Table 2 Summary of ET and environmental variables for each plot type and pond in 2014. ET shown is ␣-adjusted, based on the Priestley and Taylor equation (Eqs. (4) and (5)), and average daily rates. Air temperature was measured at 1 m above the surface (Ta-1m ) and at the near-surface (Ta-surf ; under mulch in mulched plots). Soil moisture () was averaged over a 0–5 cm depth. Water table (WT) is presented as cm below ground surface (bgs). Plot ET rates with different letters indicates a significant difference (p < 0.05). Cover Type

ET (mm)

ET rate (mm/day)

Alpha (␣)

Ta-1m (◦ C)

Ta-surf (◦ C)

 (cm3 /cm3 )

WT (cm bgs)

Pond Control Moss Moss-Mulch Seedlings Seedlings-Mulch

331 246 212 179 294 272

4.4 a 3.2 bc 2.8 bd 2.4 d 3.9 ae 3.6 ce

1.35 1.14 0.98 0.91 1.40 1.33

18.9 20.2 19.2 19.5 19.1 19.1a

22.4b 27.0 27.4 25.7 27.1 24.6

– 0.87 0.86 0.83 0.87 0.86

– 6 5 7 5 5

a b

No Ta-1m data over seedling-mulch plots; it is assumed similar to Ta over unmulched seedling plots. Pond Ta-surf is measured at 5 cm below the pond surface.

a)

250

b)

250

Q*-QG (W/m2)

Moss

Seedlings

200

200

150

150

100

100 Unmulched

50

50

0

Mulch

0 50

0

100

150

200

250

0

Bare Peat Q*-QG (W/m2)

50

100

150

200

250

Bare Peat Q*-QG (W/m2)

Fig. 3. Available energy relationships between bare peat and a) moss plots and b) seedling plots. Mulched plot available energy includes mulch heat storage (S). The solid line represents a 1:1 ratio.

350

Ponds (331)

300

EC Station (302) Seedlings (294)

250

Control (246)

Cumulative Flux (mm)

Seedling-Mulch (273)

Moss (212)

200

Moss-Mulch (179) 150 Precipitation (132) 100 50 0 150

160

170

180

190 200 Day of Year

210

220

230

Fig. 4. Cumulative evapotranspiration from the plots, ponds and EC station relative to cumulative precipitation from June 2 to August 15, 2014.

line; Fig. 3b). Moss-mulch plots consistently have lower available energy than bare peat; however, available energy over unmulched moss plots is more variable and generally equal to or greater than that over bare peat (Fig. 3a). Plot ETeq was related to ETa via the ␣ coefficients presented in Table 2. All values where ␣ > 1 means ETa exceeds ETeq for these cover types. Moss and moss-mulch were the only plots where ␣ < 1. Plot-scale ET trends followed ponds > seedlings > seedlingmulch > control > moss > moss-mulch (Table 2). Cumulative ET increased steadily over the study season (Fig. 4) and exceeded P for

all cover types after June 12 (DOY 163). P-ET ranged from −46 mm over mulch-moss plots to −199 mm over ponds. Site-scale ET measured at the EC station (302 mm) fell within the range of measured cover type ET (Fig. 4) and showed significant relationships with all cover types (p < 0.05). Seasonal EC ET most closely resembled that from ponds and seedlings (Fig. 4). However, it was significantly greater than control, moss and moss-mulch ET (ANOVA, p < 0.05) and significantly different from the cover type weighted site average (260 mm; p < 0.05).

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

205

Fig. 5. Seedling a) height, b) leaf area index (LAI) and c) stomatal resistance over the study season. Box plots represent the median, 25th and 75th percentiles and whiskers indicate the 5th and 95th percentiles. Points that extend beyond the whiskers are outliers. Means with different letters indicates a significant difference (p < 0.05).

3.3. Plot influences on evapotranspiration The greatest increase in plant height and LAI occurred between June and July (Fig. 5a & b), whereas stomatal resistance decreased in seedling plots from June to August (Fig. 5c). Decreasing stomatal resistance is consistent with an increase in mean ET rates from June to August (3.4 and 4.0 mm/day, respectively). However, there was

no significant difference between measurement periods (KruskalWallis, p > 0.05; Fig. 5c). Average water tables for each plot type only ranged from 1 to 12 cm below the ground surface and  remained > 0.69 within all plots. Therefore, there was no significant relationship between  and ET (p > 0.05; Fig. 6a) for individual plots or lumped data. There was a weak (R2 = 0.15) yet significant relationship between water table and ET (p < 0.05). However, when specific plot ET versus

206

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

Fig. 6. Relationships between daily mean ET and a) soil moisture () b) water table c) vapour pressure deficit (VPD) for each plot type. Significant relationships (p < 0.05) are indicated by trendlines and R2 values.

water table relationships are considered, only the seedling and seedling-mulch relationships were significant (R2 = 0.22 and 0.23 respectively; Fig. 6b). Differences in surface peat Tg (2.5 cm bgs) for each plot type are shown in Fig. 7b. There is greater viability early in the season (June), where Tg is noticeably lower in mulched plots. However, differences between plot Tg profiles are reduced in July and August, and thus there were no significant relationship between seasonal average Tg and ET (p > 0.05). Tg did differ significantly with depth from 2.5 to 30 cm bgs in all plot types over the study season (KruskalWallis, p < 0.05; not shown).

than over unmulched plots (1.4 kPa; paired t-test, p < 0.001). However, ET did not have a significant relationship with near-surface VPD for individual plots or lumped data (p > 0.05; Fig. 6c), or when measured 1 m above plot surface (ANOVA, p > 0.05). RH remained significantly higher under mulch for the entire season (Fig. 7a; Kruskal-Walis, p < 0.05), Tg at 2.5 cm increased significantly from June to August in mulched plots (Fig. 7b; Wilcoxon test, p < 0.05). Consequently, there was no significant difference in 2.5 cm Tg between all plots in August.

4. Discussion 3.4. Influence of mulch cover Although, moss-mulch ET rates were the lowest and significantly different from all cover types (Table 2), they did not differ significantly from unmulched moss plots. Similarly, seedlingmulch plots showed lower ET rates than unmulched seedling plots, however, they did not differ significantly from seedling or control plots (ANOVA, p > 0.05). There were no significant differences between seedling height, LAI or stomatal resistance between mulched and unmulched plots (Fig. 5a–c; Kruskal-Wallis, p > 0.05). Peak ET rates in mulched plots were 0.9 and 0.2 mm/day less than unmulched plots for moss and seedlings, respectively. Mulch increased RH at the peat surface (Fig. 7a), reducing nearsurface temperatures and VPD. Near-surface RH in a mulched-moss plot was at 100% for 23 and 51% longer than over unmulched moss and bare peat, respectively (Fig. 8). Ta at the plot surface was ∼2 ◦ C lower on average under mulch (Table 2). Consequently, near-surface VPD beneath mulch (0.6 kPa) was significantly lower

4.1. Vegetation controls on surface energy and evapotranspiration Available energy was greatest over ponds but varied minimally between vegetation types. While control and moss plots had higher Q*, these plots also had greater QG fluxes compared to seedling plots (Table 1). Differences in energy fluxes between the pond and fen were likely caused by differences in thermal regimes. Open water albedo is similar to that of wet peat, but higher Q* over the pond could be a result of reduced longwave emittance since the surface Ta of the pond was lower than that over the plots (Table 2). Larger fluxes and greater differences between day and nighttime values in pond QW compared to plot QG (Table 1) are likely due to a more rapid energy transfer within the shallow water column (∼10 cm; Petrone et al., 2007) caused by convective mixing (Oke, 1987; Petrone et al., 2007). QE constantly exceeded QH and accounted for the largest portion of surface energy fluxes over all

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

207

Fig. 7. Changes in a) surface relative humidity and b) soil temperature for each plot type over the study season. Box plots represent the median, 25th and 75th percentiles and whiskers indicate the 5th and 95th percentiles. Points that extend beyond the whiskers are outliers.

cover types, likely due to prevalently wet conditions in the fen for the entire study period. ETeq is largely controlled by available energy and therefore followed the same trends across all cover types. The ␣-values relating ETa to ETeq indicate that all cover types, except moss and mossmulch, exceeded ETeq (Table 2). Petrone et al. (2007) observed similar trends in a WBP peatland complex, with ␣-values <1 over moss-dominated surfaces and >1 over open water. Low ␣-values in moss plots can be attributed to the mosses’ limited ability to vertically transmit water under dry conditions (Price, 1997; Goetz and Price, 2015) and possibly poor connectivity to the underlying placed peat, as observed post-restoration in some harvested sites (McCarter and Price, 2014). Taylor and Price (2015) observed that the water table- relationship weakened with regenerated moss profile height, the moss layer on these plots was still thin (∼1 cm),

and probably does not yet greatly influence moisture connectivity. The poor relationship between ET and  or water table position is likely due to consistently wet conditions with a very small range in  and water table that do not result in a limitation in water availability. High ␣-values in seedling plots are likely due to dense root growth, vascular structure and relatively large LAI, which increases transpiration from these plots (Romanov, 1968; Takagi et al., 1999; Farrick and Price, 2009). Connectivity of vascular plants to the water table through rooting likely explains the stronger relationship between ET and water table position for these plots (Fig. 6b; Rezanezhad et al., 2012). Despite increased ET in seedling plots through transpiration, pond evaporation surpassed that of the adjacent vegetated fen. Price (1994) found latent heat fluxes over a Typha dominated marsh and adjacent open water were similar, with stomatal resistance of the plants (seasonal average 74 s/m)

208

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

Relative Humidity (%)

100 80 60 40

Moss-Mulch Moss

20

Bare Peat 0 0

20

40

60

80

100

% Time Exceeded Fig. 8. Exceedance probability curve for surface relative humidity. Dashed lines indicate percent time at 100% relative humidity for each plot type.

being compensated by their lower aerodynamic resistance. However at the constructed fen stomatal resistance of the sedges was substantially higher (Fig. 5c) and their low form suggests higher aerodynamic resistance, resulting in less evaporative efficiency than the pond. Lower ET rates of seedling plots can be explained by the physiological controls of vascular plants on water vapour loss. Phillips et al. (2015) found that stomatal resistance can increase under waterlogged and saline conditions. Therefore, consistently high water tables observed on the constructed fen could limit transpiration. Transpiration rates could also be affected if saline conditions develop in the future due to the presence of salts in oil sands process-affected waters. As the prevailing wind direction was from the south-southwest, the area contributing to the EC station included the SW Pond and well saturated plots (Fig. 1). Greater evaporative rates over ponds likely explain the higher site-scale ET measured at the EC station, relative to the cover type weighted site average, and stronger relationship to pond evaporation (not shown). A decrease in stomatal resistance in seedling plots over the season supports an increase in ET since leaf stomata control water loss through transpiration (Oke, 1987). The LAI and stomatal resistance relationship are similar to other studies (Schulze et al., 1995; Leuning et al., 1995), where increasing LAI indicates larger leaf surfaces contribute to canopy conductance. Furthermore, surface roughness is a function of vegetation height and LAI, which positively affects ET due to turbulent mixing as they increase over the season (Oke, 1987).

ing atmosphere-surface energy exchange to the mulch layer and reducing conductive heat loss from the peat. Furthermore, this is the first study using wood-strand mulch in peatland reclamation; previous studies used straw. Although the wood-strand mulch has a smaller air-filled porosity (60%) than straw mulch (90%, Hares and Novak, 1992), it has a much greater mass and therefore the ability to store more energy. Mulched plots had lower ET rates than the respective unmulched vegetation. However, mulched and unmulched seedling plots did not differ as greatly as mulched versus unmulched moss. ET rates from moss-mulch plots are relatively low and appear to be clustered within a small range (Fig. 6c), suggesting some mulch control on ET in moss-mulch sites. In contrast, seedling-mulch plots show high ET rates within a range of low VPD (Fig. 6c), suggesting nearsurface VPD has less influence on ET in seedling plots. Similarities between seedling plots are likely because mulch cover only limited soil evaporation, whereas transpiration rates were comparable between both plot types (Fig. 5c). VPD at 1 m did not differ between mulched and unmulched seedling plots and therefore, likely did not cause the slight differences in stomatal resistance. Lower nearsurface VPD and thus lower ET from mulch plots is a result of poor mixing within the mulch (Hares and Novak, 1992), evident by higher RH in mulch compared to the unmulched surfaces (Fig. 8). However, the importance of mulch cover lessened over the study season. The influence of mulch on Tg diminished from June to August and consequently, lower Tg under mulch likely did not affect ET rates from these plots.

4.2. Effects of mulch on evapotranspiration

4.3. Reclamation implications

Q* was consistently lower over moss-mulch plots than unmulched moss plots. Sharratt and Campbell (1994) found that straw mulch on agricultural fields had an albedo from 0.2 to 0.3, which increased reflected radiation and thus lowered Q* (Novak et al., 2000). A similar comparison between mulched and unmulched seedling plots cannot be made as there are no Q* data over seedlingmulch plots. Comparable daytime QG under all plot types shows that mulch did not have a significant effect on daytime QG fluxes, which contrasts with findings from other studies (Price et al., 1998; Petrone et al., 2004). However, lower nighttime QG losses under mulch compared to respective unmulched plots shows that mulch preferentially limited heat loss at night. Nighttime near-surface soil temperatures were on average greater in mulched versus unmulched plots (17.1 vs. 15.9 ◦ C, respectively). This is likely due to the insulating effects of the mulch (Price et al., 1998), thus limit-

In the sub-humid WBP, this study shows that mulching is an important reclamation practice for the conservation of water in peatland reclamation, especially when used with moss revegetation strategies. Borkenhagen (unpublished data) showed that mulch improved moss establishment during the first 2 years post-reclamation, with greatest percent cover when moss was combined with mulch and seedling treatments. Similarly, Price et al. (1998) and Rochefort et al. (2003) found mulch was important for protecting moss during the establishment phase of a restored peatland. Mulch reduced water loss through decreased ET and created a microclimate more favorable for peatland vegetation establishment. Malloy and Price (2014) provide evidence that a moss-sedge cover in a restored fen evaporated well below the potential rate. However, in this study, mulch did not appear to affect seedling ET. Considering the similar effects of seedlings to those of

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

mulch in regulating energy fluxes, combining seedling and mulch treatments may not be necessary in future reclamation projects depending on re-vegetation goals, especially if the transferred moss layer has a viable vascular plant seedbank. Additionally, during drying events mulch appeared to accumulate salts (salt-crusting observed on mulch), reducing soil and pore-water electrical conductivity under mulch (Scarlett, unpublished data). This may help regulate the accumulation of dissolved minerals that locally are present in the peat, through aeolian deposition and the presence of oil sands process-affected waters in groundwater discharge to the fen (Rezanezhad et al., 2012). Fen mosses are sensitive to elevated salinity (Pouliot et al., 2012; Rezanezhad et al., 2012; Pouliot et al., 2013). The importance of straw mulch decreases substantially after 3 years because it decomposes rapidly, and is associated with elevated CO2 emissions (Petrone et al., 2001; Waddington et al., 2003). However, due to the greater mass and structure of the wood-strand mulch used in this study, decomposition may be slowed and its water and energy regulation effect could persist for longer.

5. Conclusion Vegetation type plays a significant role in controlling ET from the constructed fen, where all plot types showed lower ET rates than evaporation from ponds. Seedling plots showed the greatest ET rates compared to moss and control plots, likely due to the connectivity of vascular plants to sub-surface water and their ability to transpire. Cover type and contributing area probably influenced site-scale ET measured at the EC station, resulting in a stronger relationship with pond evaporation, relative to vegetated plot ET. However, available energy alone did not provide an accurate estimate of ET. Alpha-values show that ETa > ETeq in all cover types, except in moss plots, where ␣ < 1. Mulch reduced ET in moss plots, but had less of an impact on seedling plots, as it had no noticeable effect on vascular stomatal resistance. Available energy, and thus ET, was lower in mulched plots under all energy conditions, likely due to lower Q* over mulch as a result of a higher albedo (not shown). Mulch dampened temperature fluctuation, keeping the surface and upper peat profile cooler during the day and warmer at night. Mulch also created a near-surface microclimate that increased RH and lowered VPD in the mulch layer, further mitigating ET losses. Further research is required to determine the role of mulch after the initial years post-reclamation and in the mitigation of salt accumulation at the surface. Consideration of the cumulative effect of vegetation and treatment types is required to develop recommendations for future peatland reclamation projects. Knowledge of the combined implications for the hydrology, ecology and biogeochemistry are necessary to develop future planting strategies and create a successful, self-sustaining fen ecosystem.

Acknowledgements The authors would like to thank Eric Kessel, Dylan Price, Emma Bocking, Scott Ketcheson, and James Sherwood for their assistance in the field and lab. We also thank Corey Wells and George Sutherland for their analysis of EC station data. The advice of Dr. Maria Stack on statistical analysis was greatly appreciated. Funding for this research was provided by the Natural Science and Engineering Research Council (NSERC) Collaborative Research and Development Grant program (Jonathan Price), the Canadian Oil Sands Innovation Alliance and its contributing companies, Suncor Energy, Shell Canada Energy and Imperial Oil Resources, and the Northern Scientific Training Program.

209

References Admiral, S.W., Lafleur, P.M., 2007. Partitioning of latent heat flux at a northern peatland. Aquat. Bot. 86 (2), 107–116. Admiral, S.W., Lafleur, P.M., Roulet, N.T., 2006. Controls on latent heat flux and energy partitioning at a peat bog in eastern Canada. Agric. For. Meteorol. 140, 308–321. Brigham, S.D., Pastor, J., Updegraff, K., Melterer, T.J., Johnson, K., Harth, C., Chen, J., 1999. Ecosystem control over temperature and energy flux in northern peatlands. Ecol. Appl. 9, 1345–1358. Brown, S., Petrone, R.M., Mendoza, C., Devito, K., 2010. Surface vegetation controls on evapotranspiration from a sub-humid Western Boreal Plain wetland. Hydrol. Processes 24, 1072–1085. Burba, G., 2013. Eddy Covariance Method for Scientific, Industrial, Agricultural, and Regulatory Applications: A Field Book on Measuring Ecosystem Gas Exchange and Areal Emission Rates. LI-COR Biosciences, Lincoln, NE, USA, 331 pp. Campbell Scientific Inc, 2010. NR-LITE Net Radiometer: Instruction Manual. Campbell Scientific Inc. Chee, W.-L., Vitt, D.H., 1989. The vegetation, surface water chemistry and peat chemistry of moderate-rich fens in central Alberta, Canada. Wetlands 9, 227–262. Cumulative Environmental Management Association (CEMA), 2014. Guidelines for Wetland Establishment on Reclaimed Oil Sands Leases, 3rd ed. West Hawk Associates, Fort McMurray, AB. Daly, C., Price, J., Rochefort, L., Rezanezhad, F., Poulet, R., Graf, M., 2012. Consideration for building a fen peatland in a post-mined oil sands landscape. In: Vitt, D., Bhatti, J. (Eds.), Restoration and Reclamation of Boreal Ecosystems: Attaining Sustainable Development. Cambridge University Press, Cambridge, pp. 179–198. Decagon Devices Inc, 2014. LAI: Theory and Practice (v.1.0). Pullman, Washington, USA, 11 pp. Devito, K.J., Creed, I., Gan, T., Mendoza, C., Petrone, R., Silins, U., Smerdon, B., 2005. A framework for broad-scale classification of hydrologic response units on the Boreal Plain: is topography the last thing to consider? Hydrol. Processes 19, 1705–1714. Environment Canada, 2014. Canadian Climate Normals 1981–2010: Fort McMurray Alberta, http://climate.weather.gc.ca/climate normals/index e.html (Accessed April 2014). Falge, E., et al., 2001. Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric. For. Meteorol. 107, 43–69. Farrick, K.K., Price, J.S., 2009. Ericaceous shrubs on abandoned block-cut peatlands: implications for soil water availability and Sphagnum restoration. Ecohydrology 2, 530–540. Garrigues, S., Shabanov, N.V., Swanson, K., Morisette, J.T., Baret, F., Myneni, R.B., 2008. Intercomparison and sensitivity analysis of Leaf Area Index retrievals from LAI-2000, AccuPAR, and digital hemispherical photography over croplands. Agric. For. Meteorol. 148, 1193–1209. Goetz, J.D., Price, J.S., 2015. Role of morphological structure and layering of Sphagnum and Tomenthypnum mosses on moss productivity and evaporation rates. Can. J. Soil Sci. 95, 109–124. Haapalehto, T., Kotiaho, J.S., Matilainen, R., Tahvanainen, T., 2014. The effects of long-term drainage and subsequent restoration on water table level and pore water chemistry in boreal peatlands. J. Hydrol. 519, 1493–1505. Halliwell, D.L., Rouse, W.R., 1987. Soil heat flux in permafrost: characteristics and accuracy of measurement. J. Climatol. 7, 571–584. Hares, M.A., Novak, M.D., 1992. Simulation of surface energy balance and soil temperature under strip tillage: II. Field test. Soil Sci. Soc. Am. J. 56, 29–36. Harris, M.L., 2007. Guidelines for Wetland Establishment on Reclaimed Oil Sands Leases, 2nd ed. Prepared by Lorax Environment for CEMA Wetlands and Aquatics Subgroup of the Reclamation Working Group. Fort McMurray, Alberta, Canada. Holden, J., Wallage, Z.E., Lane, S.N., McDonald, A.T., 2011. Water table dynamics in undisturbed, drained and restored blanket peat. J. Hydrol. 402, 103–114. Ingram, H.A.P., 1983. Ecosystems of the world. Volume 4A, Mires: Swamp, Bog, Fen and Moor, General Studies. In: Gore, A.J.P. (Ed.), Hydrology. Elsevier, New York, USA, pp. 67–158. Ivanov, K.E., 1981. Water Movement in Mirelands. Academic, New York, New York, USA. Ketcheson, S.J., Price, J.S., 2011. The impact of peatland restoration on the site hydrology of an abandoned block-cut bog. Wetlands 24 (1). Kucera, C.L., 1954. Some relationships of evaporation rate to vapor pressure deficit and low wind velocity. Ecology 35, 71–75. Lafleur, P.M., Hember, R.A., Admiral, S.W., Roulet, N.T., 2005. Annual and seasonal variability in evapotranspiration and water table at a shrub-covered bog in southern Ontario, Canada. Hydrol. Processes 19, 3533–3550, http://dx.doi.org/ 10.1002/hyp.5842. Leuning, R., Kelliher, F.M., Depury, D.G.G., Schulze, E.D., 1995. Leaf nitrogen, photosynthesis, conductance and transpiration – scaling from leaves to canopies. Plant Cell Environ. 18 (10), 1183–1200. Malloy, S., Price, J.S., 2014. Fen restoration on a bog harvested down t sedge peat: a hydrological assessment. Ecol. Eng. 64, 151–160. McCarter, C.P., Price, J.S., 2013. The hydrology of the Bois-des-Bel peatland restoration: 10 years post-restoration. Ecol. Eng. 55, 73–81. McCarter, C.P., Price, J.S., 2014. The hydrology of the Bois-des-Bel peatland restoration: hydrophysical properties limiting connectivity between

210

S.J. Scarlett et al. / Ecological Engineering 100 (2017) 199–210

regenerated Sphagnum and remnant vacuum harvested peat deposit. Ecohydrology 8 (2), 173–187. Novak, M.D., Chen, W., Orchansky, A.L., Ketler, R., 2000. Turbulent exchange processes within and above a straw mulch. Part II: thermal and moisture regimes. Agric. For. Meteorol. 102, 155–171. Oke, T.R., 1987. Boundary Layer Climates. Routledge Press, New York. Peichl, M., Sagerfors, J., Lindroth, A., Buffam, I., Grelle, A., Klemedtsson, L., Laudon, H., Nilsson, M.B., 2013. Energy echange and water budget partitioning in a boreal minerogenic mire. J. Geophys. Res.: Biogeosci. 118, 1–13. Petrone, R.M., Waddington, J.M., Price, J.S., 2001. Ecosystem scale evapotranspiration and net CO2 exchange from a restored peatland. Hydrol. Processes 15, 2839–2845. Petrone, R.M., Price, J.S., Waddington, M.J., von Waldow, H., 2004. Surface moisture and energy exchange from a restored peatland, Quebec, Canada. J. Hydrol. 295, 198–210. Petrone, R.M., Silins, U., Devito, K.J., 2007. Dynamics of evapotranspiration from a riparian pond complex in the Western Boreal Forest, Alberta, Canada. Hydrol. Processes 21, 1391–1401. Petrone, R.M., Chasmer, L., Hopkinson, C., Silins, U., Landhäusser, S.M., Kljun, N., Devito, K.J., 2015. Effects of harvesting and drought on CO2 and H2 O fluxes in an aspen-dominated Western Boreal Plain forest: early chronosequence recovery. Can. J. For. Res. 45 (1), 87–100, http://dx.doi.org/10.1139/cjfr-2014-0253. Phillips, T., Petrone, R.M., Wells, C.M., Price, J.S., 2016. Characterizing dominant controls governing evapotranspiration within a natural saline fen in the Athabasca Oil Sands of Alberta, Canada. Ecohydrology 9 (5), 817–829, http:// dx.doi.org/10.1002/eco.1685. Pouliot, R., Rochefort, L., Graf, M.D., 2012. Impacts of oil sands process water on fen plants: implications for plant selection in required reclamation projects. Environ. Pollut. 167, 132–137. Pouliot, R., Rochefort, L., Graf, M.D., 2013. Fen mosses can tolerate some saline conditions found in oil sands process water. Environ. Exp. Bot. 89, 44–50. Price, J.S., Rochefort, L., Quinty, F., 1998. Energy and moisture considerations on cutover peatlands: surface microtopography, mulch cover and Sphagnum regeneration. Ecol. Eng. 10 (4), 293–312. Price, J.S., McLaren, R.G., Rudolph, D.L., 2010. Landscape restoration after oil sands mining: conceptual design and hydrological modeling for fen reconstruction. Int. J. Min. Reclam. Environ. 24 (2), 109–123. Price, J.S., 1991. Evapotranspiration from a blanket bog in a foggy coastal environment. Boundary Layer Meteorol. 57, 391–406. Price, J.S., 1994. Evapotranspiration from a lakeshore Typha march on Lake Ontario. Aquat. Bot. 48, 261–272. Price, J.S., 1997. Soil moisture, water tension and water table relationship in a managed cutover bog. J. Hydrol. 202 (1–4), 21–32. Priestley, C.H.B., Taylor, R.J., 1972. On the assessment of surface heat flux and evaporation using large-scale parameters. Mon. Weather Rev. 100, 81–92. Quinty, F., Rochefort, L., 2003. Peatland Restoration Guide, 2nd ed. Canadian Sphagnum Peat Moss Association and New Brunswick Department of Natural Resources and Energy, Quebec, Quebec, Canada. R Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing (v3.0.2), Vienna, Austria, http://www.Rproject.org/ (Accessed October 2014).

Rezanezhad, F., Anderson, R., Pouliot, R., Price, J.S., Rochefort, L., Graf, M.D., 2012. How fen vegetation structure affects the transport of oil sands process-affected waters. Wetlands 32, 557–570. Rochefort, L., Quinty, F., Campeau, S., Johnson, K., Malterer, T., 2003. North American approaches to the restoration of Sphagnum dominated peatlands. Wetlands Ecol. Manage. 11, 3–20. Romanov, V.V., 1968. Hydrophysics of Bogs. Israel Program for Scientific Translations, Jerusalem, Israel. Rooney, R.C., Bayley, S.E., 2011. Setting reclamation targets and evaluating progress: submersed aquatic vegetation in natural and post-oil sands mining wetlands in Alberta, Canada. Ecol. Eng. 37 (4), 569–579. Rouse, W.R., Hardill, S.G., Lafleur, P.M., 1987. The energy balance in the coastal environment of James Bay and Hudson Bay during the growing season. J. Climatol. 7, 165–179. Runkle, B.R.K., Willie, C., Gazovic, M., Wilmking, M., Kutzbach, L., 2014. The surface energy balance and its drivers in a boreal peatland fen od northwestern Russia. J. Hydrol. 511, 359–373. Rydin, H., Jeglum, J., 2006. The Biology of Peatlands. Oxford University Press, New York, New York, USA, 64–66 pp. Scarlett, S.J., 2016. Characterizing controls on plot-scale evapotranspiration and soil water dynamics of aconstructed fen in the Athabasca Oil Sands Region. In: Alberta (MSc Thesis). UWSpace http://hdl.handle.net/10012/9747. Schulze, E.D., Leuning, R., Kelliher, F.M., 1995. Environmental regulation of surface conductance for evaporation from vegetation. Vegetatio 121 (1–2), 79–87. Sharratt, B.S., Campbell, G.S., 1994. Radiation balance of a soil-straw surface-modified by straw colour. Agron. J. 86 (1), 200–203. Strack, M., Keith, A.M., Xu, B., 2014. Growing season carbon dioxide and methane exchange at a restored peatland on the Western Boreal Plain. Ecol. Eng. 46, 231–239. Takagi, K., Tsuboya, T., Takahashi, H., Inoue, T., 1999. Effect of the invasion of vascular plants on heat and water balance in the Sarobetsu mire, northern Japan. Wetlands 19 (1), 246–254. Taylor, N., Price, J., 2015. Soil water dynamics and hydrophysical properties of regenerating Sphagnum layers in a cutover peatland. Hydrol. Processes 29 (18), 3878–3892, http://dx.doi.org/10.1002/hyp.10561. Thompson, D.K., Benscoter, B.W., Waddington, J.M., 2014. Water balance of a burned and unburned forested boreal peatland. Hydrol. Processes 28 (24), 5954–5964. Vitt, D.H., Halsey, L., Thormann, M.N., Martin, T., 1996. Peatland Inventory of Alberta. Alberta Peat Task Force, University of Alberta, Edmonton, Alberta, Canada. Vitt, D.H., Halsey, L., Bauer, I.E., Campbell, C., 2000. Spatial and temporal trends of carbon sequestration in peatlands of continental western Canada through the Holocene. Can. J. Earth Sci. 37, 683–693. Waddington, J.M., Greenwood, M.J., Petrone, R.M., Price, J.S., 2003. Mulch decomposition impedes net recovery of carbon sink. Ecol. Eng. 20, 199–210.