Carbon dioxide and methane fluxes in variably-flooded riparian forests

Geoderma 241–242 (2015) 41–50

Contents lists available at ScienceDirect

Geoderma journal homepage: www.elsevier.com/locate/geoderma

Carbon dioxide and methane fluxes in variably-flooded riparian forests P.A. Jacinthe ⁎ Department of Earth Sciences, Indiana University Purdue University, Indianapolis, IN 46202, United States

a r t i c l e

i n f o

Article history: Received 24 May 2014 Received in revised form 14 October 2014 Accepted 15 October 2014 Available online xxxx Keywords: Riparian forest Overbank flooding Methane uptake Flood protected

a b s t r a c t The water quality protection function of riparian buffers is widely recognized, but much less is known regarding the dynamics of greenhouse gases in these ecosystems. Carbon dioxide (CO2) and methane (CH4) fluxes were monitored at 6 riparian sites along a 4th-order segment of the White River (Indiana, USA) to assess the effect of vegetation and flood frequency on gas fluxes. The study sites included shrub/grass, young (b15 years) and mature (N80 years) riparian forests that were either flood-protected (FP), occasionally flooded (OF) or frequently flooded (FF). No significant effect of vegetation type on either CO2 or CH4 flux was noted. While CH4 level was sometimes high (up to 120 μL L−1) in the deep soil layers, concentration near the soil surface (1.28 μL L−1) was generally lower than in the litter layer (2.35 μL L−1). In addition to this pattern, the negative relationship (r2: 0.23, P b 0.04) between CH4 flux and soil air CH4 concentration in the 0–20 cm soil depth suggests the occurrence of a zone of active CH4 oxidation in the upper soil layers. While CO2 emission was significantly (P b 0.001) higher at the flood-impacted than at the flood-protected sites, the opposite was observed with regard to CH4 uptake. Depending on soil temperature, flood events triggered spikes in CH4 emission (up to +45.1 mg CH4-C m−2 d−1 at the FF mature forest). Among the mature forests, mean flux was +0.61, −1.57 and −3.12 mg CH4-C m−2 d−1 at the FF, OF and FP site, respectively. These results demonstrate that some riparian forests can act as strong terrestrial CH4 sinks, but that potential can be easily offset with increased frequency of flooding. Thus, a characterization of flood frequency is required for large scale assessments of CH4 fluxes in riparian ecosystems. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Riparian buffers can attenuate the transfer of pollutants to adjacent surface water bodies, and thus act as natural filters in the landscape. Depending on geomorphic settings and climate, the connection between a riparian buffer and an adjacent stream channel can be variable. The nature and degree of that connectivity can influence a range of biotic and abiotic drivers, and ultimately determine the efficiency of pollutant removal in riparian buffers. Research conducted during the last several decades has provided ample documentation of the water quality protection function of riparian ecosystems (Lowrance et al., 1997; Polyakov et al., 2005), but much less is known regarding the dynamics of greenhouse gases (GHG) in riparian buffers. Groffman et al. (1998) noted the discrepancy between the large amount of research on nutrient removal and the lack of data on trace gas dynamics in riparian ecosystems. Owing to their landscape position, riparian areas can be periodically inundated, and that can have immediate and long-term effects on biogeochemical processes in riparian soils, including the exchange of GHG with the atmosphere. Flood events can lead to O2 exclusion from soil pore space, and ultimately result in the production of methane ⁎ Department of Earth Sciences, Indiana University Purdue University Indianapolis (IUPUI), 723 W. Michigan Street, SL 118H, Indianapolis, IN 46202, United States. E-mail address: [email protected].

http://dx.doi.org/10.1016/j.geoderma.2014.10.013 0016-7061/© 2014 Elsevier B.V. All rights reserved.

(CH4) as a by-product of anaerobic metabolism. Flood events can contribute to the redistribution of nutrients and the deposition of organic debris from allochthonous sources across riparian landscapes, and such inputs can augment the supply of organic substrates available to decomposers. Flood regime can also indirectly influence soil respiration through its effect on the floristic composition of riparian vegetation communities, and the amount, distribution (shoot versus root) and chemical composition of plant tissues deposited on riparian land surface (Catford et al., 2011; Rotkin-Ellman et al., 2004; Tufekcioglu et al., 2001). The atmospheric trace gases carbon dioxide (CO2) and methane (CH4) play important roles in the chemistry and thermal balance of the earth's atmosphere, and the steady growth in their atmospheric concentration during the last 150 years (CO 2 : 280 in 1850s to 380 μL L−1 in 2005; CH4: 0.715 to 1.77 μL L−1) has been linked to the accelerated greenhouse effect (IPCC, 2007). Terrestrial ecosystems – forest soils in particular – represent an important CH4 sink with a global uptake capacity of 30–40 Tg year−1 (Mosier et al., 1997). This uptake is achieved through the oxidation of CH4 by methanotrophic bacteria in soils. Our knowledge of CH4 dynamics in terrestrial ecosystems is derived almost exclusively from research conducted in upland soils, and limited information currently exists with regard to CH4 dynamics in floodaffected riparian forest soils. This question is particularly relevant to the US Midwest region where deciduous forest is the dominant land

42

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

cover in riparian buffers (Palik et al., 2004), some of which are variably affected by flood events. McLain and Martens (2006) reported a net CH4 uptake in semi-arid riparian ecosystems when soil moisture was not severely limiting, whereas Ambus and Christensen (1995) recorded both emission and uptake upon flooding of riparian areas. Kim et al. (2010) reported an effect of vegetation cover on CH4 dynamics, with forested and grass-covered buffers acting as CH4 sinks and sources, respectively. The production of CO2 in riparian ecosystems has been investigated through modeling and laboratory experiments (Jacinthe et al., 2003; Rotkin-Ellman et al., 2004), but only few studies have reported fieldscale measurements of CO2 in riparian areas (McLain and Martens, 2006; Oberbauer et al., 1992; Tufekcioglu et al., 2001). The dynamics of CO2 and CH4 in riparian zones can be affected by bio-physical properties of soils, vegetation attributes, and hydro-climatic factors such as temperature, rainfall, and flood frequency. At the present, however, the impact of these factors on CO2 and CH4 fluxes in riparian buffers is not well characterized. Past studies have shown that flood history affect riparian vegetation community composition (Catford et al., 2011; Rotkin-Ellman et al., 2004), and the distribution of sediment, organic matter and nutrients across riparian landscapes (Blazejewski et al., 2009). Flood frequency was also found to have both short-term and long-term effects on the nitrous oxide-producing capacity of riparian soil (Jacinthe et al., 2012). In light of the foregoing, it is hypothesized that flood frequency determines the dynamics of trace gases in riparian buffers, with higher emission rates expected in flood-impacted than in flood-protected riparian areas. Therefore, the objective of this study was to assess the effect of vegetation type and flood frequency on CO2 and CH4 fluxes in riparian buffers.

2. Materials and methods 2.1. Description of study sites The study was conducted at the Lilly Arbor restored floodplain (N 39°46′, W 86°11′), Southwestway Park (N 39°39′, W 86°14′) and McCormick's Creek State Park (N 39°17′, W 86°44′) along a fourth-order stretch (70 km) of the White River in South-Central Indiana (USA), between the city of Indianapolis to the north and the town Spencer to the south (Fig. 1). The Lilly Arbor floodplain includes a shrub/grass vegetation community, and woodlots established during restoration of the site in 1999. At that location, mean (1970–2010) river discharge is 46.9 m3 s−1 and the bankfull discharge is ~ 270 m3 s−1. The site is occasionally flooded (2–3 times year−1). The Southwestway Park includes a mature forest (N80 years) and an aggrading (~ 15 years) forest stand established in the mid-1990s on farmland removed from agriculture in 1984. The mature forest is occasionally flooded (2–3 times year−1), whereas the aggrading forest is protected from flooding by a constructed levee. At the McCormick's Creek State Park, two tracts of mature forest were delineated: one tract that is flood-protected due to its position on a second terrace, and a frequently flooded (4–6 times year−1) forest stand at the confluence of the White River and McCormick's Creek. The flow of water from McCormick's Creek (a second order stream) into the White River is sometimes impeded when water level in the White River is high, causing backwater inundation of adjacent low-lying areas (Fig. 1). River discharge and water temperature data were obtained from U.S. Geological Survey (USGS; http://waterwatch.usgs.gov) gauging station 3353000

Sites S1 and S2 at Lilly Arbor

Indianapolis

Sites S3 and S5 at Southwestway Park Spencer

Sites S4 and S6 at McCormick’s Creek State Park McCormick’s Creek

Fig. 1. Location of the study sites in the White River watershed (depicted by the gray area in the Indiana map insert). The black circles indicate locations of the study sites along the White River. The triangle symbols represent the location of the weather stations in Indianapolis and Spencer, IN.

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

(N 39°46′, W 86°11′) located near Indianapolis. This information is summarized in Fig. 2. Based on flood frequency and vegetation, a total of 6 study sites were selected. They included: (S1) occasionally flooded (OF) shrub/grass; (S2) occasionally flooded (OF) aggrading forest; (S3) flood-protected (FP) aggrading forest; (S4) flood-protected (FP) mature forest; (S5) occasionally flooded (OF) mature forest; and (S6) frequently flooded (FF) mature forest. Thus, S3 and S4 are flood-protected (FP) whereas all the other sites (S1, S2, S5 and S6) are affected to a varying degree by overbank flooding. Vegetation at S1 consists of small trees, primarily mulberry (Morus alba), and Siberian elm (Ulmus pumila), and various herbaceous species including barnyard grass (Echinochloa spp.), reed canary (Phalaris arundinacea) and goldenrod (Salidago spp.). At the other sites (S2–S6), vegetation is dominated by silver maple (Acer saccharinum), white oak (Quercus alba), sycamore (Platanus occidentalis), and cottonwood (Populus deltoides). The understory consists of stinging nettle (Urtica dioica L.), Virginia wild rye (Elymus virginicus), cutleaved coneflower (Rudbeckia laciniata) and common green briar (Smilax rotundifolia). Soils at the study sites are developed from glacial outwash and/or alluvium deposits, and include the well-drained Genesee (Fluventic Eutrudepts) and Sloan (Fluvaquentic Endoaquolls) soil series. Soil samples (0–20 cm) were collected from each study site for determination of soil properties, including bulk density (cylinder method), texture (hydrometer method), pH (electrometrically in a 1:2 soil-to-water suspension with an Accumet model 25 pH/ion meter), total C and N (dry combustion at 960 °C using a Vario-Cube C-N analyzer, Elementar Americas, NJ). Gravimetric soil moisture constant was determined (drying at 105 °C for N 48 h), and all results were reported on a dry soil mass basis. Soil properties are summarized in Table 1. The region climate is temperate humid with a mean annual temperature between 11 and 11.6 °C, and rainfall between 1050 and 1140 mm in Indianapolis and Spencer, respectively (midwestern regional climate center, http://mrcc.isws.illinois.edu/). 2.2. Carbon dioxide and methane flux measurements Gas flux was measured by the static chamber method (Jacinthe and Dick, 1997; Jacinthe et al., 2012) from August 2005 to June 2007 at sites

43

S1 and S2, and from June 2006 to November 2007 at the other sites. At each site, two study areas were delineated (S2 had 4 study areas), and each study area was instrumented with a set of four static chambers. Chambers consisted of an opaque PVC pipe (H: 20 cm; diam: 15 cm) with a beveled end inserted 5-cm into the ground. During measurement, the PVC pipe was closed with a lid fitted with a gas sampling port plugged with a butyl rubber septum. Once closed, the chamber headspace was thoroughly mixed by pumping several times with a syringe inserted through the septum. Air samples (10 mL) were withdrawn from the chamber headspace (0, 30 and 60 min) and transferred into pre-evacuated glass vials (5 mL) fitted with gray butyl rubber septa (Microliter, Suwanee, GA). Sampling generally took place between 11:00 and 14:00 h local time. At each sampling occasion, surface soil temperature (0–10 cm) was measured with a portable soil thermometer (Cole Parmer, Vernon Hills, IL). At each sampling occasion, 8 to 16 composite soil samples (0–20 cm) were also collected and brought to the laboratory in plastic bags for determination of gravimetric soil moisture content (105 °C, 72 h). At the S1 and S2 sites, duplicate soil atmosphere samplers were also installed near each set of static chambers. The soil atmosphere sampler design is described in Jacinthe and Lal (2004) and consisted of a PVC rod supporting cells of silicone membrane (15-cm long, 1.6 cm id, 0.24 cm wall-thickness, Cole-Parmer H-06411-82) centered at 20, 40, 60, 80 and 100 cm depth. Our previous studies (Jacinthe and Dick, 1996; Jacinthe and Lal, 2004) have demonstrated the applicability of silicone samplers for monitoring soil atmosphere under the periodically saturated soil conditions expected in riparian zones. Soil air samples (~10 mL) were withdrawn with a syringe fitted with a stopcock and stored in evacuated glass vials until analyzed. Air samples were analyzed for CO2 and CH4 concentration on a Varian CP3800 (Palo Alto, CA) gas chromatograph equipped with thermal conductivity detector (TCD for CO2) in series with a flame ionization detector (FID for CH4). Operating conditions of the gas chromatograph were as follows: carrier gas (UHP He: 20 mL min−1), flame gases (H2: 30 mL min− 1, compressed air: 330 mL min− 1), oven temperature (90 °C), detector temperature (both TCD and FID at 150 °C). The stationary phase consisted of a pre-column (L: 0.3 m; id: 2 mm) and an analytical column (L: 1.8 m; id: 2 mm) packed with Porapak Q (80–100 mesh). Certified gas standards (100–10000 μL CO2 L−1;

800 600 400 200

8 6 4 2

2/1/06

(b)

6/1/06

10/1/06

2/1/07

6/1/07

10/1/07

25 Spring (12) Summer (4)

Fall (8) Winter (19)

o

Duration of event, days

0 10/1/05

Water temperature, C

3

Discharge, m s

-1

(a)

20 15 10 5 0

0 Fig. 2. (a) Daily discharge of the White River during the study period, and (b) seasonal variation in the duration of flooding events (bars), and water temperature (squares) during flooding events for the period 1990–2007. The number of flooding events is listed next to the bar legends. Data are from USGS gauging station 3353000 near Indianapolis.

44

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

Table 1 Physico-chemical properties of soils (0–20 cm) at the riparian sites. Values are means with standard deviations in parentheses (n = 4–8 measurements). Riparian sitesa

pH Bulk density (g cm−3) Sand (%) Clay (%) Organic carbon (g C kg−1) Total N (g N kg−1) C-to-N ratio Soil moisture (g H2O g−1 soil) Soil temperature (°C)

S1

S2

S3

S4

S5

S6

7.8 (0.4) 1 (0.1) 65.5 (3.2) 16.2 (2.4) 30.9 (3.1) 1.8 (0.1) 17.2 0.30 (0.16) 16.3 (5.4)

8.1 (0.3) 1.1 (0.1) 78.3 (4.2) 6.9 (3.1) 37.2 (3.9) 2 (0.2) 18.6 0.33 (0.14) 17.3 (5.8)

7.8 (0.2) 1.2 (0.1) 40.2 (4.8) 29.5 (3.6) 20.2 (2.1) 1.5 (0.1) 13.4 0.27 (0.09) 13.9 (5)

7.3 (0.1) 1 (0.1) 46.4 (3.5) 25.4 (3) 33.1 (9.5) 2.7 (0.7) 12.3 0.37 (0.1) 14.3 (4.9)

7.7 (0.1) 1.1 (0.1) 41.6 (23.7) 14.6 (25.2) 34.3 (9.3) 2 (0.7) 17.1 0.26 (0.05) 15.5 (5)

7.3 (0.2) 1.1 (0.2) 46.3 (20.6) 13.5 (18.7) 22.6 (1.4) 1.7 (0.2) 13.3 0.49 (0.1) 15.7 (5.3)

a S1 = occasionally-flooded, grass/shrub; S2 = occasionally-flooded aggrading forest; S3 = flood-protected aggrading forest; S4 = flood-protected mature forest; S5 = occasionally-flooded mature forest; S6 = frequently-flooded mature forest.

10 μL CH4 L−1) obtained from Matheson Tri-Gas were used for instrument calibration. Daily flux of greenhouse gas (F, mass of gas m−2 d−1) was computed as:  F¼

  ΔC V k Δt A

-

(Fig. 3g–i). Overall, soil was cooler at the flood-protected (14.1 °C) than at the flood-affected sites (16.2 °C). 3.2. Carbon dioxide fluxes

Data were analyzed using one-way analysis of variance (ANOVA) to assess the effect of vegetation and flood regime on CO2 and CH4 flux. In the analysis, vegetation type, or flood frequency (flood-protected, occasionally-, frequently-flooded) was used as the treatment factor, and study areas as pseudo-replicates of treatment. The mature forests occurred under all 3 flood regimes, whereas the aggrading forests were associated with 2 flood regimes (occasionally-flooded and floodprotected). As a result, the effect of flood regime was tested separately for the mature (S4, S5 and S6) and the aggrading forests (S2 and S3). Likewise, the effect of vegetation type was assessed using data from the occasionally-flooded sites supporting shrub/grass (S1) and forest (S2) vegetation. ANOVA was performed using the GLM (general linear modeling) procedure available in SAS 9.3 (SAS, 2002). The procedure REG (regression) was used to evaluate relationships between gas fluxes and environmental factors (soil temperature, moisture). Unless otherwise stated, statistical significance was determined at the 95% confidence level.

During the monitoring period, average daily fluxes of CO2 ranging between 0.3 and 6.2 g CO2-C m−2 d−1 were measured at the riparian sites (Fig. 3j–l). Soil respiration showed a strong seasonal pattern; it was generally lowest during the fall and winter (mean all sites: 1.54 ± 0.81 g CO2-C m−2 d−1) and highest during the spring and summer (3.69 ± 1.34 g CO2-C m−2 d−1). At all sites, maximum CO2 flux was recorded in May/June, near the peak of the growing season. Daily flux of CO2 was not significantly related to soil moisture at any of the study sites, but exhibited significant linear and exponential relationships with soil temperature at all sites (Table 2). From the exponential relationships, the temperature sensitivity of soil respiration (Q10: increase in CO2 flux per 10 °C increase in soil temperature) was assessed. The Q10 values ranged between 1.75 and 3.1 (Table 2), and were inversely related to mean annual soil temperature (y = −0.33x + 7.6, r2 = 0.67, P b 0.04) and C/N ratio of soil organic matter (y = − 0.16x + 4.9, r2 = 0.68, P b 0.04) at the study sites. Overall, vegetation type had only a marginal effect on CO2 flux. In adjacent plots under similar flood regime, CO2 emission was higher in the forested (S2: 2.83 ± 1.72 g CO2-C m−2 d−1) than in the grass-covered riparian buffer (S1: 2.13 ± 1.13 g CO2-C m−2 d−1), but difference was not statistically significant (Table 3). Regardless of the maturity of the forest stands, flood regime had a significant effect on CO2 emission. Compared to the flood-protected sites, CO 2 emission was consistently higher in the flood-affected areas (Table 3). Among the mature riparian forests, CO2 emission was significantly (P b 0.001) higher in the flood-affected (3.32 ± 0.28 g CO2-C m− 2 d− 1) than in the flood-protected stands (1.77 ± 0.14 g CO2-C m− 2 d− 1). The same trend was observed between the flood-affected (S2: 2.83 g CO 2 -C m− 2 d − 1) and flood-protected (S3: 2.11 g CO2-C m− 2 d− 1) aggrading forests (Table 3). Annual CO2 emission ranged between 6.1 and 13 Mg CO2-C ha−1 with rates being on average 1.6 times higher at the flood-affected than at the floodprotected sites.

3. Results

3.3. Methane fluxes

3.1. Soil moisture and temperature

Overall, vegetation type had only a marginal effect on CH4 flux. Although CH4 uptake was more rapid in the forested (S2: − 0.64 mg CH4-C m−2 d−1) than in the adjacent shrub/grass riparian buffer (S1: −0.28 mg CH4-C m−2 d−1, Table 3), difference was not significant, likely due to high variability in the data. Flood regime had a significant effect (P b 0.001) on CH4 fluxes in riparian forests, irrespective of the maturity of the forest stands. The flood-protected sites exhibited the highest rates of CH4 uptake measured in the study, averaging −2.22 and −3.12 mg CH4-C m−2 d−1, respectively in the aggrading (S3) and mature forest

where ΔC / Δt: rate of change of gas concentration inside the chamber during the deployment period (mass gas m−3 air min−1, obtained by linear regression); V: chamber volume (3.53 × 10−3 m3), A: area circumscribed by the chamber (1.77 × 10−2 m2), and k: time conversion factor (1440 min d−1). Gas flux was computed if the linear regression yielded r2 ≥ 0.90 for CO2 and ≥0.70 for CH4. About 10% and 18% of the CO2 and CH4 flux measurements (out of more than 340) made in the present study failed to meet these acceptability criteria, and were therefore rejected. Annual emission was computed for each sampling point by linear interpolation of daily gas fluxes between sampling dates, integrating the area under the curve using SigmaPlot 11.0 (Systat Software). 2.3. Data analysis

Soil moisture content varied throughout the monitoring period in response to rainfall, but variability was more pronounced at the floodaffected sites (Fig. 3d–f). At these sites (site 5 being the exception), instances of saturated soil conditions (0.58–0.69 g water g−1 soil) were observed after flood events. Surface soil temperature at the study sites followed a similar temporal pattern, characterized by a maximum temperature in mid/late summer and a rapid decline starting in autumn

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

Lilly Arbor

(b)

60

40

40

40

20

20

20

S3

(e)

0.6

S5

0.4

0.4

0.4

0.2

0.2

0.2

(h)

24

18

18

18

12

12

12

6

6

6

0

0

(k)

6

6

4

4

4

2

2

2

-4

-4

-4 -6 9/1/07

-2

0

11/1/07

-2

(o)

-2

7/1/07

0

5/1/07

0

(l)

2

3/1/07

2

1/1/07

2

20

(n)

11/1/06

4

9/1/06

(m)

7/1/06

4

6

12/1/07

(j)

9/1/05 11/1/05 1/1/06 3/1/06 5/1/06 7/1/06 9/1/06 11/1/06 1/1/07 3/1/07 5/1/07 7/1/07

mg CH4-C m

-2 -1 d

g CO2-C m

-2 -1 d

0

(i)

10/1/07

24

8/1/07

(g)

6/1/07

24

S4 S6

(f)

4/1/07

0.6

2/1/07

S1 S2

12/1/06

(d)

(c)

10/1/06

(a)

McCormick's Park 80

60

0.6

o Temperature, C

Moisture, g water g

Soutwestway Park 80

60

-1

Rainfall, mm

80

45

Fig. 3. Rainfall (a–c), gravimetric soil moisture (d–f), soil temperature (g–i), carbon dioxide flux (j–l), and methane flux (m–o) at the riparian sites. Errors bars represent standard deviations of the mean (N = 8–16 measurements). The star symbols (graph panels d–f) denote the flood events observed during the study. Note that CH4 fluxes are reported on different scales for sites S4 and S6. Description of riparian sites: S1 = occasionally-flooded, grass/shrub; S2 = occasionally-flooded aggrading forest; S3 = flood-protected aggrading forest; S4 = floodprotected mature forest; S5 = occasionally-flooded mature forest; S6 = frequently-flooded mature forest.

(S4) stands. In comparison, CH4 uptake in the flood-affected forest stands (S1, S2, S5, S6) was significantly lower, averaging − 0.47 mg CH4-C m−2 d−1 (Table 3). The frequently-flooded riparian forest was a net source of CH4, emitting +2.6 kg CH4-C ha−1 year−1 to the atmosphere (Table 3), 44% of which was associated with one flooding event at the end of April 2007 (Fig. 3o). All the other sites were net CH4 sinks, but the sink strength declined with increased flood frequency, averaging − 8.2 and − 2.2 kg CH4-C ha−1 in the FP and OF sites, respectively (Table 3). Under the OF flood regime, CH4 consumption in the aggrading forest (S2) was equivalent to 48% of the level in the mature forest (S5). In contrast, between the flood-protected sites, the aggrading forest (S3) consumed

CH4 at a rate equivalent to 62% of the level in the mature riparian forest (S4). At all sites, maximum rates of CH4 uptake were measured between mid-summer and autumn (July–September), and that was noticeably so in 2007 probably due to below normal rainfall and low soil moisture during that period (Fig. 3a–c). Strong pulses of CH4 emission were observed in spring 2007 at the flood-prone study sites (S1, S2, S6; Fig. 3). Sites S1 and S2 were flooded for 2 days (March 3–4, 2007), and measurements made on March 7 showed that the sites were small CH4 sources (mean: + 0.001; max: + 0.41 mg CH4-C m− 2 d−1). These study sites experienced another flood a few weeks later (March 24–27, 2007), and strong CH4 emission was recorded on March 30 (S1: +2.67; S2: +0.52 mg CH4-C m−2 d−1),

46

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

Table 2 Relationships between soil temperature (T) and carbon dioxide flux (R) at the study sites. Sitesa

Linear

Exponential

Equation S1 S2 S3 S4 S5 S6

R R R R R R

= = = = = =

0.13T 0.21T 0.22T 0.12T 0.27T 0.27T

+ − − − − −

0.27 0.38 1.24 0.01 0.91 1.12

r2

P

Equation

0.31 0.42 0.83 0.47 0.68 0.63

0.015 0.003 0.001 0.020 0.001 0.006

R R R R R R

= = = = = =

0.89exp0.056T 0.89exp0.071T 0.33exp0.114T 0.37exp0.100T 0.79exp0.087T 0.55exp0.103T

Q10 r2

P

0.30 0.41 0.85 0.57 0.72 0.64

0.019 0.003 0.001 0.011 0.001 0.005

1.75 2.02 3.1 2.7 2.38 2.8

a S1 = occasionally-flooded, grass/shrub; S2 = occasionally-flooded aggrading forest; S3 = flood-protected aggrading forest; S4 = flood-protected mature forest; S5 = occasionally-flooded mature forest; S6 = frequently-flooded mature forest.

with a peak emission of 14.5 mg CH4-C m−2 d−1 near the river margin. Likewise, backwater inundation resulted in strong CH4 emission (mean: +17.3; max: +41.5; mg CH4-C m−2 d−1) from S6 site on April 24, 2007 (Fig. 3). Peaks in CH4 emission were not noted at S5 because measurements were not possible at that location after the flood events of spring 2007 due to inability to replace, in a timely manner, the static chambers lost in floodwaters; emission spikes associated with these flood events at S5 were not captured. Relationships between CH4 flux and environmental factors varied with study sites. Significant but weak negative relationships (r2 b 0.29, P b 0.04) with soil temperature were found at sites S1, S2 and S4. Significant positive relationships (r2: 0.53–0.64; P b 0.01) between soil moisture with CH4 fluxes were found at S1, S2 and S6 but not at the other sites. 3.4. Carbon dioxide and methane concentration profiles Soil air CO2 concentration varied with season, and that was well illustrated by the progressive increase in concentration observed in the spring/summer as well as the steady decrease in concentration noted in autumn (Table 4). Averaged across sampling dates, soil depth and study sites, concentration was higher in the spring/summer (mean: 15.5 mL CO2 L−1 air) than in the fall (8.1 mL CO2 L−1 air) and winter (3.5 mL CO2 L−1 air) sites. The concentration of CO2 in soil generally increased with depth, but deviation from that trend was observed in late

Table 3 Average daily and annual fluxes of carbon dioxide and methane at the riparian sites. Vegetation and flood frequency at the study sitesa

S1 S3 S2

OF FP OF

S4 S5 S6

FP OF FF

S1 S3 S2

OF FP OF

S4 S5 S6

FP OF FF

↓

↓ ↓

↓

Shrub/grass Increased flood frequency within aggrading forests Increased flood frequency within mature forests

Shrub/grass Increased flood frequency within aggrading forests Increased flood frequency within mature forests

Gas flux Daily averageb

Annual

g CO2-C m−2 d−1 2.13 (1.17)c 2.11 (1.21) b d 2.83 (1.76) a

Mg CO2-C ha−1 7.5 (2.8) 7.8 (0.6) 9.8 (2)

1.77 (0.82) y 3.61 (1.6) x 3.05 (1.7) xy

6.1 (0.6) 13 (2.3) 9.7 (1.3)

mg CH4-C m−2 d−1 −0.28 (0.85) −2.22 (1.02) b −0.64 (0.67) a

kg CH4-C ha−1 −1.2 (2.1) −6.3 (1.2) −1.9 (1.8)

−3.12 (1.86) z −1.57 (1.24) y +0.61 (5.9) x

−10.2 (2) −3.9 (1.3) +2.6 (6.1)

spring/early summer (peak of the growing season). The highest of CO2 concentrations in soil air was recorded on July 18 in the 0–40 cm soil layers (mean: 16.6 at S1, and 22.7 mL CO2 L−1 air at S2; Table 4). Positive relationships (r2: 0.24–0.52, P b 0.04) were found between CO2 emission and CO2 concentration in soil air in the soil layers above 60 cm (Fig. 4). Methane concentration in the soil atmosphere varied to some extent with season, but primarily with soil depth (Table 5). Soil air CH4 concentration in the 0–80 cm layers varied seasonally, being generally highest in spring/summer (range: 0.63–6.26 μL CH4 L−1), moderate in autumn (0.88–3.67 μL CH4 L−1) and lowest in winter (0.58–1.7 μL CH4 L−1). In the 80–100 cm soil layer, elevated CH4 concentration (up to 120 and 9.5 μL CH4 L−1 at S1 and S2, respectively) was measured during the fall and winter seasons. At most sampling dates, the gradient was from a high CH4 concentration at depth to a low concentration near the soil surface. Methane concentration in the 0–20 cm soil layer was almost always lower than in the litter layer (2.35 ± 0.18 μL CH4 L−1, Table 5). A negative relationship (y = − 0.14x − 0.19, r2: 0.23, P b 0.04) was found between CH4 flux (y) and CH4 concentration (x) in the 0–20 cm soil layer (Fig. 4), but there was no clear trend between these variables at soil depth N40 cm. 4. Discussion 4.1. Controlling factors of soil respiration Soil respiration at the study sites was controlled by soil temperature, but not by soil moisture. Such results are typical of studies conducted in the temperate humid region (Mielnick and Dugas, 2000; Wagai et al., 1998). Despite the lack of correlation with soil moisture, CO2 efflux was significantly higher at the flood-affected than at the flood-protected sites. These results therefore suggest that the higher rate of soil respiration at the flood-affected sites was not a direct consequence of greater availability of soil moisture. Indeed, no stimulation of soil respiration was observed in the days immediately following a flood event. For example (Fig. 3l), soil respiration at S4 (flood-protected) and S6 (flooded) was in the same range (2.1 and 3.3 g CO2-C m−2 d−1, respectively) when measurements were made 2 days after the spring 2007 flood event, but 3 weeks later, a large gap in soil respiration between these 2 sites was observed (S4: 1.3; S6: 5.9 g CO2-C m−2 d−1; Fig. 3l). A similar trend was observed between sites S3 (non-flooded; 2.8 g CO2-C m−2 d−1) and S5 (flooded; 6.2 g CO2-C m−2 d−1) during that same period (Fig. 3k). These temporal patterns therefore suggest that the effect of flooding on soil respiration is probably indirect, and likely involves the deposition of nutrients and organic materials transported from upstream locations. Decomposition of these materials does not occur during the flood event since decomposition is generally depressed in wet soil environments due to limited availability of O2. However, as the floodwaters dissipate and soils become aerated, significant enhancements in soil respiration and CO2 emission occur. This interpretation would explain the delayed response of soil respiration to flood events, as well as the trend in CO2 emission in relation to flood frequency (Table 3). Similar to the trend observed in the present investigation, a study conducted in the Thur River watershed (Switzerland) found higher rates of CO2 emission in riparian buffers along flood-affected sections of the river compared to riparian areas near channelized sections of the river (Samaritani et al., 2011). 4.2. Controlling factors of CH4 fluxes

a

Abbreviations: FP = flood-protected; OF = occasionally-flooded; FF = frequentlyflooded. b Average daily flux was computed using data collected during the entire study. Annual flux was computed for the period December 2005–December 2006 at sites S1 and S2, and for the period September 2006–September 2007 at the other sites. c Standard deviation. d Within a column, consecutive numbers followed by different letters are significant different at P b 0.05.

Methane flux across the soil–atmosphere interface is the balance between CH4 formation, oxidation and transport processes, all of which can be affected to varying degree by soil moisture. In addition, CH4 formation and consumption are microbially-mediated processes and, as such, tend to respond positively to increased soil temperature (Priemé and Christensen, 1997; van Hulzen et al., 1999). However, examination

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

47

Table 4 Temporal variation in carbon dioxide concentration (mL CO2 L−1 air) in riparian soil profile as related to vegetation cover. Soil depth (cm)

Sampling date 2005

2006

10/12

10/19

11/3

11/21

1/20

2/15

2/24

4/7

04/24

07/18

Shrub/grassland Surfacea 20 40 60 80 100

0.47 8.86 8.93 10.15 10.35 11.2

0.52 7.02 7.6 9.07 9.84 10.15

0.46 7.14 8.08 8.4 9.1 9.7

0.45 6.16 6.67 7.38 6.84 11.39

0.46 3.31 2.89 4.14 3.04 3.59

0.44 3.52 3.4 3.76 3.99 4.89

0.41 3.39 3.14 3.77 4.04 4.96

0.43 7.31 7.67 7.56 7.26 6.39

0.46 16.66 11.69 11.17 8.25 7.99

0.67 15.22 17.96 21.08 20.8 21.66

Aggrading forest Surface 20 40 60 80 100

0.47 8.13 8.91 10.15 9.7 11.95

0.45 6.21 6.9 7.74 7.83 8.92

0.43 6.1 6.69 7.12 8.05 8.44

0.41 4.25 4.29 5.04 7.15 7.87

0.4 2.12 2.72 2.96 2.78 3.21

0.45 2.78 3.44 3.5 3.78 5.38

0.39 2.89 2.78 3 4.08 5.63

0.42 6.57 5.07 5.01 5.22 5.55

0.52 13.13 9.77 9.36 8.48 8.11

0.72 23.67 21.69 19.67 20.32 22.77

a

Measured at the forest floor.

-2

Carbon dioxide flux, g CO2-C m d

-1

of the CH4 flux data showed that relationships with environmental drivers were site-dependent. Because increased soil temperature can stimulate both the production and oxidation of CH4 (Avery et al., 2003; Priemé and Christensen, 1997), relationships among these variables have been inconsistent. For example, several studies reported no relationship (Kim et al., 2010; Mosier et al., 1997), while others found strong relationships between CH4 flux and soil temperature (Flessa et al., 1995; Hopfensperger et al., 2009). By showing that the relationship with soil

7 6

(a)

5 4 3 20 cm

2

40 cm

1

60 cm

0 4

8

12

16

20

24

-2

Methane flux, mg CH4-C m d

-1

Concentration in soil air, ml CO2 L-1 0.0 -0.1

(b)

-0.2 -0.3 -0.4 -0.5 -0.6 -0.7 0.5

1.0

1.5

2.0

2.5

Concentration in soil air, µl CH4 L-1 Fig. 4. Relationships between gas fluxes and the concentration CO2 (a) and CH4 (b) in soil air at the riparian sites S1 and S2.

temperature was site-dependent, this study results further underscore that complexity. More than soil temperature, soil moisture explained a greater proportion in the variability in the CH4 flux data. The strong relationships (r2: 0.53–0.64) between CH4 flux and soil moisture at S1, S2 and S6 are likely due to the wide range of soil moisture (0.13–0.69 g water g−1) measured at these flood-affected sites (Fig. 3d–f). The lack of a relationship at S5 (r2: 0.15) – which is also a flood-affected site – likely result from the failure to monitor gas fluxes at that site after the flood event at the end of March 2007 (chambers were lost in flood waters). The CH4 concentration profiles in this study resemble those observed at a deciduous forest in Ohio (Jacinthe and Lal, 2004). At most sampling dates, a progressive decrease in soil air CH4 concentration from the subsurface to the soil surface was observed. Although trends were similar, the average concentration of CH4 in soil air during the growing season was higher at the Ohio site (mean: 7.6 μL CH4 L−1 between April and August) than measured in the present study during a similar period (mean: 2.4 μL CH4 L−1, Table 5). Soil texture and drainage characteristic may have contributed to these differences. Since the Ohio forest was established on a poorly drained silt loam, gaseous diffusion restriction likely contributed to the higher CH4 concentration in soil air compared to the sandy soil texture at S1 and S2. Further, elevated CH4 concentration was repeatedly observed in the 80–100 cm soil (Table 5), suggesting that CH4 was mostly produced in the deeper section of the soil profile. However, as this CH4 migrates upward it probably undergoes bacterial consumption as suggested by the progressive decline in CH4 concentration near the soil surface. The negative relationship (Fig. 4b) between CH4 flux and CH4 concentration in the 0–20 cm soil layer also suggests the occurrence of biological CH4 oxidation in that depth range. Although additional data (13CH4 profile for example) are needed to confirm this suggestion, it is worth noting that a similar trend was observed at an Ohio forest (Jacinthe and Lal, 2004) in which increased concentration of CH4 in the upper soil layer was accompanied with increased CH4 uptake. Given the range of CH4 concentration in the upper soil layers at the study sites (0.63–2.8 μL CH4 L− 1), CH4 oxidation probably proceeds as a first-order process and therefore was CH4-limited (KM: 20–45 μL CH4 L− 1, Priemé and Christensen, 1997; Prajapati and Jacinthe, 2014). Therefore, under favorable soil moisture and temperature conditions, higher CH4 concentration in the surface soil layer could induce an increase in methanotrophic activity and consequently an enhancement in CH4 consumption. This trend was just the opposite of what was observed with CO2: increased CO2 production near soil surface (10–20 cm) generally translated into larger rates of CO2 emission (Fig. 4a). These patterns are generally similar to the results reported in previous studies (de Jong and Schappert, 1972;

48

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

Table 5 Temporal variation in methane concentration (μL CH4 L−1 air) in riparian soil profile as related to vegetation cover. Soil depth (cm)

Sampling date 2005

2006

10/12

10/19

11/3

11/21

1/20

2/15

2/24

4/7

04/24

07/18

Shrub/grassland Surfacea 20 40 60 80 100

2.39 1.1 1.29 1.14 1.1 13.9

2.39 1.69 1.72 1.72 2.56 101.5

2.2 0.91 1.05 0.89 0.94 119.1

2.19 1.33 1.73 1.56 3.67 114.5

2.53 1.13 1.39 1.18 1.21 2.54

2.23 0.79 1.12 1.05 1.06 1.48

1.95 0.66 1.31 0.92 0.94 0.93

2.06 1.08 1 1.01 1.35 3.09

2.45 1.19 1.57 1.69 1.91 2.68

2.87 1.34 1.15 0.95 0.75 1.06

Aggrading forest Surface 20 40 60 80 100

2.45 1.21 1.09 0.91 1.38 2.02

2.36 2.78 1.47 1.05 0.97 1.1

2.33 1.49 0.9 0.89 1.06 1.57

2.36 1.96 1.35 1.16 1.16 1.06

2.63 1.41 1.16 0.94 1.38 1.74

2.24 1.27 0.79 0.91 1.07 0.82

1.99 0.58 0.67 0.63 1.22 9.58

2.12 0.99 1.23 1.19 1.23 4.41

2.43 1.13 1.78 7.01 3.81 6.26

2.9 1.63 0.72 0.83 1.15 3.43

a

Measured at the forest floor.

Jacinthe and Lal, 2004; Klemedtsson and Klemedtsson, 1997). Since surface efflux is the predominant fate of the CO2 evolved in soils, linear relationships between CO2 production and surface emission are to be expected. However, with regard to CH4 the relationships between production and emission are variable and complex because increased CH4 availability can stimulate methanotrophic activity, and that can in turn lead to the consumption of a sizable portion of the CH4 produced in the soil profile.

(− 7.9; Tang et al., 2013), and deciduous forests in northwestern Pennsylvania (−8.9; Bowden et al., 2000). The flood-protected riparian forests exhibited CH4 uptake capacity that ranks among the highest reported in the literature. Therefore, these secondary-growth forests, common in the unglaciated region of Indiana, Ohio and Illinois may represent a larger CH4 sink than has been assumed so far, and therefore must be included in regional assessments of CH4 budgets. 4.4. Sensitivity of methane fluxes to flood regime

4.3. Riparian forests as strong CH4 sinks The static chamber method is the most common technique to measure GHG fluxes in terrestrial ecosystems, but pressure disturbance during chamber deployment and headspace sampling has long been regarded as a shortcoming of the method (Hutchinson and Livingston, 2001). When air samples are withdrawn from a chamber headspace, a negative pressure can be created and this could induce advective flow. For example, if CH4 concentration in the litter and soil underneath the chamber is below atmospheric level, advective flow can result in the dilution of CH4 concentration in the chamber headspace, and this dilution can erroneously be interpreted as CH4 uptake. In the present study, however, the average concentration of CH4 in the litter layer (2.35 ± 0.18 μL CH4 L−1, Table 5) was similar to the CH4 level in the chamber headspace at the time of deployment (2.47 ± 0.78 μL CH4 L−1), suggesting that the chamber method was not affected by advection. Thus, the CH4 fluxes measured in the study are reasonable estimates of CH4 uptake at the riparian sites. A surprising result of this research was the high rate of CH4 uptake measured at most of the riparian sites. With the exception of the FF site, all the riparian buffers monitored in this study were net CH4 sinks. At these sites, annual rates of CH4 uptake (− 1.2 to − 10.2 kg CH4-C ha−1 year−1, Table 3) were in the upper range reported in the literature for various world eco-regions (Dutaur and Verchot, 2007; Potter et al., 1996), as well as for other US regions including New York's Adirondack (− 2.2 kg CH4-C ha−1 year−1; Yavitt et al., 1993), Ohio (− 2.2; Jacinthe and Lal, 2004), and Iowa (− 1.7; Chan and Parkin, 2001). The magnitude of the CH4 sink in the flood-protected riparian forests was remarkably strong, averaging − 6.3 and − 10.2 kg CH4-C ha−1 year−1 in the aggrading and mature forest stands, respectively (Table 3). These rates (kg CH4-C ha−1 year−1) are comparable to those measured in the southern Ecuadorian Andes (−5.9; Wolf et al., 2012), Japanese larch (Larix kaempferi Sarg.) forest on volcanic soils (− 6.3; Ueyama et al., 2012), beech forest (Fagus sylvatica) in New Zealand (− 7.9; Price et al., 2004), desert steppe of Inner Mongolia

It is well documented that the oxidation of CH4 by methanotrophic bacteria is a process that is very sensitive to different types of disturbance including air pollution, heavy grazing, mineral N fertilizer application, and plowing of agricultural lands. Atmospheric N deposition has been blamed for observed declines in CH4 uptake in urban lawns and forests (Goldman et al., 1995; Kaye et al., 2004; Klemedtsson and Klemedtsson, 1997). The inhibitory effect of atmospheric N deposition and fertilizer application on CH4 uptake is thought to involve competition between CH4 and NH+ 4 for the CH4-monooxygenase enzyme in methanotrophic bacteria (Hanson and Hanson, 1996; Hütsch, 1998). Several studies have also documented the impact of physical land disturbance (surface mining, plowing, grazing) on CH4 uptake and the long-lasting effect of that disturbance on soil methanotrophic activity in soils (Hütsch, 1998; Tang et al., 2013). Flood events can elicit significant change in the soil redox environment, and therefore can result in enhanced CH4 emission during such events. While this immediate effect can reasonably be hypothesized, the long-term effect of flooding on the CH4 consumption capacity of riparian soils is more difficult to ascertain. To assess that impact, the data collected between July and November 2007 was used to compare the flood-protected versus flood-impacted sites. During that period, both total precipitation in Central Indiana (270 mm) and river discharge (8.44 m3 s−1) were below normal (360 mm and 17.33 m3 s−1). Despite similar soil moisture (0.30 and 0.33 g water g−1) and temperature (19.3 and 20.2 °C) during that period, CH4 uptake was significantly (P b 0.01) greater in the flood-protected riparian forests (mean of S3 and S4: −3.51 mg CH4-C m−2 d−1) than in the flood-affected stands (mean of S5 and S6: −0.88 mg CH4-C m−2 d−1; Fig. 3n–o). Thus, even during relatively dry periods, the flood-affected riparian forests consumed CH4 at much lower rates compared to their adjacent flood-protected forest stands. These results suggest that, in addition to eliciting bursts of CH4 emission, flood events negatively affect the CH4 sink strength of riparian soils, and thus should be considered a form of physical disturbance. Since the flux of CH4 measured at the soil surface is the balance between CH4 production, consumption and transport processes, any of these

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50 Table 6 Soil temperature and post-flood maximum methane emission in riparian forests in the White River watershed. Sampling date

CH4 emission (mg CH4-C m−2 d−1)

Soil temperature (°C)

Reference

March 7, 2007 March 30, 2007 April 24, 2007 April 1, 2010 July 1, 2010 Mar 24, 2011

0.41 11.59 41.51 0.25 44.52 1.07

2.4 12.2 16.3 6.8 22.5 11.9

This study (S1 and S2) This study (S1 and S2) This study (S6) Jacinthe et al. (in review) Jacinthe et al. (in review) Jacinthe et al. (in review)

factors can potentially contribute to the lower rates of CH4 uptake measured at the flood-affected riparian forests. Future studies should attempt to separate the effect of these factors, including the CH4 oxidation capacity of methanotrophs and CH4 diffusivity in flooded and nonflooded riparian soils.

4.5. Temperature dependence of flood impact of CH4 emission At a riparian forest along a similar stretch of the White River (30 km northeast of S6), a summer flooding event (July 2010) resulted in CH4 emission as high as + 44.5 mg CH4-C m− 2 d−1 (Jacinthe et al., in review). Curiously, the emission rate was of similar magnitude as the maximum post-flood emission (+41.5 mg CH4-C m−2 d−1) measured at S6 after the flood of April 2007. In addition to the flooding events noted above, the riparian areas experienced flooding on several other occasions (Table 6), but these flood events resulted in only minor increase in CH4 emission. Soil was saturated (0.58–0.69 g water g− 1 soil) during all these events, and low O2 in the soil atmosphere can be assumed. Yet, CH4 emission was low in some cases, but intense in others (Table 6). It appears that soil temperature was the key factor controlling the differential impact of flood event on CH4 emission. Indeed, exponential relationships were found between soil temperature (°C) and postflood maximum CH4 emission (mg CH4-C m−2 d−1) in White River riparian buffers (y = exp0.167T, Jacinthe et al., in review; y = exp0.227T, this study). Both the dominant pathway (acetoclastic versus CO2 reduction) and the rate of CH4 production vary with temperature (Avery et al., 2003; Itoh et al., 2008). It has also been shown that the optimum temperature for steady CH4 production is generally in the 20–30 °C range (van Hulzen et al., 1999). Thus, despite saturated soil conditions, soil temperature (range: 2.4–6.8) was probably less than optimum for methanogenesis after some of the early spring floods. An examination of the data presented in Table 6 suggests that 12 °C is the temperature threshold above which major impact of floods on CH4 emission should be expected. Therefore, soil/water temperature at the time of a flood event is an important factor to consider when assessing the impact of floods on the CH4 budget of riparian ecosystems. The White River discharge data for the period 1990–2007 was analyzed to determine the frequency of flooding events (discharge ≫ bankfull for at least 2 days). A total of 46 such events were identified over that period (average: 2.7 events year− 1), with most of them occurring in the winter and early spring when water temperature is generally cold (Fig. 2). In the summer, flood duration was generally longer and water temperature was warmer compared to the winter/early spring period (Fig. 2). Although a more complete analysis of watershed hydrology is needed, the trends described above and the data compiled in Table 6 strongly suggest that the late spring and summer flood events, though less frequent, can be the main determinants of CH4 budget in the region's riparian forests. For example, without the flood event at the end of April 2007, site S6 would have been a moderate CH4 sink (−1.3 kg CH4-C ha−1 year−1) instead of being a strong source. Thus, one single flooding event can change the direction of CH4 exchange between riparian soils and the atmosphere.

49

Several climate change models (Kothavala, 1997; Mishra et al., 2010) have predicted alteration of the hydrologic cycle in the US Midwest, including increase in total precipitation, and increased frequency of droughts and periods of excessive rainfall during the summer. If these predictions are correct, there could be an increase in overbank flooding of riparian areas throughout the region. The question then to be addressed in future studies is how these hydroclimatic alterations could affect the contribution of riparian buffers to regional GHG budget. 5. Conclusions The dynamics of trace gases in riparian soils is one of the least documented elements of riparian zone biogeochemistry. The present study found that flood regime is a strong controller of the dynamics of CO2 and CH4 in these ecosystems. Emission of CO2 was found to increase with flood frequency, being on average 1.6 higher in flood-affected than in flood-protected riparian forests, likely related to the inputs of organic materials from allochthonous sources during flood events. The flood-protected riparian forests were surprisingly strong CH4 sinks, exhibiting some of the highest rates of CH4 uptake in temperate region forests. However, with increased frequency of flooding, the direction of land–atmosphere CH4 exchange progressively changes from sink to source. These results are likely the product of both long-term and short-term impacts of flooding on CH4 dynamics. Even during nonflood periods, rate of CH4 uptake in flood-affected sites was 4-fold lower than in flood-protected area, suggesting perhaps a long-lasting effect of flood events on soil methanotrophic community. Results showed that overbank flooding could elicit hot moments of CH4 emission and, more than any other soil parameters and environmental factors, could readily convert a riparian ecosystem from a moderate sink to a strong source of CH4. Therefore, a characterization of flood frequency is needed in order to extrapolate field-based measurements to large scale assessments of CH4 fluxes in riparian ecosystems. Acknowledgments The author thanks Jonathan S. Bills, Alice Enochs, April Herman, Brandon Lewis, Andrew Schoering, and Codi Weiler who assisted with the collection and analysis of soil air samples. Special thanks to the Center for Earth and Environmental Sciences, the Indianapolis Department of Parks and Recreation, and the Indiana Department of Natural Resources for providing access to the study sites. The study was funded through a 2006 USGS 104(b) grant (Indiana Water Resources Research Center) (561-0527-01). Financial support through USDA-NRI (200935112-05241) grant is also acknowledged. References Ambus, P., Christensen, S., 1995. Spatial and seasonal nitrous oxide and methane fluxes in Danish forest ecosystems, grassland ecosystems, and agro-ecosystems. J. Environ. Qual. 24, 993–1001. Avery, G.B., Shannon, R.D., White, J.R., Martens, C.S., Alperin, M.J., 2003. Controls on methane production in a tidal freshwater estuary and a peatland: methane production via acetate fermentation and CO2 reduction. Biogeochemistry 62, 19–37. Blazejewski, G.A., Stolt, M.H., Gold, A.J., Gurwick, N., Groffman, P.M., 2009. Spatial distribution of carbon in the subsurface of riparian zones. Soil Sci. Soc. Am. J. 73, 1733–1740. Bowden, R.D., Rullo, G., Stevens, G.R., Steudler, P.A., 2000. Soil fluxes of carbon dioxide, nitrous oxide, and methane at a productive temperate deciduous forest. J. Environ. Qual. 29, 268–276. Catford, J.A., Downes, B.J., Gippel, C.J., Vesk, P.A., 2011. Flow regulation reduces native plant cover and facilitates exotic invasion in riparian wetlands. J. Appl. Ecol. 48, 432–442. Chan, A.S.K., Parkin, T.B., 2001. Effect of land use on methane flux from soil. J. Environ. Qual. 30, 786–797. de Jong, E., Schappert, H.J.V., 1972. Calculation of soil respiration and activity from CO2 profiles in the soil. Soil Sci. 113, 328–333. Dutaur, L., Verchot, L.V., 2007. A global inventory of the soil CH4 sink. Glob. Biogeochem. Cycles 21, GB4013. Flessa, H., Dorsch, P., Beese, F., 1995. Seasonal variation of N2O and CH4 fluxes in differently managed arable soils in southern Germany. J. Geophys. Res. - Atmos. 100, 23115–23124.

50

P.A. Jacinthe / Geoderma 241–242 (2015) 41–50

Goldman, M.B., Groffman, P.M., Pouyat, R.V., McDonnel, M.J., Pickett, S.T.A., 1995. CH4 uptake and N availability in forest soils along an urban to rural gradient. Soil Biol. Biochem. 27, 281–286. Groffman, P.M., Gold, A.J., Jacinthe, P.A., 1998. Nitrous oxide production in riparian zones and groundwater. Nutr. Cycl. Agroecosyst. 52, 179–186. Hanson, R.S., Hanson, T.E., 1996. Methanotrophic bacteria. Microbiol. Rev. 60, 439–476. Hopfensperger, K.N., Gault, C.M., Groffman, P.M., 2009. Influence of plant communities and soil properties on trace gas fluxes in riparian northern hardwood forests. For. Ecol. Manag. 258, 2076–2082. Hutchinson, G.L., Livingston, G.P., 2001. Vents and seals in non-steady-state chambers used for measuring gas exchange between soil and the atmosphere. Eur. J. Soil Sci. 52, 675–682. Hütsch, B.W., 1998. Tillage and land use effects on methane oxidation rates and their vertical profiles in soil. Biol. Fertil. Soils 27, 284–292. IPCC, 2007. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. Itoh, M., Ohte, N., Koba, K., Sugimoto, A., Tani, M., 2008. Analysis of methane production pathways in a riparian wetland of a temperate forest catchment, using δ13C of pore water CH4 and CO2. J. Geophys. Res. Biogeosci. 113, G03005. Jacinthe, P.A., Dick, W.A., 1996. Use of silicone tubing to sample nitrous oxide in the soil atmosphere. Soil Biol. Biochem. 28, 721–726. Jacinthe, P.A., Dick, W.A., 1997. Soil management and nitrous oxide emissions from cultivated fields in Southern Ohio. Soil Tillage Res. 41, 221–235. Jacinthe, P.A., Lal, R., 2004. Effects of soil cover on fluxes and depth concentration of trace gases. Soil Sci. 169, 243–259. Jacinthe, P.A., Groffman, P.M., Gold, A.J., 2003. Carbon dioxide production and dissolved organic carbon dynamics in a riparian aquifer: effects of hydrology and nitrate enrichment. J. Environ. Qual. 32, 1365–1374. Jacinthe, P.A., Bills, J.S., Tedesco, L.P., Barr, R.C., 2012. Nitrous oxide emission from riparian buffers in relation to vegetation and flood frequency. J. Environ. Qual. 41, 95–105. Jacinthe, P.A., Vidon, P., Fisher, K., Liu, X., Baker, M.E., 2014w. Methane and carbon dioxide fluxes in adjacent cropland and riparian buffers. Biogeochemistry (in review). Kaye, J.P., Burke, I.C., Mosier, A.R., Guerschman, J.P., 2004. Methane and nitrous oxide fluxes from urban soils to the atmosphere. Ecol. Appl. 14, 975–981. Kim, D.G., Isenhart, T.M., Parkin, T.B., Schultz, R.C., Loynachan, T.E., 2010. Methane flux in cropland and adjacent riparian buffers with different vegetation covers. J. Environ. Qual. 39, 97–105. Klemedtsson, A.K., Klemedtsson, L., 1997. Methane uptake in Swedish forest soil in relation to liming and extra N-deposition. Biol. Fertil. Soils 25, 296–301. Kothavala, Z., 1997. Extreme precipitation events and the applicability of global climate models to the study of floods and droughts. Math. Comput. Simul. 43, 261–268. Lowrance, R., Altier, L.S., Newbold, J.D., Schnabel, R.R., Groffman, P.M., Denver, J.M., Correll, D.L., Gilliam, J.W., Robinson, J.L., Brinsfield, R.B., Staver, K.W., Lucas, W., Todd, A.H., 1997. Water quality functions of riparian forest buffers in Chesapeake Bay watersheds. Environ. Manag. 21, 687–712. McLain, J.E.T., Martens, D.A., 2006. Moisture controls on trace gas fluxes in semiarid riparian soils. Soil Sci. Soc. Am. J. 70, 367–377. Mielnick, P.C., Dugas, W.A., 2000. Soil CO2 flux in a tallgrass prairie. Soil Biol. Biochem. 32, 221–228.

Mishra, K.A., Cherkauer, K.A., Shukla, S., 2010. Assessment of drought due to historic climate variability and projected future climate change in the midwestern United States. J. Hydrometeorol. 11, 46–68. Mosier, A.R., Delgado, J.A., Cochran, V.L., Valentine, D.W., Parton, W.J., 1997. Impact of agriculture on soil consumption of atmospheric CH4 and a comparison of CH4 and N2O flux in subarctic, temperate and tropical grasslands. Nutr. Cycl. Agroecosyst. 49, 71–83. Oberbauer, S.F., Gillepsie, C.T., Cheng, W., Gebauer, R., Serra, A.S., Tenhunen, J.D., 1992. Environmental effects on CO2 efflux from riparian tundra in the Northern foothills of the Brooks Range, Alaska, USA. Oecologia 92, 568–577. Palik, B., Tang, S.M., Chavez, Q., 2004. Estimating riparian area extent and land use in the Midwest. Gen. Tech. Rep. NC-248. USDA, Forest Service, North Central Research Station, St. Paul, MN. Polyakov, V., Fares, A., Ryder, M.H., 2005. Precision riparian buffers for the control of nonpoint source pollutant loading into surface water: a review. Environ. Rev. 13, 129–144. Potter, C.S., Davidson, E.A., Verchot, L.V., 1996. Estimation of global biogeochemical controls and seasonality in soil methane consumption. Chemosphere 32, 2219–2246. Prajapati, P., Jacinthe, P.A., 2014. Methane oxidation kinetics and diffusivity in soils under conventional tillage and long-term no-till. Geoderma 230, 161–170. Price, S.J., Sherlock, R.R., Kelliher, F.M., McSeveny, T.M., Tate, K.R., Condron, L.M., 2004. Pristine New Zealand forest soil is a strong methane sink. Glob. Chang. Biol. 10, 16–26. Priemé, A., Christensen, S., 1997. Seasonal and spatial variation of methane oxidation in a Danish spruce forest. Soil Biol. Biochem. 29, 1165–1172. Rotkin-Ellman, M., Addy, K., Gold, A.J., Groffman, P.M., 2004. Tree species, root decomposition and subsurface denitrification potential in riparian wetlands. Plant Soil 263, 335–344. Samaritani, E., Shrestha, J., Fournier, B., Frossard, E., Gillet, F., Guenat, C., Niklaus, P.A., Pasquale, N., Tockner, K., Mitchell, E.A.D., Luster, J., 2011. Heterogeneity of soil carbon pools and fluxes in a channelized and a restored floodplain section (Thur River, Switzerland). Hydrol. Earth Syst. Sci. 15, 1757–1769. SAS, 2002. SAS System for Windows, Version 9.3. SAS Institute Inc, Cary, NC. Tang, S., Wang, C., Wilkes, A., Zhou, P., Jiang, Y.Y., Han, G.D., Zhao, M.L., Huang, D., Schonbach, P., 2013. Contribution of grazing to soil atmosphere CH4 exchange during the growing season in a continental steppe. Atmos. Environ. 67, 170–176. Tufekcioglu, A., Raich, J.W., Isenhart, T.M., Schultz, R.C., 2001. Soil respiration within riparian buffers and adjacent crop fields. Plant Soil 229, 117–124. Ueyama, M., Hamotani, K., Nishimura, W., 2012. Continuous measurement of methane flux over a larch forest using a relaxed eddy accumulation method. Theor. Appl. Climatol. 109, 461–472. van Hulzen, J.B., Segers, R., van Bodegom, P.M., Leffelaar, P.A., 1999. Temperature effects on soil methane production: an explanation for observed variability. Soil Biol. Biochem. 31, 1919–1929. Wagai, R., Brye, K.R., Gower, S.T., Norman, J.M., Bundy, L.G., 1998. Land use and environmental factors influencing soil surface CO2 flux and microbial biomass in natural and managed ecosystems in southern Wisconsin. Soil Biol. Biochem. 30, 1501–1509. Wolf, K., Flessa, H., Veldkamp, E., 2012. Atmospheric methane uptake by tropical montane forest soils and the contribution of organic layers. Biogeochemistry 111, 469–483. Yavitt, J.B., Simmons, J.A., Fahey, T.J., 1993. Methane fluxes in a northern hardwood forest ecosystem in relation to acid precipitation. Chemosphere 26, 721–730.